JAVA / 4T RAC crash on check-out

[evolve]

All,

We have TC10.1.2.2 (64bit) and on some machines The RAC crashes when users checkout anything.

The first thing they checkout (item, item rev, dataset, ...) the RAC exits and they get this error message below.

There is nothing in any log-file, not in client log or server syslog.

Maybe 5-8 users affected, all other 70 users have no issue.

We did some tests and it doesn't look like Teamcenter or java issue. Because, on the same laptop some user has the issue and other user does not. 

Any idea's?

How can i enable more extensive logging? Is there a JAVA-log or how to enable it?

thanks

br

fred

[RandyEllsworth]

I've seen similar behavior when the OS Name (and AD) didn't match the Tc name case-wise when TcSSO was used. Windows ignores "case" but Java does not and can run into trouble when trying to write the FCC/TcRAC log. Review the TcRAC log for errors.

[rupeshs]

Hi,

There may be different reason of crash.

1) Too many items opened by user--> I obeserved users keep items opened in TC. If there are too many items opened, you may observer JAVA error.

2) Temp\cache files not cleared --> Need to delete Temp\cache files periodically(may be biweekly). You may create a batch file as attahed and provide to end users. Once you clear those, open items will also get closed.

3) JAVA heap size--> Check heap size paramaters set in otwportal.bat. We have set it to below :

-XX:PermSize=64m -XX:MaxPermSize=128m

BR,

Rupesh

[evolve]

@RandyEllsworth

We don't use SSO.It is standard login 4T. The windows username is the same as the TC username but password is managed in TC.

The TcRAC log contains some lines but no error message. Also on a working RAC we have the same lines in the log.

WARN  2018-06-26 10:10:22,636 - Windows 8 () - win32.win32.x86_64 - null - Java(TM) SE Runtime Environment (1.7.0_51-b13) - 3.8.0.v20120529-1548
WARN  2018-06-26 10:10:29,730 - Using abandoned clinetID TCLoginJob.CPCAD3.05324.02:CPCAD3.08100.01
WARN  2018-06-26 10:10:36,418 - FCC Interface initialized in 297 ms.
WARN  2018-06-26 10:10:38,683 - org.eclipse.core.runtime - org.eclipse.ui - 0 - Warnings while parsing the handlers from the 'org.eclipse.ui.handlers' extension point.
WARN  2018-06-26 10:10:38,683 - org.eclipse.core.runtime - org.eclipse.ui - 0 - There should only be one expression element: plug-in='com.teamcenter.rac.cme.collaborationcontext', id='org.eclipse.ui.edit.delete', whenElementName='enabledWhen'
WARN  2018-06-26 10:10:38,683 - org.eclipse.core.runtime - org.eclipse.ui - 0 - There should only be one when element: plug-in='com.teamcenter.rac.cme.mpp', id='org.eclipse.ui.file.print', whenElementName='enabledWhen'
WARN  2018-06-26 10:10:40,543 - org.eclipse.core.runtime - org.eclipse.osgi - 1 - NLS unused message: noContent.MSG in: com.teamcenter.rac.viewer.messages_locale
WARN  2018-06-26 10:10:40,543 - org.eclipse.core.runtime - org.eclipse.osgi - 1 - NLS missing message: UNKNOWN_NUMBER in: com.teamcenter.rac.viewer.messages_locale
WARN  2018-06-26 10:10:40,543 - org.eclipse.core.runtime - org.eclipse.osgi - 1 - NLS missing message: OBJECTS in: com.teamcenter.rac.viewer.messages_locale
WARN  2018-06-26 10:10:40,543 - org.eclipse.core.runtime - org.eclipse.osgi - 1 - NLS missing message: FOUND_MSG in: com.teamcenter.rac.viewer.messages_locale
WARN  2018-06-26 10:10:47,683 - ------ Interop Module: warn log level ------

@rupeshs

1. There are no items open by the users. everything is checked-in and closed. 
2. Cleaning all Temp/cache has no effect. This is our standard procedure when there are problems with the RAC
3. I added in portal.bat the argument "-XX:PermSize=64m" in the commandline for teamcenter.exe, but also same result.

We are looking now more in the direction of graphics drivers (it's a guess). Since the problem occurs on the same type of hardware (HP X2 1012 G2), all other RAC that are working fine have other hardware.

But this looks maybe a bit far fetched.

what you think?

br

Fred

[RandyEllsworth]

I don't think its the graphics driver since the warnings indicate an NLS issue which is really talking about *.properties files. Strange that its client specific though as an issue with properties files would normally be global. Unless you distribute custom properties files after the install (post task)?

[naveengm]

Where are they checking out, is it shared drive? Check permissions on that folder?

Try to change checkout directory for those users having problem to a different local folder. 

Edit->Options->General->CheckIn-CheckOut

Regards,

NaveenGM

[evolve]

Thank you for the replies.

@RandyEllsworth

No, we don't distribute any properties files after the install. And i checked multiple stations, we have these warnings for all users, working and crashing RAC.

I don't have much experience in this area, any idea which *.properties files i need to check?

@naveengm

The checkout preferences are locked at site level and is set to local folder.

I tested already to unlock these and disable all, but still crash.

br

Fred

[jonathan_morris]

Is FMS_HOME set correctly for the users that get the failures?

[evolve]

Yes, FMS_HOME is set in startup scripts

[RandyEllsworth]

@evolve, I'm starting to get confused if this problem is all users, specific machines, or specific users?

Please try this:
1. Log into Teamcenter as a user/machine who is known to have this problem.
2. Review Help > About > More... and record the tcserver syslog name and location.
3. Perform a checkout and observe the Tc crash.
4. Attach the tcserver syslog recorded in step 2 for our evaluation.