When a catastrophic event occurs on a DMA failover system running Cassandra. Restoring from the backup package would require reinstalling the Primary DMA with cassandra and loading the backup. The backup package would contain configuration information for both nodes in the configuration. When rejoining the failover agent to this DMA what manual actions would need to be taken to smoothly add this back into a failover state?
I would assume doing a 'nodetool remove node' of the backup node would avoid conflicts of IPs already existing in the system.peers and cassandra.yaml files?
Hi Ryan,
The current backup/restore mechanism in DataMiner takes indeed the system tables from Cassandra. Which means that the setup is also included in the backup files. As a consequence Cassandra is restored with configuration of the DB setup as well.
Towards the future this backup mechanism will change so that the backup will be independent from your DB setup (or Cassandra setup).