queuemetrics will eat all the cpus after a long running

[coolname]

in top,i see java process use 393% cpu after a long time running,should i restart queuemetrics everyday?

[patrick]

I have the same problem.  Java mem has been increased but still no change.  this happens at the same time every working day.
I have noticed that there are almost 200 sessions within mysql, these processes are usually terminated after 10 to 30 seconds but when we have 'issues' the process stack builds and builds until we have too many sessions for mysql.

What kind of diagnostics do I take? mysql, java or qm itself?

Any help would be nice.

[QueueMetrics]

Taking a thread dump is the only way to know what's actually going on.
Most likely, you are sending QM too many requests (or it's too slow to fetch data to process them, but it's the same thing) so it lags behind. Try avoiding having too many realtime screens open.

[coolname]

through threaddump，i get the below process eats too many cpus


"http-8080-4" daemon prio=10 tid=0x09e01c00 nid=0x192e runnable [0x0b969000]
   java.lang.Thread.State: RUNNABLE
   at it.loway.tpf.common.containers.LContext.eliminaPrefisso(LContext.java:237)
   at it.loway.tpf.transaction.servlets.LowayTransactionController.pulisceChiavi(LowayTransactionController.java:886)
   at it.loway.tpf.transaction.servlets.LowayTransactionController.serveRequest(LowayTransactionController.java:557)
   at it.loway.tpf.transaction.servlets.LowayTransactionController.serveRequestWrapper(LowayTransactionController.java:387)
   at it.loway.tpf.transaction.servlets.LowayTransactionController.doPost(LowayTransactionController.java:238)
   at javax.servlet.http.HttpServlet.service(HttpServlet.java:637)
   at javax.servlet.http.HttpServlet.service(HttpServlet.java:717)
   at org.apache.catalina.core.ApplicationFilterChain.internalDoFilter(ApplicationFilterChain.java:290)
   at org.apache.catalina.core.ApplicationFilterChain.doFilter(ApplicationFilterChain.java:206)
   at org.apache.catalina.core.StandardWrapperValve.invoke(StandardWrapperValve.java:233)
   at org.apache.catalina.core.StandardContextValve.invoke(StandardContextValve.java:191)
   at org.apache.catalina.core.StandardHostValve.invoke(StandardHostValve.java:127)
   at org.apache.catalina.valves.ErrorReportValve.invoke(ErrorReportValve.java:102)
   at org.apache.catalina.core.StandardEngineValve.invoke(StandardEngineValve.java:109)
   at org.apache.catalina.connector.CoyoteAdapter.service(CoyoteAdapter.java:291)
   at org.apache.coyote.http11.Http11Processor.process(Http11Processor.java:859)
   at org.apache.coyote.http11.Http11Protocol$Http11ConnectionHandler.process(Http11Protocol.java:602)
   at org.apache.tomcat.util.net.JIoEndpoint$Worker.run(JIoEndpoint.java:489)
   at java.lang.Thread.run(Thread.java:662)

[coolname]

after change into centos 6.2 64bit，mem 8GB,the problem has gone

[QueueMetrics]

Nice to see it's fixed.

[coolname]

after run about 20 days,it occured agagin

[QueueMetrics]

Do you run a programmed restart of the server?

[coolname]

no ,i manually restart the qm

[QueueMetrics]

We usually suggest doing a programmed restart just to be on the safe side of things.
BTW, in the upcoming release, we spotted a threading issue that would cause high CPU spikes; this is fixed in 12.05, so it would be wise to upgrade.

[coolname]

really? when would it realeased?

[QueueMetrics]

Next week or so. We are finalizing changes and doing final testing.