Select any two versions of this STIG to compare the individual requirements
Select any old version/release of this STIG to view the previous requirements
Check if the system requires a password for entering single-user mode. # grep ':S:' /etc/inittab If /sbin/sulogin is not listed, this is a finding.
Edit /etc/inittab and set sulogin to run in single-user mode. Example line in /etc/inittab: ~:S:wait:/sbin/sulogin Note: The first field in the /etc/inittab line may be any sequence of 1-4 characters.
Use the last command to check for multiple accesses to an account from different workstations/IP addresses. # last -w If users log directly onto accounts, rather than using the switch user (su) command from their own named account to access them, this is a finding (such as logging directly on to oracle). Verify with the SA or the IAO on documentation for users/administrators to log into their own accounts first and then switch user (su) to the account to be shared has been maintained including requirements and procedures. If no such documentation exists, this is a finding.
Use the switch user (su) command from a named account login to access shared accounts. Document requirements and procedures for users/administrators to log into their own accounts first and then switch user (su) to the account to be shared.
Check the system for duplicate account names. Example: # pwck -r If any duplicate account names are found, this is a finding.
Change user account names, or delete accounts, so each account has a unique name.
Perform the following to ensure there are no duplicate UIDs: # cut -d: -f3 /etc/passwd | uniq -d If any duplicate UIDs are found, this is a finding.
Edit user accounts to provide unique UIDs for each account.
Access the system console and make a login attempt. Check for either of the following login banners based on the character limitations imposed by the system. An exact match is required. If one of these banners is not displayed, this is a finding. You are accessing a U.S. Government (USG) Information System (IS) that is provided for USG-authorized use only. By using this IS (which includes any device attached to this IS), you consent to the following conditions: -The USG routinely intercepts and monitors communications on this IS for purposes including, but not limited to, penetration testing, COMSEC monitoring, network operations and defense, personnel misconduct (PM), law enforcement (LE), and counterintelligence (CI) investigations. -At any time, the USG may inspect and seize data stored on this IS. -Communications using, or data stored on, this IS are not private, are subject to routine monitoring, interception, and search, and may be disclosed or used for any USG-authorized purpose. -This IS includes security measures (e.g., authentication and access controls) to protect USG interests- -not for your personal benefit or privacy. -Notwithstanding the above, using this IS does not constitute consent to PM, LE or CI investigative searching or monitoring of the content of privileged communications, or work product, related to personal representation or services by attorneys, psychotherapists, or clergy, and their assistants. Such communications and work product are private and confidential. See User Agreement for details. OR I've read & consent to terms in IS user agreem't.
Edit /etc/issue and add one of the DoD login banners (based on the character limitations imposed by the system). DoD Login Banners: You are accessing a U.S. Government (USG) Information System (IS) that is provided for USG-authorized use only. By using this IS (which includes any device attached to this IS), you consent to the following conditions: -The USG routinely intercepts and monitors communications on this IS for purposes including, but not limited to, penetration testing, COMSEC monitoring, network operations and defense, personnel misconduct (PM), law enforcement (LE), and counterintelligence (CI) investigations. -At any time, the USG may inspect and seize data stored on this IS. -Communications using, or data stored on, this IS are not private, are subject to routine monitoring, interception, and search, and may be disclosed or used for any USG-authorized purpose. -This IS includes security measures (e.g., authentication and access controls) to protect USG interests- -not for your personal benefit or privacy. -Notwithstanding the above, using this IS does not constitute consent to PM, LE or CI investigative searching or monitoring of the content of privileged communications, or work product, related to personal representation or services by attorneys, psychotherapists, or clergy, and their assistants. Such communications and work product are private and confidential. See User Agreement for details. OR I've read & consent to terms in IS user agreem't.
Determine if all logon attempts are being logged. Procedure: Verify successful logins are being logged: # last -R | more If the command does not return successful logins, this is a finding. Verify if unsuccessful logons are being logged: # lastb -R | more If the command does not return unsuccessful logins, this is a finding.
Make sure the collection files exist. Procedure: If there are no successful logins being returned from the "last" command, create /var/log/wtmp: # touch /var/log/wtmp If there are no unsuccessful logins being returned from the "lastb" command, create /var/log/btmp: # touch /var/log/btmp
Check the pam_tally configuration. # more /etc/pam.d/system-auth Confirm the following line is configured, before any "auth sufficient" lines: auth required pam_tally2.so deny=3 If no such line is found, this is a finding.
By default link /etc/pam.d/system-auth points to /etc/pam.d/system-auth-ac which is the file maintained by the authconfig utility. In order to add pam options other than those available via the utility create /etc/pam.d/system-auth-local with the options and including system-auth-ac. In order to set the account lockout to three failed attempts the content should be similar to: auth required pam_access.so auth required pam_tally2.so deny=3 auth include system-auth-ac account required pam_tally2.so account include system-auth-ac password include system-auth-ac session include system-auth-ac Once system-auth-local is written reset the /etc/pam.d/system-auth to point to system-auth-local. This is necessary because authconfig writes directly to system-auth-ac so any changes made by hand will be lost if authconfig is run.
Check the value of the FAIL_DELAY variable and the ability to use it. Procedure: # grep FAIL_DELAY /etc/login.defs If the value does not exist, or is less than 4, this is a finding. Check for the use of pam_faildelay. # grep pam_faildelay /etc/pam.d/system-auth* If pam_faildelay.so module is not present, this is a finding. If pam_faildelay is present only in /etc/pam.d/system-auth-ac: ensure that /etc/pam.d/system-auth includes /etc/pam.d/system-auth-ac. #grep system-auth-ac /etc/pam.d/system-auth This should return: auth include system-auth-ac account include system-auth-ac password include system-auth-ac session include system-auth-ac /etc/pam.d/system-auth-ac should only be included by /etc/pam.d/system-auth. All other pam files should include /etc/pam.d/system-auth. If pam_faildelay is not defined in /etc/pam.d/system-auth either directly or through inclusion of system-auth-ac, this is a finding.
Add the pam_faildelay module and set the FAIL_DELAY variable. Procedure: Edit /etc/login.defs and set the value of the FAIL_DELAY variable to 4 or more. The default link /etc/pam.d/system-auth points to /etc/pam.d/system-auth-ac which is the file maintained by the authconfig utility. In order to add pam options other than those available via the utility create or modify /etc/pam.d/system-auth-local with the options and including system-auth-ac. For example: auth required pam_access.so auth optional pam_faildelay.so delay=4000000 auth include system-auth-ac account include system-auth-ac password include system-auth-ac session include system-auth-ac Once system-auth-local is written ensure the /etc/pam.d/system-auth points to system-auth-local. This is necessary because authconfig writes directly to system-auth-ac so any manual changes made will be lost if authconfig is run.
If there is an application running on the system continuously in use (such as a network monitoring application), ask the SA what the name of the application is. Verify documentation exists for the requirement and justification of the application. If no documentation exists, this is a finding. Execute "ps -ef | more" to determine which user owns the process(es) associated with the application. If the owner is root, this is a finding.
Configure the system so the owner of a session requires a continuous screen display, such as a network management display, is not root. Ensure the display is also located in a secure, controlled access area. Document and justify this requirement and ensure the terminal and keyboard for the display (or workstation) are secure from all but authorized personnel by maintaining them in a secure area, in a locked cabinet where a swipe card, or other positive forms of identification, must be used to gain entry.
Verify the system will not log in accounts with blank passwords. # grep nullok /etc/pam.d/system-auth /etc/pam.d/system-auth-ac If an entry for nullok is found, this is a finding on Linux.
Edit /etc/pam.d/system-auth and remove the "nullok" setting.
Check the system for duplicate UID 0 assignments by listing all accounts assigned UID 0. Procedure: # awk -F: '($3 == 0) { print $1 }' /etc/passwd If any accounts other than root are assigned UID 0, this is a finding.
Remove or change the UID of accounts other than root that have UID 0.
Determine if root is assigned a home directory other than / by listing its home directory. Procedure: # awk -F: '($1 == "root") { print $6 }' /etc/passwd If the root user home directory is /, this is a finding.
The root home directory should be something other than / (such as /roothome). Procedure: # mkdir /rootdir # chown root /rootdir # chgrp root /rootdir # chmod 700 /rootdir # cp -r /.??* /rootdir/. Then, edit the passwd file and change the root home directory to /rootdir. The cp -r /.??* command copies all files and subdirectories of file names beginning with "." into the new root directory, which preserves the previous root environment. Ensure you are in the "/" directory when executing the "cp" command.
Check the mode of the root home directory. Procedure: # find ~root -type d -prune -exec ls -ld {} \; If the mode of the directory is not set to 0700 or less permissive, this is a finding. If the home directory is /, this check will be marked "Not Applicable".
The root home directory will be configured to have permission set of 0700 or less permissive. Do not change the protections of the / directory. Use the following command to change protections for the root home directory: # chmod 0700 /rootdir.
To view the root user's PATH, log in as the root user, and execute: # env | grep PATH This variable is formatted as a colon-separated list of directories. Relative path entries must be documented with the ISSO. If there is an empty entry, such as a leading or trailing colon, or two consecutive colons, this is a finding. If an entry begins with a character other than a slash (/), or has not been documented with the ISSO, this is a finding.
Edit the root user's local initialization files ~/.profile,~/.bashrc (assuming root shell is bash). Remove any empty path entries. Remove any relative path entries that have not been documented with the ISSO.
Check for world-writable permissions on all directories in the root user's executable search path. Procedure: # ls -ld `echo $PATH | sed "s/:/ /g"` If any of the directories in the PATH variable are world-writable, this is a finding.
For each world-writable path in root's executable search path, do one of the following: 1. Remove the world-writable permission on the directory. Procedure: # chmod o-w <path> 2. Remove the world-writable directory from the executable search path. Procedure: Identify and edit the initialization file referencing the world-writable directory and remove it from the PATH variable.
Check /etc/securetty # more /etc/securetty If the file does not exist, or contains more than "console" or a single "tty" device this is a finding.
Create if needed and set the contents of /etc/securetty to a "console" or "tty" device. # echo console > /etc/securetty or # echo tty1 > /etc/securetty
Confirm all accounts with a GID of 499 and below are used by a system account. Procedure: List all the users with a GID of 0-499. # cut -d: -f 1,4 /etc/passwd|egrep ":[1-4][0-9]{2}$|:[0-9]{1,2}$" If a GID reserved for system accounts (0 - 499) is used by a non-system account, this is a finding.
Change the primary group GID numbers for non-system accounts with reserved primary group GIDs (those less or equal to 499).
Perform the following to ensure there are no GIDs referenced in /etc/passwd not defined in /etc/group: # pwck -r If GIDs referenced in /etc/passwd are not defined in /etc/group are returned, this is a finding.
Add a group to the system for each GID referenced without a corresponding group.
Ask the SA or ISSO if a host-based intrusion detection application is loaded on the system. Per OPORD 16-0080 the preferred intrusion detection system is McAfee HBSS available through Cybercom. If another host-based intrusion detection application is in use, such as SELinux, this must be documented and approved by the local Authorizing Official Procedure: Examine the system to see if the Host Intrusion Prevention System (HIPS) is installed # rpm -qa | grep MFEhiplsm Verify that the McAfee HIPS module is active on the system: # ps -ef | grep -i “hipclient” If the MFEhiplsm package is not installed, check for another intrusion detection system: # find / -name <daemon name> Where <daemon name> is the name of the primary application daemon to determine if the application is loaded on the system. Determine if the application is active on the system: # ps -ef | grep -i <daemon name> If the MFEhiplsm package is not installed and an alternate host-based intrusion detection application has not been documented for use, this is a finding. If no host-based intrusion detection system is installed and running on the system, this is a finding.
Install and enable the latest McAfee HIPS package, available from Cybercom. If the system does not support the McAfee HIPS package, install and enable a supported intrusion detection system application and document its use with the Authorizing Official.
Obtain the list of available package security updates from Red Hat. Check the available package security updates have been installed on the system. Use the "rpm" command to list the packages installed on the system. Example: # rpm -qa -last If updated packages are available and applicable to the system and have not been installed, this is a finding. One source for the list of Red Hat updates is available at https://access.redhat.com/security/updates/active/
Install the patches or updated packages available from the vendor.
Check system directories for uneven file permissions. Procedure: # ls -lL /etc /bin /usr/bin /usr/lbin /usr/usb /sbin /usr/sbin Uneven file permissions exist if the file owner has less permissions than the group or other user classes. If any of the files in the above listed directories contain uneven file permissions, this is a finding.
Change the mode of files with uneven permissions so owners do not have less permission than group or world users.
Check the system for files with no assigned owner. Procedure: # find / -ignore_readdir_race -nouser If any files have no assigned owner, this is a finding. Caution should be used when centralized authorization is used because valid files may appear as unowned due to communication issues.
All directories and files (executable and data) will have an identifiable owner and group name. Either trace files to an authorized user, change the file's owner to root, or delete them. Determine the legitimate owner of the files and use the chown command to set the owner and group to the correct value. If the legitimate owner cannot be determined, change the owner to root (but make sure none of the changed files remain executable because they could be Trojan horses or other malicious code). Examine the files to determine their origin and the reason for their lack of an owner/group.
Check the mode of network services daemons. # find /usr/sbin -type f -perm +022 -exec stat -c %a:%n {} \; This will return the octal permissions and name of all files that are group or world writable. If any network services daemon listed is world or group writable (either or both of the 2 lowest order digits contain a 2, 3 or 6), this is a finding. Note: Network daemons not residing in these directories (such as httpd or sshd) must also be checked for the correct permissions.
Change the mode of the network services daemon. # chmod go-w <path>
Check the mode of log files. Procedure: # find /var/log /var/log/syslog /var/adm -type f -perm -640 \! -perm 640 With the exception of /var/log/wtmp, /var/log/Xorg.0.log, and /var/log/gdm/:0.log, if any of the log files have modes more permissive than 0640, this is a finding.
Change the mode of the system log file(s) to 0640 or less permissive. Procedure: # chmod 0640 /path/to/system-log-file Note: Do not confuse system log files with audit logs.
Check skeleton files permissions. # ls -alL /etc/skel If a skeleton file has a mode more permissive than 0644, this is a finding.
Change the mode of skeleton files with incorrect mode: # chmod 0644 <skeleton file>
Perform the following to check NIS file ownership: # ls -la /var/yp/*; If the file ownership is not root, sys, or bin, this is a finding.
Change the ownership of NIS/NIS+/yp files to root, sys or bin. Procedure (example): # chown root <filename>
Perform the following to check NIS file group ownership: # ls -la /var/yp/* If the file group ownership is not root, sys, or bin, this is a finding.
Perform the following to change NIS file ownership. # chgrp root /var/yp/*
Perform the following to check NIS file permissions. # ls -la /var/yp/* If the file's mode is more permissive than 0755, this is a finding.
Change the mode of NIS/NIS+/yp command files to 0755 or less permissive. Procedure (example): # chmod 0755 <filename>
Check the mode of the manual page files. Procedure: # find /usr/share/man /usr/share/info /usr/share/infopage -type f -perm +022 -exec stat -c %a:%n {} \; |> more Note: This list only displays manual files with offending permissions. If any of the manual page files have a mode more permissive than 0644, this is a finding.
Change the mode of manual page files to 0644 or less permissive. Procedure (example): # chmod 0644 /path/to/manpage
Check the mode of library files. Procedure: # DIRS="/usr/lib /lib /usr/lib64 /lib64";for DIR in $DIRS;do find $DIR -type f -perm +022 -exec stat -c %a:%n {} \;;done This will return the octal permissions and name of all group or world writable files. If any file listed is world or group writable (either or both of the 2 lowest order digits contain a 2, 3 or 6), this is a finding.
Change the mode of library files to 0755 or less permissive. Procedure (example): # chmod go-w </path/to/library-file> Note: Library files should have an extension of ".a" or a ".so" extension, possibly followed by a version number.
Check the permissions for files in /etc, /bin, /usr/bin, /usr/lbin, /usr/usb, /sbin, and /usr/sbin. Procedure: # DIRS="/etc /bin /usr/bin /usr/lbin /usr/usb /sbin /usr/sbin";for DIR in $DIRS;do find $DIR -type f -perm +022 -exec stat -c %a:%n {} \;;done This will return the octal permissions and name of all group or world writable files. If any command file is listed and is world or group writable (either or both of the 2 lowest order digits contain a 2, 3 or 6), this is a finding. Note: Elevate to Severity Code I if any command file listed is world-writable.
Change the mode for system command files to 0755 or less permissive taking into account necessary GIUD and SUID bits. Procedure: # chmod go-w <filename>
Check the ownership of system files, programs, and directories. Procedure: # ls -lLa /etc /bin /usr/bin /usr/lbin /usr/usb /sbin /usr/sbin If any of the system files, programs, or directories are not owned by a system account, this is a finding.
Change the owner of system files, programs, and directories to a system account. Procedure: # chown root /some/system/file (A different system user may be used in place of root.)
Check the group-ownership of system files, programs, and directories. Procedure: # ls -lLa /etc /bin /usr/bin /usr/lbin /usr/usb /sbin /usr/sbin If any system file, program, or directory is not owned by a system group, this is a finding.
Change the group-owner of system files to a system group. Procedure: # chgrp root /path/to/system/file (System groups other than root may be used.)
Check the ownership of the /etc/shadow file. # ls -lL /etc/shadow If the /etc/shadow file is not owned by root, this is a finding.
Change the ownership of the /etc/shadow (or equivalent) file. # chown root /etc/shadow
Check the mode of the /etc/passwd file. Procedure: # ls -lL /etc/passwd If /etc/passwd has a mode more permissive than 0644, this is a finding.
Change the mode of the passwd file to 0644. Procedure: # chmod 0644 /etc/passwd
Check the mode of the /etc/shadow file. # ls -lL /etc/shadow If the /etc/shadow file has a mode more permissive than 0400, this is a finding.
Change the mode of the /etc/shadow (or equivalent) file. # chmod 0400 /etc/shadow
Check for the presence of aide on the system: # rpm -qa | grep aide If aide is not installed, ask the SA what file integrity tool is being used to check the system. Check the global crontabs for the presence of an "aide" job to run at least weekly, if aide is installed. Otherwise, check for the presence of a cron job to run the alternate file integrity checking application. # grep aide /etc/cron*/* If a tool is being run, then the configuration file for the appropriate tool needs to be checked for selection lines /bin, /sbin, /lib, and /usr. If the file integrity tool is set to check setuid and setgid, this is not a finding. List all setuid files on the system. Procedure: # find / -perm -4000 -exec ls -l {} \; | more Note: Executing these commands may result in large listings of files; the output may be redirected to a file for easier analysis. Ask the SA or ISSO if files with the setuid bit set have been documented. Documentation must include the owner, group-owner, mode, ACL, and location of the files. If any undocumented file has its setuid bit set, this is a finding.
Document the files with the suid bit set or unset the suid bit on the executable.
If STIGID GEN000220 is satisfied, this is not a finding. List all setgid files on the system. Procedure: # find / -perm -2000 -exec ls -l {} \; | more Note: Executing these commands may result in large listings of files; the output may be redirected to a file for easier analysis. Ask the SA or IAO if files with the setgid bit set have been documented. Documentation must include owner, group-owner, mode, ACL, and location. If any undocumented file has its setgid bit set, this is a finding. If a tool is being run then the configuration file for the appropriate tool needs to be checked for selection lines /bin, /sbin, /lib, and /usr. If a file integrity tool is set to check setuid and setgid, this is not a finding.
Document the files with the sgid bit set or unset the sgid bit on the executable.
Ask the SA for the weekly automated or manual process used to generate a list of setuid files on the system and compare it with the prior list. If no such process is in place, this is a finding. If a file integrity tool is configured to monitor setuid files weekly, this is not a finding. Review the process. If the process does not identify and report changes in setuid files, this is a finding. NOTE: The frequency may be increased to daily, if necessary, in accordance with the contingency plan.
Establish a weekly automated or manual process to generate a list of suid files on the system and compare it with the prior list. To create a list of suid files: # find / -perm -4000 > suid-file-list NOTE: The frequency may be increased to daily, if necessary, in accordance with the contingency plan.
Ask the SA if a weekly automated or manual process is used to generate a list of setgid files on the system and compare it with the prior list. If no such process is in place, this is a finding. If a file integrity tool is configured to monitor setgid files weekly, this is not a finding. NOTE: The frequency may be increased to daily, if necessary, in accordance with the contingency plan.
Establish a weekly automated or manual process to generate a list of setgid files on the system and compare it with the prior list. To create a list of setgid files: # find / -perm -2000 > setgid-file-list NOTE: The frequency may be increased to daily, if necessary, in accordance with the contingency plan.
Check /etc/mtab and verify the "nosuid" mount option is used on file systems mounted from removable media, network shares, or any other file system not containing approved setuid or setgid files. If any of these files systems do not mount with the "nosuid" option, it is a finding.
Edit /etc/fstab and add the "nosuid" mount option to all file systems mounted from removable media or network shares, and any file system not containing approved setuid or setgid files.
Check all world-writable directories have the sticky bit set. Procedure: # find / -type d -perm -002 ! -perm -1000 > wwlist If the sticky bit is not set on a world-writable directory, this is a finding.
Set the sticky bit on all public directories. Procedure: # chmod 1777 /tmp (Replace /tmp with the public directory missing the sticky bit, if necessary.)
Check the ownership of all public directories. Procedure: # find / -type d -perm -1002 -exec ls -ld {} \; If any public directory is not owned by root or an application user, this is a finding.
Change the owner of public directories to root or an application account. Procedure: # chown root /tmp (Replace root with an application user and/or /tmp with another public directory as necessary.)
NOTE: The following commands must be run in the BASH shell. Check global initialization files for the configured umask value. Procedure: # grep umask /etc/* Check local initialization files for the configured umask value. Procedure: # cut -d: -f6 /etc/passwd |xargs -n1 -IDIR find DIR -name ".*" -type f -maxdepth 1 -exec grep umask {} \; If the system and user default umask is not 077, this a finding. Note: If the default umask is 000 or allows for the creation of world-writable files this becomes a Severity Code I finding.
Edit local and global initialization files that contain "umask" and change them to use 077 instead of the current value.
Determine if default system accounts (such as those for sys, bin, uucp, nuucp, daemon, smtp) have been disabled. # cat /etc/shadow If an account's password field (which is the second field in the /etc/shadow file) is "*", "*LK*", or is prefixed with a '!', the account is locked or disabled. If there are any unlocked default system accounts this is a finding.
Lock the default system account(s). # passwd -l <user>
Determine if auditing is enabled. # ps -ef |grep auditd If the auditd process is not found, this is a finding.
Start the auditd service and set it to start on boot. # service auditd start ; chkconfig auditd on
Perform the following to determine the location of audit logs and then check the ownership. Procedure: # grep "^log_file" /etc/audit/auditd.conf|sed s/^[^\/]*//|xargs stat -c %U:%n If any audit log file is not owned by root, this is a finding.
Change the ownership of the audit log file(s). Procedure: # chown root <audit log file>
Perform the following to determine the location of audit logs and then check the mode of the files. Procedure: # grep "^log_file" /etc/audit/auditd.conf|sed s/^[^\/]*//|xargs stat -c %a:%n If any audit log file has a mode more permissive than 0640, this is a finding.
Change the mode of the audit log directories/files. # chmod 0750 <audit directory> # chmod 0640 <audit file>
Verify auditd is configured to audit failed file access attempts. There must be an audit rule for each of the access syscalls logging all failed accesses (-F success=0) or there must both an "-F exit=-EPERM" and "-F exit=-EACCES" for each access syscall. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S creat" | grep -e "-F success=0" # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S creat" | grep -e "-F exit=-EPERM" # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S creat" | grep -e "-F exit=-EACCES" If an "-S creat" audit rule with "-F success" does not exist and no separate rules containing "-F exit=-EPERM" and "-F exit=-EACCES" for "creat" exist, then this is a finding.
Edit the audit.rules file and add the following line(s) to enable auditing of failed attempts to access files and programs: either: -a exit,always -F arch=<ARCH> -S creat -F success=0 or both: -a exit,always -F arch=<ARCH> -S creat -F exit=-EPERM -a exit,always -F arch=<ARCH> -S creat -F exit=-EACCES Restart the auditd service. # service auditd restart
Check the system audit configuration to determine if file and directory deletions are audited. # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "unlink" If no results are returned, or the results do not contain "-S unlink", this is a finding.
Edit the audit.rules file and add the following line to enable auditing of deletions: -a exit,always -F arch=<ARCH> -S unlink Restart the auditd service. # service auditd restart
The message types that are always recorded to /var/log/audit/audit.log include LOGIN,USER_LOGIN,USER_START,USER_END among others and do not need to be added to audit_rules. The log files /var/log/faillog and /var/log/lastlog must be protected from tampering of the login records. Procedure: #egrep "faillog|lastlog" /etc/audit/audit.rules|grep "-p (wa|aw)" If both /var/log/faillog and /var/log/lastlog entries do not exist, this is a finding.
Ensure logins Procedure: Modify /etc/audit/audit.rules to contain: -w /var/log/faillog -p wa -w /var/log/lastlog -p wa
Check the system's audit configuration. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i " chmod " If "-S chmod" is not in the result, this is a finding
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Edit the audit.rules file and add the following lines to enable auditing of discretionary access control permissions modifications. -a exit,always -F arch=<ARCH> -S chmod Restart the auditd service. # service auditd restart
Check the owner of the xinetd configuration files. Procedure: # ls -lL /etc/xinetd.conf # ls -laL /etc/xinetd.d This is a finding if any of the above files or directories are not owned by root or bin.
Change the owner of the xinetd configuration files. # chown root /etc/xinetd.conf /etc/xinetd.d/*
Check the mode of the xinetd configuration files. Procedure: # ls -lL /etc/xinetd.conf # ls -lL /etc/xinetd.d If the mode of the file(s) is more permissive than 0640, this is a finding.
Change the mode of the xinetd configuration files. # chmod 0640 /etc/xinetd.conf /etc/xinetd.d/*
Check the ownership of the services file. Procedure: # ls -lL /etc/services If the services file is not owned by root or bin, this is a finding.
Change the ownership of the services file to root or bin. Procedure: # chown root /etc/services
Check the mode of the services file. Procedure: # ls -lL /etc/services If the services file has a mode more permissive than 0644, this is a finding
Change the mode of the services file to 0644 or less permissive. Procedure: # chmod 0644 /etc/services
Check global initialization files for the presence of "mesg -n" or "mesg n". Procedure: # grep "mesg" etc/bashrc /etc/csh.cshrc /etc/csh.login /etc/csh.logout /etc/environment /etc/ksh.kshrc /etc/profile /etc/suid_profile /etc/profile.d/* If no global initialization files contain "mesg -n" or "mesg n", this is a finding.
Edit /etc/profile or another global initialization script, and add the "mesg -n" command.
RHEL uses "cups" print service. Verify remote host access is limited. Procedure: # grep -i Listen /etc/cups/cupsd.conf The /etc/cups/cupsd.conf file must not contain a Listen *:<port> or equivalent line. If the network address of the "Listen" line is unrestricted, this is a finding. # grep -i "Allow From" /etc/cups/cupsd.conf The "Allow From" line within the "<Location />" element should limit access to the printers to @LOCAL and specific hosts. If the "Allow From" line contains "All" this is a finding
Configure cups to use only the localhost or specified remote hosts. Procedure: Modify the /etc/cups/cupsd.conf file to "Listen" only to the local machine or a known set of hosts (i.e., Listen localhost:631). Modify the /etc/cups/cupsd.conf file "<Location />" element to "Deny From All" and "Allow from 127.0.0.1" or allowed host addresses. Restart cups: # service cups restart
Check the ownership of the print service configuration file. Procedure: # ls -lL /etc/cups/printers.conf; If no print service configuration file is found, this is not applicable. If the owner of the file is not root, this is a finding
Change the owner of the /etc/cups/printers.conf to root. Procedure: # chown root /etc/cups/printers.conf
Check the mode of the print service configuration file. Procedure: # ls -lL /etc/cups/printers.conf If no print service configuration file is found, this is not applicable. If the mode of the print service configuration file is more permissive than 0664, this is a finding.
Change the mode of the /etc/cups/printers.conf file to 0664 or less permissive. Procedure: # chmod 0664 /etc/cups/printers.conf
If the "sendmail" and "postfix" packages are not installed, this is not applicable. Check the ownership of the alias files. Procedure: for sendmail: # ls -lL /etc/aliases # ls -lL /etc/aliases.db If all the files are not owned by root, this is a finding. for postfix: Verify the location of the alias file. # postconf alias maps This will return the location of the "aliases" file, by default "/etc/postfix/aliases" # ls -lL <postfix aliases file> # ls -lL <postfix aliases.db file> If all the files are not owned by root, this is a finding.
Change the owner of the /etc/aliases file to root. Procedure: for sendmail: # chown root /etc/aliases # chown root /etc/aliases.db for postfix # chown root /etc/postfix/aliases # chown root /etc/postfix/aliases.db
If the "sendmail" and "postfix" packages are not installed, this is not applicable. Check the permissions of the alias file. Procedure: for sendmail: # ls -lL /etc/aliases /etc/aliases.db If an alias file has a mode more permissive than 0644, this is a finding. for postfix: Verify the location of the alias file. # postconf alias_maps This will return the location of the "aliases" file, by default "/etc/postfix/aliases". # ls -lL <postfix aliases file> <postfix aliases.db file> If an alias file has a mode more permissive than 0644, this is a finding.
Change the mode of the alias files as needed to function. No higher than 0644. Procedure: for sendmail: # chmod 0644 /etc/aliases /etc/aliases.db for postfix (assuming the default postfix directory): # chmod 0644 /etc/postfix/aliases /etc/postfix/aliases.db
Verify the ownership of files referenced within the sendmail aliases file. Procedure: # more /etc/aliases Examine the aliases file for any utilized directories or paths. # ls -lL <directory or file path> Check the owner for any paths referenced. Check if the file or parent directory is owned by root. If not, this is a finding.
Edit the /etc/aliases file (alternatively, /usr/lib/sendmail.cf). Locate the entries executing a program. They will appear similar to the following line: Aliasname: : /usr/local/bin/ls (or some other program name) Ensure root owns the programs and the directory(ies) they reside in by using the chown command to change owner to root. Procedure: # chown root <file or directory name>
If the "sendmail" package is not installed, this is not applicable. Examine the contents of the /etc/aliases file. Procedure: # more /etc/aliases Examine the aliases file for any referenced programs, which are specified with the pipe (|) symbol. # ls -lL <file referenced from aliases> Check the permissions for any paths referenced. If any file referenced from the aliases file has a mode more permissive than 0755, this is a finding.
Use the chmod command to change the access permissions for files executed from the alias file. For example: # chmod 0755 filename
If the "sendmail" package is not installed, this is not applicable. Check if sendmail logging is set to level nine: Procedure: for sendmail: # grep "O L" /etc/mail/sendmail.cf OR # grep LogLevel /etc/mail/sendmail.cf If logging is set to less than nine, this is a finding. for Postfix: This rule is not applicable to postfix which does not use "log levels" in the same fashion as sendmail.
Edit the sendmail.cf file, locate the "O L" or "LogLevel" entry and change it to 9.
