Details
Description
with configuration of:
thread_handling = pool-of-threads
a cluster node may remain hanging during last steps of handling normal shutdown.
There must be active connections in the node, while shutdown is started, and the shutdown operator apparently believes that there are still active connections, although all connections have been closed already.
Attachments
Issue Links
- blocks
-
MDEV-35777 MariaDB 10.11 Hangs during Shutdown
-
- Closed
-
- duplicates
-
MDEV-35777 MariaDB 10.11 Hangs during Shutdown
-
- Closed
-
Reason for the hanging during shutdown is now known and a potential fix has been developed. Next step is to prepare a deterministic test for the issue and run full cluster testing on the fix proposal