How to control Stuck Thread behavior in Jedox

This site uses cookies. By continuing to browse this site, you are agreeing to our Cookie Policy.

  • How to control Stuck Thread behavior in Jedox

    I had written a Php code which has FOR loops 4 levels deep for generating sample data by randomizing data within certain boundaries.

    It usually takes 4 to 5 seconds to finish.

    Sometimes it gets stuck in infinite loop, where I see command status still showing executing; also I see one of my CPU is consuming 99% of its resource.

    During this period the Jedox web responds like a zombie, There is no graceful exit and only way is to kill the Palo.exe from taskmanager and start the PaloMolapService

    This is a clear behavior of a stuck Thread,

    So my question is how can I control stuck thread behavior in Jedox/Palo - Do I have a way to kill or suspend it ?



    Best regards,

    Noel.

    JEDOX 4 SR1
  • Thanks Jiri,

    If I am able to reproduce this issue, I shall try to close the session and see if it gets terminated.

    BTW, Do you witness these kinds of issues often ? what would be the way to avoid getting the script stuck, it scares me because I am heavily relying on Php functions for my model.
    Can I have a way to make a system level setting of making my Php script DIE if its been running past a time interval - say 10 mins ?

    Best regards,
    Noel.
  • Hi,

    depends what is your code actually doing. And what process is busy palo.exe or apache? I understood that your php code is filling some cube with data randomly. How big dimensions are? One possibility is that it sometimes generates splash requests on large consolidations in more than one dimension? It can make palo.exe busy for some time and use lot of memory. Look at the process monitor to see what process uses CPU and how much memory it holds.

    Jiri
  • I am population only base elements, Estimated 13440 cells will get populated by the script with random data.
    Cube is pretty simple having 7 dimensions with dimensions participating in this code having 12, 56, 4, 4 base elements respectively.
    My system is above average, 4 cores 3.2 GHZ clock speed, 8 GB RAM.

    Yesterday I faced another issue when running a PHP script which calls a function which calls a function which calls a function = 3 functions deep,
    every time I ran the script the PALO CORE server just crashed.
    Solution was to merge the 2 functions to one and call the 3rd function, pretty strange.
    I was unable to get the error by making a simple test case, if I find some bandwidth I shall persue in making a simple test case and report it.


    Best regards,
    Noel.
  • The scenario seems to be simple and shouldn't cause any problems in OLAP. I'm doing such things quite often. I think it could be connected to the second issue with CORE component you described. One of the results of scripts that CORE is not able to handle could be the slow run described at the beginning of this thread in different situation the same problem can cause crash.
    If possible please report the crash. Simple tests case would be best, but even crash dump file can help sometimes.