Check the syslog configuration file for mail.crit logging configuration. Depending on what system is used for log processing either /etc/syslog.conf or /etc/rsyslog.conf will be the logging configuration file. Procedure: # grep "mail\." /etc/syslog.conf Or: #grep "mail\." /etc/syslog.conf If syslog is not configured to log critical sendmail messages ("mail.crit" or "mail.*"), this is a finding.
Edit the syslog.conf or rsyslog.conf file and add a configuration line specifying an appropriate destination for "mail.crit" or "mail.*" syslog messages.
Locate any mail log files by checking the syslog configuration file. Procedure: The check procedure is the same for both sendmail and Postfix. Identify any log files configured for the "mail" service (excluding mail.none) at any severity level and check the ownership. Depending on what system is used for log processing, either /etc/syslog.conf or /etc/rsyslog.conf will be the logging configuration file. For syslog: # egrep "mail\.[^n][^/]*" /etc/syslog.conf|sed 's/^[^/]*//'|xargs ls -lL For rsyslog: # egrep "mail\.[^n][^/]*" /etc/rsyslog.conf|sed 's/^[^/]*//'|xargs ls -lL If any mail log file is not owned by root, this is a finding.
Change the ownership of the sendmail log file. Procedure: The fix procedure is the same for both sendmail and Postfix. # chown root <sendmail log file>
Check the mode of the SMTP service log file. Procedure: The check procedure is the same for both sendmail and Postfix. Identify any log files configured for the "mail" service (excluding mail.none) at any severity level and check the permissions. Depending on what system is used for log processing, either /etc/syslog.conf or /etc/rsyslog.conf will be the logging configuration file. For syslog: # egrep "mail\.[^n][^/]*" /etc/syslog.conf|sed 's/^[^/]*//'|xargs ls -lL For rsyslog: # egrep "mail\.[^n][^/]*" /etc/rsyslog.conf|sed 's/^[^/]*//'|xargs ls -lL If the log file permissions are greater than 0644, this is a finding.
Change the mode of the SMTP service log file. Procedure: The fix procedure is the same for both sendmail and Postfix. # chmod 0644 <sendmail log file>
Check for the existence of the ftpusers file. Procedure: For gssftp: # ls -l /etc/ftpusers For vsftp: # ls -l /etc/vsftpd.ftpusers or # ls -l /etc/vsftpd/ftpusers If the appropriate ftpusers file for the running FTP service does not exist, this is a finding.
Create an ftpusers file appropriate for the running FTP service. For gssftp: Create an /etc/ftpusers file containing a list of accounts not authorized for FTP. For vsftp: Create an /etc/vsftpd.ftpusers or /etc/vsftpd/ftpusers (as appropriate) file containing a list of accounts not authorized for FTP.
Check the contents of the ftpusers file. For gssftp: # more /etc/ftpusers For vsftp: # more /etc/vsftpd.ftpusers /etc/vfsftpd/ftpusers If the system has accounts not allowed to use FTP and not listed in the ftpusers file, this is a finding.
For gssftp: Add accounts not allowed to use FTP to the /etc/ftpusers file. For vsftp: Add accounts not allowed to use FTP to the /etc/vsftpd.ftpusers or /etc/vsftpd/ftpusers file (as appropriate).
Check the ownership of the ftpusers file. Procedure: For gssftp: # ls -l /etc/ftpusers For vsftp: # ls -l /etc/vsftpd.ftpusers /etc/vsftpd/ftpusers If the ftpusers file is not owned by root, this is a finding
Change the owner of the ftpusers file to root. For gssftp: # chown root /etc/ftpusers For vsftp: # chown root /etc/vsftpd.ftpusers /etc/vsftpd/ftpusers
Check the permissions of the ftpusers file. Procedure: For gssftp: # ls -l /etc/ftpusers For vsftp: # ls -l /etc/vsftpd.ftpusers /etc/vsftpd/ftpusers If the ftpusers file has a mode more permissive than 0640, this is a finding.
Change the mode of the ftpusers file to 0640. Procedure: For gssftp: # chmod 0640 /etc/ftpusers For vsftp: # chmod 0640 /etc/vsftpd.ftpusers /etc/vsftpd/ftpusers
Find if logging is applied to the ftp daemon. The procedure depends on the implementation of ftpd used by the system. Procedures: For vsftpd: If vsftpd is started by xinetd: #grep vsftpd /etc/xinetd.d/* This will indicate the xinetd.d startup file #grep server_args <vsftpd xinetd.d startup file> This will indicate the vsftpd config file used when starting through xinetd. If the line is missing then "/etc/vsftpd/vsftpd.conf", the default config file, is used. #grep xferlog_enable <vsftpd config file> If "xferlog_enable" is missing or is not set to "yes", this is a finding. If vsftp is not started by xinetd: #grep xferlog_enable /etc/vsftpd/vsftpd.conf If "xferlog_enable" is missing or is not set to "yes", this is a finding. For gssftp: Find if the -l option will be applied when xinetd starts gssftp # grep server_args /etc/xinetd.d/gssftp If the line is missing or does not contain at least one -l, this is a finding.
Enable logging by changing ftpd startup or config files. Procedure: The procedure depends on the implementation of ftpd used by the system. For vsftpd: Ensure the server settings in "/etc/vsftpd.conf" (or other configuration file specified by the vaftpd xinetd.d startup file) contains: xferlog_enable = yes For gssftp: If the "disable" server setting is missing or set to "no" in "/etc/xinetd.d/gssftp" then ensure the server settings in "/etc/xinetd.d/gssftp" contains: server_args = -l The -l option may be added up to three times. Each -l will provide increasing verbosity on the log. Refer to the main page for ftpd for more information. For both if started using xinetd: If the "disable" server setting is missing or set to "no" in the /etc/xinetd.d startup file then ensure the server settings contains: log_on_success += DURATION USERID This will log the startup and shutdown of the daemon. log_on_failure += HOST USERID
Attempt to log into this host with a user name of anonymous and a password of guest (also try the password of guest@mail.com). If the logon is successful and the use of anonymous ftp has not been documented and approved by the IAO, this is a finding. Procedure: # ftp localhost Name: anonymous 530 Guest login not allowed on this machine.
Configure the FTP service to not permit anonymous logins.
# grep server_args /etc/xinetd.d/tftp If the "-s" parameter is not specified, this is a finding.
Edit /etc/xinetd.d/tftp file and specify the "-s" parameter in server_args.
Check the mode of the TFTP daemon. Procedure: # grep "server " /etc/xinetd.d/tftp # ls -lL <in.tftpd binary> If the mode of the file is more permissive than 0755, this is a finding.
Change the mode of the TFTP daemon. Procedure: # chmod 0755 <in.tftpd binary>
Check the /etc/passwd file to determine if TFTP is configured properly. Procedure: Check if TFTP if used. # grep disable /etc/xinetd.d/tftp If the file does not exist or the returned line indicates "yes", then this is not a finding. Otherwise, if the returned line indicates "no" then TFTP is enabled and must use a dedicated "tftp" user. # grep user /etc/xinetd.d/tftp If the returned line indicates a user other than the dedicated "tftp" user, this is a finding. # grep tftp /etc/passwd If a "tftp" user account does not exist and TFTP is active, this is a finding. Check the user shell for the "tftp" user. If it is not /bin/false or equivalent, this is a finding. Check the home directory assigned to the "tftp" user. If no home directory is set, or the directory specified is not dedicated to the use of the TFTP service, this is a finding.
Configure TFTP to use a dedicated "tftp" user. Procedure: Create a dedicated "tftp" user account if none exists. Assign a non-login shell to the "tftp" user account, such as /bin/false. Assign a home directory to the "tftp" user account. Edit /etc/xinetd.d/tftp to have "tftp" as the value of the "user" parameter.
Check for .Xauthority or .xauth files being utilized by looking for such files in the home directory of a user. Procedure: # find / -name '.xauth*' | more If no .xauth files are found in a user's home directory, ensure that Xwindows is not active on the system by performing the command: # ps -ef | grep X If Xwindows is not running, this rule is not applicable. If the .Xauthority or .xauth (followed by apparently random characters) files do not exist, ask the SA if the user is using Xwindows. If the user is utilizing Xwindows and none of these files exist, this is a finding.
Ensure the X Windows host is configured to write .Xauthority files into user home directories. Edit the Xaccess file. Ensure the line writing the .Xauthority file is uncommented.
Perform the following to determine if NIS is active on the system: # ps -ef | grep ypbind If NIS is found active on the system, this is a finding.
Disable the use of NIS/NIS+. Use as a replacement Kerberos or LDAP.
Use pwck to verify home directory assignments are present. # pwck If any user is not assigned a home directory, this is a finding.
Assign a home directory to any user without one.
Use pwck to verify assigned home directories exist. # pwck If any user's assigned home directory does not exist, this is a finding.
If a user has no home directory, determine why. If possible, delete accounts without a home directory. If the account is valid, then create the home directory using the appropriate system administration utility or manually. For instance: mkdir directoryname; copy the skeleton files into the directory; chown accountname for the new directory and the skeleton files. Document all changes.
Check the home directory mode of each user in /etc/passwd. Procedure: # cut -d: -f6 /etc/passwd|sort|uniq|xargs -n1 ls -ld If a user home directory's mode is more permissive than 0750, this is a finding. Note: Application directories are allowed and may need 0755 permissions (or greater) for correct operation.
Change the mode of user home directories to 0750 or less permissive. Procedure (example): # chmod 0750 <home directory> Note: Application directories are allowed and may need 0755 permissions (or greater) for correct operation.
Check the ownership of each user home directory listed in the /etc/passwd file. Procedure: # cut -d : -f 6 /etc/passwd | xargs -n1 ls -ld If any user home directory is not owned by the assigned user, this is a finding.
Change the owner of a user's home directory to its assigned user. Procedure: # chown <user> <home directory>
Check the group ownership for each user in the /etc/passwd file. Procedure: # cut -d : -f 6 /etc/passwd | xargs -n1 ls -ld If any user home directory is not group-owned by the assigned user's primary group, this is a finding. Home directories for application accounts requiring different group ownership must be documented using site-defined procedures.
Change the group-owner for user home directories to the primary group of the assigned user. Procedure: Find the primary group of the user (GID) which is the fourth field of the user entry in /etc/passwd. # chgrp <GID> <user home directory> Document all changes.
NOTE: The following commands must be run in the BASH shell. Check the ownership of local initialization files. Procedure: # ls -al /<usershomedirectory>/.login # ls -al /<usershomedirectory>/.cshrc # ls -al /<usershomedirectory>/.logout # ls -al /<usershomedirectory>/.profile # ls -al /<usershomedirectory>/.bash_profile # ls -al /<usershomedirectory>/.bashrc # ls -al /<usershomedirectory>/.bash_logout # ls -al /<usershomedirectory>/.env # ls -al /<usershomedirectory>/.dtprofile # ls -al /<usershomedirectory>/.dispatch # ls -al /<usershomedirectory>/.emacs # ls -al /<usershomedirectory>/.exrc # find /<usershomedirectory>/.dt ! -fstype nfs ! -user <username> -exec ls -ld {} \; If local initialization files are not owned by the home directory's user, this is a finding.
Change the ownership of the startup and login files in the user's directory to the user or root, as appropriate. Examine each user's home directory and verify all filenames beginning with "." are owned by the owner of the directory or root. If they are not, use the chown command to change the owner to the user and research the reasons why the owners were not assigned as required. Procedure: # chown username .filename Document all changes.
Check the modes of local initialization files. Procedure: # ls -al /<usershomedirectory>/.bashrc # ls -al /<usershomedirectory>/.bash_login # ls -al /<usershomedirectory>/.bash_logout # ls -al /<usershomedirectory>/.bash_profile # ls -al /<usershomedirectory>/.cshrc # ls -al /<usershomedirectory>/.kshrc # ls -al /<usershomedirectory>/.login # ls -al /<usershomedirectory>/.logout # ls -al /<usershomedirectory>/.profile # ls -al /<usershomedirectory>/.tcshrc # ls -al /<usershomedirectory>/.env # ls -al /<usershomedirectory>/.dtprofile (permissions should be 0755) # ls -al /<usershomedirectory>/.dispatch # ls -al /<usershomedirectory>/.emacs # ls -al /<usershomedirectory>/.exrc # find /<usershomedirectory>/.dt ! -fstype nfs \( -perm -0002 -o -perm -0020 \) -exec ls -ld {} \; (permissions not to be more permissive than 0755) If local initialization files are more permissive than 0740 or the .dt directory is more permissive than 0755 or the .dtprofile file is more permissive than 0755, this is a finding.
Ensure user startup files have permissions of 0740 or more restrictive. Examine each user's home directory and verify all file names beginning with "." have access permissions of 0740 or more restrictive. If they do not, use the chmod command to correct the vulnerability. Procedure: # chmod 0740 .filename Note: The period is part of the file name and is required.
Check run control script modes. # cd /etc # ls -lL rc* # cd /etc/init.d # ls -l If any run control script has a mode more permissive than 0755, this is a finding.
Ensure all system startup files have mode 0755 or less permissive. Examine the "rc" files, and all files in the rc1.d (rc2.d, and so on) directories, and in the /etc/init.d directory to ensure they are not world-writable. If they are world-writable, use the chmod command to correct the vulnerability and research why they are world-writable. Procedure: # chmod 755 <startup file>
Verify run control scripts' library search paths. # grep -r '\bPATH\b' /etc/rc* /etc/init.d This variable is formatted as a colon-separated list of directories. Relative path entries must be documented with the ISSO. If there is an empty entry, such as a leading or trailing colon, or two consecutive colons, this is a finding. If an entry begins with a character other than a slash (/), or has not been documented with the ISSO, this is a finding.
Edit the run control script and remove any relative path entries from the executable search path variable that are not documented with the ISSO. Edit the run control script and remove any empty entry that is defined.
Check the permissions on the files or scripts executed from system startup scripts to see if they are world-writable. Create a list of all potential run command level scripts. # ls -l /etc/init.d/* | tr '\011' ' ' | tr -s ' ' | cut -f 9,9 -d " " OR # ls -l /sbin/init.d/* | tr '\011' ' ' | tr -s ' ' | cut -f 9,9 -d " " Create a list of world writeable files. # find / -perm -002 -type f >> worldWriteableFileList Determine if any of the world writeable files in worldWriteableFileList are called from the run command level scripts. Note: Depending upon the number of scripts vs world writeable files, it may be easier to inspect the scripts manually. # more `ls -l /etc/init.d/* | tr '\011' ' ' | tr -s ' ' | cut -f 9,9 -d " "` OR # more `ls -l /sbin/init.d/* | tr '\011' ' ' | tr -s ' ' | cut -f 9,9 -d " "` If any system startup script executes any file or script that is world-writable, this is a finding.
Remove the world-writable permission from programs or scripts executed by run control scripts. Procedure: # chmod o-w <program or script executed from run control script>
Check the system for the existence of any .netrc files. Procedure: # find / -name .netrc If any .netrc file exists, this is a finding.
Remove the .netrc file(s). Procedure: # find / -name .netrc # rm <.netrc file>
For each user in the /etc/passwd file, check for the presence of files and directories within the user's home directory not owned by the home directory owner. Procedure: # find /<usershomedirectory> ! -fstype nfs ! -user <username> ! \( -name .bashrc -o -name .bash_login -o -name .bash_logout -o -name .bash_profile -o -name .cshrc -o -name .kshrc -o -name .login -o -name .logout -o -name .profile -o -name .tcshrc -o -name .env -o -name .dtprofile -o -name .dispatch -o -name .emacs -o -name .exrc \) -exec ls -ld {} \; If user home directories contain files or directories not owned by the home directory owner, this is a finding.
Change the ownership of files and directories in user home directories to the owner of the home directory. Procedure: # chown accountowner filename
For each user in the /etc/passwd file, check for files and directories with a mode more permissive than 0750. Procedure: # find /<usershomedirectory> ! -fstype nfs ! \( -name .bashrc -o -name .bash_login -o -name .bash_logout -o -name .bash_profile -o -name .cshrc -o -name .kshrc -o -name .login -o -name .logout -o -name .profile -o -name .tcshrc -o -name .env -o -name .dtprofile -o -name .dispatch -o -name .emacs -o -name .exrc \) \( -perm -0001 -o -perm -0002 -o -perm -0004 -o -perm -0020 -o -perm -2000 -o -perm -4000 \) -exec ls -ld {} \; If user home directories contain files or directories more permissive than 0750, this is a finding.
Change the mode of files and directories within user home directories to 0750. Procedure: # chmod 0750 filename Document all changes.
Verify /etc/shells exists. # ls -l /etc/shells If the file does not exist, this is a finding.
Create a /etc/shells file containing a list of valid system shells. Consult vendor documentation for an appropriate list of system shells. Procedure: # echo "/bin/bash" >> /etc/shells # echo "/bin/csh" >> /etc/shells (Repeat as necessary for other shells.)
Confirm the login shells referenced in the /etc/passwd file are listed in the /etc/shells file. Procedure: Determine which shells are permitted for use by users: # more /etc/shells Note: /usr/bin/false, /bin/false, /dev/null, /sbin/nologin, /bin/sync, /sbin/halt, /sbin/shutdown, (and equivalents) cannot be placed in the /etc/shells file. Determine which shells are being used: # more /etc/passwd (optionally shells found in /etc/passwd can be grepped for in /etc/shells) If any shells are found that are not in /etc/shells, or if false shells are found in /etc/shells, then this is a finding.
Use the "chsh" utility or edit the /etc/passwd file and correct the error by changing the default shell of the account in error to an acceptable shell name contained in the /etc/shells file. Example: # chsh -s /bin/bash testuser
Indications of inactive accounts are those that have no entries in the "last" log. Check the date in the "last" log to verify it is within the last 35 days or the maximum numbers of days set by the site if more restrictive. If an inactive account is not disabled via an entry in the password field in the /etc/passwd or /etc/shadow (or equivalent), check the /etc/passwd file to check if the account has a valid shell. The passwd command can also be used to list a status for an account. For example, the following may be used to provide status information on each local account: NOTE: The following must be done in the BASH shell. # cut -d: -f1 /etc/passwd | xargs -n1 passwd -S If an inactive account is found not disabled, this is a finding.
All inactive accounts will have /sbin/nologin (or an equivalent), as the default shell in the /etc/passwd file and have the password disabled. Examine the user accounts using the "last" command. Note the date of last login for each account. If any (other than system and application accounts) exceed 35 days or the maximum number of days set by the site, not to exceed 35 days, then disable the accounts using system-config-users tool. Alternately place a shell field of /sbin/nologin /bin/false or /dev/null in the passwd file entry for the account.
Check the ownership of the system shells. # cat /etc/shells | xargs -n1 ls -l If any shell is not owned by root or bin, this is a finding.
Change the ownership of the shell with incorrect ownership. # chown root <shell>
If /etc/shells exists, check the group ownership of each shell referenced. # cat /etc/shells | xargs -n1 ls -l Otherwise, check any shells found on the system. # find / -name "*sh" | xargs -n1 ls -l If a shell has a mode more permissive than 0755, this is a finding.
Change the mode of the shell. # chmod 0755 <shell>
Determine if there are any device files outside of /dev: # find / -type b -o -type c |more Check for the presence of an aide on the system: # rpm -qa | grep aide If aide is not installed, ask the SA what file integrity tool is being used to check the system. Check the global crontabs for the presence of an "aide" job to run at least weekly, if aide is installed. Otherwise, check for the presence of a cron job to run the alternate file integrity checking application. # grep aide /etc/cron*/* If a tool is being run, then the configuration file for the appropriate tool needs to be checked for selection lines for /dev and any other directories/subdirectories that contain device files. Review the process to determine if the system is checked for extraneous device files on a weekly basis. If no weekly automated or manual process is in place, this is a finding. If the process is not identifying extraneous device files, this is a finding.
Establish a weekly automated or manual process to create a list of device files on the system and determine if any files have been added, moved, or deleted since the last list was generated. A list of device files can be generated with this command: # find / -type b -o -type c > device-file-list
Find all world-writable device files existing anywhere on the system. Procedure: # find / -perm -2 -a \( -type b -o -type c \) > devicelist Check the permissions on the directories above subdirectories containing device files. If any of the device files or their parent directories are world-writable, excepting device files specifically intended to be world-writable such as /dev/null, this is a finding. These world-writable files on installation are intended to be world-writable: /dev/full /dev/null /selinux/null /dev/ptmx /dev/random /dev/tty /dev/vsock /dev/zero /dev/log
Remove the world-writable permission from the device file(s). Procedure: # chmod o-w <device file> Document all changes.
Check the system for world-writable device files. Procedure: # find / -perm -2 -a \( -type b -o -type c \) -exec ls -ld {} \; Ask the SA to identify any device files used for backup purposes. If any device file(s) used for backup are writable by users other than root or the designated backup user, this is a finding.
Use the chmod command to remove the world-writable bit from the backup device files. Procedure: # chmod o-w <back device filename> Document all changes.
Check the owner of the exports file. Example: # ls -lL /etc/exports If the export configuration file is not owned by root, this is a finding.
Change the owner of the exports file to root. Example: # chown root /etc/exports
# ls -lL /etc/exports If the file has a mode more permissive than 0644, this is a finding.
# chmod 0644 /etc/exports
Check for NFS exported file systems. Procedure: # cat /etc/exports For each file system displayed, check the ownership. # ls -lLa <exported file system path> If the files and directories are not owned by root, this is a finding.
Change the ownership of exported file systems not owned by root. Procedure: # chown root <path>
Check if the 'anonuid' and 'anongid' options are set correctly for exported file systems. List exported filesystems: # exportfs -v Each of the exported file systems should include an entry for the 'anonuid=' and 'anongid=' options set to "-1" or an equivalent (60001, 65534, or 65535). If appropriate values for 'anonuid' or 'anongid' are not set, this is a finding.
Edit "/etc/exports" and set the "anonuid=-1" and "anongid=-1" options for exports lacking it. Re-export the filesystems.
Check the permissions on exported NFS file systems. Procedure: # exportfs -v If the exported file systems do not contain the 'rw' or 'ro' options specifying a list of hosts or networks, this is a finding.
Edit /etc/exports and add ro and/or rw options (as appropriate) specifying a list of hosts or networks which are permitted access. Re-export the file systems.
List the exports. # cat /etc/exports If any export contains "no_root_squash" or does not contain "root_squash" or "all_squash", this is a finding.
Edit the "/etc/exports" file and add "root_squash" (or "all_squash") and remove "no_root_squash".
Check the system for NFS mounts not using the "nosuid" option. Procedure: # mount -v | grep " type nfs " | egrep -v "nosuid" If the mounted file systems do not have the "nosuid" option, this is a finding.
Edit "/etc/fstab" and add the "nosuid" option for all NFS file systems. Remount the NFS file systems to make the change take effect.
The tcp_wrappers package is provided with the RHEL distribution. Other access control programs may be available but will need to be checked manually. Determine if tcp_wrappers is installed. # rpm -qa | grep tcp_wrappers If no package is listed, this is a finding.
Install and configure the tcp_wrappers package.
The tcp_wrappers package is provided with the RHEL distribution. Other access control programs may be available but will need to be checked manually. Depending on what system is used for log processing either /etc/syslog.conf or /etc/rsyslog.conf will be the logging configuration file. Normally, tcpd logs to the mail facility in "/etc/syslog.conf" or “/etc/rsyslog.conf”. Determine if syslog or rsyslog is configured to log events by tcpd. Procedure: # grep -E “(\*.info|\*.debug|authpriv.info|authpriv.debug|authpriv.\*)” /etc/syslog.conf | grep –v ‘#’ Or: # grep -E “(\*.info|\*.debug|authpriv.info|authpriv.debug|authpriv.\*)” /etc/rsyslog.conf | grep –v ‘#’ If no entries exist, this is a finding. If there are no “authpriv.info”, “authpriv.debug”, “authpriv.*” or “*.info” or “*.debug” not followed by “authpriv.none”, this is a finding. If an alternate access control program is used and it does not provide logging of access attempts, this is a finding.
Configure the access restriction program to log every access attempt. Ensure the implementation instructions for tcp_wrappers are followed so system access attempts are recorded to the system log files. If an alternate application is used, it must support this function.
This check is not applicable if only the root user is permitted to use cron. Check for the existence of the cron.allow and cron.deny files. # ls -lL /etc/cron.allow # ls -lL /etc/cron.deny If neither file exists, this is a finding.
Create /etc/cron.allow and/or /etc/cron.deny with appropriate content and reboot the system to ensure no lingering cron jobs are processed.
Check mode of the cron.allow file. Procedure: # ls -lL /etc/cron.allow If the file has a mode more permissive than 0600, this is a finding.
Change the mode of the cron.allow file to 0600. Procedure: # chmod 0600 /etc/cron.allow
List all cronjobs on the system. Procedure: # ls /var/spool/cron # ls /etc/cron.d /etc/crontab /etc/cron.daily /etc/cron.hourly /etc/cron.monthly /etc/cron.weekly or # ls /etc/cron*|grep -v deny If cron jobs exist under any of the above directories, use the following command to search for programs executed by cron: # more <cron job file> Perform a long listing of each program file found in the cron file to determine if the file is group-writable or world-writable. # ls -la <cron program file> If cron executes group-writable or world-writable files, this is a finding.
Remove the world-writable and group-writable permissions from the cron program file(s) identified. # chmod go-w <cron program file>
List all cronjobs on the system. Procedure: # ls /var/spool/cron # ls /etc/cron.d /etc/crontab /etc/cron.daily /etc/cron.hourly /etc/cron.monthly /etc/cron.weekly or # ls /etc/cron*|grep -v deny If cron jobs exist under any of the above directories, use the following command to search for programs executed by at: # more <cron job file> Perform a long listing of each directory containing program files found in the cron file to determine if the directory is world-writable. # ls -ld <cron program directory> If cron executes programs in world-writable directories, this is a finding.
Remove the world-writable permission from the cron program directories identified. Procedure: # chmod o-w <cron program directory>
Check the mode of the crontab files. # ls -lL /var/spool/cron/ # ls -lL /etc/cron.d/ # ls -lL /etc/crontab If any crontab file has a mode more permissive than 0600, this is a finding.
Change the mode of the crontab files. # chmod 0600 /var/spool/cron/* /etc/cron.d/* /etc/crontab
Check the mode of the crontab directories. Procedure: # ls -ld /var/spool/cron # ls -ld /etc/cron.d /etc/crontab /etc/cron.daily /etc/cron.hourly /etc/cron.monthly /etc/cron.weekly or # ls -ld /etc/cron*|grep -v deny If the mode of any of the crontab directories is more permissive than 0755, this is a finding.
Change the mode of the crontab directories. # chmod 0755 <crontab directory>
Check the owner of the crontab directories. Procedure: # ls -ld /var/spool/cron # ls -ld /etc/cron.d /etc/crontab /etc/cron.daily /etc/cron.hourly /etc/cron.monthly /etc/cron.weekly or # ls -ld /etc/cron*|grep -v deny If the owner of any of the crontab directories is not root or bin, this is a finding.
Change the mode of the crontab directories. # chown root <crontab directory>
Check the group owner of cron and crontab directories. Procedure: # ls -ld /var/spool/cron # ls -ld /etc/cron.d /etc/crontab /etc/cron.daily /etc/cron.hourly /etc/cron.monthly /etc/cron.weekly or # ls -ld /etc/cron*|grep -v deny If a directory is not group-owned by root, sys, bin, or cron, this is a finding.
Change the group owner of cron and crontab directories. # chgrp root <crontab directory>
Depending on what system is used for log processing either /etc/syslog.conf or /etc/rsyslog.conf will be the logging configuration file. # grep cron /etc/syslog.conf Or: # grep cron /etc/rsyslog.conf If cron logging is not configured, this is a finding. Check the configured cron log file found in the cron entry of /etc/syslog.conf or /etc/rsyslog.conf (normally /var/log/cron). # ls -lL /var/log/cron If this file does not exist, or is older than the last cron job, this is a finding.
Edit /etc/syslog.conf or /etc/rsyslog.conf and setup cron logging
Check the mode of the cron log file. Procedure: Depending on what system is used for log processing either /etc/syslog.conf or /etc/rsyslog.conf will be the logging configuration file. Check the configured cron log file found in the cron entry in /etc/syslog.conf or /etc/rsyslog.conf (normally /var/log/cron). # grep cron /etc/syslog.conf Or: # grep cron /etc/rsyslog.conf # ls -lL /var/log/cron If the mode is more permissive than 0600, this is a finding.
Change the mode of the cron log file. # chmod 0600 /var/log/cron
If the "at" package is not installed, this is not applicable. Check for the existence of at.allow and at.deny files. # ls -lL /etc/at.allow # ls -lL /etc/at.deny If neither file exists, this is a finding.
Create at.allow and/or at.deny files containing appropriate lists of users to be allowed or denied access to the "at" daemon.
# more /etc/at.deny If the at.deny file exists and is empty, this is a finding.
Add appropriate users to the at.deny file, or remove the empty at.deny file if an at.allow file exists.
# more /etc/at.allow If default accounts (such as bin, sys, adm, and others) are listed in the at.allow file, this is a finding.
Remove the default accounts (such as bin, sys, adm, and others, traditionally UID less than 500) from the at.allow file.
Check the mode of the at.allow file. # ls -lL /etc/at.allow If the at.allow file has a mode more permissive than 0600, this is a finding.
Change the mode of the at.allow file. # chmod 0600 /etc/at.allow
List the "at" jobs on the system. Procedure: # ls -la /var/spool/at For each "at" job file, determine which programs are executed. Procedure: # more <at job file> Check the each program executed by "at" for group- or world-writable permissions. Procedure: # ls -la <at program file> If "at" executes group or world-writable programs, this is a finding.
Remove group-write and world-write permissions from files executed by at jobs. Procedure: # chmod go-w <file>
List any "at" jobs on the system. Procedure: # ls /var/spool/at For each "at" job, determine which programs are executed by "at." Procedure: # more <at job file> Check the directory containing each program executed by "at" for world-writable permissions. Procedure: # ls -la <at program file directory> If "at" executes programs in world-writable directories, this is a finding.
Remove the world-writable permission from directories containing programs executed by "at". Procedure: # chmod o-w <at program directory>
Check the SNMP configuration for default passwords. Procedure: Examine the default install location /etc/snmp/snmpd.conf or: # find / -name snmpd.conf # more <snmpd.conf file> Identify any community names or user password configuration. If any community name or password is set to a default value such as "public", "private", "snmp-trap", or "password", or any value which does not meet DISA password requirements, this is a finding.
Change the default passwords. To change them, locate the file snmpd.conf. Edit the file. Locate the line system-group-read-community which has a default password of "public" and make the password something more secure and less guessable. Do the same for the lines reading system-group-write-community, read-community, write-community, trap and trap-community. Read the information in the file carefully. The trap is defining who to send traps to, for instance, by default. It is not a password, but the name of a host.
Check the mode of the SNMP daemon configuration file. Procedure: Examine the default install location /etc/snmp/snmpd.conf or: # find / -name snmpd.conf # ls -lL <snmpd.conf file> If the snmpd.conf file has a mode more permissive than 0600, this is a finding.
