VMware vSphere 7.0 Virtual Machine Security Technical Implementation Guide

Shrinking a virtual disk reclaims unused space in it. If there is empty space in the disk, this process reduces the amount of space the virtual disk occupies on the host drive. Normal users and processes (those without root or administrator privileges) within virtual machines have the capability to invoke this procedure. However, if this is done repeatedly, the virtual disk can become unavailable while this shrinking is being performed, effectively causing a denial of service. In most datacenter environments, disk shrinking is not done, so this feature must be disabled. Repeated disk shrinking can make a virtual disk unavailable. The capability to shrink is available to nonadministrative users operating within the VM's guest operating system.

Shrinking and wiping (erasing) a virtual disk reclaims unused space in it. If there is empty space in the disk, this process reduces the amount of space the virtual disk occupies on the host drive. Normal users and processes (those without root or administrator privileges) within virtual machines have the capability to invoke this procedure. However, if this is done repeatedly, the virtual disk can become unavailable while this shrinking is being performed, effectively causing a denial of service. In most datacenter environments, disk shrinking is not done, so this feature must be disabled. Repeated disk shrinking can make a virtual disk unavailable. The capability to wipe (erase) is available to nonadministrative users operating within the VM's guest operating system.

The security issue with nonpersistent disk mode is that successful attackers, with a simple shutdown or reboot, might undo or remove any traces they were ever on the machine. To safeguard against this risk, production virtual machines should be set to use persistent disk mode; additionally, ensure activity within the VM is logged remotely on a separate server, such as a syslog server or equivalent Windows-based event collector. Without a persistent record of activity on a VM, administrators might never know whether they have been attacked or hacked. There can be valid use cases for these types of disks, such as with an application presentation solution where read-only disks are desired, and such cases should be identified and documented.

Setting "isolation.tools.hgfsServerSet.disable" to "true" disables registration of the guest's HGFS server with the host. Application Programming Interfaces (APIs) that use HGFS to transfer files to and from the guest operating system, such as some VIX commands, will not function. An attacker could use this to transfer files inside the guest operating system.

Ensure no device is connected to a virtual machine if it is not required. For example, floppy, serial, and parallel ports are rarely used for virtual machines in a data center environment, and CD/DVD drives are usually connected only temporarily during software installation.

Ensure no device is connected to a virtual machine if it is not required. For example, floppy, serial, and parallel ports are rarely used for virtual machines in a data center environment, and CD/DVD drives are usually connected only temporarily during software installation.

Ensure no device is connected to a virtual machine if it is not required. For example, floppy, serial, and parallel ports are rarely used for virtual machines in a datacenter environment, and CD/DVD drives are usually connected only temporarily during software installation.

Ensure no device is connected to a virtual machine if it is not required. For example, floppy, serial, and parallel ports are rarely used for virtual machines in a data center environment, and CD/DVD drives are usually connected only temporarily during software installation.

By default, more than one user at a time can connect to remote console sessions. When multiple sessions are activated, each terminal window receives a notification about the new session. If an administrator in the VM logs in using a VMware remote console during their session, a nonadministrator in the VM might connect to the console and observe the administrator's actions. Also, this could result in an administrator losing console access to a VM. For example, if a jump box is being used for an open console session and the administrator loses connection to that box, the console session remains open. Allowing two console sessions permits debugging via a shared session. For the highest security, allow only one remote console session at a time.

In a virtual machine, users and processes without root or administrator privileges can connect or disconnect devices, such as network adaptors and CD-ROM drives, and can modify device settings. Use the virtual machine settings editor or configuration editor to remove unneeded or unused hardware devices. To use the device again, prevent a user or running process in the virtual machine from connecting, disconnecting, or modifying a device from within the guest operating system. By default, a rogue user with nonadministrator privileges in a virtual machine can: 1. Connect a disconnected CD-ROM drive and access sensitive information on the media left in the drive. 2. Disconnect a network adaptor to isolate the virtual machine from its network, which is a denial of service. 3. Modify settings on a device.

If enabled, a VM can obtain detailed information about the physical host. The default value for the parameter is FALSE. This setting should not be TRUE unless a particular VM requires this information for performance monitoring. An adversary could use this information to inform further attacks on the host.

The VM console enables a connection to the console of a virtual machine, in effect seeing what a monitor on a physical server would show. The VM console also provides power management and removable device connectivity controls, which could allow a malicious user to bring down a VM. In addition, it impacts performance on the service console, especially if many VM console sessions are open simultaneously.

When accessing the VM console, the guest operating system must be locked when the last console user disconnects, limiting the possibility of session hijacking. This setting only applies to Windows-based VMs with VMware tools installed.

vMotion migrations in vSphere 6.0 and earlier transferred working memory and CPU state information in clear text over the vMotion network. As of vSphere 6.5, this transfer can be transparently encrypted using 256-bit AES-GCM with negligible performance impact. vSphere enables encrypted vMotion by default as "Opportunistic", meaning that encrypted channels are used where supported but the operation will continue in plain text where encryption is not supported. For example, when vMotioning between two hosts, encryption will always be used. However, because 6.0 and earlier releases do not support this feature, vMotion from a 7.0 host to a 6.0 host would be allowed but would not be encrypted. If the encryption is set to "Required", vMotions to unsupported hosts will fail. This must be set to "Opportunistic" or "Required".

The ESXi hypervisor maintains logs for each individual VM by default. These logs contain information including but not limited to power events, system failure information, tools status and activity, time sync, virtual hardware changes, vMotion migrations and machine clones. Due to the value these logs provide for the continued availability of each VM and potential security incidents, these logs must be enabled.

The ESXi hypervisor maintains logs for each individual VM by default. These logs contain information including but not limited to power events, system failure information, tools status and activity, time sync, virtual hardware changes, vMotion migrations, and machine clones. By default, the size of these logs is unlimited, and they are only rotated on vMotion or power events. This can cause storage issues at scale for VMs that do not vMotion or power cycle often.

The ESXi hypervisor maintains logs for each individual VM by default. These logs contain information including but not limited to power events, system failure information, tools status and activity, time sync, virtual hardware changes, vMotion migrations, and machine clones. By default, 10 of these logs are retained. This is normally sufficient for most environments, but this configuration must be verified and maintained.

VMDirectPath I/O (PCI passthrough) enables direct assignment of hardware PCI functions to VMs. This gives the VM access to the PCI functions with minimal intervention from the ESXi host. This is a powerful feature for legitimate applications such as virtualized storage appliances, backup appliances, dedicated graphics, etc., but it also allows a potential attacker highly privileged access to underlying hardware and the PCI bus.

Fault Tolerance log traffic can be encrypted. This could contain sensitive data from the protected machine's memory or CPU instructions. vSphere Fault Tolerance performs frequent checks between a primary VM and secondary VM so the secondary VM can quickly resume from the last successful checkpoint. The checkpoint contains the VM state that has been modified since the previous checkpoint. When Fault Tolerance is turned on, FT encryption is set to "Opportunistic" by default, which means it enables encryption only if both the primary and secondary host are capable of encryption.