Details
-
Improvement
-
Status: Resolved
-
Low
-
Resolution: Fixed
-
None
-
None
Description
In trunk 1.2.x, NEWS.txt says the following :
- The hints schema was changed from 1.1 to 1.2. Cassandra automatically snapshots and then truncates the hints column family as part of starting up 1.2 for the first time. Additionally, upgraded nodes will not store new hints destined for older (pre-1.2) nodes. It is therefore recommended that you perform a cluster upgrade when all nodes are up.
But this means that :
1) as you do a rolling restart upgrade, your new nodes don't store hints for old nodes as they restart for the upgrade.
2) the last old node being upgraded will be not have any hints stored for it while it is upgrading.
These two facts mean that you need to run a cluster-wide repair after upgrading in order to be as consistent as you would expect if hinted handoff were working.
I propose modifying NEWS.txt to read :
- The hints schema was changed from 1.1 to 1.2. Cassandra automatically snapshots and then truncates the hints column family as part of starting up 1.2 for the first time. Additionally, upgraded nodes will not store new hints destined for older (pre-1.2) nodes. It is therefore recommended that you perform a cluster upgrade when all nodes are up. Because hints *will* be lost, a cluster-wide repair (with -pr) is required after upgrade of all nodes.