Change the mode of the SNMP daemon configuration file to 0600. Procedure: # chmod 0600 <snmpd.conf>
Check the modes for all Management Information Base (MIB) files on the system. Procedure: # find / -name *.mib # ls -lL <mib file> Any file returned with a mode 0640 or less permissive is a finding.
Change the mode of MIB files to 0640. Procedure: # chmod 0640 <mib file>
Check the system for world-writable files. Procedure: # find / -perm -2 -a \( -type d -o -type f \) -exec ls -ld {} \; If any world-writable files are located, except those required for system operation such as /tmp and /dev/null, this is a finding.
Remove or change the mode for any world-writable file on the system not required to be world-writable. Procedure: # chmod o-w <file> Document all changes
The /etc/xinetd.conf file and each file in the /etc/xinetd.d directory file should be examined for the following: Procedure: log_type = SYSLOG authpriv log_on_success = HOST PID USERID EXIT log_on_failure = HOST USERID If xinetd is running and logging is not enabled, this is a finding.
Edit each file in the /etc/xinetd.d directory and the /etc/xinetd.conf file to contain: log_type = SYSLOG authpriv log_on_success = HOST PID USERID EXIT log_on_failure = HOST USERID The /etc/xinetd.conf file contains default values that will hold true for all services unless individually modified in the service's xinetd.d file. To make the new settings effective, restart the xinetd service: # service xinetd restart
Determine if the system is configured to boot from devices other than the system startup media. If so, this is a finding.
Configure the system to only boot from system startup media. Procedure: On systems with a BIOS or system controller use the BIOS interface at startup to remove all but the proper boot device from the boot device list.
Verify the options of the running Xwindows server are correct. Procedure: Get the running xserver information # ps -ef |grep X If the response contains /usr/bin/Xorg:0 /usr/bin/Xorg:0 -br -audit 0 -auth /var/gdm/:0.Xauth -nolisten tcp vt7 this is indicative of Xorg starting through gdm. This is the default on RHEL. Examine the Xorg line: If the "-auth" option is missing this would be a finding. If the "-audit" option is missing or not set to 4, this is a finding. If the "-s" option is missing or greater than 15, this is a finding. If the response to the grep contains X:0 /usr/bin/X:0 this indicates the X server was started with the xinit command with no associated .xserverrc in the home directory of the user. No options are selected by default. This is a finding. Otherwise if there are options on the X:0 line: If the "-auth" option is missing this is a finding If the "-audit" option is missing or not set to 4, this is a finding. If the "-s" option is missing or greater than 15, this is a finding.
Enable the following options: -audit (at level 4), -auth and -s with 15 minutes as the timeout value. Procedure for gdm: Edit /etc/gdm/custom.conf and add the following: [server-Standard] name=Standard server command=/usr/bin/Xorg -br -audit 4 -s 15 chooser=false handled=true flexible=true priority=0 Procedure for xinit: Edit or create a .xserverrc file in the users home directory containing the startup script for xinit. This script must have an exec line with at least these options: exec /usr/bin/X -audit 4 -s 15 -auth <Xauth file> & The <Xauth file> is created using the "xauth" command and is customarily located in the users home directory with the name ".Xauthority".
If the "xorg-x11-server-Xorg" package is not installed, this is not applicable. Verify the options of the running Xwindows server are correct. Procedure: Get the running xserver information # ps -ef |grep X If the response contains /usr/bin/Xorg:0 /usr/bin/Xorg:0 -br -audit 0 -auth /var/gdm/:0.Xauth -nolisten tcp vt7 this is indicative of Xorg starting through gdm. This is the default window manager on RHEL. If the "-ac" option is found, this is a finding. If the "-core" option is found, this is a finding. If the "-nolock" option is found, this is a finding. If the response to the grep contains X:0 /usr/bin/X:0 Examine the X:0 line: If the "-ac" option is found, this is a finding. If the "-core" option is found, this is a finding. If the "-nolock" option is found, this is a finding.
Disable the unwanted options: Procedure: For gdm: Remove the -ac, -core and -nolock options by creating a "command" entry in the /etc/gdm/custom.conf file with the options removed. For Xwindows started by xinit: Create or modify the .xserverrc script in the users home directory to remove the -ac, -core and -nolock options from the exec /usr/bin/X command.
# ps -ef | egrep "innd|nntpd" If an Internet Network News server is running, this is a finding.
Disable the INN server.
Check access configuration ownership: # ls -lL /etc/security/access.conf If this file exists and is not owned by root, this is a finding.
Follow the correct configuration parameters for access configuration file. Use the chown command to configure it properly. (for example: # chown root /etc/security/access.conf ).
SWAT is a tool for configuring Samba and should only be found on a system with a requirement for Samba. If SWAT is used, it must be utilized with SSL to ensure a secure connection between the client and the server. Procedure: # grep -H "bin/swat" /etc/xinetd.d/*|cut -d: -f1 |xargs grep "only_from" If the value of the "only_from" line in the "xinetd.d" file which starts "/usr/sbin/swat" is not "localhost" or the equivalent, this is a finding.
Disable SWAT or require SWAT is only accessed via SSH. Procedure: If SWAT is not needed for operation of the system remove the SWAT package: # rpm -qa|grep swat Remove "samba-swat" or "samba3x-swat" depending on which one is installed # rpm --erase samba-swat or # rpm --erase samba3x-swat If SWAT is required but not at all times disable it when it is not needed. Modify the /etc/xinetd.d file for "swat" to contain a "disable = yes" line. To access using SSH: Follow vendor configuration documentation to create an stunnel for SWAT.
Check the ownership of the /etc/samba/smb.conf file. Procedure: # ls -l /etc/samba/smb.conf If an smb.conf file is not owned by root, this is a finding.
Change the ownership of the smb.conf file. Procedure: # chown root smb.conf
Check the mode of the smb.conf file. Procedure: # ls -lL /etc/samba/smb.conf If the "smb.conf" has a mode more permissive than 0644, this is a finding.
Change the mode of the smb.conf file to 0644 or less permissive. Procedure: # chmod 0644 smb.conf.
Check the ownership of the "smbpasswd" file. # ls -l /etc/samba/passdb.tdb /etc/samba/secrets.tdb If the "smbpasswd" file is not owned by root, this is a finding.
Use the chown command to configure the files maintained by smbpasswd. For instance: # chown root /etc/samba/passdb.tdb /etc/samba/secrets.tdb
Examine the "smb.conf" file. # more /etc/samba/smb.conf If the "hosts" option is not present to restrict access to a list of authorized hosts and networks, this is a finding.
Edit the "/etc/samba/smb.conf" file and set the "hosts" option to permit only authorized hosts to access Samba.
Check the minimum time period between password changes for each user account is 1 day. # cat /etc/shadow | cut -d ':' -f 4 | grep -v 1 If any results are returned, this is a finding.
Change the minimum time period between password changes for each user account to 1 day. # passwd -n 1 <user name>
Determine if root has logged in over an unencrypted network connection. Examine /etc/syslog.conf to confirm the location to which "authpriv" messages are being sent. # grep authpriv.* /etc/syslog.conf Once the file is determined, perform the following command: # grep password <file> | more Look for any lines that do not have sshd as the associated service. If root has logged in over the network and sshd is not running, this is a finding.
Enable SSH on the system and use it for all remote connections used to attain root access
Determine if the SSH daemon is configured to permit root logins. Procedure: # grep -v "^#" /etc/ssh/sshd_config | grep -i permitrootlogin If the PermitRootLogin entry is not found or is not set to "no", this is a finding.
Edit the sshd_config file and set the PermitRootLogin option to "no".
Check the mode of audio devices. # ls -lL /dev/audio* /dev/snd/* If the mode of audio devices are more permissive than 660, this is a finding.
Change the mode of audio devices. # chmod 0660 <audio device>
Check the owner of audio devices. # ls -lL /dev/audio* /dev/snd/* If the owner of any audio device file is not root, this is a finding.
Edit the /etc/security/console.perms.d/50-default.perms file and comment the following line: <console> 0600 <sound> 0660 root.audio
Check access configuration group ownership: # ls -lL /etc/security/access.conf If this file exists and has a group-owner that is not a privileged user, this is a finding.
Use the chgrp command to ensure the group owner is root, sys, or bin. (for example: # chgrp root /etc/security/access.conf ).
Check access configuration mode: # ls -lL /etc/security/access.conf If this file exists and has a mode more permissive than 0640, this is a finding.
Use the chmod command to set the permissions to 0640. (for example: # chmod 0640 /etc/security/access.conf ).
Check the group ownership of the "smb.conf" file. Procedure: # ls -lL /etc/samba/smb.conf If the "smb.conf" file is not group-owned by root, bin, sys, or system, this is a finding.
Change the group owner of the smb.conf file. Procedure: # chgrp root smb.conf
Check "smbpasswd" ownership: # ls -lL /etc/samba/passdb.tdb /etc/samba/secrets.tdb If the "smbpasswd" file is not group-owned by root, this is a finding.
Use the chgrp command to ensure that the group owner of the smbpasswd file is root. For instance: # chgrp root /etc/samba/passdb.tdb /etc/samba/secrets.tdb
Check the mode of files maintained using "smbpasswd". Procedure: # ls -lL /etc/samba/passdb.tdb /etc/samba/secrets.tdb If a "smbpasswd" maintained file has a mode more permissive than 0600, this is a finding.
Change the mode of the files maintained through smbpasswd to 0600. Procedure: # chmod 0600 /etc/samba/passdb.tdb /etc/samba/secrets.tdb
Check the group-owner of audio devices. Procedure: # ls -lL /dev/audio* /dev/snd/* If the group-owner of an audio device is not root, sys, bin, system, or audio this is a finding.
Change the group-owner of the audio device. Procedure: # chgrp <root, sys, bin, system, audio> <audio device>
Determine if root's shell executable resides on a dedicated file system. Procedure: Find the location of the root user's shell # grep "^root" /etc/passwd|cut -d: -f7|cut -d/ -f2 The result is the top level directory under / where the shell resides (e.g., usr) Check if it is on a dedicated file system. # grep /<top level directory> /etc/fstab If /<top level directory> is on a dedicated file system, this is a finding.
Change the root account's shell to one present on the / file system. Procedure: Edit /etc/passwd and change the shell for the root account to one present on the / file system (such as /bin/sh, assuming /bin is not on a separate file system). If the system does not store shell configuration in the /etc/passwd file, consult vendor documentation for the correct procedure for the system.
If the "xorg-x11-server-Xorg" package is not installed, this is not applicable. For the Gnome screen saver, check the idle_activation_enabled flag. Procedure: # gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --get /apps/gnome-screensaver/idle_activation_enabled If this does not return "true" and a documented exception has not been made by the IAO, this is a finding.
For the Gnome screen saver, set the idle_activation_enabled flag. Procedure: # gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --type bool --set /apps/gnome-screensaver/idle_activation_enabled true
# ls /etc/security/opasswd If /etc/security/opasswd does not exist, then this is a finding. # grep password /etc/pam.d/system-auth| egrep '(pam_pwhistory.so|pam_unix.so|pam_cracklib.so)' | grep remember If the "remember" option in /etc/pam.d/system-auth is not 5 or greater, this is a finding. Check for system-auth-ac inclusions. # grep -c system-auth-ac /etc/pam.d/* If the system-auth-ac file is included anywhere # more /etc/pam.d/system-auth-ac | grep password | egrep '(pam_pwhistory.so|pam_unix.so|pam_cracklib.so)' | grep remember If in /etc/pam.d/system-auth-ac is referenced by another file and the "remember" option is not set to 5 or greater this is a finding.
Create the password history file. # touch /etc/security/opasswd # chown root:root /etc/security/opasswd # chmod 0600 /etc/security/opasswd Enable password history. If /etc/pam.d/system-auth references /etc/pam.d/system-auth-ac refer to the man page for system-auth-ac for a description of how to add options not configurable with authconfig. Edit /etc/pam.d/system-auth to include the remember option on any "password pam_unix" or "password pam_history" lines set to at least 5.
Determine the world writable files on the system (Note: ignore all files under /proc): # find / -perm -002 -a -type f -exec ls -ld {} \; | <more or redirect the output to a file> # find / -perm -002 -a -type d -exec ls -ld {} \; | <more or redirect the output to a file> View the password file to determine where the home directories for users are: # more /etc/passwd Once the directory for the human users is determined, grep for the lists of world writable files and directories within the users’ home directories. An example would be: # grep /opt/app/bin/daemon /home/*/.* where /home is the directory for the human users on the system and /opt/app/bin/daemon is a world writable file.
Remove the world-writable permission of files referenced by local initialization scripts, or remove the references to these files in the local initialization scripts.
Check run control scripts' ownership. # ls -lL /etc/rc* /etc/init.d Alternatively: # find /etc -name "[SK][0-9]*"|xargs stat -L -c %U:%n If any run control script is not owned by root or bin, this is a finding.
Change the ownership of the run control script(s) with incorrect ownership. # find /etc -name "[SK][0-9]*"|xargs stat -L -c %U:%n|egrep -v "^root:"|cut -d: -f2|xargs chown root
Check run control scripts' group ownership. Procedure: # ls -lL /etc/rc* /etc/init.d Alternatively: # find /etc -name "[SK][0-9]*"|xargs stat -L -c %G:%n|egrep -v "^(root|sys|bin|other):" If any run control script is not group-owned by root, sys, bin, or other system groups, this is a finding.
Change the group ownership of the run control script(s) with incorrect group ownership. Procedure: # chgrp root <run control script> # find /etc -name "[SK][0-9]*"|xargs stat -L -c %G:%n|egrep -v "^(root|sys|bin|other):"|cut -d: -f2|xargs chgrp root
Determine the programs executed by system start-up files. Determine the ownership of the executed programs. # cat /etc/rc*/* /etc/init.d/* | more # ls -l <executed program> Alternatively: # for FILE in `egrep -r "/" /etc/rc.* /etc/init.d|awk '/^.*[^\/][0-9A-Za-z_\/]*/{print $2}'|egrep "^/"|sort|uniq`;do if [ -e $FILE ]; then stat -L -c '%U:%n' $FILE;fi;done This provides a list of files referenced by initialization scripts and their associated UIDs. If any file is run by an initialization file and is not owned by root, sys, bin, or in rare cases, an application account, this is a finding.
Change the ownership of the file executed from system startup scripts to root, bin, sys, or other. # chown root <executed file>
On systems with a BIOS or system controller, verify a supervisor or administrator password is set. If a password is not set, this is a finding. If the BIOS or system controller supports user-level access in addition to supervisor/administrator access, determine if this access is enabled. If so, this is a finding.
Access the system's BIOS or system controller. Set a supervisor/administrator password if one has not been set. Disable a user-level password if one has been set.
Ask the SA if the system uses removable media for the boot loader. If it does, this is a finding.
Configure the system to use a bootloader installed on fixed media.
Determine if the system uses the GRUB boot loader; # ls -l /boot/grub/grub.conf If no grub.conf file exists, and the bootloader on the system has not been authorized, justified, and documented, this is a finding.
Configure the system to use the GRUB bootloader or document, justify, and authorize the alternate bootloader.
Check the "/boot/grub/grub.conf" or "/boot/grub/menu.lst" files. # more /boot/grub/menu.lst Check for a password configuration line, such as: password --md5 <password-hash> This line should be just below the line beginning with "timeout". Please note <password-hash> will be replaced by the actual MD5 encrypted password. If the password line is not in either of the files, this is a finding. For any bootloader other than GRUB which has been authorized, justified and documented for use on the system refer to the vendor documentation on password support. If the bootloader does not support encrypted passwords, this is a finding.
The GRUB console boot loader can be configured to use an MD5 encrypted password by adding password --md5 password-hash to the "/boot/grub/grub.conf" file. Use "/sbin/grub-md5-crypt" to generate MD5 passwords from the command line.
Check /boot/grub/grub.conf permissions: # ls -lL /boot/grub/grub.conf If /boot/grub/grub.conf has a mode more permissive than 0600, then this is a finding. For any bootloader other than GRUB which has been authorized, justified and documented for use on the system refer to the vendor documentation for the location of the configuration file. If the bootloader configuration file has a mode more permissive than 0600, this is a finding.
Change the mode of the grub.conf file to 0600. # chmod 0600 /boot/grub/grub.conf
Ask the SA if the system boots from removable media. If so, ask if the boot media is stored in a secure container when not in use. If it is not, this is a finding.
Store the system boot media in a secure container when not in use.
Perform the following to check for unnecessary privileged accounts: # grep "^shutdown" /etc/passwd # grep "^halt" /etc/passwd # grep "^reboot" /etc/passwd If any unnecessary privileged accounts exist this is a finding.
Remove any special privilege accounts, such as shutdown and halt, from the /etc/passwd and /etc/shadow files using the "userdel" or "system-config-users" commands.
Check the system for unnecessary user accounts. Procedure: # more /etc/passwd Obtain a list of authorized accounts from the IAO. If any unnecessary accounts are found on the system, this is a finding.
Remove all unnecessary accounts from the /etc/passwd file before connecting a system to the network. Other accounts that are associated with a service not in use should also be removed.
RHEL uses the InternetNewsDaemon (innd) news server. The file corresponding to "/etc/news/hosts.nntp" is "/etc/news/incoming.conf". Check the permissions for "/etc/news/incoming.conf". # ls -lL /etc/news/incoming.conf If "/etc/news/incoming.conf" has a mode more permissive than 0600, this is a finding.
Change the mode of the "/etc/news/incoming.conf" file to 0600. # chmod 0600 /etc/news/incoming.conf
RHEL uses the InternetNewsDaemon (innd) news server. The file that corresponds to "/etc/news/hosts.nntp.nolimit" is "/etc/news/infeed.conf". Check the permissions for "/etc/news/infeed.conf". # ls -lL /etc/news/infeed.conf If "/etc/news/infeed.conf" has a mode more permissive than 0600, this is a finding.
Change the mode of "/etc/news/infeed.conf" to 0600. # chmod 0600 /etc/news/infeed.conf
Check the permissions for "/etc/news/readers.conf". # ls -lL /etc/news/readers.conf If /etc/news/readers.conf has a mode more permissive than 0600, this is a finding.
Change the mode of the /etc/news/readers.conf file to 0600. # chmod 0600 /etc/news/readers.conf
Check "/etc/news/passwd.nntp" permissions: # ls -lL /etc/news/passwd.nntp If "/etc/news/passwd.nntp" has a mode more permissive than 0600, this is a finding.
Change the mode of the "/etc/news/passwd.nntp" file. # chmod 0600 /etc/news/passwd.nntp
Check the ownership of the files in "/etc/news". Procedure: # ls -al /etc/news If any files are not owned by root or news, this is a finding.
Change the ownership of the files in "/etc/news" to root or news. Procedure: # chown root /etc/news/*
Check "/etc/news" files group ownership: Procedure: # ls -al /etc/news If "/etc/news" files are not group-owned by root or news, this is a finding.
Change the group-owner of the files in "/etc/news" to root or news. Procedure: # chgrp root /etc/news/*
Locate the sshd_config file: # more /etc/ssh/sshd_config Examine the file. If the variables 'Protocol 2,1' or 'Protocol 1' are defined on a line without a leading comment, this is a finding. If the SSH server is F-Secure, the variable name for SSH 1 compatibility is 'Ssh1Compatibility', not 'protocol'. If the variable 'Ssh1Compatiblity' is set to 'yes', then this is a finding.
Edit the sshd_config file and set the "Protocol" setting to "2". If using the F-Secure SSH server, set the "Ssh1Compatibility" setting to "no". Restart the SSH daemon. # /sbin/service sshd restart
Check /etc/securetty # more /etc/securetty If the file does not exist, or contains more than "console" or a single "tty" device this is a finding.
Create if needed and set the contents of /etc/securetty to a "console" or "tty" device. # echo console > /etc/securetty or # echo tty1 > /etc/securetty
Check if NTP running: # ps -ef | egrep "xntpd|ntpd" Check if "ntpd -qg" scheduled to run: # grep "ntpd -qg" /var/spool/cron/* # grep "ntpd -qg" /etc/cron.d/* # grep "ntpd -qg" /etc/cron.daily/* # grep "ntpd -qg" /etc/cron.hourly/* # grep "ntpd -qg" /etc/cron.monthly/* # grep "ntpd -qg" /etc/cron.weekly/* If NTP is running or "ntpd -qg" is found: # more /etc/ntp.conf Confirm the timeservers and peers or multicast client (as applicable) are local or authoritative U.S. DoD sources appropriate for the level of classification which the network operates. If a non-local/non-authoritative time-server is used, this is a finding.
Use an authoritative local time server or a time server operated by the U.S. government. Ensure all systems in the facility feed from one or more local time servers which feed from the authoritative U.S. government time server.
Logging should be enabled for those types of file systems not turning on logging by default. Procedure: # mount JFS, VXFS, HFS, XFS, reiserfs, EXT3 and EXT4 all turn logging on by default and will not be a finding. The ZFS file system uses other mechanisms to provide for file system consistency, and will not be a finding. For other file systems types, if the root file system does not support journaling this is a finding. If the 'nolog' option is set on the root file system that does support journaling, this is a finding.
Implement file system journaling for the root file system, or use a file system with other mechanisms to ensure file system consistency. If the root file system supports journaling, enable it. If the file system does not support journaling or another mechanism to ensure file system consistency, a migration to a different file system will be necessary.
Check the system for a running Samba server. Procedure: # ps -ef |grep smbd If the Samba server is running, ask the SA if the Samba server is operationally required. If it is not, this is a finding.
If there is no functional need for Samba and the daemon is running, disable the daemon by killing the process ID as noted from the output of ps -ef |grep smbd. The samba package should also be removed or not installed if there is no functional requirement. Procedure: rpm -qa |grep samba This will show whether "samba" or "samba3x" is installed. To remove: rpm --erase samba or rpm --erase samba3x
Check /etc/sysctl.conf ownership. # ls -lL /etc/sysctl.conf If /etc/sysctl.conf is not owned by root, this is a finding.
Use the chown command to change the owner of /etc/sysctl.conf to root: # chown root /etc/sysctl.conf
Check /etc/sysctl.conf group ownership: # ls -lL /etc/sysctl.conf If /etc/sysctl.conf is not group-owned by root, this is a finding.
Use the chgrp command to change the group owner of /etc/sysctl.conf to root: # chgrp root /etc/sysctl.conf
Check /etc/sysctl.conf permissions: # ls -lL /etc/sysctl.conf If /etc/sysctl.conf has a mode more permissive than 0600, this is a finding.
Use the chmod command to change the mode of the /etc/sysctl.conf file. # chmod 0600 /etc/sysctl.conf
Determine if an NFS server is running on the system by: # ps -ef |grep nfsd If an NFS server is running, confirm it is not configured with the insecure_locks option by: # exportfs -v The example below would be a finding: /misc/export speedy.example.com(rw,insecure_locks)
Remove the "insecure_locks" option from all NFS exports on the system. Procedure: Edit /etc/exports and remove all instances of the insecure_locks option. Re-export the file systems to make the setting take effect. # exportfs -a
Verify that reboot using the CTRL-ALT-DELETE key sequence has been disabled by performing: # grep ctrlaltdel /etc/inittab If the line returned does not specify "/usr/bin/logger", or is not commented out, this is a finding.
Ensure the CTRL-ALT-DELETE key sequence has been disabled and attempts to use the sequence are logged. In the /etc/inittab file replace: ca::ctrlaltdel:/sbin/shutdown -t3 -r now with ca:nil:ctrlaltdel:/usr/bin/logger -p security.info "Ctrl-Alt-Del was pressed"
Ensure the pam_console.so module is not configured in any files in /etc/pam.d by: # cd /etc/pam.d # grep pam_console.so * Or # ls -la /etc/security/console.perms If either the pam_console.so entry or the file /etc/security/console.perms is found then this is a finding.
Configure PAM to not grant sole access of administrative privileges to the first user logged in at the console. Identify any instances of pam_console. # cd /etc/pam.d # grep pam_console.so * For any files containing an un-commented reference to pam_console.so, edit the file and remove or comment out the reference. Remove the console.perms file if it exists: # rm /etc/security/console.perms
Check for any crontab entries that rotate audit logs. Procedure: # crontab -l If such a cron job is found, this is not a finding. Otherwise, query the SA. If there is a process automatically rotating audit logs, this is not a finding. If the SA manually rotates audit logs, this is a finding, because if the SA is not there, it will not be accomplished. If the audit output is not archived daily, to tape or disk, this is a finding. This can be ascertained by looking at the audit log directory and, if more than one file is there, or if the file does not have today's date, this is a finding.
Configure a cron job or other automated process to rotate the audit logs on a daily basis.
Check the mode of the cron.deny file. # ls -lL /etc/cron.deny If the cron.deny file does not exist this is not a finding. If the cron.deny file exists and the mode is more permissive than 0600, this is a finding.
Change the mode of the cron.deny file. # chmod 0600 /etc/cron.deny
Determine if there are any crontabs by viewing a long listing of the directory. If there are crontabs, examine them to determine what cron jobs exist. Check for any programs specifying an umask more permissive than 077: Procedure: # ls -lL /var/spool/cron # ls -lL /etc/cron.d /etc/cron.daily /etc/cron.hourly /etc/cron.monthly /etc/cron.weekly or # ls -lL /etc/cron.*|grep -v deny # cat <crontab file> # grep umask <cron program> If there are no cron jobs present, this vulnerability is not applicable. If any cron job contains an umask more permissive than 077, this is a finding.
Edit cron script files and modify the umask to 077.
# ls -lL /etc/cron.allow If the cron.allow file is not owned by root, sys, or bin, this is a finding.
# chown root /etc/cron.allow
Check the mode of the "at" directory. Procedure: # ls -ld /var/spool/at If the directory mode is more permissive than 0755, this is a finding.
Change the mode of the "at" directory to 0755. Procedure: # chmod 0755 <at directory>
Check the ownership of the "at" directory: Procedure: # ls -ld /var/spool/at If the directory is not owned by root, sys, bin, daemon, or cron, this is a finding.
Change the owner of the "at" directory to root, bin, sys, or system. Procedure: # chown <root or other system account> <"at" directory>
Determine what "at" jobs exist on the system. Procedure: # ls /var/spool/at If there are no "at" jobs present, this is not applicable. Determine if any of the "at" jobs or any scripts referenced execute the "umask" command. Check for any umask setting more permissive than 077. # grep umask <at job or referenced script> If any "at" job or referenced script sets umask to a value more permissive than 077, this is a finding.
Edit "at" jobs or referenced scripts to remove "umask" commands that set umask to a value less restrictive than 077.
# ls -lL /etc/at.allow If the at.allow file is not owned by root, sys, or bin, this is a finding.
Change the owner of the at.allow file. # chown root /etc/at.allow
# ls -lL /etc/at.deny If the at.deny file is not owned by root, sys, or bin, this is a finding.
Change the owner of the at.deny file. # chown root /etc/at.deny
# ls -lL /bin/traceroute If the traceroute command is not owned by root, this is a finding.
Change the owner of the traceroute command to root. Example: # chown root /bin/traceroute
Check the group ownership of the traceroute file. Procedure: # ls -lL /bin/traceroute If the traceroute command is not group-owned by root, sys, bin, or system, this is a finding.
Change the group-owner of the traceroute command to root. Procedure: # chgrp root /bin/traceroute
# ls -lL /bin/traceroute If the traceroute command has a mode more permissive than 0700, this is a finding.
Change the mode of the traceroute command. # chmod 0700 /bin/traceroute
Look in the root account home directory for a .mozilla directory. If none exists, this is not a finding. If there is one, verify with the root users and the IAO the intent of the browsing. If the browsing is not limited to authorized local services administration, this is a finding.
Enforce policy requiring administrative accounts use web browsers only for local service administration.
To check for the version of either sendmail or Postfix being displayed in the greeting: # telnet localhost 25 If a version number is displayed, this is a finding.
Ensure sendmail or Postfix has been configured to mask the version information. Procedure for sendmail: Edit the /etc/mail/sendmail.mc file to mask the veresion number by editing the line with "dnl" as follows: define(`confSMTP_LOGIN_MSG', ` Mail Server Ready ; $b')dnl rebuild the sendmail.cf file. for Postfix: Examine the "smtpd_banner" line of /etc/postfix/main.conf and remove any "$mail_version" entry on it or comment the entire "smtpd_banner" line to use the default value which does not display the version information.
Check forwarding capability from sendmail. Procedure: grep "O ForwardPath" /etc/mail/sendmail.cf If the entry contains a file path, this is a finding. Search for any .forward in users home directories on the system by: # for pwline in `cut -d: -f1,6 /etc/passwd`; do homedir=`echo ${pwline}|cut -d: -f2`;username=`echo ${pwline} | cut -d: -f1`;echo $username `stat -c %n $homedir/.forward 2>null`; done|egrep "\.forward" If any users have a .forward file in their home directory, this is a finding.
Disable forwarding for sendmail and remove .forward files from the system Procedure: Edit the /etc/mail/sendmail.mc file to change the ForwardPath entry to a null path by adding the line define(`confFORWARD_PATH',`') rebuild the sendmail.cf file. If the /etc/mail/sendmail.mc file does not exist, the sendmail.cf file should be updated directly. Remove all .forward files on the system # find / -name .forward -delete
Check the shell for the anonymous FTP account. Procedure: # cat /etc/passwd | grep -i "^ftp:" | awk -F":" '{print$7}' If the seventh field is empty (the entry ends with a ':') or if the seventh field does not contain one of the following, this is a finding: /bin/false /dev/null /usr/bin/false /bin/true /sbin/nologin
Configure anonymous FTP accounts to use a non-functional shell. If necessary, edit the /etc/passwd file to remove any functioning shells associated with the ftp account and replace them with non-functioning shells, such as /dev/null.
Ask the SA if this is an NMS server. If it is an NMS server, then ask what other applications run on it. If there is anything other than network management software and DBMS software used only for the storage and inquiry of NMS data, this is a finding.
Ensure only authorized software is loaded on a designated NMS server. Authorized software is limited to the NMS software itself, a database management system for the NMS server if necessary, and network management software.
Check /etc/syslog.conf or /etc/rsyslog.conf ownership: For syslog: # ls -lL /etc/syslog.conf For rsyslog: # ls -lL /etc/rsyslog.conf If /etc/syslog.conf or /etc/rsyslog.conf is not owned by root, this is a finding.
Use the chown command to set the owner to root. # chown root /etc/syslog.conf Or: # chown root /etc/rsyslog.conf
Check /etc/syslog.conf or /etc/rsyslog.conf group ownership. Procedure: # ls -lL /etc/syslog.conf Or: # ls -lL /etc/syslog.conf If /etc/syslog.conf or /etc/rsyslog.conf is not group owned by root, sys, bin, or system, this is a finding.
Procedure: # chgrp root /etc/syslog.conf Or # chgrp root /etc/rsyslog.conf
Examine the syslog.conf or rsyslog.conf file for any references to remote log hosts. # grep -v "^#" /etc/syslog.conf | grep '@' Or: # grep -v "^#" /etc/rsyslog.conf | grep '@' Destination locations beginning with an '@' represent log hosts. If the log host name is a local alias such as "loghost", consult the /etc/hosts or other name databases as necessary to obtain the canonical name or address for the log host. Determine if the host referenced is a log host documented using site-defined procedures. If an undocumented log host is referenced, this is a finding.
