Hi,
If a failover (backup) agent has been down for a few weeks, is there something one need to take into consideration before bringing it back up again? Will the backup agent receive all necessary data from the main agent when the backup agent comes back online again and sync connection is established again?
Running DataMiner 10.5 with a self-hosted Cassandra/OpenSearch cluster.
Second question: Does the backup agent communicate with the Cassandra cluster or is it just the main servers that does ?
Hi,
It depends a bit on what you mean by self-hosted Cassandra/OpenSearch.
If you're using dedicated clustered storage (shared Cassandra and OpenSearch clusters), then the database synchronization is handled by Cassandra/OpenSearch themselves, not by the DataMiner failover synchronization.
In that case, I'd expect the backup agent to reconnect normally after being brought back online. Once it's up:
- Wait for a midnight sync to occur.
- Check the Failover status in DataMiner Cube.
- Review the logs for any SLErrors related to Failover or synchronization.
- Verify there are no Cassandra/OpenSearch connectivity or health issues.
Regarding your second question, it depends on the database architecture:
If you have a single Cassandra cluster (single DC), both the active and backup agent will connect directly to that cluster.
If you have a multi-DC Cassandra deployment, each agent will typically connect to its preferred DC, while the Cassandra clusters handle replication/synchronization between the DCs themselves.
So, in short: if you're using shared clustered storage, the databases should take care of catching up themselves. After bringing the backup agent online, I'd mainly validate that the midnight sync completes successfully, the Failover status is healthy, and there are no sync-related errors in the logs