BGMS Restarting Continuesly & Publishing Queue are Stuck.

[Shashwath]

I'm using Windchill 11.1 M020 CPS 30.

Any Suggestions / Solution for below Issue

Case:CS54402 (2nd Point)

One or more transactions execute a large volume of SQL retrieving small numbers of rows each time, but cumulatively results in a large amount of Method Server memory being used

• A log file will show a series, usually dozens of 'StatementCache' lines like the one below for the SocketThread12 thread.
• The thread name will vary but will usually be named SocketThread##, RMI TCP Connection(###) or SimpleTaskDispatcher###.  'hits' in the line below is a rough estimate of the number of SQL statements sent to the database for processing
  • 'RMI TCP Connection': generally a user UI action
  • SimpleTaskDispatcher: often an EPM related operation
  • SocketThread: usually an InfoEngine task
• Example of output: [SocketThread12] wt.pom.statementCache - StatementCache: wt.util.Cache%618734817 [size=50, count=48, hits=239727, misses=84, aged=0]
  • If there an OutOfMemoryError generated in the log and thread which writes the stack matches the thread generating a large volume of StatementCache lines, the stack is often the root cause of the method server outage
  • A series of Thread dumps when the MethodServer is struggling with memory is very helpful in aiding the diagnosis of root cause (regardless of the suspected root cause)

[HelesicPetr]

Hi @Shashwath 

What is processed by queue ? Publishing ? many publish jobs?

Then increase the memory for the background method server

Or if the issue is cased by publish scheduler plan then schedule the job to non working hours.

PetrH. 

[Shashwath · Accepted solution]

Hello @HelesicPetr 

"DoJobs" are processed under PublisherQueue1 , 2 and so on. 

We have increased the memory too.

Main issue is that the workers are not getting jobs from PublisherQueue & BGMS-WVS is restarting.

[Shashwath]

I'm trying to understand the solution for the case number mentioned above for the 2nd point.

[HelesicPetr]

Hi @Shashwath 

There is no general solution.

You need to investigate what is casing the issue from the logs.

BGMethod server log, and BG GC.log.

The article describes that there should be interesting information that can help you to solve your issue,.

PetrH