Remove or document the referenced undocumented log host.
Check the system for an IPv4 default route. If the system is a VM host and acts as a router solely for the benefit of its client systems, then this rule is not applicable. Procedure: # netstat -r |grep default If a default route is not defined, this is a finding.
Set a default gateway for IPv4.
If the system is a VM host and acts as a router solely for the benefit of its client systems, then this rule is not applicable. Check to see if the system is a router: # chkconfig --list | grep :on | egrep '(ospf|route|bgp|zebra|quagga)' If the system is running a routing service, it is a router. If it is not, this is not applicable. Check the system for non-routing network services. Procedure: # netstat -a | grep -i listen # ps -ef If non-routing services, including Web servers, file servers, DNS servers, or applications servers, but excluding management services such as SSH and SNMP, are running on the system, this is a finding.
Ensure only authorized software is loaded on a designated router. Authorized software will be limited to the most current version of routing protocols and SSH for system administration purposes.
If the system does not use NIS or NIS+, this is not applicable. Check if NIS or NIS+ is implemented using UDP. Procedure: # rpcinfo -p | grep yp | grep udp If NIS or NIS+ is implemented using UDP, this is a finding.
Configure the system to not use UDP for NIS and NIS+. Consult vendor documentation for the required procedure.
Locate and examine all r-commands access control files. Procedure: # find / -name .rhosts # more /<directorylocation>/.rhosts # find / -name .shosts # more /<directorylocation>/.shosts # find / -name hosts.equiv # more /<directorylocation>/hosts.equiv # find / -name shosts.equiv # more /<directorylocation>/shosts.equiv If any .rhosts, .shosts, hosts.equiv, or shosts.equiv file contains other than host-user pairs, this is a finding.
If possible, remove the .rhosts, .shosts, hosts.equiv, and shosts.equiv files. If the files are required, remove any content from the files except for necessary host-user pairs.
Procedure: # ls -l /etc/hosts.equiv # ls -l /etc/ssh/shosts.equiv # find / -name .rhosts # ls -al <home directory>/.rhosts # find / -name .shosts # ls -al <home directory>/.shosts # find / -name .netrc # ls -al <home directory>/.netrc If the .rhosts, .shosts, hosts.equiv, or shosts.equiv files have permissions greater than 600, then this is a finding. If the /etc/hosts.equiv, or /etc/ssh/shosts.equiv files are not owned by root, this is a finding. Any .rhosts, .shosts and .netrc files outside of home directories have no meaning and are not subject to this rule If the ~/.rhosts or ~/.shosts are not owned by the owner of the home directory where they are immediately located or by root, this is a finding.
Ensure the permission for these files is set to 600 or more restrictive and their owner is root or the same as the owner of the home directory in which they reside. Procedure: # chmod 600 /etc/hosts.equiv # chmod 600 /etc/ssh/shosts.equiv # chown root /etc/hosts.equiv # chown root /etc/ssh/shosts.equiv # find / -name .rhosts # chmod 600 /<home directory>/.rhosts # chown <home directory owner> <home directory>/.rhosts # find / -name .shosts # chmod 600 <directory location>/.shosts # chown <home directory owner> <home directory>/.shosts # find / -name .netrc # chmod 600 <directory location>/.netrc # chown <home directory owner> <home directory>/.netrc
# ls -lL /etc/cron.deny If the cron.deny file is not owned by root, sys, or bin, this is a finding.
# chown root /etc/cron.deny
Check to see if rshd is configured to run on startup. Procedure: # grep disable /etc/xinetd.d/rsh If /etc/xinetd.d/rsh exists and rsh is found to be enabled, this is a finding.
Edit /etc/xinetd.d/rsh and set "disable=yes".
# grep disable /etc/xinetd.d/rexec If the service file exists and is not disabled, this is a finding.
Edit /etc/xinetd.d/rexec and set "disable=yes"
Determine the version of the SMTP service software. Procedure: #rpm -q sendmail RedHat sendmail 8.13.8-8 is the latest required version. If the RedHat sendmail is installed and the version is not at least 8.13.8-8, this is a finding. #rpm -q postfix RedHat postfix-2.3.3-6-el5 is the latest required version. If the postfix is installed and the version is not at least postfix-2.3.3-6-el5, this is a finding.
Obtain and install a newer version of the SMTP service software (sendmail or Postfix) from RedHat.
Check for an enabled "debug" command provided by the SMTP service. Procedure: # telnet localhost 25 debug If the command does not return a 500 error code of "command unrecognized" or a 550 error code of "access denied", this is a finding. The RHEL distribution ships with sendmail Version 8.13.8 which is not vulnerable. This should never be a finding.
Obtain and install a newer version of the SMTP service software (sendmail or Postfix) from RedHat.
Check the SMTP service for an active "decode" command. Procedure: # telnet localhost 25 decode If the command does not return a 500 error code of "command unrecognized", this is a finding.
Disable mail aliases for decode and uudecode. If the /etc/aliases or /usr/lib/aliases (mail alias) file contains entries for these programs, remove them or disable them by placing "#" at the beginning of the line, and then executing the new aliases command. For more information on mail aliases, refer to the man page for aliases. Disabled aliases would be similar to these examples: # decode: |/usr/bin/uudecode # uudecode: |/usr/bin/uuencode -d
This vulnerability is applicable only to sendmail. If Postfix is the SMTP service for the system this will never be a finding. Procedure: Determine if EXPN is disabled. # grep -v "^#" /etc/mail/sendmail.cf |grep -i PrivacyOptions If nothing is returned or the returned line does not contain "noexpn", this is a finding.
Rebuild /etc/mail/sendmail.cf with the "noexpn" Privacy Flag set. Procedure: Edit /etc/mail/sendmail.mc resetting the Privacy Flags to the default: define('confPRIVACYFLAGS', 'authwarnings,novrfy,noexpn,restrictqrun')dnl Rebuild the sendmail.cf file with: # make -C /etc/mail Restart the sendmail service. # service sendmail restart
Determine if VRFY is disabled. Procedure: # telnet localhost 25 vrfy root If the command does not return a 500 error code of "command unrecognized", this is a finding. or: # grep -v "^#" /etc/mail/sendmail.cf |grep -i vrfy Verify the VRFY command is disabled with an entry in the sendmail.cf file. The entry could be any one of "Opnovrfy", "novrfy", or "goaway", which could also have other options included, such as "noexpn". The "goaway" argument encompasses many things, such as "novrfy" and "noexpn". If no setting to disable VRFY is found, this is a finding.
Add the "novrfy" flag to your sendmail in /etc/mail/sendmail.cf. Procedure: Edit the definition of "confPRIVACY_FLAGS" in /etc/mail/sendmail.mc to include "novrfy". Rebuild the sendmail.cf file with: # make -C /etc/mail Restart the sendmail service. # service sendmail restart
Log into the sendmail server with telnet and test the "wiz" command. Procedure: # telnet localhost 25 Trying 127.0.0.1... Connected to locahost.localdomain (127.0.0.1). Escape character ... Once the telnet greeting is complete type: wiz If you do not get a "Command unrecognized: " message, this is a finding.
If the WIZ command exists on sendmail then the version of sendmail is archaic and should be replaced with the latest version from RedHat. WIZ is not available on any sendmail distribution of RHEL. However, if the WIZ command is enabled on sendmail, it should be disabled by adding this line to the sendmail.cf configuration file (note that it must be typed in uppercase): OW* For the change to take effect, kill the sendmail process, refreeze the sendmail.cf file, and restart the sendmail process.
Determine if the TFTP daemon is active. # chkconfig --list | grep tftp If TFTP is found enabled ("on") and not documented using site-defined procedures, it is a finding.
Document or Disable the TFTP daemon. If the TFTP daemon is necessary on the system, document and justify its usage for approval from the IAO. If the TFTP daemon is not necessary on the system, turn it off. # chkconfig tftp off # service xinetd restart
# service uucp status if UUCP is "running", this is a finding.
# chkconfig uucp off # service uucp stop # service xinetd restart
If Xwindows is not used on the system, this is not applicable. Check the output of the "xhost" command from an X terminal. Procedure: # xhost If the output reports access control is enabled (and possibly lists the hosts able to receive X window logins), this is not a finding. If the xhost command returns a line indicating access control is disabled, this is a finding. Note: It may be necessary to define the display if the command reports it cannot open the display. Procedure: $ DISPLAY=MachineName:0.0; export DISPLAY MachineName may be replaced with an Internet Protocol Address. Repeat the check procedure after setting the display.
If using an xhost-type authentication the "xhost -" command can be used to remove current trusted hosts and then selectively allow only trusted hosts to connect with "xhost +" commands. A cryptographically secure authentication, such as provided by the xauth program, is always preferred. Refer to your X11 server's documentation for further security information.
# grep disable /etc/xinetd.d/finger If the finger service is not disabled, this is a finding.
Edit /etc/xinetd.d/finger and set "disable=yes"
Use the command "ftp" to connect the system's FTP service. Attempt to log into this host with a user name of anonymous and a password of guest (also try the password of guest@mail.com). If the logon is not successful, this check is Not Applicable. Ask the SA if the system is located on a DMZ network. If the system is not located on a DMZ network, this is a finding.
Remove anonymous ftp capability or move the system to a DMZ network.
Check the version of the operating system. Example: # cat /etc/redhat-release Vendor End-of-Support Information: Red Hat Enterprise 5: 31 Mar 2017 Check with the vendor for additional information. If the version installed is not supported, this is a finding.
Upgrade to a supported version of the operating system.
Check for the presence of an aide on the system: # rpm –qa | grep aide If aide is not installed, ask the SA what file integrity tool is being used to check the system. Check the global crontabs for the presence of an "aide" job to run at least weekly, if aide is installed. Otherwise, check for the presence of a cron job to run the alternate file integrity checking application. # grep aide /etc/cron*/* If a tool is being run then the configuration file for the appropriate tool needs to be checked for selection lines /bin, /sbin, /lib, and /usr. Procedure: Check the root crontab (crontab -l) and the global crontabs in /etc/crontab, /etc/cron.d/* for the presence of an "aide" job to run at least weekly, which should have asterisks (*) in columns 3, 4, and 5. Check the weekly cron directory (/etc/cron.weekly) for any script running "aide --check" or "aide -C" or simply "aide". If there is not, this is a finding. NOTE: The frequency may be increased to daily, if necessary, in accordance with the contingency plan.
Establish an automated job, scheduled to run weekly or more frequently, to run "aide --check" which is the file integrity tool to check for unauthorized system libraries or binaries. NOTE: The frequency may be increased to daily, if necessary, in accordance with the contingency plan.
Check the UID assignments for all accounts. # cut -d: -f 1,3 /etc/passwd | egrep ":[1-4][0-9]{2}$|:[0-9]{1,2}$" Confirm all accounts with a UID of 499 and below are used by a system account. If a UID reserved for system accounts (0 - 499) is used by a non-system account, then this is a finding.
Change the UID numbers for non-system accounts with reserved UIDs (those less or equal to 499).
Check the system password length setting. Procedure: Check the password minlen option # grep pam_cracklib.so /etc/pam.d/system-auth Confirm the minlen option is set to at least 15 as in the example below: password required pam_cracklib.so minlen=15 There may be other options on the line. If no such line is found, or the minlen is less than 15, this is a finding. # grep PASS_MIN_LEN /etc/login.defs Confirm the PASS_MIN_LEN option is set to at least 15 as in the example below: PASS_MIN_LEN 15
Edit "/etc/pam.d/system-auth" to include the line: password required pam_cracklib.so minlen=15 prior to the "password include system-auth-ac" line. Edit /etc/login.defs to include the line: PASS_MIN_LEN 15
Check the ucredit setting. # grep ucredit /etc/pam.d/system-auth If ucredit is not set to -1, this is a finding.
Edit "/etc/pam.d/system-auth" to include the line: password required pam_cracklib.so ucredit=-1 prior to the "password include system-auth-ac" line.
Check the dcredit setting. Procedure: Check the password dcredit option # grep pam_cracklib.so /etc/pam.d/system-auth Confirm the dcredit option is set to -1 as in the example: password required pam_cracklib.so dcredit=-1 There may be other options on the line. If no such line is found, or the dcredit option is not -1 this is a finding.
Edit "/etc/pam.d/system-auth" to include the line: password required pam_cracklib.so dcredit=-1 prior to the "password include system-auth-ac" line.
Check the ocredit setting. Procedure: Check the password ocredit option # grep pam_cracklib.so /etc/pam.d/system-auth Confirm the ocredit option is set to -1 as in the example: password required pam_cracklib.so ocredit=-1 There may be other options on the line. If no such line is found, or the ocredit is not -1 this is a finding.
Edit "/etc/pam.d/system-auth" to include the line: password required pam_cracklib.so ocredit=-1 prior to the "password include system-auth-ac" line.
Check the maxrepeat setting. Procedure: Check the password maxrepeat configuration # grep pam_cracklib.so /etc/pam.d/system-auth If the maxrepeat option is missing, this is a finding. If the maxrepeat option is set to more than 3, this is a finding.
Edit "/etc/pam.d/system-auth" to include the line: password required pam_cracklib.so maxrepeat=3 prior to the "password include system-auth-ac" line.
Check the max days field (the 5th field) of /etc/shadow. # more /etc/shadow If the max days field is equal to 0 or greater than 60 for any user, this is a finding.
Set the max days field to 60 for all user accounts. # passwd -x 60 <user>
Ask the SA if there are any automated processing accounts on the system. If there are automated processing accounts on the system, ask the SA if the passwords for those automated accounts are changed at least once a year or are locked. If SA indicates passwords for automated processing accounts are not changed once per year or are not locked, this is a finding.
Implement or establish procedures to change the passwords of automated processing accounts at least once per year or lock them.
Check if root is used for direct logins. Procedure: # last root | grep -v reboot Direct logins are indicated by the presence of a terminal or pseudo-terminal ID and/or X display name in the output of the last command. If any direct login records for root are listed, this is a finding.
Enforce policy requiring all root account access is attained by first logging into a user account and then becoming root preferably through the use of "sudo" which provides traceability to the command level. If that is not workable then using "su" to access the root account will provide traceability to the login user.
Check the log files to determine if access to the root account is being logged. Procedure: Depending on what system is used for log processing either /etc/syslog.conf or /etc/rsyslog.conf will be the logging configuration file. Examine /etc/syslog.conf or /etc/rsyslog.conf to confirm the location to which "authpriv" messages will be directed. The default syslog.conf or rsyslog.conf uses /var/log/messages and /var/log/secure but this needs to be confirmed. # grep @ /etc/syslog.conf Or: # grep @ /etc/rsyslog.conf If a line starting with "*.*" is returned then all syslog messages will be sent to system whose address appears after the "@". In this case syslog may or may not be configured to also log "authpriv" messages locally. # grep authpriv /etc/syslog.conf Or: # grep authpriv /etc/rsyslog.conf If any lines are returned which do not start with "#" the "authpriv" messages will be sent to the indicated files or remote systems. Try to "su -" and enter an incorrect password. If there are no records indicating the authentication failure, this is a finding.
Troubleshoot the system logging configuration to provide for logging of root account login attempts. Procedure: Edit /etc/syslog.conf or /etc/rsyslog.conf to make sure "authpriv.*" messages are directed to a file or remote system. Examine /etc/audit/audit.rules to ensure user authentication messages have not been specifically excluded. Remove any entries that correspond to: -a exclude,never -Fmsgtype=USER_START -a exclude,never -Fmsgtype=USER_LOGIN -a exclude,never -Fmsgtype=USER_AUTH -a exclude,never -Fmsgtype=USER_END -a exclude,never -Fmsgtype=USER_ACCT
Check global initialization files permissions: # ls -l /etc/bashrc # ls -l /etc/csh.cshrc # ls -l /etc/csh.login # ls -l /etc/csh.logout # ls -l /etc/environment # ls -l /etc/ksh.kshrc # ls -l /etc/profile # ls -l /etc/suid_profile # ls -l /etc/profile.d/* If global initialization files are more permissive than 0644, this is a finding.
Change the mode of the global initialization file(s) to 0644. # chmod 0644 <global initialization file>
Check the ownership of global initialization files. Procedure: # ls -lL etc/bashrc /etc/csh.cshrc /etc/csh.login /etc/csh.logout /etc/environment /etc/ksh.kshrc /etc/profile /etc/suid_profile /etc/profile.d/* This should show information for each file. Examine to ensure the owner is always root or: # ls etc/bashrc /etc/csh.cshrc /etc/csh.login /etc/csh.logout /etc/environment /etc/ksh.kshrc /etc/profile /etc/suid_profile /etc/profile.d/* 2>null|xargs stat -L -c %U:%n|egrep -v "^root" This will show you only the owner and filename of files not owned by root. If any global initialization file is not owned by root, this is a finding.
Change the ownership of global initialization files with incorrect ownership. Procedure: # chown root <global initialization files> or: # ls etc/bashrc /etc/csh.cshrc /etc/csh.login /etc/csh.logout /etc/environment /etc/ksh.kshrc /etc/profile /etc/suid_profile /etc/profile.d/* 2>null|xargs stat -L -c %U:%n|egrep -v "^root"|cut -d: -f2|xargs chown root will set the owner of all files not currently owned by root to root.
Check the group ownership of global initialization files. Procedure: # ls -lL etc/bashrc /etc/csh.cshrc /etc/csh.login /etc/csh.logout /etc/environment /etc/ksh.kshrc /etc/profile /etc/suid_profile /etc/profile.d/* This should show information for each file. Examine to ensure the group is always root or: # ls -lL etc/bashrc /etc/csh.cshrc /etc/csh.login /etc/csh.logout /etc/environment /etc/ksh.kshrc /etc/profile /etc/suid_profile /etc/profile.d/* 2>null|sed "s/^[^\/]*//"|xargs stat -L -c %G:%n|egrep -v "^(root|sys|bin|other):" will show you only the group and filename of files not owned by one of the approved groups. If any global initialization file is not group-owned by root, sys, bin, other, system, or the system default, this is a finding.
Change the group ownership of the global initialization file(s) with incorrect group ownership. Procedure: # chgrp root <global initialization file> or: # ls -lL /etc/bashrc /etc/csh.cshrc /etc/csh.login /etc/csh.logout /etc/environment /etc/ksh.kshrc /etc/profile /etc/suid_profile /etc/profile.d/* 2>null|sed "s/^[^\/]*//"|xargs stat -L -c %G:%n|egrep -v "^(root|sys|bin|other):"|cut -d: -f2|xargs chgrp root will set the group of all files not currently owned by an approved group to root.
Check skeleton files ownership. # ls -alL /etc/skel If a skeleton file is not owned by root or bin, this is a finding.
Change the ownership of skeleton files with incorrect mode: # chown root <skeleton file> or # ls -L /etc/skel|xargs stat -L -c %U:%n|egrep -v "^(root|bin):"|cut -d: -f2|chown root will change all files not owned by root or bin to root.
Check the global initialization files' executable search paths. Procedure: # grep PATH /etc/bashrc /etc/csh.cshrc /etc/csh.login /etc/csh.logout /etc/environment /etc/ksh.kshrc /etc/profile /etc/suid_profile /etc/profile.d/* This variable is formatted as a colon-separated list of directories. If there is an empty entry, such as a leading or trailing colon, or two consecutive colons, this is a finding. If an entry begins with a character other than a slash (/), or has not been documented with the ISSO, this is a finding.
Edit the global initialization file(s) with PATH variables containing relative paths and remove any relative path from the PATH variables that have not been authorized by the ISSO. Edit the global initialization file(s) and remove any empty entry that is defined.
Verify local initialization files have executable search paths containing only absolute paths or relative paths that have been documented by the ISSO. Procedure: NOTE: This must be done in the BASH shell. # cut -d: -f6 /etc/passwd |xargs -n1 -IDIR find DIR -name ".*" -type f -maxdepth 1 -exec grep -l PATH {} \; This variable is formatted as a colon-separated list of directories. If there is an empty entry, such as a leading or trailing colon, or two consecutive colons, this is a finding. If an entry begins with a character other than a slash (/), or has not been documented with the ISSO, this is a finding.
Edit the local initialization file and remove the relative path entry from the executable search path variable. If this is not feasible, justify and document the necessity of having the relative path for a specific application with the ISSO. Edit the local initialization file and remove any empty entry that is defined.
Check system configuration files for plus (+) entries. Procedure: # find / -name .rhosts # grep + /<directorylocation>/.rhosts # find / -name .shosts # grep + /<directorylocation>/.shosts # find / -name hosts.equiv # grep + /<directorylocation>/hosts.equiv # find / -name shosts.equiv # grep + /<directorylocation>/shosts.equiv # grep + /etc/passwd # grep + /etc/shadow # grep + /etc/group If the .rhosts, .shosts, hosts.equiv, shosts.equiv, /etc/passwd, /etc/shadow, and/or /etc/group files contain a plus (+) and do not define entries for NIS+ netgroups, this is a finding.
Edit the .rhosts, .shosts, hosts.equiv, shosts.equiv, /etc/passwd, /etc/shadow, and/or /etc/group files and remove entries containing a plus (+).
Check for the existence of the files. # find / -name .rhosts # find / -name .shosts # find / -name hosts.equiv # find / -name shosts.equiv If .rhosts, .shosts, hosts.equiv, or shosts.equiv are found and their use has not been documented and approved by the IAO, this is a finding.
Remove all the r-commands access control files. Procedure: # find / -name .rhosts -exec rm {} \; # find / -name .shosts -exec rm {} \; # find / -name hosts.equiv -exec rm {} \; # find / -name shosts.equiv -exec rm {} \;
Check the PAM configuration for rhosts_auth. Example: # grep rhosts_auth /etc/pam.d/* If a rhosts_auth entry is found, this is a finding.
Edit the file(s) in /etc/pam.d referencing the rhosts_auth module, and remove the references to the rhosts_auth module.
Check the group-ownership of public directories. Procedure: # find / -type d -perm -1002 -exec ls -ld {} \; If any public directory is not group-owned by root, sys, bin, or an application group, this is a finding.
Change the group-ownership of the public directory. Procedure: # chgrp root /tmp (Replace root with a different system group and/or /tmp with a different public directory as necessary.)
List all crontabs on the system. # ls -lL /var/spool/cron # ls -lL /etc/cron.d /etc/crontab /etc/cron.daily /etc/cron.hourly /etc/cron.monthly /etc/cron.weekly or # ls -lL /etc/cron*|grep -v deny If any crontab is not owned by root or the creating user, this is a finding.
Change the crontab owner to root or the crontab creator. # chown root <crontab file> or # chown <user> <crontab file>
Check the cron.allow and cron.deny files for the system. # more /etc/cron.allow # more /etc/cron.deny If a default system account (such as bin, sys, adm, or others, traditionally UID less than 500) is listed in the cron.allow file, or not listed in the cron.deny file and if no cron.allow file exists, this is a finding.
Remove default system accounts (such as bin, sys, adm, or others, traditionally UID less than 500) from the cron.allow file if it exists, or add those accounts to the cron.deny file.
# ulimit -Hc If the above command does not return 0 and the enabling of core dumps has not been documented and approved by the IAO, this a finding.
Edit /etc/security/limits.conf and set a hard limit for "core" to 0 for all users. A new login will be required for the changes to take effect.
Verify the location of the kernel core dump data directory: # grep “path” /etc/kdump.conf Note: If this setting is not configured or commented out, the default kernel dump data directory is /var/crash. Check the ownership of the dump data directory: # ls –ld <path to kernel core dump data directory> If the kernel core dump data directory is not owned by root, this is a finding.
Change the owner of the kernel core dump data directory to root. # chown root <path to kernel core dump data directory>
Verify "exec_shield" and "randomize_va_space" have not been changed from the default "1" settings. Procedure: #sysctl kernel.exec-shield If the return value is not: kernel.exec-shield = 1 this is a finding. #sysctl kernel.randomize_va_space If the return value is not: kernel.randomize_va_space = 1 this is a finding.
Examine /etc/sysctl.conf for "kernel.exec-shield" and "kernel.randomize_va_space" entries and if found remove them. The system default of "1" enables these modules.
Verify the system does not accept source-routed IPv4 packets. Procedure: # grep [01] /proc/sys/net/ipv4/conf/*/accept_source_route|egrep "default|all" If all of the returned lines do not end with 0, this is a finding. Note: The same setting is used by Linux for both the local acceptance and forwarding of source-routed IPv4 packets.
Configure the system to not accept source-routed IPv4 packets. Edit /etc/sysctl.conf and add a setting for "net.ipv4.conf.all.accept_source_route=0" and "net.ipv4.conf.default.accept_source_route=0". Reload the sysctls. Procedure: # sysctl -p
Determine if the /home path is a separate filesystem. # grep "/home " /etc/fstab If no result is returned, /home is not on a separate filesystem this is a finding.
Migrate the /home (or equivalent) path onto a separate file system.
Depending on what system is used for log processing either /etc/syslog.conf or /etc/rsyslog.conf will be the logging configuration file. Check /etc/syslog.conf or /etc/rsyslog.conf and verify the authpriv facility is logging both the "notice" and "info" priority messages. Procedure: For a given action all messages of a higher severity or "priority" are logged. The three lowest priorities in ascending order are "debug", "info" and "notice". A priority of "info" will include "notice". A priority of "debug" includes both "info" and "notice". Enter/Input for syslog: # grep "authpriv.debug" /etc/syslog.conf # grep "authpriv.info" /etc/syslog.conf # grep "authpriv\.\*" /etc/syslog.conf Enter/Input for rsyslog: # grep "authpriv.debug" /etc/rsyslog.conf # grep "authpriv.info" /etc/rsyslog.conf # grep "authpriv\.\*" /etc/rsyslog.conf If an "authpriv.*", "authpriv.debug", or "authpriv.info" entry is not found, this is a finding.
Edit /etc/syslog.conf or /etc/rsyslog.conf and add local log destinations for "authpriv.*", "authpriv.debug" or "authpriv.info".
# ps -ef |grep xinetd If xinetd is not running, this check is not a finding. # grep -v "^#" /etc/xinetd.conf # grep disable /etc/xinetd.d/* |grep no If no active services are found, and the inetd daemon is running, this is a finding.
# service xinetd stop ; chkconfig xinetd off
Check if Help is disabled. This rule is for "sendmail" only and not applicable to "Postfix". Procedure: # telnet localhost 25 > help If the help command returns any sendmail version information, this is a finding. If sendmail is not installed this check is not applicable.
To disable the SMTP HELP command, clear the Sendmail help file. # echo -n > /etc/mail/helpfile
Perform the following to determine if unencrypted FTP is enabled: # chkconfig --list gssftp # chkconfig --list vsftpd If any of these services are found, ask the SA if these services are encrypted. If they are not, this is a finding.
Disable the FTP daemons. Procedure: # chkconfig gssftp off # chkconfig vsftpd off
Check the umask setting for FTP users. Procedure: For gssftp: Assuming an anonymous ftp user has been defined with no user initialization script invoked to change the umask # ftp localhost Name: (localhost:root): anonymous Password: anything ftp>umask If the umask value returned is not 077, this is a finding. or: # grep "server_args" /etc/xinetd.d/gssftp The default umask for FTP is "023" if the server _args entry does not contain "-u 077" this is a finding. For vsftp: # grep "_mask" /etc/vsftpd/vsftpd.conf The default "local_umask" setting is 077. If this has been changed, or the "anon_umask" setting is not 077, this is a finding.
Edit the initialization files for the ftp user and set the umask to 077. Procedure: For gssftp: Modify the /etc/xinetd.d/gssftp file adding "-u 077" to the server_args entry. For vsftp: Modify the "/etc/vsftpd/vsftpd.conf" setting "local_umask" and "anon_umask" to 077.
Check the file permissions for the .Xauthority files. Procedure: # ls -la |egrep "(\.Xauthority|\.xauth)" If the file mode is more permissive than 0600, this is finding.
Change the mode of the .Xauthority files. Procedure: # chmod 0600 .Xauthority
Determine if the X server is running. Procedure: # ps -ef |grep X Determine if xauth is being used. Procedure: # xauth xauth> list If the above command sequence does not show any host other than the localhost, then xauth is not being used. Search the system for an X*.hosts file, where "*" is a display number used to limit X window connections. If no files are found, X*.hosts files are not being used. If the X*.hosts files contain any unauthorized hosts, this is a finding. If both xauth and X*.hosts files are not being used, this is a finding.
Create an X*.hosts file, where "*" is a display number used to limit X window connections. Add the list of authorized X clients to the file.
Check the X window system access is limited to authorized clients. Procedure: # xauth xauth> list Ask the SA if the clients listed are authorized. If any are not, this is a finding.
Remove unauthorized clients from the xauth configuration. # xauth remove <display name>
Determine if the X window system is running. Procedure: # ps -ef |grep Xorg Ask the SA if the X window system is an operational requirement. If it is not, this is a finding.
Disable the X Windows server on the system.
Determine the owner of the SNMP configuration file. Procedure: Find the snmpd.conf file. The default install location is /etc/snmp/snmpd.conf but may be different depending on the SNMP agent installed. # find / -name snmpd.conf # ls -lL <snmpd.conf> If the snmpd.conf file is not owned by root, this is a finding.
Change the owner of the snmpd.conf file to root. Procedure: # chown root <snmpd.conf file>
Ask the SA if the loghost server is collecting data for hosts outside the local enclave. If it is, this is a finding.
Configure the hosts outside of the local enclave to not log to this system.
Ask the SA if the system is an authorized syslog server. If the system is an authorized syslog server, this is not applicable. Determine if the system's syslog service is configured to accept remote messages. # ps -ef | grep syslogd If the '-r' option is present, the system is configured to accept remote syslog messages, and this is a finding.
Edit /etc/sysconfig/syslog to removing the '-r' in SYSLOGD_OPTIONS. Restart the syslogd service.
Check the TCP wrappers configuration files to determine if sshd is configured to use TCP wrappers. Procedure: # grep sshd /etc/hosts.deny # grep sshd /etc/hosts.allow If no entries are returned, the TCP wrappers are not configured for sshd, this is a finding.
Add appropriate IP restrictions for SSH to the /etc/hosts.deny and/or /etc/hosts.allow files.
Check if the system is configured for IPv4 forwarding. If the system is a VM host and acts as a router solely for the benefits of its client systems, then this rule is not applicable. Procedure: # cat /proc/sys/net/ipv4/ip_forward If the value is set to "1", IPv4 forwarding is enabled this is a finding.
Edit "/etc/sysctl.conf" and set net.ipv4.ip_forward to "0". Restart the system or run "sysctl -p" to make the change take effect.
If an IM client is installed, ask the SA if it has access to any public domain IM servers. If it does have access to public servers, this is a finding.
Uninstall the IM client from the system, or configure the client to only connect to DoD-approved IM services.
Ask the SA if any peer-to-peer file-sharing applications are installed. Some examples of these applications include: - Napster - Kazaa - ARES - Limewire - IRC Chat Relay - BitTorrent If any of these applications are installed, this is a finding.
Uninstall the peer-to-peer file sharing application(s) from the system.
Check the domain name for NIS maps. Procedure: # domainname If the name returned is simple to guess, such as the organization name, building or room name, etc., this is a finding. If the system does not use NIS, this is not applicable.
