For demonstrating and verifying the capabilities of PowerFlex replication with SRM for disaster recovery, a two node Oracle RAC cluster running at the primary/protected site is failed over to the recovery site.
Figure 11. Oracle RAC disaster recovery solution architecture
The tests were conducted in the following two scenarios:
In both cases, after the recovery plan execution is complete, the Oracle VMs came up automatically at the recovery site and the database state was also verified. In the planned case, PowerFlex and the SRA will finish all pending replication syncs and then power down the VMs. In an unplanned case, PowerFlex aims to have the latest set of data that is consistent across the volumes in the RCG. It is important to note that the process aims at crash consistency at the storage level but not necessarily application consistency.