Nodes must be restored in a specific order and some restoration scenarios may require manual configuration changes.
Depending on the tool used for restoring, you can restore the virtual machines to the same host, a different host on the same data center, or a different host on a target remote data center. See Changing Configurations After Restoration
Verify that the restored nodes are in the powered off state.
Verify that the cluster instances are powered off before restoring the cluster to a new site.
Verify that no split-brain behavior occurs when the same IP addresses and FQDNs are used on the recovery site.
Verify that no one is accidentally using a partially working cluster on the primary site.
- Restore the master node first before restoring worker nodes.
- Restore worker nodes in any order.
- (Optional) Restore the forwarders if configured.
Be sure the vRealize Log Insight server (the master node and all the worker nodes in a cluster setup) are restored before restoring the forwarders.
- Restore any recovered agents.
What to do next
When restoring a vRealize Log Insight cluster, if the same IP addresses are used, verify that all restored node IP addresses and FQDNs are associated with their original counterparts.
For example, the following scenario would fail. In a three-node cluster with nodes A, B, and C, node A is restored with IP address B, node B is restored with IP address C, and node C is restored with IP address A.
If the same IP addresses are used for only a subset of restored nodes, verify that for these nodes, all restored images are associated with their original IP addresses.
Most backup restoration and disaster recovery tools provide a monitoring view for watching the progress of the restoration operations for failures or warnings. Take appropriate actions on any identified problems.
If manual configuration changes are required before the site can be fully restored, follow the guidelines in the Changing Configurations After Restoration.
When a successful restoration is finished, perform a spot check of the cluster that was restored.