changing the geo-replication crawl from changelog to hybrid
Issue
Changelog crawl is going into FAULTY state because of the wrong and invalid operations that have been saved in the changelog (such as creation of entries with NULL gfid).
The geo-replication daemon constantly tries to sync such invalid operations to the slave. But since the brick process does not allow creation of a entry with NULL gfid, sync fails. It was introduced as part of the bug #bug1474178
And geo-replication goes to FAULTY state.
Environment
- RHGS-3.1.x
- RHS-3.0
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.