Change the NIS domainname to a value difficult to guess. Consult vendor documentation for the required procedure.
For each security tool on the system, determine if the tool is configured to notify the IAO and SA of any detected security problem. If such notifications are not configured, this is a finding.
Configure the security tools on the system to notify the IAO and SA when any security issues are detected.
Check for the existence of the "/etc/hosts.allow" and "/etc/hosts.deny" files. Procedure: # ls -la /etc/hosts.allow # ls -la /etc/hosts.deny If either file does not exist, this is a finding. Check for the presence of a "default deny" entry. Procedure: # grep "ALL: ALL" /etc/hosts.deny If the "ALL: ALL" entry is not present the "/etc/hosts.deny" file, any TCP service from a host or network not matching other rules will be allowed access. If the entry is not in "/etc/hosts.deny", this is a finding.
Edit the "/etc/hosts.all" and "/etc/hosts.deny" files to configure access restrictions.
Check /etc/securetty group ownership: # ls -lL /etc/securetty If /etc/securetty is not group owned by root, sys, or bin, then this is a finding.
Change the group-owner of /etc/securetty to root, sys, or bin. Example: # chgrp root /etc/securetty
Check /etc/securetty ownership. Procedure: # ls -lL /etc/securetty If /etc/securetty is not owned by root, this is a finding.
Change the owner of the /etc/securetty file to root. Procedure: # chown root /etc/securetty
Check /etc/securetty permissions. Procedure: # ls -lL /etc/securetty If /etc/securetty has a mode more permissive than 0600, this is a finding.
Change the mode of the /etc/securetty file to 0600. Procedure: # chmod 0600 /etc/securetty
Determine if any network analysis tools are installed. Procedure: # find / -name ethereal # find / -name wireshark # find / -name tshark # find / -name nc # find / -name tcpdump # find / -name snoop If any network analysis tools are found, this is a finding
Remove each network analysis tool binary from the system. Remove package items with a package manager, others remove the binary directly. Procedure: Find the binary file: # find / -name <Item to be removed> Find the package, if any, to which it belongs: # rpm -qf <binary file> Remove the package if it does not also include other software: # rpm -e <package name> or # yum remove <package name> If the item to be removed is not in a package, or the entire package cannot be removed because of other software it provides, remove the item's binary file. # rm <binary file>
Check for the existence of a cron job to execute a DoD-approved command-line scan tool daily. Other tools may be available but will have to be manually reviewed if they are installed. In addition, the definitions files should not be older than 7 days. Check if DoD-approved command-line scan tool is scheduled to run: # grep [scan tool] /var/spool/cron/* /etc/cron.d/* /etc/cron.daily/* /etc/cron.hourly/* /etc/cron.monthly/* /etc/cron.weekly/* If a virus scanner is not being run daily and an exception has not been documented with the IAO, this is a finding. Perform the following command to ensure the virus definition signature files are not older than 7 days. # cd <scan tool install directory> # ls -la *.dat If the virus definitions are older than 7 days, this is a finding.
Install a DoD-approved command-line virus scan tool, or an appropriate alternative. Ensure the virus signature definition files are no older than 7 days. Configure the system to run a virus scan on altered files dynamically or daily. If daily scans impede operations, justify, document, and obtain IAO approval for alternate scheduling.
Check to see if ntp is running in continuous mode. # ps -ax | grep ntp If the process is found, then check the ntp.conf file for the maxpoll option setting. # grep maxpoll /etc/ntp.conf If the option is set to 17 or is not set, this is a finding.
Enable the NTP daemon for continuous synchronization. # service ntpd start ; chkconfig ntpd on Edit the "/etc/ntp.conf" file and add or update an entry to define "maxpoll" to "10".
Check the root crontab (crontab -l) and the global crontabs in /etc/crontab, /etc/cron.d/*, or scripts in the /etc/cron.daily directory for the presence of an "ntpd -qg" job. If the "ntpd -qg" command is not invoked with at least two external NTP servers listed, this is a finding. Check the NTP daemon configuration for at least two external servers. # grep ^server /etc/ntp.conf | egrep -v '(127.127.1.0|127.127.1.1)' If less than two servers or external reference clocks (127.127.x.x other than 127.127.1.0 or 127.127.1.1) are listed, this is a finding.
If using "ntpd -qg", add additional NTP servers to the cron job running "ntpd -qg". If using the NTP daemon, add an additional "server" line to /etc/ntp.conf for each additional NTP server.
Check the root crontab (crontab -l) and the global crontabs in /etc/crontab, /etc/cron.d/*, or scripts in the /etc/cron.daily directory for the presence of an "ntpd -qg" job. If the "ntpd -qg" command is invoked with NTP servers outside of the enclave, this is a finding. Check the NTP daemon configuration for NTP servers. # grep ^server /etc/ntp.conf | grep -v 127.127.1.1 If an NTP server is listed outside of the enclave, this is a finding.
If using "ntpd -qg", remove NTP servers external to the enclave from the cron job running "ntpd -qg". If using the NTP daemon, remove the "server" line from /etc/ntp.conf for each NTP server external to the enclave.
Check the ownership of the NTP configuration file. # ls -l /etc/ntp.conf If the owner is not root, this is a finding.
Change the owner of the NTP configuration file. # chown root /etc/ntp.conf
Check the group ownership of the NTP configuration file. Procedure: # ls -lL /etc/ntp.conf If the group owner is not root, bin, or sys, this is a finding.
Change the group-owner of the NTP configuration file. Procedure: # chgrp root /etc/ntp.conf
Check the mode for the NTP configuration file is not more permissive than 0640. # ls -l /etc/ntp.conf If the mode is more permissive than 0640, this is a finding.
Change the mode of the NTP configuration file to 0640 or more restrictive. # chmod 0640 /etc/ntp.conf
Check the NTP configuration file has no extended ACL. # ls -l /etc/ntp.conf If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the NTP configuration file. # setfacl --remove-all /etc/ntp.conf
Check for a default maxlogins line in the /etc/security/limits.conf and /etc/security/limits.d/* files. Procedure: #grep maxlogins /etc/security/limits.conf /etc/security/limits.d/* The default maxlimits should be set to a max of 10 or a documented site defined number: * - maxlogins 10 If no such line exists, this is a finding.
Add a "maxlogins" line such as "* hard maxlogins 10" to /etc/security/limits.conf or a file in /etc/security/limits.d. The enforced maximum should be defined by site requirements and policy.
Check that pam_lastlog is used and not silent, or that the SSH daemon is configured to display last login information. # grep pam_lastlog /etc/pam.d/sshd If pam_lastlog is present, and does not have the "silent" option, this is not a finding. # grep -i PrintLastLog /etc/ssh/sshd_config If PrintLastLog is not present in the configuration, this is not a finding. This is the default setting. If PrintLastLog is present in the configuration and set to "yes" (case insensitive), this is not a finding. Otherwise, this is a finding.
Implement pam_lastlog, or enable PrintLastLog in the SSH daemon. To enable pam_lastlog, add a line such as "session required pam_lastlog.so" to /etc/pam.d/sshd. To enable PrintLastLog in the SSH daemon, remove any lines disabling this option from /etc/ssh/sshd_config.
Determine if a publicly-viewable pattern is displayed during a session lock. Some screensaver themes available but not included in the RHEL distribution use a snapshot of the current screen as a graphic. This theme does not qualify as a publicly-viewable pattern. If the session lock pattern is not publicly-viewable this is a finding.
Configure the system to display a publicly-viewable pattern during a session lock. This is done graphically by selecting a screensaver theme using gnome-screensaver-preferences command. Any of the themes distributed with RHEL may be used including "Blank Screen".
Verify no valid password hash in /etc/passwd or /etc/shadow begins with a character other than an underscore (_) or dollar sign ($). # cut -d ':' -f2 /etc/passwd # cut -d ':' -f2 /etc/shadow If any valid password hash is present that does not have an initial underscore (_) or dollar sign ($) character, this is a finding. Note: Locked accounts are indicated by a leading exclamation point (!). System accounts, other than "root", may have an asterisk (*) in the password field. On systems utilizing shadow passwords, the password field in /etc/passwd will be a single "x".
Change the passwords for all accounts using non-compliant password hashes. (This requires GEN000590 is already met.)
Verify the algorithm used for password hashing is of the SHA-2 family. # egrep "password .* pam_unix.so" /etc/pam.d/system-auth-ac # egrep "ENCRYPT_METHOD" /etc/login.defs # egrep "crypt_style" /etc/libuser.conf If any output indicates the hash algorithm is not set to sha256 or sha512, this is a finding.
Change the default password algorithm. # authconfig --passalgo=sha512 --update NOTE: Executing the above command will also update the required parameters in /etc/login.defs and /etc/libuser.conf
Check all password hashes in /etc/passwd or /etc/shadow begin with '$5$' or '$6$'. Procedure: # cut -d ':' -f2 /etc/passwd # cut -d ':' -f2 /etc/shadow Any password hashes present not beginning with '$5$' or, '$6$' is a finding. Any entries showing only NP, LK, or x are not findings.
Change the passwords for all accounts using non-compliant password hashes. (This requires GEN000590 is already met.)
Check /etc/pam.d/system-auth for lcredit setting. Procedure: Check the password lcredit option # grep pam_cracklib.so /etc/pam.d/system-auth Confirm the lcredit option is set to -1 as in the example: password required pam_cracklib.so lcredit=-1 There may be other options on the line. If no such line is found, or the lcredit is not -1 this is a finding.
Edit "/etc/pam.d/system-auth" to include the line: password required pam_cracklib.so lcredit=-1 prior to the "password include system-auth-ac" line.
Check /etc/pam.d/system-auth for a pam_cracklib parameter difok. Procedure: # grep difok /etc/pam.d/system-auth If difok is not present, or has a value less than 8, this is a finding. Check for system-auth-ac inclusions. # grep -c system-auth-ac /etc/pam.d/* If the system-auth-ac file is included anywhere # more /etc/pam.d/system-auth-ac | grep difok If system-auth-ac is included anywhere and difok is not present, or has a value less than 8, this is a finding. Ensure the passwd command uses the system-auth settings. # grep system-auth /etc/pam.d/passwd If a line "password include system-auth" is not found then the password checks in system-auth will not be applied to new passwords and this is a finding.
If /etc/pam.d/system-auth references /etc/pam.d/system-auth-ac refer to the man page for system-auth-ac for a description of how to add options not configurable with authconfig. Edit /etc/pam.d/system-auth and add or edit a pam_cracklib entry with an difok parameter set equal to or greater than 8.
Check /etc/pam.d/system-auth for pam_cracklib configuration. Procedure: # grep pam_cracklib /etc/pam.d/system-auth* If pam_cracklib is not present. This is a finding. If pam_cracklib is present only in /etc/pam.d/system-auth-ac: ensure that /etc/pam.d/system-auth includes /etc/pam.d/system-auth-ac. #grep system-auth-ac /etc/pam.d/system-auth This should return: auth include system-auth-ac account include system-auth-ac password include system-auth-ac session include system-auth-ac /etc/pam.d/system-auth-ac should only be included by /etc/pam.d/system-auth. All other pam files should include /etc/pam.d/system-auth. If pam_cracklib is not defined in /etc/pam.d/system-auth either directly or through inclusion of system-auth-ac, this is a finding. Ensure the passwd command uses the system-auth settings. # grep system-auth /etc/pam.d/passwd If a line "password include system-auth" is not found then the password checks in system-auth will not be applied to new passwords, this is a finding.
If /etc/pam.d/system-auth references /etc/pam.d/system-auth-ac refer to the man page for system-auth-ac for a description of how to add options not configurable with authconfig. Edit /etc/pam.d/system-auth and configure pam_cracklib by adding a line such as "password required pam_cracklib.so"
Check /etc/pam.d/su uses pam_wheel. # grep pam_wheel /etc/pam.d/su If pam_wheel is not present, or is commented out, this is a finding.
Edit /etc/pam.d/su and uncomment or add a line such as "auth required pam_wheel.so". If necessary, create a "wheel" group and add administrative users to the group.
Check the root account's home directory has no extended ACL. # find ~root -type d -prune -exec ls -ld {} \; If the permissions include a '+' the directory has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the root account's home directory. # setfacl --remove-all <root home directory>
Check the LD_LIBRARY_PATH environment variable is empty or not defined for the root user. # echo $LD_LIBRARY_PATH If a path list is returned, this is a finding.
Edit the root user initialization files and remove any definition of LD_LIBRARY_PATH.
Check the LD_PRELOAD environment variable is empty or not defined for the root user. # echo $LD_PRELOAD If a path list is returned, this is a finding.
Edit the root user initialization files and remove any definition of LD_PRELOAD.
Search the system for files without a valid group-owner. # find / -ignore_readdir_race -nogroup If any files are found, this is a finding.
Change the group-owner for each file without a valid group-owner. # chgrp avalidgroup /tmp/a-file-without-a-valid-group-owner
Check network services daemon files have no extended ACLs. # ls -la /usr/sbin If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding. Note: Network daemons not residing in these directories must also be checked.
Remove the extended ACL from the file. # setfacl --remove-all /usr/sbin/*
Check all system command files have no extended ACLs. # ls -lL /etc /bin /usr/bin /usr/lbin /usr/usb /sbin /usr/sbin If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all [file with extended ACL]
Verify system log files have no extended ACLs. Procedure: # ls -lL /var/log If the permissions include a '+', the file has an extended ACL. If an extended ACL exists, verify with the SA if the ACL is required to support authorized software and provides the minimum necessary permissions. If an extended ACL exists providing access beyond the needs of authorized software, this is a finding.
Remove the extended ACL from the file. Procedure: # setfacl --remove-all [file with extended ACL]
Verify all manual page files have no extended ACLs. # ls -lLR /usr/share/man /usr/share/info /usr/share/infopage If the permissions include a '+', the file has an extended ACL this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /usr/share/man/* /usr/share/info/* /usr/share/infopage/*
Verify system libraries have no extended ACLs. # ls -lLR /usr/lib/ /lib/ /usr/lib64 /lib64 | grep "+ " If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and has not been approved by the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /usr/lib/* /lib/*
Verify NIS/NIS+/yp files have no extended ACLs. # ls -lL /var/yp/* If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /var/yp/*
Verify the /etc/resolv.conf file is owned by root. # ls -l /etc/resolv.conf If the file is not owned by root, this is a finding.
Change the owner of the /etc/resolv.conf file to root. # chown root /etc/resolv.conf
Check the group ownership of the resolv.conf file. Procedure: # ls -lL /etc/resolv.conf If the file is not group-owned by root, bin, or sys, this is a finding.
Change the group-owner of the /etc/resolv.conf file to root, bin, or sys. Procedure: # chgrp root /etc/resolv.conf
Check the mode of the /etc/resolv.conf file. # ls -l /etc/resolv.conf If the file mode is not 0644, this is a finding.
Change the mode of the /etc/resolv.conf file to 0644. # chmod 0644 /etc/resolv.conf
Verify /etc/resolv.conf has no extended ACL. # ls -l /etc/resolv.conf If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/resolv.conf
Verify the /etc/hosts file is owned by root. # ls -l /etc/hosts If the file is not owned by root, this is a finding.
Change the owner of the /etc/hosts file to root. # chown root /etc/hosts
Check the /etc/hosts file's group ownership. Procedure: # ls -lL /etc/hosts If the file is not group-owned by root, bin, or sys, this is a finding.
Change the group-owner of the /etc/hosts file to root, sys, or bin. Procedure: # chgrp root /etc/hosts
Check the mode of the /etc/hosts file. # ls -l /etc/hosts If the file mode is not 0644, this is a finding.
Change the mode of the /etc/hosts file to 0644. # chmod 0644 /etc/hosts
Verify /etc/hosts has no extended ACL. # ls -l /etc/hosts If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/hosts
Verify the /etc/nsswitch.conf file is owned by root. # ls -l /etc/nsswitch.conf If the file is not owned by root, this is a finding.
Change the owner of the /etc/nsswitch.conf file to root. # chown root /etc/nsswitch.conf
Check the group ownership of the nsswitch.conf file. Procedure: # ls -lL /etc/nsswitch.conf If the file is not group-owned by root, bin or sys, this is a finding.
Change the group-owner of the /etc/nsswitch.conf file to root, bin or sys. Procedure: # chgrp root /etc/nsswitch.conf
Check the mode of the /etc/nsswitch.conf file. # ls -l /etc/nsswitch.conf If the file mode is not 0644, this is a finding.
Change the mode of the /etc/nsswitch.conf file to 0644 or less permissive. # chmod 0644 /etc/nsswitch.conf
Verify /etc/nsswitch.conf has no extended ACL. # ls -l /etc/nsswitch.conf If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/nsswitch.conf
Determine if DNS is enabled on the system. # grep dns /etc/nsswitch.conf If no line is returned, or any returned line is commented out, the system does not use DNS, and this is not applicable. Determine the name servers used by the system. # grep nameserver /etc/resolv.conf If less than two lines are returned that are not commented out, this is a finding.
Edit /etc/resolv.conf and add additional "nameserver" lines until at least two are present.
Verify the /etc/passwd file is owned by root. # ls -l /etc/passwd If the file is not owned by root, this is a finding.
Change the owner of the /etc/passwd file to root. # chown root /etc/passwd
Check the group ownership of the passwd file. Procedure: # ls -lL /etc/passwd If the file is not group-owned by root, bin or sys, this is a finding.
Change the group-owner of the /etc/passwd file to root, bin or sys. Procedure: # chgrp root /etc/passwd
Verify /etc/passwd has no extended ACL. # ls -l /etc/passwd If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/passwd
Verify the /etc/group file is owned by root. # ls -l /etc/group If the file is not owned by root, this is a finding.
Change the owner of the /etc/group file to root. # chown root /etc/group
Check the group ownership of the /etc/group file. Procedure: # ls -lL /etc/group If the file is not group-owned by root, bin or sys, this is a finding.
Change the group-owner of the /etc/group file. Procedure: # chgrp root /etc/group
Check the mode of the /etc/group file. # ls -l /etc/group If the file mode is more permissive than 0644, this is a finding.
Change the mode of the /etc/group file to 0644 or less permissive. # chmod 0644 /etc/group
Verify /etc/group has no extended ACL. # ls -l /etc/group If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/group
Check the ownership of the /etc/shadow file. Procedure: # ls -lL /etc/shadow If the file is not group-owned by root, bin, or sys, this is a finding.
Change the group-owner of the /etc/shadow file. Procedure: # chgrp root /etc/shadow
Verify /etc/shadow has no extended ACL. # ls -l /etc/shadow If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/shadow
Check the /etc/gshadow file is owned by root. # ls -l /etc/gshadow If the file is not owned by root, this is a finding.
Change the owner of the /etc/gshadow file to root. # chown root /etc/gshadow
Check the /etc/gshadow file is group-owned by root. # ls -l /etc/gshadow If the file is not group-owned by root, this is a finding.
Change the group-owner of the /etc/gshadow file to root. # chgrp root /etc/gshadow
Check the mode of the /etc/gshadow file. # ls -l /etc/gshadow If the file mode is more permissive than 0400, this is a finding.
Change the mode of the /etc/gshadow file to 0400 or less permissive. # chmod 0400 /etc/gshadow
Check /etc/gshadow has no extended ACL. # ls -l /etc/gshadow If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/gshadow
Verify no password hashes are present in /etc/passwd. # cut -d : -f 2 /etc/passwd | egrep -v '^(x|\*)$' If any password hashes are returned, this is a finding.
Migrate /etc/passwd password hashes to /etc/shadow. # pwconv
Check the /etc/group file for password hashes. # cut -d : -f 2 /etc/group | egrep -v '^(x|!)$' If any password hashes are returned, this is a finding.
Edit /etc/group and change the password field to an exclamation point (!) to lock the group password.
Check the /etc/gshadow file for password hashes. # cut -d : -f 2 /etc/gshadow | egrep -v '^(x|!!)$' If any password hashes are returned, this is a finding.
Edit /etc/gshadow and change the password field to an exclamation point (!) to lock the group password.
Verify user home directories have no extended ACLs. # cut -d : -f 6 /etc/passwd | xargs -n1 ls -ld If the permissions include a '+', the file has an extended ACL this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all [user home directory with extended ACL]
Check the contents of user home directories for files group-owned by a group of which the home directory's owner is not a member. 1. List the user accounts. # cut -d : -f 1 /etc/passwd 2. For each user account, get a list of GIDs for files in the user's home directory. # find ~username -printf %G\\n | sort | uniq 3. Obtain the list of GIDs where the user is a member. # id -G username 4. Check the GID lists. If there are GIDs in the file list not present in the user list, this is a finding.
Change the group of a file not group-owned by a group of which the home directory's owner is a member. # chgrp <group with user as member> <file with bad group ownership> Document all changes.
Check the contents of user home directories for files with extended ACLs. # cut -d : -f 6 /etc/passwd | xargs -n1 -IDIR ls -alLR DIR If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all <user file with extended ACL>
Verify run control scripts have no extended ACLs. # ls -lL /etc/rc* /etc/init.d If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all <run control script with extended ACL>
Verify run control scripts' library search paths. # grep -r LD_LIBRARY_PATH /etc/rc* /etc/init.d This variable is formatted as a colon-separated list of directories. If there is an empty entry, such as a leading or trailing colon, or two consecutive colons, this is a finding. If an entry begins with a character other than a slash (/), or has not been documented with the ISSO, this is a finding.
Edit the run control script and remove the relative path entries from the library search path variables that are not documented with the ISSO. Edit the run control script and remove any empty entry that is defined.
Verify run control scripts' library preload list. # grep -r LD_PRELOAD /etc/rc* /etc/init.d This variable is formatted as a colon-separated list of paths. If there is an empty entry, such as a leading or trailing colon, or two consecutive colons, this is a finding. If an entry begins with a character other than a slash (/), or has not been documented with the ISSO, this is a finding.
Edit the run control script and remove the relative path entries from the library preload variables that are not documented with the ISSO. Edit the run control script and remove any empty entry that is defined.
Check global initialization files for extended ACLs: # ls -l /etc/bashrc /etc/csh.cshrc /etc/csh.login /etc/csh.logout /etc/environment /etc/ksh.kshrc /etc/profile /etc/suid_profile /etc/profile.d/* 2>null|grep "\+ " If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # ls -l etc/bashrc /etc/csh.cshrc /etc/csh.login /etc/csh.logout /etc/environment /etc/ksh.kshrc /etc/profile /etc/suid_profile /etc/profile.d/* 2>null|grep "\+ "|sed "s/^.* \///g"|xargs setfacl --remove-all
Check skeleton files for extended ACLs: # ls -alL /etc/skel If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all [skeleton file with extended ACL] or: # ls -lL /etc/skel|grep "\+ "|sed "s/^.* \//|xargs setfacl --remove-all will remove all ACLs from the files.
Verify the skeleton files are group-owned by root. Procedure: # ls -alL /etc/skel If a skeleton file is not group-owned by root, bin, sys, system, or other this is a finding.
Change the group-owner of the skeleton file to root, bin, sys, system, or other. Procedure: # chgrp <group> /etc/skel/[skeleton file] or: # ls -L /etc/skel|xargs stat -L -c %G:%n|egrep -v "^(root|bin|sy|sytem|other):"|cut -d: -f2|chgrp root will change the group of all files not already one of the approved group to root.
Check the global initialization files' library search paths. Procedure: # grep LD_LIBRARY_PATH /etc/bashrc /etc/csh.cshrc /etc/csh.login /etc/csh.logout /etc/environment /etc/ksh.kshrc /etc/profile /etc/suid_profile /etc/profile.d/* This variable is formatted as a colon-separated list of directories. If there is an empty entry, such as a leading or trailing colon, or two consecutive colons, this is a finding. If an entry begins with a character other than a slash (/), or has not been documented with the ISSO, this is a finding.
Edit the global initialization file and remove the relative path entries from the library search path variables that have not been documented with the ISSO. Edit the global initialization file(s) and remove any empty entry that is defined.
Check the global initialization files' library preload list. # grep -r LD_PRELOAD /etc/bashrc /etc/csh.cshrc /etc/csh.login /etc/csh.logout /etc/environment /etc/ksh.kshrc /etc/profile /etc/suid_profile /etc/profile.d/* This variable is formatted as a colon-separated list of paths. If there is an empty entry, such as a leading or trailing colon, or two consecutive colons, this is a finding. If an entry begins with a character other than a slash (/), or has not been documented with the ISSO, this is a finding.
Edit the global initialization file and remove the relative path entry from the library preload variable that has not been authorized by the ISSO. Edit the global initialization file(s) and remove any empty entry that is defined.
Check user home directories for local initialization files group-owned by a group other than the user's primary group or root. Procedure: # FILES=" .login .cshrc .logout .profile .bash_profile .bashrc .bash_logout .env .dtprofile .dispatch .emacs .exrc"; # for PWLINE in `cut -d: -f4,6 /etc/passwd`; do HOMEDIR=$(echo ${PWLINE}|cut -d: -f2);GROUP=$(echo ${PWLINE} | cut -d: -f1);for INIFILE in $FILES;do stat -c %g/%G:%n ${HOMEDIR}/${INIFILE} 2>null|egrep -v "${GROUP}";done;done If any file is not group-owned by root or the user's primary GID, this is a finding.
Change the group-owner of the local initialization file to the user's primary group, or root. # chgrp <user's primary GID> <user's local initialization file> Procedure: # FILES=".bashrc .bash_login .bash_logout .bash_profile .cshrc .kshrc .login .logout .profile .tcshrc .env .dtprofile .dispatch .emacs .exrc"; # for PWLINE in `cut -d: -f4,6 /etc/passwd`; do HOMEDIR=$(echo ${PWLINE}|cut -d: -f2);GROUP=$(echo ${PWLINE} | cut -d: -f1);for INIFILE in $FILES;do MATCH=$(stat -c %g/%G:%n ${HOMEDIR}/${INIFILE} 2>null|egrep -c -v "${GROUP}");if [ $MATCH != 0 ] ; then chgrp ${GROUP} ${HOMEDIR}/${INIFILE};fi;done;done
Check user home directories for local initialization files with extended ACLs. # cut -d : -f 6 /etc/passwd | xargs -n1 -IDIR ls -alL DIR/.bashrc DIR/.bash_login DIR/.bash_logout DIR/.bash_profile DIR/.cshrc DIR/.kshrc DIR/.login DIR/.logout DIR/.profile DIR/.env DIR/.dtprofile DIR/.dispatch DIR/.emacs DIR/.exrc If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all <local initialization file with extended ACL>
Verify local initialization files have library search paths containing only authorized paths. Procedure: NOTE: This must be done in the BASH shell. # cut -d: -f6 /etc/passwd |xargs -n1 -IDIR find DIR -name ".*" -type f -maxdepth 1 -exec grep -H LD_LIBRARY_PATH {} \; This variable is formatted as a colon-separated list of directories. If there is an empty entry, such as a leading or trailing colon, or two consecutive colons, this is a finding. If an entry begins with a character other than a slash (/), or has not been documented with the ISSO, this is a finding.
Edit the local initialization file and remove any relative path entries that have not been documented with the ISSO. Edit the local initialization file and remove any empty entry that is defined.
Verify local initialization files have a library preload list containing only authorized paths. NOTE: The following must be done in the BASH shell. Procedure: # cut -d: -f6 /etc/passwd |xargs -n1 -IDIR find DIR -name ".*" -type f -maxdepth 1 -exec grep -H LD_PRELOAD {} \; This variable is formatted as a colon-separated list of paths. If there is an empty entry, such as a leading or trailing colon, or two consecutive colons, this is a finding. If an entry begins with a character other than a slash (/), or has not been documented with the ISSO, this is a finding.
Edit the local initialization file and remove any relative path entry from the library preload variable that has not been authorized by the ISSO. Edit the local initialization file and remove any empty entry that is defined.
If /etc/shells exists, check the group ownership of each shell referenced. Procedure: # cat /etc/shells | xargs -n1 ls -l Otherwise, check any shells found on the system. Procedure: # find / -name "*sh" | xargs -n1 ls -l If a shell is not group-owned by root, bin, sys, or system, this is a finding.
Change the group-owner of the shell to root, bin, sys, or system. Procedure: # chgrp root <shell>
If /etc/shells exists, check the permissions of each shell referenced. # cat /etc/shells | xargs -n1 ls -lL Otherwise, check any shells found on the system. # find / -name "*sh" | xargs -n1 ls -lL If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all [shell]
Check the permissions of audio devices. # ls -lL /dev/audio* /dev/snd/* If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all [device file]
Check /etc/mtab and verify the "nodev" mount option is used on any filesystems mounted from removable media or network shares. If any filesystem mounted from removable media or network shares does not have this option, this is a finding.
Edit /etc/fstab and add the "nodev" option to any filesystems mounted from removable media or network shares.
Check the system audit log files for extended ACLs. Procedure: # grep "^log_file" /etc/audit/auditd.conf|sed s/^[^\/]*//|xargs ls -l If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the system audit file(s).
Verify the audit tool executables are owned by root. # ls -l /sbin/auditctl /sbin/auditd /sbin/ausearch /sbin/aureport /sbin/autrace /sbin/audispd If any listed file is not owned by root, this is a finding.
Change the owner of the audit tool executable to root. # chown root [audit tool executable]
Verify the audit tool executables are group-owned by root, bin, sys, or system. Procedure: # ls -lL /sbin/auditctl /sbin/auditd /sbin/ausearch /sbin/aureport /sbin/autrace /sbin/audispd If any listed file is not group-owned by root, bin, sys, or system, this is a finding.
Change the group-owner of the audit tool executable to root, bin, sys, or system. Procedure: # chgrp root <audit tool executable>
Check the mode of audit tool executables. # ls -l /sbin/auditctl /sbin/auditd /sbin/ausearch /sbin/aureport /sbin/autrace /sbin/audispd If any listed file has a mode more permissive than 0750, this is a finding.
Change the mode of the audit tool executable to 0750, or less permissive. # chmod 0750 [audit tool executable]
Check the permissions of audit tool executables. # ls -l /sbin/auditctl /sbin/auditd /sbin/ausearch /sbin/aureport /sbin/autrace /sbin/audispd If the permissions include a '+' the file has an extended ACL, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all [audit file]
Verify the /etc/audit/auditd.conf has the disk_full_action and disk_error_action parameters set. Procedure: # grep disk_full_action /etc/audit/auditd.conf If the disk_full_action parameter is missing or set to "suspend" or "ignore" this is a finding. # grep disk_error_action /etc/audit/auditd.conf If the disk_error_action parameter is missing or set to "suspend" or "ignore" this is a finding.
Edit /etc/audit/auditd.conf and set the disk_full_action and/or disk_error_action parameters to a valid setting of "syslog", "exec", "single" or "halt", adding the parameters if necessary.
Check /etc/audit/auditd.conf for the space_left_action and action_mail_accnt parameters. If the space_left_action or the action_mail_accnt parameters are set to blanks, this is a finding. If the space_left_action is set to "syslog", the system logs the event, this is not a finding. If the space_left_action is set to "exec", the system executes a designated script. If this script informs the SA of the event, this is not a finding. If the space_left_action parameter is missing, this is a finding. If the space_left_action parameter is set to "ignore" or "suspend" no logging would be performed after the event, this is a finding. If the space_left_action parameter is set to "single" or "halt" this effectively stops the system causing a Denial of Service, this is a finding. If the space_left_action is set to "email" and the action_mail_acct parameter is not set to the e-mail address of the system administrator, this is a finding. The action_mail_acct parameter, if missing, defaults to "root". Note that if the email address of the system administrator is on a remote system "sendmail" must be available.
