RQM NoClassDefFoundError
Hello,
We are currently migrated from 4.0.2 CLM to 4.0.6 on WAS 8.5 and DB2.
We are facing an issue with QM, where QM application is unreachable with java lang NoClassDefFoundError
entry in the qm.log. This happens very frequently.
[qm: AsynchronousTaskRunner-3 @@ 21:04] ERROR com.ibm.team.datawarehouse - java.lang.NoClassDefFoundError: com.ibm.team.datawarehouse.common.internal
And other errors too
[ WebContainer : 27] ERROR net.jazz.ajax/AjaxFramework - java.lang.Exception: Evaluation failed for dojo message bundle: dojo.cldr.nls.en.number
Clearing qm.war from WAS temp directory and WAS restart will bring back the application.
We did some test runs to understand the problem in test env , we restarted WAS and observed the QM application status..(3 to 4 restarts) observed that qm is unreachable twice and available after a restart again with out clearing qm.war.
And in second trial cleared qm.war and redeployed . Did 3 ,4 restarts to check the qm status.
But still the issue remains the same....un reachable once.
So what could be the possible cause for this issue. Even we ug graded WAS feom 8.5 to 8.5.5.2 which did not help us to resolve the issue.
Kindly let us know resolutions if any body come across the issue....
Thanks in advance...
Regards
Ashwath Garugula
We are currently migrated from 4.0.2 CLM to 4.0.6 on WAS 8.5 and DB2.
We are facing an issue with QM, where QM application is unreachable with java lang NoClassDefFoundError
entry in the qm.log. This happens very frequently.
[qm: AsynchronousTaskRunner-3 @@ 21:04] ERROR com.ibm.team.datawarehouse - java.lang.NoClassDefFoundError: com.ibm.team.datawarehouse.common.internal
And other errors too
[ WebContainer : 27] ERROR net.jazz.ajax/AjaxFramework - java.lang.Exception: Evaluation failed for dojo message bundle: dojo.cldr.nls.en.number
Clearing qm.war from WAS temp directory and WAS restart will bring back the application.
We did some test runs to understand the problem in test env , we restarted WAS and observed the QM application status..(3 to 4 restarts) observed that qm is unreachable twice and available after a restart again with out clearing qm.war.
And in second trial cleared qm.war and redeployed . Did 3 ,4 restarts to check the qm status.
But still the issue remains the same....un reachable once.
So what could be the possible cause for this issue. Even we ug graded WAS feom 8.5 to 8.5.5.2 which did not help us to resolve the issue.
Kindly let us know resolutions if any body come across the issue....
Thanks in advance...
Regards
Ashwath Garugula
Accepted answer
Hi Ashwath,
It took a while, but we had to collaborate with a few Java teams to pin this down.
We have now documented the problem in tech note: http://www-01.ibm.com/support/docview.wss?uid=swg21684531
There are many symptoms, but the key check would be to analyse the following logs for the following strings:
It took a while, but we had to collaborate with a few Java teams to pin this down.
We have now documented the problem in tech note: http://www-01.ibm.com/support/docview.wss?uid=swg21684531
There are many symptoms, but the key check would be to analyse the following logs for the following strings:
-
Check any application server logs or application logs: jts.log, qm.log, ccm.log, SystemOut.log, or catalina.out
-
Check for a
NullPointerException
, where the callstack contains any class in the package:
org.eclipse.emf.ecore.impl
-
Check native_stderr.log, a Tomcat output or a Javacore, for the Java version.
The string: 26sr2 , indicates Java 6.0.1 SR 2. - The string: jvmxa6460sr16 indicates Java 6.0 SR 16.
For WebSphere 8.5:
-
Java 7.0.0 SR6
PI19107: SHIP SDK 7 SR7 FOR WSAS V8.5.0.X AND V8.5.5.X
-
Java 6.0.1 SR8:
PI19108: SHIP SDK 626 SR8 FOR WSAS V8.5.0.X AND V8.5.5.X
2 other answers
Hi Paul,
Thanks for all for writing your valuable solution....We have an open PMR on this issue where the RQM application outage with intermittent errors like NoClassDefFoundError, NullPointerException, IntilaizationException errors.
And Rational support team investigated the problem in various verticals(Java debugging...etc) and concluded for one exception error NullPointerException is problem with EMF module. and as you suggested they asked us to apply an JAVA 7.0 SR6 IFIX for WAS 8.5.
And seen this error is resolved after fix apply. But still not sure the root cause for NoClassDefFound error yet.
I will update the solution moment we get the solution for support.
Once again thank you for your analysis and time on JAVA debugging and other dependencies.
Ashwath........
Thanks for all for writing your valuable solution....We have an open PMR on this issue where the RQM application outage with intermittent errors like NoClassDefFoundError, NullPointerException, IntilaizationException errors.
And Rational support team investigated the problem in various verticals(Java debugging...etc) and concluded for one exception error NullPointerException is problem with EMF module. and as you suggested they asked us to apply an JAVA 7.0 SR6 IFIX for WAS 8.5.
And seen this error is resolved after fix apply. But still not sure the root cause for NoClassDefFound error yet.
I will update the solution moment we get the solution for support.
Once again thank you for your analysis and time on JAVA debugging and other dependencies.
Ashwath........
Comments
Donald Nong
Aug 22 '14, 12:56 a.m.There is not much information for others to comment on. What does it mean "unreachable"? What errors do you get on the client side? Are you certain of the cause-effect relationship between the unreachableness and the errors that you quoted? At any rate, it looks like the QM workspace gets corrupted given the information above (by a long shot). If it's indeed the case, that will likely be a system level (file system) issue rather than application level.
Paul Ellis
JAZZ DEVELOPER Aug 22 '14, 3:57 a.m.Hi Ashwat,
I am currently investigating a very similar issue for another client. As soon as I get to the root cause I shall update this forum post.
If you can contact Rational Client Support though, I would recommend it, as I think this issue is going to require additional work that is less appropriate for a forum exchange.
Cheers,
Paul