In rare scenarios, database does not restart after a node failure
Details
| Detail name | Value |
|---|---|
| Changelog Number | 14682 |
| Type | Bug |
| Status | Open |
| Affected Versions |
Background
When a failed node is detected, the node failover process starts:
- A database that was running on the failed node is stopped.
- A reserve node replaces the failed node.
- Afterward, the database is automatically restarted and is online again.
Then the failed node can be fixed without impacting the running database.
Description
In rare cases, the automatic node failover mechanism fails to replace the failed node with the reserve node. This happens when the node failure cannot be propagated to the other nodes in the cluster in time.
Workaround
Manually restart the database
Fix
The automatic failover mechanism handles such scenarios.