Edit /etc/audit/auditd.conf and set the space_left_action parameter to a valid setting other than "ignore". If the space_left_action parameter is set to "email" set the action_mail_acct parameter to an e-mail address for the system administrator.
Determine if execution of the useradd and groupadd executable are audited. # auditctl -l | egrep '(useradd|groupadd)' If either useradd or groupadd are not listed with a permissions filter of at least 'x', this is a finding. Determine if /etc/passwd, /etc/shadow, /etc/group, and /etc/gshadow are audited for appending. # auditctl -l | egrep '(/etc/passwd|/etc/shadow|/etc/group|/etc/gshadow)' If any of these are not listed with a permissions filter of at least 'a', this is a finding.
Configure execute auditing of the useradd and groupadd executables. Add the following to audit.rules: -w /usr/sbin/useradd -p x -k useradd -w /usr/sbin/groupadd -p x -k groupadd Configure append auditing of the passwd, shadow, group, and gshadow files. Add the following to audit.rules: -w /etc/passwd -p a -k passwd -w /etc/shadow -p a -k shadow -w /etc/group -p a -k group -w /etc/gshadow -p a -k gshadow Restart the auditd service.
Determine if execution of the usermod and groupmod executable are audited. # auditctl -l | egrep '(usermod|groupmod)' If either useradd or groupadd are not listed with a permissions filter of at least 'w', this is a finding. Determine if /etc/passwd, /etc/shadow, /etc/group, and /etc/gshadow are audited for writing. # auditctl -l | egrep '(/etc/passwd|/etc/shadow|/etc/group|/etc/gshadow)' If any of these are not listed with a permissions filter of at least 'w', this is a finding.
Configure execute auditing of the usermod and groupmod executables. Add the following to the audit.rules file: -w /usr/sbin/usermod -p x -k usermod -w /usr/sbin/groupmod -p x -k groupmod Configure append auditing of the passwd, shadow, group, and gshadow files. Add the following to the audit.rules file: -w /etc/passwd -p w -k passwd -w /etc/shadow -p w -k shadow -w /etc/group -p w -k group -w /etc/gshadow -p w -k gshadow Restart the auditd service.
Determine if execution of the passwd executable is audited. # auditctl -l | grep /usr/bin/passwd If passwd is not listed with a permissions filter of at least 'x', this is a finding.
Configure execute auditing of the passwd executable. Add the following to the audit.rules file: -w /usr/bin/passwd -p x -k passwd Restart the auditd service.
Determine if execution of the userdel and groupdel executable are audited. # auditctl -l | egrep '(userdel|groupdel)' If either userdel or groupdel are not listed with a permissions filter of at least 'x', this is a finding.
Configure execute auditing of the userdel and groupdel executables. Add the following to the audit.rules file: -w /usr/sbin/userdel -p x -w /usr/sbin/groupdel -p x Restart the auditd service.
Determine if the init_module syscall is audited. # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "init_module" If the result does not contain "-S init_module", this is a finding.
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Configure auditing of the init_module syscalls. Add the following to the "etc/audit/audit.rules" or "etc/audit.rules" file: -a exit,always -S init_module Restart the auditd service. # service auditd restart
Check the permissions of the cron.allow file. # ls -l /etc/cron.allow If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/cron.allow
Check the group ownership of the crontab files. Procedure: # ls -lL /var/spool/cron # ls -lL /etc/cron.d /etc/crontab /etc/cron.daily /etc/cron.hourly /etc/cron.monthly /etc/cron.weekly or # ls -lL /etc/cron*|grep -v deny If the group owner is not root or the crontab owner's primary group, this is a finding.
Change the group owner of the crontab file to root, cron, or the crontab's primary group. Procedure: # chgrp root [crontab file]
Check the permissions of the crontab files. Procedure: # ls -lL /var/spool/cron # ls -lL /etc/cron.d /etc/crontab /etc/cron.daily /etc/cron.hourly /etc/cron.monthly /etc/cron.weekly or # ls -lL /etc/cron*|grep -v deny If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all [crontab file]
Check the permissions of the crontab directories. Procedure: # ls -ld /var/spool/cron # ls -ld /etc/cron.d /etc/crontab /etc/cron.daily /etc/cron.hourly /etc/cron.monthly /etc/cron.weekly or # ls -ld /etc/cron*|grep -v deny If the permissions include a '+' the directory has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the directory. # setfacl --remove-all <crontab directory>
Check the permissions of the file. Procedure: Depending on what system is used for log processing either /etc/syslog.conf or /etc/rsyslog.conf will be the logging configuration file. Check the configured cron log file found in the cron entry in /etc/syslog.conf or /etc/rsyslog.conf(normally /var/log/cron). # grep cron /etc/syslog.conf Or: # grep cron /etc/rsyslog.conf # ls -lL /var/log/cron If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /var/log/cron
Check the permissions of the file. # ls -lL /etc/cron.deny If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/cron.deny
Check the permissions of the file. # ls -lL /etc/at.allow If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/at.allow
Check the group ownership of the file. Procedure: # ls -lL /etc/cron.allow If the file exists and is not group-owned by root, bin, sys or cron, this is a finding.
Change the group ownership of the file. Procedure: # chgrp root /etc/cron.allow
Check the permissions of the file. # ls -lL /etc/at.deny If the file has a mode more permissive than 0600, this is a finding.
Change the mode of the file. # chmod 0600 /etc/at.deny
Check the permissions of the file. # ls -lL /etc/at.deny If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/at.deny
Check the group ownership of the file. Procedure: # ls -lL /etc/cron.deny If the file is not group-owned by root, bin, or sys, this is a finding.
Change the group ownership of the file. # chgrp root /etc/cron.deny
Check the permissions of the directory. # ls -lLd /var/spool/at If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the directory. # setfacl --remove-all /var/spool/at
Check the group ownership of the file. Procedure: # ls -lL /var/spool/at If the file is not group-owned by root, bin, sys, daemon or cron, this is a finding.
Change the group ownership of the file to root, bin, sys, daemon or cron. Procedure: # chgrp <root or other system group> <"at" directory>
Check the group ownership of the file. Procedure: # ls -lL /etc/at.allow If the file is not group-owned by root, bin, sys, or cron, this is a finding.
Change the group ownership of the file. Procedure: # chgrp root /etc/at.allow
Check the group ownership of the file. Procedure: # ls -lL /etc/at.deny If the file is not group-owned by root, bin, sys, or cron, this is a finding.
Change the group ownership of the at.deny file to root, sys, bin, or cron. Procedure: # chgrp root /etc/at.deny
Verify the kdump service is not running. Procedure: # service kdump status If "Kdump is operational" is returned, this is a finding.
Disable kdump. # service kdump stop # chkconfig kdump off
Determine the kernel core dump data directory and check its ownership. Procedure: Examine /etc/kdump.conf. The "path" parameter, which defaults to /var/crash, determines the path relative to the crash dump device. The crash device is specified with a filesystem type and device, such as "ext3 /dev/sda2". Using this information, determine where this path is currently mounted on the system. # ls -ld <kernel dump data directory> If the directory is not group-owned by root, bin, sys, or system, this is a finding.
Change the group-owner of the kernel core dump data directory. # chgrp root <kernel core dump data directory>
Verify the location of the kernel core dump data directory: # grep “path” /etc/kdump.conf Note: If this setting is not configured or commented out, the default kernel dump data directory is /var/crash. Check the permissions of the dump data directory: # ls -ld <path to kernel core dump data directory> If the directory has a mode more permissive than 0700, this is a finding.
Set the permissions on the kernel core dump data directory to 0700. # chmod 0700 <kernel core dump data directory>
Determine the kernel core dump data directory and check its permissions. Procedure: Verify the location of the kernel core dump data directory: # grep “path” /etc/kdump.conf Note: If this setting is not configured or commented out, the default kernel dump data directory is /var/crash. Check the permissions of the dump data directory: # ls -ld <path to kernel core dump data directory> If the permissions include a '+', the directory has an extended ACL. If the directory has an extended ACL and it has not been documented with the ISSO, this is a finding.
Remove the extended ACL from the directory. # setfacl --remove-all <path to kernel core dump data directory>
Check the system for user-controlled network interfaces. # grep -l '^USERCTL=yes' /etc/sysconfig/network-scripts/ifcfg* If any results are returned, this is a finding.
Edit the configuration for the user-controlled interface and remove the "USERCTL=yes" configuration line or set to "USERCTL=no".
Verify the system does not respond to ICMP TIMESTAMP-REQUESTs Procedure: # grep "timestamp" /etc/sysconfig/iptables This should return entries for "timestamp-reply" and "timestamp-request". Both should end with "-j DROP'. If either does not exist or does not "DROP" the message, this is a finding.
Configure the system to not respond to ICMP TIMESTAMP-REQUESTs. This is done by rejecting ICMP type 13 and 14 messages at the firewall. Procedure: Edit /etc/sysconfig/iptables to add: -A RH-Firewall-1-INPUT -p ICMP --icmp-type timestamp-request -j DROP -A RH-Firewall-1-INPUT -p ICMP --icmp-type timestamp-reply -j DROP Restart the firewall: # service iptables restart
Verify the system does not respond to ICMP ECHO_REQUESTs set to broadcast addresses. Procedure: # cat /proc/sys/net/ipv4/icmp_echo_ignore_broadcasts If the result is not 1, this is a finding.
Configure the system to not respond to ICMP ECHO_REQUESTs sent to broadcast addresses. Edit /etc/sysctl.conf and add a setting for "net.ipv4.icmp_echo_ignore_broadcasts=1" and reload the sysctls. Procedure: # echo "net.ipv4.icmp_echo_ignore_broadcasts=1" >> /etc/sysctl.conf # sysctl -p
Verify the system does not respond to ICMP TIMESTAMP_REQUESTs set to broadcast addresses. Procedure: # cat /proc/sys/net/ipv4/icmp_echo_ignore_broadcasts If the result is not 1, this is a finding. Note: The same parameter controls both ICMP ECHO_REQUESTs and TIMESTAMP_REQUESTs.
Configure the system to not respond to ICMP TIMESTAMP_REQUESTs sent to broadcast addresses. Edit /etc/sysctl.conf and add a setting for "net.ipv4.icmp_echo_ignore_broadcasts=1" and reload the sysctls. Procedure: # echo "net.ipv4.icmp_echo_ignore_broadcasts=1" >> /etc/sysctl.conf # sysctl -p
Verify the system does not accept source-routed IPv4 packets. Procedure: # grep [01] /proc/sys/net/ipv4/conf/*/accept_source_route|egrep "default|all" If all of the resulting lines do not end with "0", this is a finding.
Configure the system to not accept source-routed IPv4 packets. Edit /etc/sysctl.conf and add a setting for "net.ipv4.conf.all.accept_source_route=0" and "net.ipv4.conf.default.accept_source_route=0". Reload the sysctls. Procedure: # sysctl -p
Verify the system does not use proxy ARP. # grep [01] /proc/sys/net/ipv4/conf/*/proxy_arp|egrep "default|all" If all of the resulting lines do not end with "0", this is a finding.
Configure the system to not use proxy ARP. Edit /etc/sysctl.conf and add a setting for "net.ipv4.conf.all.proxy_arp=0" and "net.ipv4.conf.default.proxy_arp=0". # sysctl -p
Verify the system does not accept IPv4 ICMP redirect messages. # grep [01] /proc/sys/net/ipv4/conf/*/accept_redirects|egrep "default|all" If all of the resulting lines do not end with "0", this is a finding.
Configure the system to not accept IPv4 ICMP redirect messages. Edit /etc/sysctl.conf and add a setting for "net.ipv4.conf.all.accept_redirects=0" and "net.ipv4.conf.default.accept_redirects=0". # sysctl -p
Verify the system does not send IPv4 ICMP redirect messages. # grep [01] /proc/sys/net/ipv4/conf/*/send_redirects|egrep "default|all" If all of the resulting lines do not end with "0", this is a finding.
Configure the system to not send IPv4 ICMP redirect messages. Edit /etc/sysctl.conf and add a setting for "net.ipv4.conf.all.send_redirects=0" and "net.ipv4.conf.default.send_redirects=0". # sysctl -p
Verify the system logs martian packets. # grep [01] /proc/sys/net/ipv4/conf/*/log_martians|egrep "default|all" If all of the resulting lines do not end with "1", this is a finding.
Configure the system to log martian packets. Edit /etc/sysctl.conf and add a setting for "net.ipv4.conf.all.log_martians=1" and "net.ipv4.conf.default.log_martians=1". Reload the sysctls. Procedure: # sysctl -p
Verify the system configured to use TCP syncookies when experiencing a TCP SYN flood. # cat /proc/sys/net/ipv4/tcp_syncookies If the result is not "1", this is a finding.
Configure the system to use TCP syncookies when experiencing a TCP SYN flood. Edit /etc/sysctl.conf and add a setting for "net.ipv4.tcp_syncookies=1". # sysctl -p
Verify the system is not configured for bridging. # ls /proc/sys/net/bridge If the directory exists, this is a finding. # lsmod | grep '^bridge ' If any results are returned, this is a finding. Fix Text: Configure the system to not use bridging.
Configure the system to not use bridging. # rmmod bridge Edit /etc/modprobe.conf and add a line such as "install bridge /bin/false" to prevent the loading of the bridge module.
Verify local filesystems use journaling. # mount | grep '^/dev/' | egrep -v 'type (ext3|ext4|jfs|reiserfs|xfs|iso9660|udf)' If a mount is listed, this is a finding.
Convert local file systems to use journaling or another mechanism ensuring file system consistency.
Check the group ownership of the xinetd configuration files and directories. Procedure: # ls -alL /etc/xinetd.conf /etc/xinetd.d If a file or directory is not group-owned by root, bin, sys, or system, this is a finding.
Change the group-owner of the xinetd configuration files and directories. Procedure: # chgrp -R root /etc/xinetd.conf /etc/xinetd.d
Check the permissions of the xinetd configuration files. Procedure: # ls -alL /etc/xinetd.conf If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/xinetd.conf
Check the permissions of the xinetd configuration directories. # ls -dlL /etc/xinetd.d If the mode of the directory is more permissive than 0755, this is a finding.
Change the mode of the directory. # chmod 0755 /etc/xinetd.d
Check the permissions of the xinetd configuration files and directories. # ls -alL /etc/xinetd.conf /etc/xinetd.d If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/xinetd.d
Check the group ownership of the services file. Procedure: # ls -lL /etc/services If the file is not group-owned by root or bin, this is a finding.
Change the group-owner of the services file. Procedure: # chgrp root /etc/services
Check the permissions of the /etc/services file. # ls -lL /etc/services If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/services
Check the status of the portmap service. # service portmap status If the service is running, this is a finding.
Shutdown and disable the portmap service. # service portmap stop; chkconfig portmap off
Check if the portmap package is installed. # rpm -qa | grep portmap If a package is found, this is a finding.
Remove the portmap package. # rpm -e portmap or # yum remove portmap
Check if the rsh-server package is installed. Procedure: # rpm -qa | grep rsh-server If a package is found, this is a finding.
Remove the rsh-server package. Procedure: # rpm -e rsh-server
Check the rlogind configuration. # cat /etc/xinetd.d/rlogin If the file exists and does not contain "disable = yes" this is a finding.
Remove or disable the rlogin configuration and restart xinetd. # rm /etc/xinetd.d/rlogin ; service xinetd restart
Check if the rsh-server package is installed. Procedure: # rpm -qa | grep rsh-server If a package is found, this is a finding.
Remove the rsh-server package. Procedure: # rpm -e rsh-server
Check if the rsh-server package is installed. Procedure: # rpm -qa | grep rsh-server If a package is found, this is a finding.
Remove the rsh-server package. Procedure: # rpm -e rsh-server
Check the group ownership of the /etc/cups/printers.conf file. Procedure: # ls -lL /etc/cups/printers.conf If the file is not group-owned by lp, this is a finding.
Change the group-owner of the printers.conf file. Procedure: # chgrp lp /etc/cups/printers.conf
Check the permissions of the /etc/cups/printers.conf file. # ls -lL /etc/cups/printers.conf If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/cups/printers.conf
Check the permissions of the /bin/traceroute file. # ls -lL /bin/traceroute If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /bin/traceroute
If the "sendmail" and "postfix" packages are not installed, this is not applicable. Check the group ownership of the alias files. Procedure: for sendmail: # ls -lL /etc/aliases If the files are not group-owned by root, this is a finding. # ls -lL /etc/aliases.db If the file is not group-owned by the same system group as sendmail, which is smmsp by default, this is a finding. for postfix: Verify the location of the alias file. # postconf alias maps This will return the location of the "aliases" file, by default "/etc/postfix/aliases" # ls -lL <postfix aliases file> If the files are not group-owned by root, this is a finding. # ls -lL <postfix aliases.db file> If the file is not group-owned by root, this is a finding.
Change the group-owner of the /etc/aliases file. Procedure: for sendmail: # chgrp root /etc/aliases # chgrp smmsp /etc/aliases.db The aliases.db file must be owned by the same system group as sendmail, which is smmsp by default. for postfix # chgrp root /etc/postfix/aliases # chgrp root /etc/postfix/aliases.db
If the "sendmail" and "postfix" packages are not installed, this is not applicable. Check the permissions of the alias file. Procedure: for sendmail: # ls -lL /etc/aliases /etc/aliases.db If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding. for postfix: Verify the location of the alias file. # postconf alias maps This will return the location of the "aliases" file, by default "/etc/postfix/aliases" # ls -lL <postfix aliases file> <postfix aliases.db file> If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended permissions from the alias files. Procedure: for sendmail: # setfacl --remove-all /etc/aliases /etc/aliases.db for postfix (assuming the default postfix directory): # setfacl --remove-all /etc/postfix/aliases /etc/postfix/aliases.db
Examine the contents of the /etc/aliases file. Procedure: # more /etc/aliases Examine the aliases file for any utilized directories or paths. # ls -lL <file referenced from aliases> Check the permissions for any paths referenced. If the group owner of any file is not root, bin, sys, or system, this is a finding.
Change the group ownership of the file referenced from /etc/aliases. Procedure: # chgrp root <file referenced from aliases>
Examine the contents of the /etc/aliases file. Procedure: # more /etc/aliases Examine the aliases file for any utilized directories or paths. # ls -lL <file referenced from aliases> Check the permissions for any paths referenced. If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all <file referenced from aliases>
Depending on what system is used for log processing, either /etc/syslog.conf or /etc/rsyslog.conf will be the logging configuration file. Examine /etc/syslog.conf or /etc/rsyslog.conf and determine the log file(s) receiving logs for "mail.crit", "mail.debug", mail.*, or "*.crit". Procedure: This check is applicable to both Postfix and sendmail servers. Check the permissions on these log files.Identify any log files configured for "*.crit" and the "mail" service (excluding mail.none) and at any severity level. For syslog: # egrep "(\*.crit|mail\.[^n][^/]*)" /etc/syslog.conf|sed 's/^[^/]*//'|xargs ls -lL For rsyslog: # egrep "(\*.crit|mail\.[^n][^/]*)" /etc/rsyslog.conf|sed 's/^[^/]*//'|xargs ls -lL If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the ISSO, this is a finding.
This fix is applicable to both Postfix and sendmail servers. Remove the extended ACL from the file. # setfacl --remove-all <log file>
Check the group ownership of the ftpusers file. Procedure: # ls -lL /etc/ftpusers /etc/vsftpd.ftpusers /etc/vsftpd/ftpusers If the file is not group-owned by root, bin, sys, or system, this is a finding.
Change the group owner of the ftpusers file. Procedure: # chgrp root /etc/ftpusers /etc/vsftpd.ftpusers /etc/vsftpd/ftpusers
Check the permissions of the /etc/ftpusers file. # ls -lL /etc/ftpusers /etc/vsftpd.ftpusers /etc/vsftpd/ftpusers If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/ftpusers /etc/vsftpd.ftpusers /etc/vsftpd/ftpusers
Check the file permissions for the .Xauthority files. These files will be located in user home directories. Procedure: # ls -la ~username |egrep "(\.Xauthority|\.xauth)" If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all .Xauthority
Check the SNMP daemon is not configured to use the v1 or v2c security models. Procedure: Examine the default install location /etc/snmpd.conf or: # find / -name snmpd.conf # grep -E '(v1|v2c|community|com2sec)' <snmp.conf file> | grep -v '^#' If any configuration is found, this is a finding.
Edit /etc/snmpd.conf and remove references to the "v1", "v2c", "community", or "com2sec". Restart the SNMP service. # service snmpd restart
Verify the SNMP daemon uses SHA for SNMPv3 users. Procedure: Examine the default install location /etc/snmp/snmpd.conf or: # find / -name snmpd.conf # grep -v '^#' <snmpd.conf file> | grep -i createuser | grep -vi SHA If any line is present this is a finding.
Edit /etc/snmp/snmpd.conf and add the SHA keyword for any create user statement without one. Restart the SNMP service. # service snmpd restart
Verify the SNMP daemon uses AES for SNMPv3 users. Procedure: Examine the default install location /etc/snmp/snmpd.conf or: # find / -name snmpd.conf # grep -v '^#' <snmpd.conf file> | grep -i createuser | grep -vi AES If any line is present this is a finding.
Edit /etc/snmp/snmpd.conf and add the AES keyword for any create user statement without one. Restart the SNMP service. # service snmpd restart
Check the file permissions for the MIB files. # find / -name *.mib # ls -lL <mib file> If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all <mib file>
Check the group ownership of the SNMP configuration file. Procedure: Examine the default install location /etc/snmp/snmpd.conf or: # find / -name snmpd.conf # ls -lL <snmpd.conf> If the file is not group-owned by root, bin, sys, or system, this is a finding.
Change the group ownership of the SNMP configuration file. Procedure: # chgrp root <snmpd.conf>
Check the permissions of the SNMP configuration file. Procedure: Examine the default install location /etc/snmp/snmpd.conf or: # find / -name snmpd.conf # ls -lL <snmpd.conf> If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all <snmpd.conf file>
Check the permissions of the syslog configuration file. Depending on what system is used for log processing either /etc/syslog.conf or /etc/rsyslog.conf will be the logging configuration file. # ls -lL /etc/syslog.conf Or: # ls -lL /etc/rsyslog.conf If the mode of the file is more permissive than 0640, this is a finding.
Change the permissions of the syslog or rsyslog configuration file. # chmod 0640 /etc/syslog.conf Or: # chmod 0640 /etc/rsyslog.conf
Check the permissions of the syslog configuration file. Depending on what system is used for log processing either /etc/syslog.conf or /etc/rsyslog.conf will be the logging configuration file. # ls -lL /etc/syslog.conf Or: # ls -lL /etc/rsyslog.conf If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/syslog.conf Or: # setfacl -- remove-all /etc/rsyslog.conf
Check the syslog configuration file for remote syslog servers. Depending on what system is used for log processing either /etc/syslog.conf or /etc/rsyslog.conf will be the logging configuration file. # grep '@' /etc/syslog.conf | grep -v '^#' Or: # grep '@' /etc/rsyslog.conf | grep -v '^#' If no line is returned, this is a finding.
Edit the syslog or rsyslog configuration file and add an appropriate remote syslog server.
Check the SSH client configuration for allowed protocol versions. # grep -i protocol /etc/ssh/ssh_config | grep -v '^#' If the returned protocol configuration allows versions less than 2, this is a finding.
Edit the /etc/ssh/ssh_config file and add or edit a "Protocol" configuration line not allowing versions less than 2.
Ask the SA to identify which interfaces on the system are designated for management traffic. If all interfaces on the system are authorized for management traffic, this is not applicable. Check the SSH daemon configuration for listening network addresses. # grep -i Listen /etc/ssh/sshd_config | grep -v '^#' If no configuration is returned, or if a returned 'Listen' configuration contains addresses not designated for management traffic, this is a finding.
Edit the SSH daemon configuration to specify listening network addresses designated for management traffic. Restart the SSH daemon. # /sbin/service sshd restart
Check the SSH daemon configuration for allowed ciphers. # grep -i ciphers /etc/ssh/sshd_config | grep -v '^#' If no lines are returned, or the returned ciphers list contains any cipher not starting with "3des" or "aes", this is a finding
Edit the SSH daemon configuration and remove any ciphers not starting with "3des" or "aes" and remove any ciphers ending with "cbc". If necessary, add a "Ciphers" line. Ciphers aes256-ctr,aes192-ctr,aes128-ctr Restart the SSH daemon. # /sbin/service sshd restart
Check the SSH daemon configuration for allowed ciphers. # grep -i ciphers /etc/ssh/sshd_config | grep -v '^#' If no lines are returned, or the returned ciphers list contains any cipher ending with cbc, this is a finding.
Edit /etc/ssh/sshd_config and add or edit the "Ciphers" line. Only include ciphers that start with "3des" or "aes" and do not contain "cbc". For the list of available ciphers for the particular version of your software, consult the sshd_config manpage. Restart the SSH daemon.
Check the SSH daemon configuration for allowed MACs. Procedure: # grep -i macs /etc/ssh/sshd_config | grep -v '^#' If no lines are returned, or the returned MACs list contains any MAC that is not hmac-sha1 or a better hmac algorithm that is on the FIPS 140-2 approved list, this is a finding.
Edit the SSH daemon configuration and remove any MACs that are not hmac-sha1 or a better hmac algorithm that is on the FIPS 140-2 approved list. If necessary, add a MACs line. Restart the SSH daemon. # /sbin/service sshd restart
Check the SSH client configuration for allowed ciphers. # grep -i ciphers /etc/ssh/ssh_config | grep -v '^#' If no lines are returned, or the returned ciphers list contains any cipher not starting with "3des" or "aes", this is a finding.
Edit the SSH client configuration and remove any ciphers not starting with "3des" or "aes" and remove any ciphers ending with "cbc". If necessary, add a "Ciphers" line.
Check the SSH client configuration for allowed ciphers. # grep -i ciphers /etc/ssh/ssh_config | grep -v '^#' If no lines are returned, or the returned ciphers list contains any cipher ending with "cbc", this is a finding.
Edit the SSH client configuration and remove any ciphers not starting with "3des" or "aes" and remove any ciphers ending with "cbc". If necessary, add a "Ciphers" line.
Check the SSH client configuration for allowed MACs. Procedure: # grep -i macs /etc/ssh/ssh_config | grep -v '^#' If no lines are returned, or the returned MACs list contains any MAC that is not hmac-sha1 or a better hmac algorithm that is on the FIPS 140-2 approved list, this is a finding.
Edit the SSH client configuration and remove any MACs that are not hmac-sha1 or a better hmac algorithm that is on the FIPS 140-2 approved list. If necessary, add a MACs line.
There are two ways in which access to SSH may restrict users or groups. Check if /etc/pam.d/sshd is configured to require daemon style login control. # grep pam_access.so /etc/pam.d/sshd|grep "required"|grep "account"| grep -v '^#' If no lines are returned, sshd is not configured to use pam_access. Check the SSH daemon configuration for the AllowGroups setting. # egrep -i "AllowGroups|AllowUsers" /etc/ssh/sshd_config | grep -v '^#' If no lines are returned, sshd is not configured to limit access to users/groups. If sshd is not configured to limit access either through pam_access or the use "AllowUsers" or "Allowgroups", this is a finding.
Edit the SSH daemon configuration and add an "AllowGroups" or "AllowUsers" directive specifying the groups and users allowed to have access. Restart the SSH daemon. # /sbin/service sshd restart Alternatively, modify the /etc/pam.d/sshd file to include the line account required pam_access.so accessfile=<path to access.conf for sshd> If the "accessfile" option is not specified the default "access.conf" file will be used. The "access.conf" file must contain the user restriction definitions.
Check the permissions for SSH public host key files. # ls -lL /etc/ssh/*key.pub If any file has a mode more permissive than 0644, this is a finding.
Change the permissions for the SSH public host key files. # chmod 0644 /etc/ssh/*key.pub
Check the permissions for SSH private host key files. # ls -lL /etc/ssh/*key If any file has a mode more permissive than 0600, this is a finding.
Change the permissions for the SSH private host key files. # chmod 0600 /etc/ssh/*key
Ask the SA if GSSAPI authentication is used for SSH authentication to the system. If so, this is not applicable. Check the SSH daemon configuration for the GSSAPIAuthentication setting. # grep -i GSSAPIAuthentication /etc/ssh/sshd_config | grep -v '^#' If no lines are returned, or the setting is set to "yes", this is a finding.
Edit the SSH daemon configuration and set (add if necessary) a "GSSAPIAuthentication" directive set to "no". Restart the SSH daemon. # /sbin/service sshd restart
The default setting for GSSAPIAuthentication is "no" . Check for a change from the default. # grep -i GSSAPIAuthentication /etc/ssh/ssh_config | grep -v '^#' If the setting is "yes" this is a finding.
Edit the SSH client configuration and set the GSSAPIAuthentication" directive set to "no".
Ask the SA if Kerberos authentication is used by the system. If it is, this is not applicable. Check the SSH daemon configuration for the KerberosAuthentication setting. # grep -i KerberosAuthentication /etc/ssh/sshd_config | grep -v '^#' If no lines are returned, or the setting is set to "yes", this is a finding.
Edit the SSH daemon configuration and set (add if necessary) the "KerberosAuthentication" directive set to "no". Restart the SSH daemon. # /sbin/service sshd restart
Check the SSH daemon configuration for the StrictModes setting. # grep -i StrictModes /etc/ssh/sshd_config | grep -v '^#' If the setting is not present, or not set to "yes", this is a finding.
Edit the SSH daemon configuration and add or edit the "StrictModes" setting value to "yes". Restart the SSH daemon. # /sbin/service sshd restart
Check the SSH daemon configuration for the UsePrivilegeSeparation setting. # grep -i UsePrivilegeSeparation /etc/ssh/sshd_config | grep -v '^#' If the setting is not present, or not set to "yes", this is a finding.
Edit the SSH daemon configuration and add or edit the "UsePrivilegeSeparation" setting value to "yes". Restart the SSH daemon. # /sbin/service sshd restart
Check the SSH daemon configuration for the RhostsRSAAuthentication setting. # grep -i RhostsRSAAuthentication /etc/ssh/sshd_config | grep -v '^#' If the setting is set to "yes", this is a finding.
Edit the SSH daemon configuration and add or edit the "RhostsRSAAuthentication" setting value to "no". Restart the SSH daemon. # /sbin/service sshd restart
Check the SSH daemon configuration for the compression setting. # grep -i Compression /etc/ssh/sshd_config | egrep "no|delayed" If the setting is missing or is commented out, this is a finding. If the setting is present but is not set to "no" or "delayed", this is a finding.
