Table Lock on SIBOWNER & SIBOWNERO causing all JVMs in the cluster to stop responding

Body

From time to time, a Maximo environment stops responding. When the environment is reviewed, the database has locks on SIB-related tables. Here, the issue is not primarily because of SIB tables. It occurs because the SIB tables are the victim of Maximo system freezing and holding locks on the SIB tables.

If Maximo integration components are writing to JMS Queues that use database storage, then SIB tables are created and used in the database. Whenever Maximo writes data or reads data from the JMS queues, internally, the SIB tables are used by the JMS engine.

Maximo system lockup situations are mostly due to JVM running out of memory or being low on memory and the garbage collection cycle is working hard to release the memory. Now, if a system lockup situation is present, then the JVM is paused. Locks held on the SIB tables remain locked until the JVM releases them, which can take several minutes. Sometimes when the JVM crashes, locks on the SIB tables are not released until the database lock timeout occurs.  Hence when the Maximo system appears to be hanging or crashing, the SIB table locks are visible to the DBA.  But, in fact the SIB table locks are a victim of the JVM freeze or crash.

In most cases, JVM freezes and crashes are due to memory leaks or high usage memory issues. The system needs to be analyzed and fixed to resolve the issue and once the system returns to normal operation, the SIB locking situation goes away.
Here is a useful link on how the SIB tables and locks work.

Webcast replay: WebSphere Application Server - Service Integration Bus Messaging Engine Data Store Connectivity Problems and Solutions