Follow these steps if a node is offline and data is missing/corrupt:

************This example is from dbcoma01aprd to dbcoma01bprd

** If you are using this process to fix a results table, you need to stop rcvMonRes.sh after the resource group is frozen.

1. Force the cluster resource group "pgmaster01" to the appropriate database node if necessary:

2. Freeze the group on the machine because we will be stopping services:

3. Now stop bucardo so replication doesn't try to take place while we dump

4. Disable cleanup cron jobs in /etc/crontab

5. Now dump any tables that are latent:

6. Copy the dump to the destination server:

7. Import the data:

8. Delete the dump

9. On the target db server, cleanout the bucardo delta table for the gnw database

10. Start bucardo and verify it is working

11. Enable the bucardo clean cronjobs in /etc/crontab listed in step 4

12. Un-freeze the cluster

*** Make sure all dump files get deleted to conserve space

Data Inconsistency Recovery (last edited 2011-08-15 20:42:09 by Bryce Camp)