Call Monitoring Kernel Panics the PBX Server

Discussion in 'General' started by izrunas, Sep 24, 2009.

  1. izrunas

    Joined:
    Mar 4, 2009
    Messages:
    94
    Likes Received:
    0
    I am having two ongoing issues:

    1) Call monitoring does not record all calls.

    2) During high call volume loads, the machine will kernel panic and die.

    We determined through test-and-check that the call monitoring was the only difference between when the system would and would not crash during routine operation.

    We determined that the system will crash under a load of 18-23 simultaneous calls (all of which are set to record). (System connects to a single PRI circuit)

    We previously installed a separate physical hard drive "just for recordings" to decrease demands on the regular hard drive.

    The server runs an average memory usage of 400-600 Mb, with 4Gb total on-board memory.

    CPU load does not generally exceed 25% and swap is never used.

    Elastix on that machine has been updated via yum to current levels.


    Anyone have an idea?

    Thanks,
    Eriks
     
  2. juanchonk

    Joined:
    Jun 8, 2009
    Messages:
    16
    Likes Received:
    0
    Hi, I had a problem very similar to yours, it was that the hard drive was failing, to detect it I used nmon and checked the writing on the Hard Drive when the call load was high (before Panic, obviously).
     
  3. Mirko87

    Joined:
    Oct 20, 2008
    Messages:
    128
    Likes Received:
    0
    It could be the HD... If it doesn't work good, it could cause Kernel Panic...
    Try to change it... or make up a raid system to be more safe...

    Regards,
    Mirko
     
  4. izrunas

    Joined:
    Mar 4, 2009
    Messages:
    94
    Likes Received:
    0
    The first hard drive still checks out fine (the one we used to use) and then the new hard drive we added (the one where recordings go now) also checks out fine. They're both SATA drives. They're both brand new.

    Original configuration: 1 SATA HDD

    Current configuration: 2 SATA HDDs, new/bigger one mounts at /var/spool/
     
  5. Chilling_Silence

    Joined:
    Sep 23, 2008
    Messages:
    488
    Likes Received:
    0
    What filesystem is the 2nd HDD?
    So it happened when there was only 1 HDD?
    What mount options are you using ?
     
  6. izrunas

    Joined:
    Mar 4, 2009
    Messages:
    94
    Likes Received:
    0
    So we tried turning off the call monitoring/recording. The system still crashed this morning. The only other change we made was to remove a cron job we had running to clear the cache periodically which we subsequently re-added.

    The machine was reformatted between the one-drive-only config and the one where the second drive was mounted at /var/spool.

    Here's the screen shot of the panic (a stitch-together of two images from the same screen): [​IMG]
     
  7. Mirko87

    Joined:
    Oct 20, 2008
    Messages:
    128
    Likes Received:
    0
    Mmmhmh... I think this is an HD problem...

    Have you solved it?

    Regards,

    Mirko
     
  8. izrunas

    Joined:
    Mar 4, 2009
    Messages:
    94
    Likes Received:
    0
    We have not yet solved it. The drives are both Western Digital and are brand new. Any chance it could be a SATA driver issue? I've seen where others have issues with Asus hardware and SATA drives when running Linux.
     
  9. Mirko87

    Joined:
    Oct 20, 2008
    Messages:
    128
    Likes Received:
    0
    Mmm... At home I have 2 PC, a Desktop and a Notebook... Both two are Asus, but I'never had problems during the last year... I hope to continue on this trend... :)
     

Share This Page