Deadlock in the main cluster caused by unexpected termination/crash of a worker cluster
Details
| Detail name | Value |
|---|---|
| Changelog Number | 31433 |
| Type | Bug |
| Status | Open |
| Affected Versions | Exasol 8.0.0 |
Description
If a worker cluster crashes while holding a FlushLock, the lock can remain stale on the main cluster, blocking transaction cleanup and any subsequent database operations indefinitely.
Workaround
No workaround is available. Shut down the worker cluster, then the main cluster, and restart them; this is the only way to release the lock.
Fix
The fix ensures exception-safe release and removes locks held by dead worker clusters.
