Problem Summary: How to Recover if a Target Database Fails?
Answer:
Depending on the type of failure and/or whether the target database is a clustered database (e.g. Oracle RAC) the following actions need to be performed:
1.When the target database is a clustered database (e.g. Oracle RAC), the target database will fail-over to another node and HVR should recover automatically. This requires that HVR is set-up properly for cluster fail-over, but this is described elsewhere. Depending on HVR's activity at this time an error message may be written to the log file which will then be reported by an HVRMaint email as well.
2.When (the connection to) the target database gets restored and the target database did not did not lose any data, HVR will recover automatically using a retry. An error message will be written to the log file which will then be reported by an HVRMaint email as well.
3.The target database comes back on-line and has lost data but gets restored to a point a recent point in time (the archive / journal files containing these changes are still available). HVR will not be able to recover. A manual step is necessary: Run HVR Initialize with a capture rewind until a time stamp from before the crash. This will cause HVR to replay the changes from that time stamp against the target database
4.The target database is completely lost or it is restored to a point which is too far back in time. HVR will not be able to recover. A manual step is necessary: Perform an Online Refresh to re-sync source and target database. After the refresh has completed successfully the replication jobs can be started.