hardware issues on license node can cause DB to stop and fail to restart

Details

Detail name Value
Changelog Number 9629
Type Bug
Status Open
Affected Versions Exasol 6.1.0, Exasol 6.2.0, Exasol 7.0.0

Issue

When the license node of a cluster hangs for a short amount of time, it is in some circumstances possible that

  1. A database running on the other nodes is stopped erroneously.
  2. The state of the database is mistakenly shown as running despite the above.
  3. The database can not be restarted via the regular interface.

This bug has been observed when a hard drive on the license server is hanging for a short amount of time but might also be triggered by other circumstances.

Indication

In the case of the symptoms described above being caused by this issue, the database/DWAd log service service in EXAoperation's Monitoring tab should show messages such as this under the NOTICE priority:

NOTICE: Got synchronized event (…) from node with ID 1 that has source revision 2222, but my revision is 1111.

Workaround

Restarting the Cluster Services on all nodes via the Nodes tab in EXAoperation should restore normal operation and allow a restart of the database.