Managing Protection During VMware Host Maintenance
When a host machine requires VMware maintenance, at least for the duration of the maintenance, to ensure continuous protection:
For a host machine on the protected site – Remove affinity rules for protected virtual machines on the host that requires maintenance and vMotion these machines to any other host with a VRA installed. Shut down the VRA before starting the host maintenance. After the host maintenance power on the VRA.
For a host machine on the recovery site – VRA data and recovery volumes maintained by the VRA on the host should be moved to another machine, by changing the destination host for all the virtual machines being recovered to that host, as described in
Changing a Recovery VRA. Shut down the VRA before starting the host maintenance. After the host maintenance power on the VRA.
Note: Changing the VPG default host does not change the hosts for virtual machines already included in the VPG. These virtual machine hosts must be changed directly by configuring the individual virtual machines in the VPG definition.
To enable VMware host maintenance for a VRA both protecting and recovering virtual machines:
1. Remove affinity rules for protected virtual machines on the host that requires maintenance and vMotion these machines to any other host with a VRA installed.
2. Change the host for all virtual machines in VPGs recovering to this VRA to another host as described in
Changing a Recovery VRA.
3. Wait for any synchronization to complete.
4. Shutdown the VRA on the host manually in order to enable the host to enter VMware Maintenance.
5. Enter VMware Maintenance for the host.
Note: Do not migrate powered-off virtual machines, if prompted to.
6. Remove the host from the cluster: Place it under the datacenter entity rather than the cluster entity.
7. Perform required maintenance, for example, upgrading the host.
8. Exit VMware maintenance mode.
9. Power on the VRA.
10. Wait for the Zerto Virtual Manager to connect to the local VRAs. You can monitor the alerts to determine when the connections have been established.
11. Add the host back in to the cluster.