Details
-
Bug
-
Status: Open (View Workflow)
-
Major
-
Resolution: Unresolved
-
10.6.18
-
None
-
None
-
None
-
Debian 11.9
Description
Hello,
We have a 3 nodes galera cluster. I would like to optimize all database .
One my first node i changed the wsrep_OSU_method to RSU ( SET GLOBAL wsrep_OSU_method="RSU"; ) .
The node as attempted turned into "Desynced" status.
During the optimize, we noticed in our app logs hundreds of " java.sql.BatchUpdateException: (conn=326928) Duplicate entry '10873291-1' for key 'PRIMARY'
at org.mariadb.jdbc.MariaDbStatement.executeBatchExceptionEpilogue(MariaDbStatement.java:327) "
In our monitoring tool, the graph confirmed the node in RSU mode hadn't received queries during the RSU mode => Normal behaviour ( between 13h40 and 14h10 ) => 14h10 i manually stopped the optimize.
So, why the cluster generates hundred of duplicate entries while the optimize is only local ( on the first node) . Our users were impacted, it's not just logs.
I read official documentation and others... but don"t find any potential explanation
Thank you