Edit the SSH daemon configuration and add or edit the "Compression" setting value to "no" or "delayed". Restart the SSH daemon. # /sbin/service sshd restart
Verify the SSH daemon is configured for logon warning banners. Procedure: An exact match is required to have a valid warning banner. Check for the following login banner. You are accessing a U.S. Government (USG) Information System (IS) that is provided for USG-authorized use only. By using this IS (which includes any device attached to this IS), you consent to the following conditions: -The USG routinely intercepts and monitors communications on this IS for purposes including, but not limited to, penetration testing, COMSEC monitoring, network operations and defense, personnel misconduct (PM), law enforcement (LE), and counterintelligence (CI) investigations. -At any time, the USG may inspect and seize data stored on this IS. -Communications using, or data stored on, this IS are not private, are subject to routine monitoring, interception, and search, and may be disclosed or used for any USG-authorized purpose. -This IS includes security measures (e.g., authentication and access controls) to protect USG interests--not for your personal benefit or privacy. -Notwithstanding the above, using this IS does not constitute consent to PM, LE or CI investigative searching or monitoring of the content of privileged communications, or work product, related to personal representation or services by attorneys, psychotherapists, or clergy, and their assistants. Such communications and work product are private and confidential. See User Agreement for details. OR I've read & consent to terms in IS user agreem't.
Edit /etc/issue and the DoD login banner. DoD Login Banners: You are accessing a U.S. Government (USG) Information System (IS) that is provided for USG-authorized use only. By using this IS (which includes any device attached to this IS), you consent to the following conditions: -The USG routinely intercepts and monitors communications on this IS for purposes including, but not limited to, penetration testing, COMSEC monitoring, network operations and defense, personnel misconduct (PM), law enforcement (LE), and counterintelligence (CI) investigations. -At any time, the USG may inspect and seize data stored on this IS. -Communications using, or data stored on, this IS are not private, are subject to routine monitoring, interception, and search, and may be disclosed or used for any USG-authorized purpose. -This IS includes security measures (e.g., authentication and access controls) to protect USG interests- -not for your personal benefit or privacy. -Notwithstanding the above, using this IS does not constitute consent to PM, LE or CI investigative searching or monitoring of the content of privileged communications, or work product, related to personal representation or services by attorneys, psychotherapists, or clergy, and their assistants. Such communications and work product are private and confidential. See User Agreement for details. OR I've read & consent to terms in IS user agreem't. Find the location of the banner file for sshd and examine the content: # grep -i banner /etc/ssh/sshd_config | grep -v '^#' # cat Edit the SSH daemon configuration and add or edit a "Banner" setting referencing a file containing a logon warning banner. Restart the SSH daemon. # /sbin/service sshd restart
Check for a default route for IPv6. If the system is a VM host and acts as a router solely for the benefit of its client systems, then this rule is not applicable. # ip -6 route list | grep default If the system uses IPv6, and no results are returned, this is a finding.
Add a default route for IPv6. Edit /etc/sysconfig/network-scripts/ifcfg-eth0 (substitute interface as appropriate). Add an IPV6_DEFAULTGW=<gateway> configuration setting. Restart the interface. # ifdown eth0; ifup eth0
Check if the system is configured for IPv6 forwarding. # grep [01] /proc/sys/net/ipv6/conf/*/forwarding|egrep "default|all" If the /proc/sys/net/ipv6/conf/*/forwarding entries do not exist because of compliance with GEN007720, this is not a finding. If all of the resulting lines do not end with 0, this is a finding.
Disable IPv6 forwarding. Edit /etc/sysctl.conf and add a setting for "net.ipv6.conf.all.forwarding=0" and "net.ipv6.conf.default.forwarding=0". Reload the sysctls. Procedure: # sysctl -p
Check the group ownership of the NFS export configuration file. Procedure: # ls -lL /etc/exports If the file is not group-owned by root, bin, sys, or system, this is a finding.
Change the group ownership of the NFS export configuration file. Procedure: # chgrp root /etc/exports
Check the permissions of the NFS export configuration file. # ls -lL /etc/exports If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/exports
List the exports. # cat /etc/exports For each file system displayed, check the ownership. # ls -ldL <exported file system path> If the directory is not group-owned by root, bin, sys, or system, this is a finding.
Change the group owner of the export directory. # chgrp root <export>
Check the permissions of the Samba configuration file. # ls -lL /etc/samba/smb.conf If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/samba/smb.conf
Check the permissions of the Samba password files. Procedure: # ls -lL /etc/samba/passdb.tdb /etc/samba/secrets.tdb If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/samba/passdb.tdb /etc/samba/secrets.tdb
Check the security mode of the Samba configuration. # grep -i security /etc/samba/smb.conf If the security mode is "share", this is a finding.
Edit the "/etc/samba/smb.conf" file and change the "security" setting to "user" or another valid setting other than "share".
If the "samba-common" package is not installed, this is not applicable. Check the encryption setting of Samba. # grep -i 'encrypt passwords' /etc/samba/smb.conf If the setting is not present, or not set to 'yes', this is a finding.
Edit the "/etc/samba/smb.conf" file and change the "encrypt passwords" setting to "yes".
Check the access to shares for Samba. # grep -i 'guest ok' /etc/samba/smb.conf If the setting exists and is set to 'yes', this is a finding.
Edit the "/etc/samba/smb.conf" file and change the "guest ok" setting to "no".
Check the permissions of the file. # ls -lL /etc/news/incoming.conf If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/news/incoming.conf
Check the permissions for "/etc/news/hosts.nntp.nolimit". # ls -lL /etc/news/hosts.nntp.nolimit If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/news/hosts.nntp.nolimit
Check the permissions of the file. # ls -lL /etc/news/nnrp.access If the permissions include a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/news/nnrp.access
Check the permissions of the file. # ls -lL /etc/news/passwd.nntp If the mode includes a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/news/passwd.nntp
Check the root crontab (crontab -l) and the global crontabs in "/etc/crontab", "/etc/cron.*" for the presence of an rpm verification command such as: rpm -qVa | awk '$2!="c" {print $0}' If no such cron job is found, this is a finding. If the result of the cron job indicates packages which do not pass verification exist, this is a finding unless the changes were made due to another STIG entry.
Add a cron job to run an rpm verification command such as: rpm -qVa | awk '$2!="c" {print $0}' For packages which failed verification: If the package is not necessary for operations, remove it from the system. If the package is necessary for operations, re-install the package.
If using an Advanced Intrusion Detection Environment (AIDE), verify that the configuration contains the "ACL" option for all monitored files and directories. Procedure: Check for the default location /etc/aide/aide.conf or: # find / -name aide.conf # egrep "[+]?acl" <aide.conf file> If the option is not present. This is a finding. If using a different file integrity tool, check the configuration per tool documentation.
If using AIDE, edit the configuration and add the "ACL" option for all monitored files and directories. If using a different file integrity tool, configure ACL checking per the tool's documentation.
If using an Advanced Intrusion Detection Environment (AIDE), verify the configuration contains the "xattrs" option for all monitored files and directories. Procedure: Check for the default location /etc/aide/aide.conf or: # find / -name aide.conf # egrep "[+]?xattrs" <aide.conf file> If the option is not present. This is a finding. If using a different file integrity tool, check the configuration per tool documentation.
If using AIDE, edit the configuration and add the "xattrs" option for all monitored files and directories. If using a different file integrity tool, configure extended attributes checking per the tool's documentation.
If using an Advanced Intrusion Detection Environment (AIDE), verify the configuration contains the "sha256" or "sha512" options for all monitored files and directories. Procedure: Check for the default location /etc/aide/aide.conf or: # find / -name aide.conf # egrep "[+]?(sha256|sha512)" <aide.conf file> If the option is not present. This is a finding. If one of these options is not present. This is a finding. If using a different file integrity tool, check the configuration per tool documentation.
If using AIDE, edit the configuration and add the "sha512" option for all monitored files and directories. If using a different file integrity tool, configure FIPS 140-2 approved cryptographic hashes per the tool's documentation.
Verify the SCTP protocol handler is prevented from dynamic loading. # grep 'install sctp /bin/true' /etc/modprobe.conf /etc/modprobe.d/* If no result is returned, this is a finding.
Prevent the SCTP protocol handler for dynamic loading. # echo "install sctp /bin/true" >> /etc/modprobe.conf
Verify the DCCP protocol handler is prevented from dynamic loading. # grep 'install dccp /bin/true' /etc/modprobe.conf /etc/modprobe.d/* If no result is returned, this is a finding. # grep 'install dccp_ipv4 /bin/true' /etc/modprobe.conf /etc/modprobe.d/* If no result is returned, this is a finding. # grep 'install dccp_ipv6 /bin/true' /etc/modprobe.conf /etc/modprobe.d/* If no result is returned, this is a finding.
Prevent the DCCP protocol handler for dynamic loading. # echo "install dccp /bin/true" >> /etc/modprobe.conf # echo "install dccp_ipv4 /bin/true" >> /etc/modprobe.conf # echo "install dccp_ipv6 /bin/true" >> /etc/modprobe.conf
Verify the AppleTalk protocol handler is prevented from dynamic loading. # grep 'install appletalk' /etc/modprobe.conf /etc/modprobe.d/* If anything is returned check that appletalk is disabled by having the executable set to '/bin/true'. If an uncommented line containing "appletalk" is found which has not been disabled, this is a finding.
Prevent the AppleTalk protocol handler for dynamic loading. # echo "install appletalk /bin/true" >> /etc/modprobe.conf
Ask the SA if RDS is required by application software running on the system. If so, this is not applicable. Verify the RDS protocol handler is prevented from dynamic loading. # grep 'install rds /bin/true' /etc/modprobe.conf /etc/modprobe.d/* If no result is returned, this is a finding.
Prevent the RDS protocol handler for dynamic loading. # echo "install rds /bin/true" >> /etc/modprobe.conf
Verify the TIPC protocol handler is prevented from dynamic loading. # grep 'install tipc /bin/true' /etc/modprobe.conf /etc/modprobe.d/* If no result is returned, this is a finding.
Prevent the TIPC protocol handler for dynamic loading. # echo "install tipc /bin/true" >> /etc/modprobe.conf
Verify the Bluetooth protocol handler is prevented from dynamic loading. # grep 'install bluetooth /bin/true' /etc/modprobe.conf /etc/modprobe.d/* If no result is returned, this is a finding.
Prevent the Bluetooth protocol handler for dynamic loading. # echo "install bluetooth /bin/true" >> /etc/modprobe.conf
Check the system for any active 6to4 tunnels without specific remote addresses. # ip tun list | grep "remote any" | grep "ipv6/ip" If any results are returned the "tunnel" is the first field. If any results are returned, this is a finding.
Disable the active 6to4 tunnel. # ip link set <tunnel> down Add this command to a startup script, or remove the configuration creating the tunnel.
Verify the Miredo service is not running. # ps ax | grep miredo | grep -v grep If the miredo process is running, this is a finding.
Edit startup scripts to prevent the service from running on startup.
Check for any IP tunnels. # ip tun list # ip -6 tun list If any tunnels are listed, this is a finding.
Remove the tunnels. # ip tun del <tunnel> Edit system startup scripts to prevent tunnel creation on startup.
Verify no interface is configured to use DHCP. # grep -i bootproto=dhcp /etc/sysconfig/network-scripts/ifcfg-* If any configuration is found, this is a finding.
Edit the "/etc/sysconfig/network-scripts/ifcfg-*" file(s) and change the "bootproto" setting to "static".
If the "dhclient" package is not installed, this is not applicable. Verify the DHCP client is configured to not send dynamic DNS updates. Procedure: # grep do-forward-updates /etc/dhclient.conf If the file is not present, does not contain this configuration, or has the setting set to "true", this is a finding.
Edit or add the "/etc/dhclient.conf" file and add or edit the "do-forward-updates" setting to false. Procedure: # echo "do-forward-updates false;" >> /etc/dhclient.conf
Verify the system is configured to ignore IPv6 ICMP redirect messages. # cat /proc/sys/net/ipv6/conf/all/accept_redirects If the /proc/sys/net/ipv6/conf/all/accept_redirects entry does not exist because of compliance with GEN007720, this is not a finding. If the returned value is not "0", this is a finding.
Configure the system to ignore IPv6 ICMP redirect messages. Edit "/etc/sysctl.conf" and add a settings for "net.ipv6.conf.default.accept_redirects=0" and "net.ipv6.conf.all.accept_redirects=0". Restart the system for the setting to take effect.
Determine if the system is configured to forward IPv6 source-routed packets. Procedure: # egrep "net.ipv6.conf.*forwarding" /etc/sysctl.conf If there are no entries found or the value of the entries is not = "0", this is a finding.
Configure the system to not forward IPv6 source-routed packets. Procedure: Edit the /etc/sysctl.conf file to include: net.ipv6.conf.all.forwarding = 0 net.ipv6.conf.default.forwarding = 0 Reload the kernel parameters: # sysctl -p
To check to see if the system is an LDAP server, verify LDAP is running on the system: # ps -ef | grep ldap Find out which LDAP is used (if not determined via the command above). # rpm -qa | grep ldap If using nssldap: # grep base /etc/ldap.conf Check to see if the base is set to something besides the default of "dc=example,dc=com". If using openldap: # grep suffix /etc/openldap/slapd.conf Check whether the system is an LDAP client: # grep server /etc/ldap.conf # grep server /etc/openldap/ldap.conf Check whether the server option has an address other than the loopback, then check the nsswitch.conf file. # grep ldap /etc/nsswitch.conf Look for the following three lines: passwd: files ldap shadow: files ldap group: files ldap If all three files are not configured to look for an LDAP source, then the system is not using LDAP for authentication. If the system is not using LDAP for authentication, this is not applicable. Check if NSS LDAP is using TLS. # grep '^ssl start_tls' /etc/ldap.conf If no lines are returned, this is a finding. Check if NSS LDAP TLS is using only FIPS 140-2 approved cryptographic algorithms. # grep '^tls_ciphers' /etc/ldap.conf If the line is not present or contains ciphers not approved by FIPS 140-2, this is a finding. FIPS-approved ciphers include 3DES and AES. FIPS-approved hashes include the SHA hash family.
Edit "/etc/ldap.conf" and add a "ssl start_tls" and "tls_ciphers" options with only FIPS 140-2 approved ciphers.
Verify the source of the LDAP certificates Check if the system is using NSS LDAP. # grep -v '^#' /etc/nsswitch.conf | grep ldap If no lines are returned, this vulnerability is not applicable. Verify with the SA that the system is connected to the GIG. If the system part of a stand alone network which is not connected to the GIG this vulnerability is not applicable. Verify a certificate is used for client authentication to the server. # grep -i '^tls_cert' /etc/ldap.conf If no line is found, this is a finding. List the certificate issuer. # openssl x509 -text -in <cert> If the certificate is not issued by DoD PKI or a DoD-approved external PKI, this is a finding.
Edit "/etc/ldap.conf" and add or edit the 'tls_cert' setting to reference a file containing a client certificate issued by DoD PKI or a DoD-approved external PKI.
Check if the system is using NSS LDAP. # grep -v '^#' /etc/nsswitch.conf | grep ldap If no lines are returned, this vulnerability is not applicable. Verify a server certificate is required and verified by the NSS LDAP configuration. # grep -i '^tls_checkpeer' /etc/ldap.conf If no line is returned, or the value is not "yes", this is a finding.
Edit "/etc/ldap.conf" and add or set the "tls_checkpeer" setting to "yes".
Check if the system is using NSS LDAP. # grep -v '^#' /etc/nsswitch.conf | grep ldap If no lines are returned, this vulnerability is not applicable. Verify the NSS LDAP client is configured to check certificates against a certificate revocation list. # grep -i '^tls_crlcheck' /etc/ldap.conf If the setting does not exist, or the value is not "all", this is a finding.
Edit "/etc/ldap.conf" and add or set the "tls_crlcheck" setting to "all".
Check the permissions of the file. # ls -lL /etc/ldap.conf If the mode of the file is more permissive than 0644, this is a finding.
Change the permissions of the file. # chmod 0644 /etc/ldap.conf
Check the ownership of the file. # ls -lL /etc/ldap.conf If the file is not owned by root, this is a finding.
Change the owner of the file. # chown root /etc/ldap.conf
Check the group ownership of the file. Procedure: # ls -lL /etc/ldap.conf If the file is not group-owned by root, bin, sys, or system, this is a finding.
Change the group owner of the file to root, bin, sys, or system. Procedure: # chgrp root /etc/ldap.conf
Check the permissions of the file. # ls -lL /etc/ldap.conf If the mode includes a '+', the file has an extended ACL. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the "/etc/ldap.conf" file. # setfacl --remove-all /etc/ldap.conf
Determine if LDAP is used for account information on the system. To check to see if the system is an LDAP server, verify LDAP is running on the system: # ps -ef | grep ldap Find out which LDAP is used (if not determined via the command above). # rpm -qa | grep ldap If using nssldap: # grep base /etc/ldap.conf Check to see if the base is set to something besides the default of "dc=example,dc=com". If using openldap: # grep suffix /etc/openldap/slapd.conf Check whether the system is an LDAP client: # grep server /etc/ldap.conf # grep server /etc/openldap/ldap.conf Check whether the server option has an address other than the loopback, then check the nsswitch.conf file. # grep ldap /etc/nsswitch.conf Look for the following three lines: passwd: files ldap shadow: files ldap group: files ldap If no un-commented reference to "ldap" is identified, LDAP is not used for account information on the system, and this is not applicable. Determine the certificate authority file and/or directory. # grep -i '^tls_cacert' /etc/ldap.conf For each file or directory returned, check the ownership. # ls -lLd <certpath> If the owner of any file or directory is not root, this is a finding.
Change the ownership of the file or directory. # chown root <certpath>
Determine the certificate authority file and/or directory. # grep -i '^tls_cacert' /etc/ldap.conf For each file or directory returned, check the group ownership. # ls -lLd <certpath> If the group-owner of any file or directory is not root, bin, sys, or system, this is a finding.
Change the group ownership of the file or directory. # chgrp root <certpath>
Determine the certificate authority file and/or directory. Procedure: # grep -i '^tls_cacert' /etc/ldap.conf For each file or directory returned, check the permissions. Procedure: # ls -lLd <certpath> If the mode of the file is more permissive than 0644 (or 0755 for directories), this is a finding.
Change the mode of the file or directory. File Procedure: # chmod 0644 <certpath> Directory Procedure: # chmod 0755 <certpath>
Determine the certificate authority file and/or directory. # grep -i '^tls_cacert' /etc/ldap.conf For each file or directory returned, check the permissions. # ls -lLd <certpath> If the mode of the file or directory contains a '+', an extended ACL is present. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the certificate file. Procedure: For each certificate file found remove all extended permissions # setfacl --remove-all <certpath>
Determine the certificate file. # grep -i '^tls_cert' /etc/ldap.conf Check the ownership. # ls -lL <certpath> If the owner of the file is not root, this is a finding.
Change the ownership of the file. # chown root <certpath>
Determine the certificate file. Procedure: # grep -i '^tls_cert' /etc/ldap.conf Check the group ownership. Procedure: # ls -lL <certpath> If the group owner of the file is not root, bin, sys, or system, this is a finding.
Change the group ownership of the file. Procedure: # chgrp root <certpath>
Determine the certificate file. # grep -i '^tls_cert' /etc/ldap.conf Check the permissions. # ls -lL <certpath> If the mode of the file is more permissive than 0644, this is a finding.
Change the mode of the file. # chmod 0644 <certpath>
Determine the certificate file. # grep -i '^tls_cert' /etc/ldap.conf Check the permissions. # ls -lL <certpath> If the mode of the file contains a '+', an extended ACL is present. This is a finding.
Remove the extended ACL from the certificate file. Procedure: For each certificate file found remove all extended permissions. # setfacl --remove-all <certpath>
Determine the key file. # grep -i '^tls_key' /etc/ldap.conf Check the ownership. # ls -lL <keypath> If the owner of the file is not root, this is a finding.
Change the ownership of the file. # chown root <keypath>
Determine the key file. # grep -i '^tls_key' /etc/ldap.conf Check the group ownership. # ls -lL <keypath> If the file is not group owned by root, bin, or sys, this is a finding.
Change the group ownership of the file. # chgrp root <keypath>
Determine the key file. # grep -i '^tls_key' /etc/ldap.conf Check the permissions. # ls -lL <keypath> If the mode of the file is more permissive than 0600, this is a finding.
Change the mode of the file. # chmod 0600 <keypath>
Determine the key file. # grep -i '^tls_key' /etc/ldap.conf Check the permissions. # ls -lL <keypath> If the permissions of the file contains a '+', an extended ACL is present. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the key file. Procedure: For each key file found remove all extended permissions. # setfacl --remove-all <keypath>
Verify exec-shield is enabled if present. # cat /proc/sys/kernel/exec-shield If the file is present and contains a value of "0", this is a finding.
Edit the kernel boot parameters, or "/etc/sysctl.conf", and set exec-shield to "1". Reboot the system.
If the autofs service is needed, this vulnerability is not applicable. Check if the autofs service is running. # service autofs status If the service is running, this is a finding.
Stop and disable the autofs service. # service autofs stop # chkconfig autofs off
If the system needs USB, this vulnerability is not applicable. Check if the directory "/proc/bus/usb" exists. If so, this is a finding.
Edit the grub bootloader file "/boot/grub/grub.conf" or "/boot/grub/menu.lst" by appending the "nousb" parameter to the kernel boot line.
If the system needs USB storage, this vulnerability is not applicable. Check if usb-storage is prevented from loading. # grep 'install usb-storage /bin/true' /etc/modprobe.conf /etc/modprobe.d/* If no results are returned, this is a finding.
Prevent the usb-storage module from loading. # echo 'install usb-storage /bin/true' >> /etc/modprobe.conf
If the system needs IEEE 1394 (Firewire), this is not applicable. Check if the firewire module is not disabled. # grep 'install ieee1394 /bin/true' /etc/modprobe.conf /etc/modprobe.d/* If no results are returned, this is a finding.
Prevent the system from loading the firewire module. # echo 'install ieee1394 /bin/true' >> /etc/modprobe.conf
Determine if the system is using a local firewall. # chkconfig --list iptables If the service is not "on" in the standard runlevel (ordinarily 3 or 5), this is a finding.
Enable the system's local firewall. # chkconfig iptables on # service iptables start
Check the firewall rules for a default deny rule. # iptables --list Example of a rule meeting this criteria: REJECT all -- anywhere anywhere reject-with icmp-host-prohibited A rule using DROP is also acceptable. The default rule should be the last rule of a table and match all traffic. If there is no default deny rule, this is a finding.
Edit "/etc/sysconfig/iptables" and add a default deny rule. An example of a default deny rule: -A RH-Firewall-1-INPUT -j REJECT --reject-with icmp-host-prohibited Restart the iptable service. # service iptables restart
Check if SELinux is enabled with at least a "targeted" policy. # grep ^SELINUX /etc/sysconfig/selinux If the SELINUX option is not set to "enforcing", this is a finding. If the SELINUXTYPE option is not set to "targeted" or "strict", this is a finding. If the use of the system is incompatible with the confines of SELinux this rule may be waived.
Enable one of the SELinux policies. Edit /etc/sysconfig/selinux and set the value of the SELINUX option to "enforcing" and SELINUXTYPE to "targeted" or "strict". Restart the system.
Check the permissions of the file. # ls -lL /boot/grub/grub.conf If the permissions of the file or directory contains a '+', an extended ACL is present. This is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/grub.conf
Check the ownership of the file. # ls -lLd /boot/grub/grub.conf If the owner of the file is not root, this is a finding.
Change the ownership of the file. # chown root /boot/grub/grub.conf
Check the group ownership of the file. # ls -lLd /boot/grub/grub.conf If the group-owner of the file is not root, bin, sys, or system this is a finding.
Change the group ownership of the file. # chgrp root /boot/grub/grub.conf
Verify YUM signature validation is not disabled. # grep gpgcheck /etc/yum.conf /etc/yum.repos.d/* If no results are returned, or the returned “gpgcheck” settings are not equal to “1”, this is a finding.
Edit the YUM configuration containing "gpgcheck=0" and set the value to "1".
Verify the YUM service is enabled. # service yum-updatesd status If the service is enabled, this is a finding.
Disable the yum service. # chkconfig yum-updatesd off ; service yum-updatesd stop
Check the permissions of the file. # ls -lL /etc/security/access.conf If the permissions of the file or directory contain a '+', an extended ACL is present. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/security/access.conf
Check the permissions of the file. # ls -lL /etc/sysctl.conf If the permissions of the file or directory contain a '+', an extended ACL is present. If the file has an extended ACL and it has not been documented with the IAO, this is a finding.
Remove the extended ACL from the file. # setfacl --remove-all /etc/sysctl.conf
Check for the audit=1 kernel parameter. # grep 'audit=1' /proc/cmdline If no results are returned, this is a finding.
Edit the grub bootloader file /boot/grub/grub.conf or /boot/grub/menu.lst by appending the "audit=1" parameter to the kernel boot line. Reboot the system for the change to take effect.
Check for any running routing protocol daemons. If the system is a VM host and acts as a router solely for the benefits of its client systems, then this rule is not applicable. # chkconfig --list |grep :on|egrep '(ospf|route|bgp|zebra|quagga)' If any routing protocol daemons are listed, this is a finding.
Disable any routing protocol daemons.
Check the group ownership of the audit logs. Procedure: # grep "^log_file" /etc/audit/auditd.conf|sed s/^[^\/]*//|xargs stat -c %G:%n If any audit log file is not group-owned by root, bin, sys, or system, this is a finding.
Change the group ownership of the audit log file(s). Procedure: # chgrp root <audit log file>
FTP to the system. # ftp localhost Check for either of the following login banners based on the character limitations imposed by the system. An exact match is required. If one of these banners is not displayed, this is a finding. If the system does not run the FTP service, this is not applicable. DoD Login Banners: You are accessing a U.S. Government (USG) Information System (IS) that is provided for USG-authorized use only. By using this IS (which includes any device attached to this IS), you consent to the following conditions: -The USG routinely intercepts and monitors communications on this IS for purposes including, but not limited to, penetration testing, COMSEC monitoring, network operations and defense, personnel misconduct (PM), law enforcement (LE), and counterintelligence (CI) investigations. -At any time, the USG may inspect and seize data stored on this IS. -Communications using, or data stored on, this IS are not private, are subject to routine monitoring, interception, and search, and may be disclosed or used for any USG-authorized purpose. -This IS includes security measures (e.g., authentication and access controls) to protect USG interests--not for your personal benefit or privacy. -Notwithstanding the above, using this IS does not constitute consent to PM, LE or CI investigative searching or monitoring of the content of privileged communications, or work product, related to personal representation or services by attorneys, psychotherapists, or clergy, and their assistants. Such communications and work product are private and confidential. See User Agreement for details. OR I've read & consent to terms in IS user agreem't.
Provide the proper text for the DoD banner to be presented by the FTP server to the user. For vsftp: Examine the /etc/vsftp.conf file for the "banner_file" entry. (i.e. banner_file = /etc/banner/vsftp) For gssftp: Examine the /etc/xinetd.d/gssftp file for the "banner" entry. (i.e. banner = /etc/banner/gssftp) For both: Add the banner entry if one is not found. Modify or create the referenced banner file to contain one of the following DoD login banners (based on the character limitations imposed by the system). DoD Login Banners: You are accessing a U.S. Government (USG) Information System (IS) that is provided for USG-authorized use only. By using this IS (which includes any device attached to this IS), you consent to the following conditions: -The USG routinely intercepts and monitors communications on this IS for purposes including, but not limited to, penetration testing, COMSEC monitoring, network operations and defense, personnel misconduct (PM), law enforcement (LE), and counterintelligence (CI) investigations. -At any time, the USG may inspect and seize data stored on this IS. -Communications using, or data stored on, this IS are not private, are subject to routine monitoring, interception, and search, and may be disclosed or used for any USG-authorized purpose. -This IS includes security measures (e.g., authentication and access controls) to protect USG interests--not for your personal benefit or privacy. -Notwithstanding the above, using this IS does not constitute consent to PM, LE or CI investigative searching or monitoring of the content of privileged communications, or work product, related to personal representation or services by attorneys, psychotherapists, or clergy, and their assistants. Such communications and work product are private and confidential. See User Agreement for details. OR I've read & consent to terms in IS user agreem't.
Determine if the /var path is a separate filesystem. # grep /var /etc/fstab If no result is returned, /var is not on a separate filesystem this is a finding.
Migrate the /var path onto a separate file system.
Determine if the /var/log/audit path is a separate filesystem. # grep /var/log/audit /etc/fstab If no result is returned, /var/log/audit is not on a separate filesystem this is a finding.
Migrate the /var/log/audit path onto a separate filesystem.
Determine if the /tmp path is a separate filesystem. # egrep "[\t ]/tmp[\t ]" /etc/fstab If no result is returned, /tmp is not on a separate filesystem this is a finding.
Migrate the /tmp path onto a separate file system.
# cat /proc/sys/net/ipv4/tcp_max_syn_backlog If the result is not 1280 or greater, this is a finding.
Edit /etc/sysctl.conf and add a setting for "net.ipv4.tcp_max_syn_backlog=1280". Procedure: # sysctl -p
If the system uses sendmail examine the configuration files. Determine if sendmail only binds to loopback addresses by examining the "DaemonPortOptions" configuration options. Procedure: # grep -i "O DaemonPortOptions" /etc/mail/sendmail.cf If there are uncommented DaemonPortOptions lines, and all such lines specify system loopback addresses, this is not a finding. Otherwise, determine if sendmail is configured to allow open relay operation. Procedure: # grep -i promiscuous_relay /etc/mail/sendmail.mc If the promiscuous relay feature is enabled, this is a finding. If the system uses Postfix, locate the main.cf file. Procedure: # find / -name main.cf Determine if Postfix only binds to loopback addresses by examining the "inet_interfaces" line. Procedure: # grep inet_interfaces </path/to/main.cf> If "inet_interfaces" is set to "loopback-only" or contains only loopback addresses such as 127.0.0.1 and [::1], Postfix is not listening on external network interfaces, and this is not a finding. Otherwise, determine if Postfix is configured to restrict clients permitted to relay mail by examining the "smtpd_client_restrictions" line. Procedure: # grep smtpd_client_restrictions </path/to/main.cf> If the "smtpd_client_restrictions" line is missing, or does not contain "reject", this is a finding. If the line contains "permit" before "reject", this is a finding. If the system is using other SMTP software, consult the software's documentation for procedures to verify mail relaying is restricted.
If the system uses sendmail, edit the sendmail.mc file and remove the "promiscuous_relay" configuration. Rebuild the sendmail.cf file from the modified sendmail.mc and restart the service. If the system does not need to receive mail from external hosts, add one or more DaemonPortOptions lines referencing system loopback addresses (such as "O DaemonPortOptions=Addr=127.0.0.1,Port=smtp,Name=MTA") and remove lines containing non-loopback addresses. Restart the service. If the system uses Postfix, edit the main.cf file and add or edit the "smtpd_client_restrictions" line to have contents "permit_mynetworks, reject" or a similarly restrictive rule. If the system does not need to receive mail from external hosts, add or edit the "inet_interfaces" line to have contents "loopback-only" or a set of loopback addresses for the system. Restart the service. If the system is using other SMTP software, consult the software's documentation for procedures to restrict mail relaying.
