Failure to automatically restart database
Details
| Detail name | Value |
|---|---|
| Changelog Number | 10925 |
| Type | Bug |
| Status | Open |
| Affected Versions |
Background
The database is in a healthy state if all nodes are visible to one another. If this not the case, a node may not be visible because of network issues or a node has crashed. If network issues are responsible, the node should re-join the cluster once connectivity is established.
Description
In some cases, small network outages are misinterpreted as a crashed node, which causes the node not to rejoin the cluster. The effect is that the database does not restart automatically like expected.
Workaround
If this happens, the database has to be started manually after the network interruption is resolved.
Fix
Update to a version of Exasol that includes the fixes EXASOL-2670 and EXASOL-2802, which include improvements that make it less likely that this issue is triggered.
