As a provider, you restore VMware Cloud Director Availability appliances in the cloud site on appliances with the same version, appliance role and network settings and by using a single appliance backup .enc file, extracted from the locally downloaded backup archive.
In the management interface of a restored Replicator Service instance, on the System Tasks page the Reload destination tasks run indefinitely or fail with a Lock acquisition timed out for object: 'H4-id' for each replication that is not present since restoring the backup. To manually delete these replications, click the Emergency Recovery page, select them then click Delete.
- To restore a single appliance in the cloud site, depending on whether restoring on a newly deployed, or restoring in-place over the backed-up appliance:
- When deploying a new appliance, power off the existing backed-up appliance in the site with the same role.
- When restoring in-place over the backed-up appliance, without deploying a new appliance, do not power it off.
- To restore several of the appliances in the cloud site, but not all, follow the same restore order as with restoring the entire site. For more information, see Backing up and restoring in the Cloud Director site.
- To restore an entire backed-up cloud site, restore the same number of appliances with matching appliance roles. For example, to restore a site consisting of a Tunnel Appliance, a Cloud Director Replication Management Appliance, and a couple of Replicator Appliance instances, you must restore a Tunnel Appliance, a Cloud Director Replication Management Appliance, and a couple of Replicator Appliance instances by following the restore order bellow.
- One tunnel-backup_id.tar.bz2.enc appliance backup file for firstly restoring the Tunnel Appliance.
- One cloud-backup_id.tar.bz2.enc appliance backup file for then restoring the Cloud Director Replication Management Appliance.
- One or more replicator-backup_id-IP_Address.tar.bz2.enc appliance backup files for lastly restoring each Replicator Appliance instance in the site.
To restore multiple appliances, repeat this procedure multiple times, according to the restore order and restore the appliances in the cloud site by using the appropriate appliance backup file for each appliance role, as extracted from the backup archive.
Prerequisites
- Verify that VMware Cloud Director Availability 4.3 or later is installed in the cloud site for in-place restore and for restore of a single or several cloud appliances, but not all appliances in the site.
- Verify that you have the backup password and that you locally extract the backup archive cloud-backup-product.version.build-site_name-date-timestampUTC.tar.bz2 file, resulting in several .enc appliance backup files, listed below.
- Verify that the following settings of the cloud appliance for restoring exactly match the backed-up appliance.
- Version
- Appliance role
- Network settings
- Verify that before restoring on a newly deployed appliance, the existing backed-up appliance in the site with the same role is powered off.
Caution: Restoring on a newly deployed appliance while the existing backed-up appliance in the site with the same role is operational might corrupt the replications.However, for VMware Cloud Director Availability 4.3 or later, when restoring an appliance in-place, where the backed-up appliance matches the appliance on which you restore, do not power it off.
Procedure
Results
- The Tunnel Appliance is restored from the backup.
- The Cloud Director Replication Management Appliance is restored from the backup.
- All of the Replicator Appliance instances are restored from the backup.
- After restore, there might be a misalignment between the replication settings stored in the database and the ones loaded from the backup file. As a result, you might see RPO violations, differing numbers of instances, and others, that you can resolve by reconfiguring the affected replications for reapplying their replication settings.
- After restoring, if an RPO violation is present, the replication might be missing from the source Replicator Service. This situation might happen when the source site is restored to a point in time when the replication is not yet started, leading to not working synchronization. As a workaround, you can attempt manually synchronizing the replication. If the synchronization task fails with SourceReplicationNotFound, fail over the replication, stop the replication, then deactivate the replication services for that virtual machine in the source ESXi host, see KB https://kb.vmware.com/s/article/2106946. Finally, start a new replication with a seed virtual machine.
- After evacuating a datastore, all backups taken priorly cannot restore the replications. Take a backup every time the replications are moved from one datastore to another to ensure restoring is successful. For information about datastore evacuation, see Evacuate the replications data from a datastore.
What to do next
You can perform replication workflows. After confirming that the restored appliances are operational, if you restored on a newly deployed appliances, you can decommission the backed-up appliances that are powered off.