4 Replies Latest reply: Nov 14, 2012 4:16 PM by 866532 RSS

    EAS crashes regularly and calcs, outlines etc. very slow to open!!

    mike.r
      Since upgrading to Oracle EPM version 11.1.2.1 we have experienced multiple EAS crashes per day. Also, calc scripts, Outlines (especially in view mode), filters, report scripts take a very long time to open. As the day goes on they get slower and slower. We have seen some calc scripts take up to 3 minutes to open!!!

      Here are some examples of the errors we receive (I replaced the server name with the word SERVER):

      <1350997718151> <BEA-000000> <Internal error: java.net.SocketException: socket write error: Connection aborted by peer
      (I can post the whole thing if you would like to see it)

      ####<Oct 23, 2012 7:08:38 AM MDT> <Error> <HTTP> <SERVER> <EssbaseAdminServices0> <[ACTIVE] ExecuteThread: '20' for queue: 'weblogic.kernel.Default (self-tuning)'> <<WLS Kernel>> <> <> <1350997718167> <BEA-101020> <[ServletContext@11309671[app:EAS module:easconsole path:/easconsole spec-version:2.5 version:11.1.2.0]] Servlet failed with Exception
      java.lang.IllegalStateException: Response already committed

      ####<Oct 23, 2012 7:08:38 AM MDT> <Notice> <Diagnostics> <SERVER> <EssbaseAdminServices0> <[ACTIVE] ExecuteThread: '18' for queue: 'weblogic.kernel.Default (self-tuning)'> <<WLS Kernel>> <> <> <1350997718198> <BEA-320068> <Watch 'UncheckedException' with severity 'Notice' on server 'EssbaseAdminServices0' has triggered at Oct 23, 2012 7:08:38 AM MDT. Notification details:
      WatchRuleType: Log
      WatchRule: (SEVERITY = 'Error') AND ((MSGID = 'WL-101020') OR (MSGID = 'WL-101017') OR (MSGID = 'WL-000802') OR (MSGID = 'BEA-101020') OR (MSGID = 'BEA-101017') OR (MSGID = 'BEA-000802'))
      WatchData: DATE = Oct 23, 2012 7:08:38 AM MDT SERVER = EssbaseAdminServices0 MESSAGE = [ServletContext@11309671[app:EAS module:easconsole path:/easconsole spec-version:2.5 version:11.1.2.0]] Servlet failed with Exception
      java.lang.IllegalStateException: Response already committed


      I have contacted Oracle support and they offered 3 solutions:

      1) Archive the files in the tmp folder as specified for the WebLogic EAS. Sometimes it runs out of space. We have 500 GB available space. (E:\Oracle\Middleware\user_projects\domains\EPMSystem\servers\EssbaseAdminServices0\tmp) NEXT.

      2) Increase the Heap settings .xmx values in the registry to 768m. We also spoke to Interrel and Oracle at the Calgary road show and they recommended setting both the .xms and .xmx values to 1024. Specifically JVMOption42, 43, 45 and 46 under the HyS9eas registy tree. (HKey_Local_machine\softwware\Hyperion Solutions\EssbaseAdminServices0\HyS9eas) We did this. Stopped and started EAS webservice. Still getting EAS crashes. Still getting slower and slower calc openings etc as the day goes on. NEXT.

      3) Last recommendation was to get up to patch level .104 for EAS server, Essbase RTC, Essbase server, Studio console, Studio server and APS. This is a lot of work for us. Must first apply to TQA. Setup VM's for the clients to test with (assume client versions of xl addin, smartview etc. will need to be updated). Complete UAT. Apply patches to PRD, Rescript the clients for desktop deployment etc. etc. Before we go down this road, there must be other people that have experienced this same problem. Did you have to do this or were there other solutions that fix the problem? If this was the solution I would like confirmation from others that it did work, before we put all our users through this.

      Thanks so much,
      Mike

      Edited by: Mike on Nov 8, 2012 4:06 PM