Cognos Analytics

Cognos Analytics

Connect, learn, and share with thousands of IBM Cognos Analytics users! 

 View Only
  • 1.  Lot's of errors in cognosserver.log

    Posted Sat September 21, 2019 04:32 AM
    Hi All,

    When I look into the cognosserver.log file I see a lot of error's:

    Index dir 'D:\Program Files\ibm\cognos\analytics\data\search\collections\cm\data\index/' of core 'cm' is already locked.
    The most likely cause is another Solr server (or another solr core in this server) also configured to use this directory;
    other possible causes may be specific to lockType: simple

    What is this error and what should I do to fix this ?

    Chris



    ------------------------------
    Chris Bos
    ------------------------------

    #CognosAnalyticswithWatson


  • 2.  RE: Lot's of errors in cognosserver.log

    Posted Mon September 23, 2019 07:42 AM
    It would help to know what version of CA you have.
    I would try rebuilding the indexes
    1. Stop the services
    2. Delete the files, subdirs of
      • 11.x <install location> \ data \ search \ indices
      • 11.1.2+ <install location> \ data \ search \
    3. Restart services


    ------------------------------
    NORBERT Bracke
    ------------------------------



  • 3.  RE: Lot's of errors in cognosserver.log

    Posted Mon September 23, 2019 08:23 AM
    Hi Norbert,

    Sorry, your right, it helps.
    It's IBM Cognos Analytics 11.1 R3.
    I had a issue that the search wasn't working, so I deleted the '<install location> \ data \ search \ indices' files/subdirs.
    Restarted the machine.
    Search is now working, but I see still the errors.

    Kind Regards,

    Chris


    ------------------------------
    Chris Bos
    ------------------------------



  • 4.  RE: Lot's of errors in cognosserver.log

    Posted Mon September 23, 2019 08:33 AM

    You do not mention stopping the server before deleting the directory. It is possible that the problematic state was reached again before you restarted the machine.

    We started hitting this problem with 11.1.1 quite a bit, but never figured out exactly what was happening.
    Our deployment pipeline in the affected environments (DEV/TEST) would always go through a cycle of uninstallation and reinstallation of the Cognos server, with the timestamp of the messages starting exactly at that time, making us suspect a bug where the server was told to stop and not quite stopping before we started the new installation.

    Our "solution"/workaround was to add a step after stopping the server to detect and forcibly "kill -9 ..." any eventual straggling Cognos process before starting the reinstallation process.



    ------------------------------
    Denilson Nastacio
    ------------------------------



  • 5.  RE: Lot's of errors in cognosserver.log

    Posted Mon September 23, 2019 12:28 PM
    Hi Denilson ,

    True, but for me it's a default process to stop Cognos always whatever I am doing.
    Before modify any file and or delete any directory, I always stop Cognos as step one.

    My current client is Using Windows, so I always stop Cognos Configuration before I change anything and reboot the Windows Server before start Cognos again.

    Still I see the error messages in the cognosserver.log and I would like to understand why and what is going on.

    Thanks.

    Chris







    ------------------------------
    Chris Bos
    ------------------------------



  • 6.  RE: Lot's of errors in cognosserver.log

    Posted Mon September 23, 2019 03:18 PM
    Edited by System Admin Fri January 20, 2023 04:50 PM

    To be clear, the issue we believe to have found is with Cognos stop not actually stopping every Cognos process, but reporting back to the operating system as stopped.

    We run on Linux, so it was easy enough to list all processes owned by Cognos a few seconds after issuing:

    ${COGNOS_BIN_PATH}/cogbootstrapservice" -stop ...

    and then issued `kill -9 ...` against the process identifier of any cognos process still running.

    We also logged the situations where processes were still found after requesting Cognos to stop, and noticed it would not happen every time.

    And to reemphasize, we are not sure these straggler processes were the actual root cause of the Solr error messages, but the messages went away after implementing this change. Maybe we just changed the timing of the restarts long enough to prevent the actual root cause.



    ------------------------------
    Denilson Nastacio
    ------------------------------



  • 7.  RE: Lot's of errors in cognosserver.log

    Posted Tue September 24, 2019 04:51 PM
    Hi Denilson,

    Thanks again for this reply.

    I think there are now 2 issues mixed-up within this post:

    1: I see a lot of same errors (-some folder index locks-) within the cognosserver.log file 
    2: You faced an issue Cognos is not completely stopped

    My question if someone can explaine my error lines within the cognosserver.log file is still valid.

    1 - My client had an issue: the search was not working anymore
    2 - I found some solutions how to solve this
    3 - I stopped Cognos with the Cognos Configutation
    4 - I rebooted the Server (only Cognos Analytics was installed and was NOT started by default, but manualy)
    5 - I deleted the folders / files what was told in the forum posts (Cognos Analytics search is not working anymore)
    6 - I rebooted the Server again to be sure nothing is running (background or not) on the Server
    7 - I started Cognos Analytics using Cognos Configuration
    8 - The Search is working again ( needed some time to update the internal index update)
    9 - I looked into the cognosserver.log file

    My Client search issue is solved, but I would like to understand the error lines.
    (I rebooted the Server. so all Cognos processes are killed and there are no locks in my case)

    Kind Regards,

    Chris Bos











    ------------------------------
    Chris Bos
    ------------------------------



  • 8.  RE: Lot's of errors in cognosserver.log

    Posted Thu September 26, 2019 08:20 AM

    I don't know the exact cause. My theory for our problem with similar log entries at the time was the combination of (1) server processes not completely stopped and (2) our deployment pipeline reinstalling the product while those straggling processes were still running.

    It seems possible that the messages have more than one combination of causes. At this point you may be better off reporting the issue to Cognos support since they can open what they call a "collaboration case" and get the development team involved in those unusual cases.



    ------------------------------
    Denilson Nastacio
    ------------------------------