Performing the resync operation for virtual machines
When disaster strikes on a production data center, the recover operation is invoked to start the resiliency groups on the recovery data center.
Since the production data center is not working, the data replication between the two sites does not happen. After the production site is back up and running, you need to prepare the production site for the next failover or for a migration operation. This preparation includes cleaning up any residue and resuming the replication from the recovery to the production site.
Use the Resync operation on the Resiliency Platform console to automate these steps for the required resiliency groups. This operation cleans up the residue which includes stopping physical and virtual machines, unregistering virtual machines, unmounting file systems, datastores, etc.
In Microsoft Failover Cluster environments, the Resync operation may fail in the first step to cleanup the virtual machine residue. You can manually cleanup the virtual machine residue and proceed.
See Manually cleaning up virtual machines.
Consider the following if you have configured your assets for recovery to vCloud Director without adding the VMware vCenter server or Hyper-V server:
You need to perform the resync operation after Migrate or Recover. In the Activities panel, if the workflow is in Paused state for "Resync Replication" subtask, then you need to manually start the physical and virtual machines. Ensure that the physical and virtual machines boot from PXE OS of the replication gateway that is configured as PXE server. You can verify the virtual machines boot progress from vCenter or Hyper-V console by checking the virtual machine console. When the boot is complete, click so that the workflow proceeds with synchronzing data from the disks on recovery datacenter to those on the on-premises data center. After Resync operation is complete, do not shut down the virtual machines, otherwise the subsequent Migrate or Recover operations fail.
Performing the resync operation
- Prerequisites
If the target (recovery) data center is on-premises, and the last performed operation was recover, then you may need to restart the Hyper-V server or the ESX server. Although the Resync operation cleans up any residue on the source data center before resuming replication, there could be some residues that can be cleaned up only by restarting the hypervisors.
For the replication technology HPE 3PAR Remote Copy, ensure that for VMware virtual machines the config.vpxd.filter.hostRescanFilter value is set to false.
If the status of the virtual machine on the recovery data center is not correctly displayed, then you need to refresh the cloud discovery or the virtualization server discovery.
Prerequisites for resync operation of physical workloads
The resiliency group must be configured for DR and migrated to the recovery datacenter.
Ensure that PXE boot server is configured on the on-premises Replication Gateway.
Ensure to set network boot (PXE boot) as the first boot priority in the system BIOS.
If you are not using 3rd party DHCP server, then configure DHCP server on the PXE Boot server that is configured on the on-premises Replication Gateway.
- Navigate to Assets (navigation pane) > Resiliency Group(s) tab
- Double-click the resiliency group to view the details page. Click Resync.
- In the Resync panel, select the production data center name from the drop-down list, and click Next.
If the Resync operation fails, check to know the reason and fix it. You can then launch the operation. The operation restarts the resync workflow, it skips the steps that were successfully completed and retries those that had failed.
Do not restart the workflow service while any workflow is in running state, otherwise the operation may not work as expected.
After completing recover from cloud and resync, the resiliency group details page shows entries for deleted or unavailable virtual machines on cloud data center. To remove these stale entries, after resync is complete, edit the resiliency group with Edit Configuration intent. You may submit the wizard without making any changes.