Database Deadlock in SAP Commerce 1811 with Percona XtraDB Cluster

[blakeberg]

Hello experts,

Our long running problem is that we have an issue with percona xtradb cluster and SAP Commerce 181100P_1 with deadlock occuring at initialization. The log is

 Caused by: org.springframework.dao.CannotAcquireLockException: PreparedStatementCallback; SQL [UPDATE tasks SET p_runningonclusternode = ? WHERE p_runningonclusternode = ? AND PK = ? AND hjmpTS = ? ]; Lock wait timeout exceeded; try restarting transaction; nested exception is com.mysql.jdbc.exceptions.jdbc4.MySQLTransactionRollbackException: Lock wait timeout exceeded; try restarting transaction

This happens on our test machine (actually one node hybris with mysql connector 5.1.47, three mysql 5.7.24 node in percona xtradb cluster) for our new SAP Commerce 1811 development. I have found issues in hybris experts, sap knowledge base and hybris jira patch / ecp that are similar but not the same and solution to our problem.

Have anyone the same or similar problem with deadlocks especially at initialization or deadlocks comming from update task statements?

Best regards.

[blakeberg · Accepted solution]

Ok, the official sap support has a solution for this Thank you for your information.

We have identified the exception to be related with a feature in newer versions that enables jdbc interceptors. As a workaround for the initialization, please set the property jdbc.interceptor.enabled=false in your local.properties, then restart hybris & retry the initialization.

Related links: https://jira.hybris.com/browse/ECP-3542 https://jira.hybris.com/browse/ECP-3500