Before you shut down, reboot, or disconnect a host that is a member of a Virtual SAN cluster, you must put the host in maintenance mode.

When working with maintenance mode, consider the following guidelines:

  • When you place an ESXi host in maintenance mode, you must select a data evacuation mode, such as Ensure accessibility or Full data migration.

  • When any member host of a Virtual SAN cluster enters maintenance mode, the cluster capacity automatically reduces as the member host no longer contributes storage to the cluster.

  • A virtual machine's compute resources might not reside on the host that is being placed in maintenance mode, and the storage resources for virtual machines might be located anywhere in the cluster.

  • The Ensure accessibility mode is faster than the Full data migration mode because the Ensure accessibility migrates only the components from the hosts that are essential for running the virtual machines. When in this mode, if you encounter a failure, the availability of your virtual machine is affected. Selecting the Ensure accessibility mode does not reprotect your data during failure and you might experience unexpected data loss.

  • When you select the Full data migration mode, your data is automatically reprotected against a failure, if the resources are available and the Number of failures to tolerate set to 1 or more. When in this mode, all components from the host are migrated and, depending on the amount of data you have on the host, the migration could take longer. With Full data migration mode, your virtual machines can tolerate failures, even during planned maintenance.

  • When working with a three-host cluster, you cannot place a server in maintenance mode with Full data migration. Consider designing a cluster with four or more hosts for maximum availability.

Before you place a host in maintenance mode, you must verify the following:

  • If you are using Full data migration mode, verify that you have enough hosts and capacity available in the cluster to meet the Number of failures to tolerate policy requirements.

  • Verify that you have enough flash capacity on the remaining hosts to handle any flash read cache reservations. You can run the vsan.whatif_host_failures RVC command to analyze the current capacity utilization per host and whether a single host failure could make the cluster run out of space, and impact the cluster capacity, cache reservation, and cluster components. For information about the RVC commands, see the RVC Command Reference Guide.

  • Verify that you have enough capacity devices in the remaining hosts to handle stripe width policy requirements, if selected.

  • Make sure you have enough free capacity on the remaining hosts to handle the amount of data that must be migrated from the host entering maintenance mode.