We are using Golden Gate for replication between our Primary Database server and our DR database server. We are having to shutdown our DR server for a few hours as the SAN is being retired. In this context, I am trying to figure out the sequence of steps to follow before the DR shutdown and after DR restart to maintain proper Golden Gate replication without ABENDING the Extracts/Replicat or any other issues on the Primary/DR linux servers. Any input would be appreciated.
Asked
Active
Viewed 2,586 times
1 Answers
0
Step 1: Extract's need not be touched on primary database (Since nothing is changing on primary database extract's should be up and extracting all data changes to the trail files while DR activity is going on).
Step 2: Make sure that there are no long running session(usually batch processes, which introduce lag into the golden gate processes). This is more of a sanity check to make sure that processes are not abending or need special attention once DR activity is done and processes are being brough up).
Step 3: Shutdown pump on primary database site. If there are multiple pumps pumping the data to multiple target databases make sure that you shutdown only the pumps pointing to the target database where DR activity is happening.
Step 4: Verify that pumps are down and all the in flight data(data changes that are pending to be applied since pumps were shutdown are fully applied on target database). You can use LAG command to verify if there is any data pending to be applied on DR site.
Step 5: Shutdown replicat process on the DR database site.
Step 6: Turnover of DR maintenance.
Once DR activity is complete and DR database is up follow these steps.
Step 1: Start replicats on DR site. Step 2: Start pumps on primary site. Step 3: Verify that the lag goes down to zero. Monitor ggserr.log for any issues until lag goes down zero.
Thanks Rahul
Rahul Malireddy
- 31
- 8