Details
-
Bug
-
Status: Resolved
-
Normal
-
Resolution: Duplicate
-
None
-
None
-
None
-
Normal
Description
When the AntiEntropService sends the snapshot repair request, it sets up a callback in an ExpiringMap. If the time it takes for the snapshot exceeds the RPC timeout, the callback will expire from the map and the snapshot responses will be dropped. The repair then gets stuck forever blocking at the snapshotLatch. It's not even possible to kill the repair with forceTerminateAllRepairSessions()
This is likely fixed in 2.0 since that part of the code is completely rewritten.
Attachments
Issue Links
- duplicates
-
CASSANDRA-7560 'nodetool repair -pr' leads to indefinitely hanging AntiEntropySession
- Resolved
-
CASSANDRA-6747 MessagingService should handle failures on remote nodes.
- Resolved