Product Documentation

Test Failover

Oct 16, 2015

Failover testing is an essential component in disaster recovery planning. You can use the Disaster Recovery wizard to perform non-disruptive testing of your disaster recovery system. During a test failover operation, all the steps are the same as for failover, but instead of be started after they have been recovered to the DR site, the VMs and vApps are placed in a paused state. At the end of a test failover operation, all VMs, vApps and storage recreated on the DR site are automatically removed.

After initial DR configuration, and after you make significant configuration changes in a DR-enabled pool, we recommend that you verify that failover still works correctly by performing a test failover.

To perform a test failover of VMs and vApps to a secondary site:

  1. In XenCenter, select the secondary pool, and on the Pool menu, click Disaster Recovery to open the Disaster Recovery wizard.
  2. Select Test Failover and then click Next.
    Note: If you use Fibre Channel shared storage with LUN mirroring to replicate the data to the secondary site, before you attempt to recover data, mirroring must be broken so that the secondary site has Read/Write access.
  3. Select the storage repositories (SRs) containing the pool metadata for the VMs and vApps that you want to recover. By default, the list on this wizard page shows all SRs that are currently attached within the pool. To scan for more SRs, choose Find Storage Repositories and then select the storage type to scan for:
    • To scan for all the available Hardware HBA SRs, select Find Hardware HBA SRs.
    • To scan for software iSCSI SRs, select Find Software iSCSI SRs and then enter the target host, IQN and LUN details in the dialog box.
    When you have selected the required SRs in the wizard, click Next to continue.
  4. Select the VMs and vApps that you wish to recover then click Next to progress to the next wizard page and begin failover prechecks.
  5. Before beginning the test failover process, the wizard performs a number of pre-checks, for example, to ensure that all the storage required by the selected VMs and vApps is available.
    1. Check that storage is available. If any storage is missing, you can click Attach SR on this page to find and attach the relevant SR.
    2. Check that HA is not enabled on the target DR pool. To avoid having the same VMs running on both the primary and DR pools, HA must be disabled on the secondary pool to ensure that the recovered VMs and vApps are not started up automatically by HA after recovery. To disable HA on the secondary pool, you can simply click Disable HA on the this page. (If HA is disabled at this point, it will be enabled again automatically at the end of the test failover process.)
    Resolve any issues on the pre-checks page, and then click Failover to begin the test failover.
  6. A progress page is displayed showing whether recovery was successful for each VM and vApp. Failover may take some time depending on the number of VMs and vApps you are recovering, as the metadata for the VMs and vApps are recovered from the replicated storage, the VMs and vApps are re-created in the DR pool, the SRs containing the virtual disks are attached to the re-created VMs. The recovered VMs are placed in a paused state: they will not be started up on the secondary site during a test failover.
  7. After you are satisfied that the test failover was performed successfully, click Next in the wizard to have the wizard clean up on the DR site:
    • VMs and vApps that were recovered during the test failover will be removed.
    • Storage that was recovered during the test failover will be detached.
    • If HA on the DR pool was disabled at the prechecks stage to allow the test failover to take place, it will be enabled again automatically.
    The progress of the cleanup process is displayed in the wizard.
  8. Click Finish to close the wizard.