3 Replies Latest reply: Mar 5, 2013 6:38 AM by Per Lindström RSS

    Tuxedo server process getting restarted


      Tuxedo server process are getting restarted while processing the request and it is not logging any errors while restarting.


      135847 TuxGeneralServer.16370.1510907088.0: LIBTUX_CAT:476: WARN: Server 800/8111: client process 22062: lost message
      135847TuxGeneralServer.16370.1510907088.0: LIBTUX_CAT:477: WARN: SERVICE=TadN2Tux MSG_ID=0 REASON=server died
      135847TuxShmServer.25760.3494362320.0: INFO: DebugRegServer(): Pid=16370, SrvId=8111, SrvName=TuxGeneralServer, SrvGrp=APPGRP, GrpNo=800

      Could you please help me to find the cause for restart of the process. Is there any way to take that perticular process dump or if we enable complete tuxedo server logs ,will it log any details?

      Thanks in adavance.


      Edited by: R828 on Jan 20, 2013 2:30 PM
        • 1. Re: Tuxedo server process getting restarted
          Bobfinan - Oracle-Oracle
          The reason the server process is restarted is because you have it configured to be in the ubb configuration file with the RESTART=Y paramater setting:
          "RESTART = {Y | N}
          Specifies whether or not AOUT is restartable. The default is N. If server migration is specified, RESTART must be set to Y. Note that a server terminated with a SIGTERM signal cannot be restarted; it must be rebooted."

          Bob Finan

          To find out the reason your application server process died you will need to debug the core file created. If no core was generated you will need to check the OS system limits and allow the core to be generated.
          Additionally you should check the error messages in the ULOG, the server stdout, and the server stderr prior to the crash.
          • 2. Re: Tuxedo server process getting restarted
            Hi Bob Finan,

            Tanks a lot for your reply.

            Yes .RESTART falg is set to Y in ubb.

            No core files were generated this time. Nothing was reported in ulogs apart from the lines which i copied in my post.
            OS settings are fine , i can see core files for other process. How can we log the process dump and suggest me are there any other ways to find the reason for process died?

            • 3. Re: Tuxedo server process getting restarted
              Per Lindström
              Any chance of a call to exit() somewhere in the code executed by that particular service?

              Calling exit() from within code executed by a Tuxedo service is a no-no. In these cases you really want to do tpreturn(TPEXIT,...) instead.

              Sometimes code you call from a library may contain calls to exit() you're not aware of.

              Hope this helps,