Check the system for the 'ldd' binary. Procedure: # ls -lL /usr/bin/ldd If the 'ldd' binary has any executable permissions bits set, this is a finding.
Remove the execute permissions from the 'ldd' executable. Procedure: # chmod a-x /usr/bin/ldd
Check for an iptables rule that drops inbound IPv6 ICMP ECHO_REQUESTs sent to the all-hosts multicast address. Procedure: # less /etc/sysconfig/ip6tables Check for a rule in, or referenced by, the INPUT chain such as: -A INPUT -p icmpv6 -d ff02::1 --icmpv6-type 128 -j DROP If such a rule does not exist, this is a finding.
Add an iptables rule that drops inbound IPv6 ICMP ECHO_REQUESTs sent to the all-hosts multicast address. Edit /etc/sysconfig/ip6tables and add a rule in, or referenced by, the INPUT chain such as: -A INPUT -p icmpv6 -d ff02::1 --icmpv6-type 128 -j DROP Reload the iptables rules. Procedure: # service ip6tables restart
Access the graphical desktop environment(s) provided by the system and attempt to log in. Check for either of the following login banners based on the character limitations imposed by the system. An exact match is required. If one of these banners is not displayed, this is a finding. You are accessing a U.S. Government (USG) Information System (IS) that is provided for USG-authorized use only. By using this IS (which includes any device attached to this IS), you consent to the following conditions: -The USG routinely intercepts and monitors communications on this IS for purposes including, but not limited to, penetration testing, COMSEC monitoring, network operations and defense, personnel misconduct (PM), law enforcement (LE), and counterintelligence (CI) investigations. -At any time, the USG may inspect and seize data stored on this IS. -Communications using, or data stored on, this IS are not private, are subject to routine monitoring, interception, and search, and may be disclosed or used for any USG-authorized purpose. -This IS includes security measures (e.g., authentication and access controls) to protect USG interests- -not for your personal benefit or privacy. -Notwithstanding the above, using this IS does not constitute consent to PM, LE or CI investigative searching or monitoring of the content of privileged communications, or work product, related to personal representation or services by attorneys, psychotherapists, or clergy, and their assistants. Such communications and work product are private and confidential. See User Agreement for details. OR I've read & consent to terms in IS user agreem't.
Configure the system to display one of the DoD login banners prior to, or as part of, the graphical desktop environment login process. Procedure: Modify /usr/share/gdm/themes/RHEL/RHEL.xml by adding the following xml after the first two "pixmap" entries. <item type="rect" id="custom-dod-banner"> <pos anchor="nw" x="20%" y="10" width="80%" height="100%"/> <box> <item type="label"> <normal font="Sans Bold 9" color="#ffffff"/> <text> Insert the "approved text" here based on the character limitations imposed by the system. </text> </item> </box> </item> Approved text: DoD Login Banners: You are accessing a U.S. Government (USG) Information System (IS) that is provided for USG-authorized use only. By using this IS (which includes any device attached to this IS), you consent to the following conditions: -The USG routinely intercepts and monitors communications on this IS for purposes including, but not limited to, penetration testing, COMSEC monitoring, network operations and defense, personnel misconduct (PM), law enforcement (LE), and counterintelligence (CI) investigations. -At any time, the USG may inspect and seize data stored on this IS. -Communications using, or data stored on, this IS are not private, are subject to routine monitoring, interception, and search, and may be disclosed or used for any USG-authorized purpose. -This IS includes security measures (e.g., authentication and access controls) to protect USG interests--not for your personal benefit or privacy. -Notwithstanding the above, using this IS does not constitute consent to PM, LE or CI investigative searching or monitoring of the content of privileged communications, or work product, related to personal representation or services by attorneys, psychotherapists, or clergy, and their assistants. Such communications and work product are private and confidential. See User Agreement for details. OR I've read & consent to terms in IS user agreem't.
Consult vendor documentation to determine if the system is capable of CAC authentication. If it is not, this is not applicable. Interview the SA to determine if all accounts not exempted by policy are using CAC authentication. If non-exempt accounts are not using CAC authentication, this is a finding.
Consult vendor documentation to determine the procedures necessary for configuring CAC authentication. Configure all accounts required by policy to use CAC authentication.
Verify the system is configured to forward all audit records to a remote server. If the system is not configured to provide this function, this is a finding. Procedure: Ensure the audit option for the kernel is enabled. # grep "audit" /boot/grub/grub.conf | grep -v "^#" If the kernel does not have the "audit=1" option specified, this is a finding. Ensure the kernel auditing is active. # grep "active" /etc/audisp/plugins.d/syslog.conf | grep -v "^#" If the "active" setting is either missing or not set to "yes", this is a finding. Ensure all audit records are forwarded to a remote server. # grep "\*.\*" /etc/syslog.conf |grep "@" | grep -v "^#" (for syslog) or: # grep "\*.\*" /etc/rsyslog.conf | grep "@" | grep -v "^#" (for rsyslog) If neither of these lines exist, it is a finding.
Configure the system to send audit records to a remote server. Procedure: These instructions assume a known remote audit server is available to this system. Modify /etc/syslog.conf or /etc/rsyslog.conf to contain a line sending all audit records to a remote audit server. The server is specified by placing an "@" before the DNS name or IP address in the line. *.* @<remote audit server> Edit the "active" line in /etc/audisp/plugins.d/syslog.conf so it shows "active = yes". Restart audit and syslog: # service auditd restart # service syslog restart Or: # service rsyslog restart
Verify LDAP is running on the system. To check to see if the system is an LDAP server, run: # ps -ef | grep ldap Find out which LDAP is used (if not determined via the command above). # rpm -qa | grep ldap If using nssldap: # grep base /etc/ldap.conf Check to see if the base is set to something besides the default of "dc=example,dc=com". If using openldap: # grep suffix /etc/openldap/slapd.conf Check whether the system is an LDAP client: # grep server /etc/ldap.conf # grep server /etc/openldap/ldap.conf Check whether the server option has an address other than the loopback, then check the nsswitch.conf file: # grep ldap /etc/nsswitch.conf Look for the following three lines: passwd: files ldap shadow: files ldap group: files ldap If all three files are not configured to look for an LDAP source, then the system is not using LDAP for authentication. If the system is not using LDAP for authentication, this is not applicable. Check for the "bindpw" option being used in the "/etc/ldap.conf" file. # grep bindpw /etc/ldap.conf If an uncommented "bindpw" option is returned, then a cleartext password is in the file, and this is a finding.
Edit the "/etc/ldap.conf" file to use anonymous binding by removing the "bindpw" option.
The telnet service included in the RHEL distribution is part of krb5-workstation. There are two versions of telnetd server provided. The xinetd.d file ekrb5-telnet allows only connections authenticated through kerberos. The xinetd.d krb5-telnet allows normal telnet connections as well as kerberized connections. Both are set to "disable = yes" by default. Ensure that neither is running. Procedure: Check if telnetd is running: # ps -ef |grep telnetd If the telnet daemon is running, this is a finding. Check if telnetd is enabled on startup: # chkconfig --list|grep telnet If an entry with "on" is found, this is a finding.
Identify the telnet service running and disable it. Procedure: Disable the telnet server. # chkconfig telnet off Verify the telnet daemon is no longer running. # ps -ef |grep telnet
Check GRUB for password configuration. Procedure: Check the /boot/grub/grub.conf or /boot/grub/menu.lst files. # grep "password" /boot/grub/grub.conf /boot/grub/menu.lst Check for a password configuration line, such as: password --md5 <password-hash> If the boot loader passwords are not protected using an MD5 hash or stronger, this is a finding.
Consult vendor documentation for procedures concerning the system's boot loader. Configure the boot loader to hash boot loader passwords using MD5 or a stronger hash.
Verify a system integrity baseline exists. The Advanced Intrusion Detection Environment (AIDE) is included in the distribution of RHEL. Other host intrusion detection system (HIDS) software is available but must be checked manually. Procedure: # grep DBDIR /etc/aide.conf If /etc/aide.conf does not exist AIDE has not been installed. Unless another HIDS is used on the system, this is a finding. Examine the response for "database" this indicates the location of the system integrity baseline database used as input to a comparison. # ls -la <DBDIR> If no "database" file as defined in /etc/aide.conf exists a system integrity baseline has not been created, this is a finding. Examine /etc/aide.conf to ensure some form of cryptographic hash (i.e. md5,rmd160,sha256) is used for files. In the default /etc/aide.conf the "NORMAL" or "LSPP" rules which are used for virtually all files DO include some form of cryptographic hash. If the site has defined rules to replace the functionality provided by the default "NORMAL" and "LSPP" rules but DOES NOT include cryptographic hashes, this is a finding. Otherwise, if any element used to define the "NORMAL" and "LSPP" rules has been modified resulting in cryptographic hashes not being used, this is a finding. If any other modification to the default /etc/aide.conf file have been made resulting in rules which do not include cryptographic hashes on appropriate files, this is a finding.
Use AIDE to create a file integrity baseline, including cryptographic hashes, for the system. Configure the /etc/aide.conf file to ensure some form of cryptographic hash (e.g., md5,rmd160,sha256) is used for files. In the default /etc/aide.conf the "NORMAL" or "LSPP" rules which are used for virtually all files DO include some form of cryptographic hash.
Verify a system integrity baseline is maintained. The baseline has been updated to be consistent with the latest approved system configuration changes. The Advanced Intrusion Detection Environment (AIDE) is included in the distribution of RHEL-5. Other host intrusion detection system (HIDS) software is available but must be checked manually. Procedure: # grep DBDIR /etc/aide.conf If /etc/aide.conf does not exist AIDE has not been installed. Unless another HIDS is used on the system, this is a finding. Examine the response for "database" indicates the location of the system integrity baseline database used as input to a comparison. # ls -la <DBDIR> If the no "database" file as defined in /etc/aide.conf a system integrity baseline has not been created, this is a finding. Ask the SA when the last approved system configuration changes occurred. If the modification date of the AIDE database is prior to the last approved configuration change, this is a finding.
Regularly rebuild the integrity baseline, including cryptographic hashes, for the system to be consistent with the latest approved system configuration. Procedure: After an approved modification to the system configuration has been made perform: # aide -u This will update the database.
Check the system for the unnecessary "news" accounts. Procedure: # rpm -q inn If the "inn" is installed the "news" user is necessary and this is not a finding. # grep ^news /etc/passwd If this account exists and "inn" is not installed, this is a finding.
Remove the "news" account from the /etc/passwd file before connecting a system to the network.
Check the system for the unnecessary "gopher" accounts. Procedure: # grep ^gopher /etc/passwd If this account exists, it is a finding.
Remove the "gopher" account from the /etc/passwd file before connecting a system to the network.
Check the system for the unnecessary "ftp" accounts. Procedure: # rpm -q krb5-workstation An ftp server is part of "krb5-workstation". If it is installed the "ftp" user is necessary and this is not a finding. # rpm -q vsftp If the "vsftp" ftp server is installed the "ftp" user is necessary and this is not a finding. # grep ^ftp /etc/passwd If this account exists and no ftp server is installed which requires it, this is a finding.
Remove the "ftp" account from the /etc/passwd file before connecting a system to the network.
If the "xorg-x11-server-Xorg" package is not installed, this is not applicable. For the Gnome screen saver, check the idle_delay setting. Procedure: # gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --get /apps/gnome-screensaver/idle_delay If this does not return 15 or less, this is a finding.
For the Gnome screen saver, set idle_delay to 15. Procedure: # gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --type int --set /apps/gnome-screensaver/idle_delay 15
If the "xorg-x11-server-Xorg" package is not installed, this is not applicable. For the Gnome screen saver, check the lock_enabled flag. Procedure: # gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --get /apps/gnome-screensaver/lock_enabled If this does not return "true", this is a finding.
For the Gnome screen saver, set the lock_enabled flag. Procedure: # gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --type bool --set /apps/gnome-screensaver/lock_enabled true
Verify the system-auth settings are being applied. Procedure: Verify the additional pam.d requirements are in use. The file "/etc/pam.d/system-auth-ac" is auto generated by "authconfig". Any manual changes made to it will be lost next time "authconfig" is run. Check to see if the systems default of the symlink "/etc/pam.d/system-auth" pointing to "/etc/pam.d/system-auth-ac" has been changed. # ls -l /etc/pam.d/system-auth If the symlink points to "/etc/pam.d/system-auth-ac", manual changes cannot be protected. This is a finding. # grep system-auth-ac /etc/pam.d/system-auth The local system-auth file pointed to by "/etc/pam.d/system-auth" must contain "/etc/pam.d/system-auth-ac" for the auth, account, password, and session lines. If it does not then the parameters maintained by "authconfig" will not be applied, this is a finding.
In the default distribution of RHEL "/etc/pam.d/system-auth" is a symlink "/etc/pam.d/system-auth-ac" which is an autogenerated file. When a site adds password requirements a new system-auth-local file must be created with only the additional requirements and includes for auth, account, passwd and session pointing to "/etc/pam.d/system-auth-ac". Then the symlink "/etc/system-auth" is modified to point to "/etc/pam.d/system-auth-local". This way any changes made do not get lost when "/etc/pam.d/system-auth-ac" is regenerated and each program's pam.d definition file need only have "include system-auth" for auth, account, passwd and session, as needed, in order to assure the password requirements will be applied to it.
Check that auditd is configured to audit failed file access attempts. There must be an audit rule for each of the access syscalls that logs all failed accesses (-F success=0) or there must both an "-F exit=-EPERM" and "-F exit=-EACCES" for each access syscall. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S open" | grep -e "-F success=0" # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S open" | grep -e "-F exit=-EPERM" # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S open" | grep -e "-F exit=-EACCES" If an "-S open" audit rule with "-F success" does not exist and no separate rules containing "-F exit=-EPERM" and "-F exit=-EACCES" for "open" exist, then this is a finding.
Edit the audit.rules file and add the following line(s) to enable auditing of failed attempts to access files and programs: either: -a exit,always -F arch=<ARCH> -S open -F success=0 or both: -a exit,always -F arch=<ARCH> -S open -F exit=-EPERM -a exit,always -F arch=<ARCH> -S open -F exit=-EACCES Restart the auditd service. # service auditd restart
Verify auditd is configured to audit failed file access attempts. There must be an audit rule for each of the access syscalls logging all failed accesses (-F success=0) or there must both an "-F exit=-EPERM" and "-F exit=-EACCES" for each access syscall. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S openat" | grep -e "-F success=0" # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S openat" | grep -e "-F exit=-EPERM" # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S openat" | grep -e "-F exit=-EACCES" If an "-S openat" audit rule with "-F success" does not exist and no separate rules containing "-F exit=-EPERM" and "-F exit=-EACCES" for "openat" exist, then this is a finding.
Edit the audit.rules file and add the following line(s) to enable auditing of failed attempts to access files and programs: either: -a exit,always -F arch=<ARCH> -S openat -F success=0 or both: -a exit,always -F arch=<ARCH> -S openat -F exit=-EPERM -a exit,always -F arch=<ARCH> -S openat -F exit=-EACCES Restart the auditd service. # service auditd restart
Verify auditd is configured to audit failed file access attempts. There must be an audit rule for each of the access syscalls logging all failed accesses (-F success=0) or there must both an "-F exit=-EPERM" and "-F exit=-EACCES" for each access syscall. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S truncate" | grep -e "-F success=0" # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S truncate" | grep -e "-F exit=-EPERM" # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S truncate" | grep -e "-F exit=-EACCES" If an "-S truncate" audit rule with "-F success" does not exist and no separate rules containing "-F exit=-EPERM" and "-F exit=-EACCES" for "truncate" exist, then this is a finding.
Edit the audit.rules file and add the following line(s) to enable auditing of failed attempts to access files and programs: either: -a exit,always -F arch=<ARCH> -S truncate -F success=0 or both: -a exit,always -F arch=<ARCH> -S truncate -F exit=-EPERM -a exit,always -F arch=<ARCH> -S truncate -F exit=-EACCES Restart the auditd service. # service auditd restart
Verify auditd is configured to audit failed file access attempts. There must be an audit rule for each of the access syscalls logging all failed accesses (-F success=0) or there must both an "-F exit=-EPERM" and "-F exit=-EACCES" for each access syscall. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S ftruncate" | grep -e "-F success=0" # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S ftruncate" | grep -e "-F exit=-EPERM" # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -e "-S ftruncate" | grep -e "-F exit=-EACCES" If an "-S ftruncate" audit rule with "-F success" does not exist and no separate rules containing "-F exit=-EPERM" and "-F exit=-EACCES" for "ftruncate" exist, then this is a finding.
Edit the audit.rules file and add the following line(s) to enable auditing of failed attempts to access files and programs: either: -a exit,always -F arch=<ARCH> -S ftruncate -F success=0 or both: -a exit,always -F arch=<ARCH> -S ftruncate -F exit=-EPERM -a exit,always -F arch=<ARCH> -S ftruncate -F exit=-EACCES Restart the auditd service. # service auditd restart
Check the system audit configuration to determine if file and directory deletions are audited. # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "rmdir" If no results are returned, or the results do not contain "-S rmdir", this is a finding.
Edit the audit.rules file and add the following line to enable auditing of deletions: -a exit,always -S rmdir Restart the auditd service. # service auditd restart
Check the auditing configuration of the system. Procedure: # cat /etc/audit/audit.rules | grep -i "audit.rules" If no results are returned, or the line does not start with "-w", this is a finding.
The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Add the following lines to the audit.rules file to enable auditing of administrative, privileged, and security actions: -w /etc/audit/audit.rules Restart the auditd service. # service auditd restart
Check the auditing configuration of the system. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "adjtimex" If the result does not contain "-S adjtimex", this is a finding.
The "-F arch=<ARCH>"restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>"restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Add the following lines to the audit.rules file to enable auditing of administrative, privileged, and security actions: -a exit,always -F arch=<ARCH> -S adjtimex Restart the auditd service. # service auditd restart
Check the auditing configuration of the system. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "settimeofday" If the result does not contain "-S settimeofday", this is a finding.
The "-F arch=<ARCH>"restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>"restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Add the following lines to the audit.rules file to enable auditing of administrative, privileged, and security actions: -a exit,always -F arch=<ARCH> -S settimeofday Restart the auditd service. # service auditd restart
Check the auditing configuration of the system. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "stime" # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "settimeofday" If the result does not contain "-S stime" or "-S settimeofday ", this is a finding. The "stime" keyword is only required on systems using an i386 architecture.
The "-F arch=<ARCH>"restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>"restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Add the following lines to the audit.rules file to enable auditing of administrative, privileged, and security actions: -a exit,always -F arch=<ARCH> -S stime (only used for systems using an i386 architecture) -a exit,always -F arch=<ARCH> -S settimeofday (used on all non-i386 architectures such as b64 and x86_64) Restart the auditd service. # service auditd restart
Check the auditing configuration of the system. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "clock_settime" If the result does not contain "-S clock_settime", this is a finding.
The "-F arch=<ARCH>"restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>"restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Add the following lines to the audit.rules file to enable auditing of administrative, privileged, and security actions: -a exit,always -F arch=<ARCH> -S clock_settime Restart the auditd service. # service auditd restart
Check the auditing configuration of the system. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "sethostname" If the result does not contain "-S sethostname", this is a finding.
The "-F arch=<ARCH>"restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>"restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Add the following lines to the audit.rules file to enable auditing of administrative, privileged, and security actions: -a exit,always -F arch=<ARCH> -S sethostname Restart the auditd service. # service auditd restart
Check the auditing configuration of the system. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "setdomain" If the result does not contain "-S setdomain", this is a finding.
The "-F arch=<ARCH>"restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>"restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Add the following lines to the audit.rules file to enable auditing of administrative, privileged, and security actions: -a exit,always -F arch=<ARCH> -S setdomainname Restart the auditd service. # service auditd restart
Check the auditing configuration of the system. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "sched_setparam" If the result does not contain "-S sched_setparam", this is a finding.
The "-F arch=<ARCH>"restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>"restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: A Real Time Operating System (RTOS) provides specialized system scheduling which causes an inordinate number of messages to be produced when the sched_setparam and set_setscheduler are audited. This not only may degrade the system speed to an unusable level but obscures any forensic information which may otherwise have been useful. Unless the operating system is a Red Hat 5 based RTOS (including MRG and AS5300) the following should also be present in /etc/audit/audit.rules -a exit,always -F arch=<ARCH> -S sched_setparam Restart the auditd service. # service auditd restart
Check the auditing configuration of the system. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "sched_setscheduler" If the result does not contain "-S sched_setscheduler", this is a finding.
The "-F arch=<ARCH>"restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>"restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: A Real Time Operating System (RTOS) provides specialized system scheduling which causes an inordinate number of messages to be produced when the sched_setparam and set_setscheduler are audited. This not only may degrade the system speed to an unusable level but obscures any forensic information which may otherwise have been useful. Unless the operating system is a Red Hat 5 based RTOS (including MRG and AS5300) the following should also be present in /etc/audit/audit.rules -a exit,always -F arch=<ARCH> -S sched_setscheduler Restart the auditd service. # service auditd restart
Check the system's audit configuration. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "fchmod" If "-S fchmod" is not in the result, this is a finding
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Edit the audit.rules file and add the following lines to enable auditing of discretionary access control permissions modifications. -a exit,always -F arch=<ARCH> -S fchmod Restart the auditd service. # service auditd restart
Check the system's audit configuration. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "fchmodat" If "-S fchmodat" is not in the result, this is a finding.
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Edit the audit.rules file and add the following lines to enable auditing of discretionary access control permissions modifications. -a exit,always -F arch=<ARCH> -S fchmodat Restart the auditd service. # service auditd restart
Check the system's audit configuration. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "chown" If "-S chown" is not in the result, this is a finding. Additionally, the following rule is required in systems supporting the 32-bit syscall table (such as i686 and x86_64): # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "chown32" If "-S chown32" is not in the result, this is a finding.
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Edit the audit.rules file and add the following lines to enable auditing of discretionary access control permissions modifications. -a exit,always -F arch=<ARCH> -S chown Additionally, the following rule is required in systems supporting the 32-bit syscall table (such as i686 and x86_64): -a exit,always -F arch=<ARCH> -S chown32 Restart the auditd service. # service auditd restart
Check the system's audit configuration. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "fchown" If "-S fchown" is not in the result, this is a finding. Additionally, the following rule is required in systems supporting the 32-bit syscall table (such as i686 and x86_64): # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "fchown32" If "-S fchown32" is not in the result, this is a finding.
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Edit the audit.rules file and add the following lines to enable auditing of discretionary access control permissions modifications. -a exit,always -F arch=<ARCH> -S fchown Additionally, the following rule is required in systems supporting the 32-bit syscall table (such as i686 and x86_64): -a exit,always -F arch=<ARCH> -S fchown32 Restart the auditd service. # service auditd restart
Check the system's audit configuration. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "fchownat" If "-S fchownat" is not in the result, this is a finding.
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Edit the audit.rules file and add the following lines to enable auditing of discretionary access control permissions modifications. -a exit,always -F arch=<ARCH> -S fchownat Restart the auditd service. # service auditd restart
Check the system's audit configuration. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "lchown" If "-S lchown" is not in the result, this is a finding. Additionally, the following rule is required in systems supporting the 32-bit syscall table (such as i686 and x86_64): # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "lchown32" If "-S lchown32" is not in the result, this is a finding.
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Edit the audit.rules file and add the following lines to enable auditing of discretionary access control permissions modifications. -a exit,always -F arch=<ARCH> -S lchown Additionally, the following rule is required in systems supporting the 32-bit syscall table (such as i686 and x86_64): -a exit,always -F arch=<ARCH> -S lchown32 Restart the auditd service. # service auditd restart
Check the system's audit configuration. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "setxattr" If "-S setxattr" is not in the result, this is a finding.
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Edit the audit.rules file and add the following lines to enable auditing of discretionary access control permissions modifications. -a exit,always -F arch=<ARCH> -S setxattr Restart the auditd service. # service auditd restart
Check the system's audit configuration. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "lsetxattr" If "-S lsetxattr" is not in the result, this is a finding.
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Edit the audit.rules file and add the following lines to enable auditing of discretionary access control permissions modifications. -a exit,always -F arch=<ARCH> -S lsetxattr Restart the auditd service. # service auditd restart
Check the system's audit configuration. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "fsetxattr" If "-S fsetxattr" is not in the result, this is a finding.
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Edit the audit.rules file and add the following lines to enable auditing of discretionary access control permissions modifications. -a exit,always -F arch=<ARCH> -S fsetxattr Restart the auditd service. # service auditd restart
Check the system's audit configuration. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "removexattr" If "-S removexattr" is not in the result, this is a finding.
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Edit the audit.rules file and add the following lines to enable auditing of discretionary access control permissions modifications. -a exit,always -F arch=<ARCH> -S removexattr Restart the auditd service. # service auditd restart
Check the system's audit configuration. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "lremovexattr" If "-S lremovexattr" is not in the result, this is a finding.
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Edit the audit.rules file and add the following lines to enable auditing of discretionary access control permissions modifications. -a exit,always -F arch=<ARCH> -S lremovexattr Restart the auditd service. # service auditd restart
Check the system's audit configuration. Procedure: # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "fremovexattr" If "-S fremovexattr" is not in the result, this is a finding.
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Edit the audit.rules file and add the following lines to enable auditing of discretionary access control permissions modifications. -a exit,always -F arch=<ARCH> -S fremovexattr Restart the auditd service. # service auditd restart
Determine if the delete_module syscall is audited. # cat /etc/audit/audit.rules | grep -e "-a exit,always" | grep -i "delete_module" If the result does not contain "-S delete_module", this is a finding.
The "-F arch=<ARCH>" restriction is required on dual-architecture systems (such as x86_64). On dual-architecture systems, two separate rules must exist - one for each architecture supported. Use the generic architectures "b32" and "b64" for specifying these rules. On single architecture systems, the "-F arch=<ARCH>" restriction may be omitted, but if present must match either the architecture of the system or its corresponding generic architecture. The architecture of the system may be determined by running "uname -m". See the auditctl(8) manpage for additional details. Any restrictions (such as with "-F") beyond those provided in the example rules are not in strict compliance with this requirement, and are a finding unless justified and documented appropriately. The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Configure auditing of the delete_module syscalls. Add the following to the "etc/audit/audit.rules" or "etc/audit.rules" file: -a exit,always -S delete_module Restart the auditd service. # service auditd restart
Determine if /sbin/insmod is audited. # cat /etc/audit/audit.rules | grep "/sbin/insmod" If the result does not start with "-w" and contain "-p x", this is a finding.
The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Configure auditing of the /sbin/insmod, files. Add the following to the "etc/audit/audit.rules" or "etc/audit.rules" file: -w /sbin/insmod -p x Restart the auditd service. # service auditd restart
Determine if the /sbin/modprobe file is audited. # cat /etc/audit/audit.rules | grep "/sbin/modprobe" If the result does not start with "-w" and contain "-p x", this is a finding.
The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: -w /sbin/modprobe -p x Restart the auditd service. # service auditd restart
Determine if the /sbin/rmmod file is audited. # cat /etc/audit/audit.rules | grep "/sbin/rmmod" If the result does not start with "-w" and contain "-p x", this is a finding.
The use of audit keys consistent with the provided example is encouraged to provide for uniform audit logs, however omitting the audit key or using an alternate audit key is not a finding. Procedure: Configure auditing of the /sbin/rmmod file. Add the following to the "etc/audit/audit.rules" or "etc/audit.rules" file: -w /sbin/rmmod -p x Restart the auditd service. # service auditd restart
Check the mode of scripts in cron job directories. # ls -lL /etc/cron.daily/ # ls -lL /etc/cron.hourly/ # ls -lL /etc/cron.monthly/ # ls -lL /etc/cron.weekly/ If any cron script has a mode more permissive than 0700, this is a finding.
Change the mode of the cron scripts. # chmod 0700 /etc/cron.daily/* /etc/cron.hourly/* /etc/cron.monthly/* /etc/cron.weekly/*
Check the system for the unnecessary "games" accounts. Procedure: # grep ^games /etc/passwd If this account exists, it is a finding.
Remove the "games" account. Procedure: # userdel games
To verify the "INACTIVE" setting, run the following command: grep "INACTIVE" /etc/default/useradd The output should indicate the "INACTIVE" configuration option is set to an appropriate integer as shown in the example below: # grep "INACTIVE" /etc/default/useradd INACTIVE=35 If it does not, this is a finding.
To specify the number of days after a password expires (which signifies inactivity) until an account is permanently disabled, add or correct the following lines in "/etc/default/useradd", substituting "[NUM_DAYS]" appropriately: INACTIVE=[NUM_DAYS] A value of 35 is recommended. If a password is currently on the verge of expiration, then 35 days remain until the account is automatically disabled. However, if the password will not expire for another 60 days, then 95 days could elapse until the account would be automatically disabled. See the "useradd" man page for more information. Determining the inactivity timeout must be done with careful consideration of the length of a "normal" period of inactivity for users in the particular environment. Setting the timeout too low incurs support costs and also has the potential to impact availability of the system to legitimate users.
To determine how the SSH daemon's "HostbasedAuthentication" option is set, run the following command: # grep -i HostbasedAuthentication /etc/ssh/sshd_config If no line, a commented line, or a line indicating the value "no" is returned, then the required value is set. If the required value is not set, this is a finding.
SSH's cryptographic host-based authentication is more secure than ".rhosts" authentication since hosts are cryptographically authenticated. However, it is not recommended that hosts unilaterally trust one another, even within an organization. To disable host-based authentication, add or correct the following line in "/etc/ssh/sshd_config": HostbasedAuthentication no
Verify neither the "NOPASSWD" option nor the "!authenticate" option is configured for use in "/etc/sudoers" and associated files. Note that the "#include" and "#includedir" directives may be used to include configuration data from locations other than the defaults enumerated here. # egrep '^[^#]*NOPASSWD' /etc/sudoers /etc/sudoers.d/* # egrep '^[^#]*!authenticate' /etc/sudoers /etc/sudoers.d/* If the "NOPASSWD" or "!authenticate" options are configured for use in "/etc/sudoers" or associated files, this is a finding.
Update the "/etc/sudoers" or other sudo configuration files to remove or comment out lines utilizing the "NOPASSWD" and "!authenticate" options. # visudo # visudo -f [other sudo configuration file]
This is N/A for systems that do not have wireless network adapters. Verify that there are no wireless interfaces configured on the system: # ifconfig -a eth0 Link encap:Ethernet HWaddr b8:ac:6f:65:31:e5 inet addr:192.168.2.100 Bcast:192.168.2.255 Mask:255.255.255.0 inet6 addr: fe80::baac:6fff:fe65:31e5/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:2697529 errors:0 dropped:0 overruns:0 frame:0 TX packets:2630541 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:2159382827 (2.0 GiB) TX bytes:1389552776 (1.2 GiB) Interrupt:17 lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 inet6 addr: ::1/128 Scope:Host UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:2849 errors:0 dropped:0 overruns:0 frame:0 TX packets:2849 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:2778290 (2.6 MiB) TX bytes:2778290 (2.6 MiB) If a wireless interface is configured, it must be documented and approved by the local Authorizing Official. If a wireless interface is configured and has not been documented and approved, this is a finding.
Configure the system to disable all wireless network interfaces.