Database unexpectedly not working after cluster split and succeeding cluster merge
Details
| Detail name | Value |
|---|---|
| Changelog Number | 6291 |
| Type | Bug |
| Status | Resolved |
| Affected Versions | EXASOL 6.0.0, EXASolution 5.0.0 |
| Fix Versions | Exasol 6.1.3 |
| Resolution Date | 2019-03-13 |
Background
A cluster split is caused by a network outage that separates an Exasol cluster into two or more separated segments. Quorum mechanics are used to prevent having multiple "live" database instances in such a case. When the network problem is fixed, the cluster should self-repair and merge the segments again.
Problem
Under rare - and unspecific - circumstances, the following messages could occur endlessly in the monitoring logs after a cluster split (i.e. network outage between cluster nodes) and a succeeding cluster merge:
Process exited: SqlProcess ... exited after signal 6.
At the same time, the database system would be unusable for issuing further SQL statements.
Required Action
To get back into working condition, the database must be restarted manually via EXAoperation.
