Can’t start monitor daemon

Clovertech Forums Read Only Archives Cloverleaf Cloverleaf Can’t start monitor daemon

  • Creator
    Topic
  • #52816
    Jim Beall
    Participant

      5.6 on HP UX

      I’ve managed to bung up a non-production site to the point where I can’t get the monitord to start.  I’m not sure exactly what I did but I couldn’t get a process to start.  In the midst of attempting to get that resolved, I’ve gotten this site to the point where the monitord won’t start; I’m seeing this in hcimonitord.err:

      [msi :msi :ERR /0:  hcimonitord:11/22/2011 15:33:15] oilMsemInit: semget failed, errno = 28

      Semaphore file : /quovadx/qdx5.6/integrator/certsite/exec/sem_cac9a000  Key Generated : 373645472

      [msi :msi :ERR /0:  hcimonitord:11/22/2011 15:33:15] msiRegionInit: Can’t init semaphore cac9a000: No space left on device 28

      [msi :msi :ERR /0:  hcimonitord:11/22/2011 15:33:15] oilMsemRemove: Can’t determine semaphore id (0xcac9a000): semget failed, errno = 2

      Semaphore file : /quovadx/qdx5.6/integrator/certsite/exec/sem_cac9a000  Key Generated : 373645472

      [msi :msi :ERR /0:  hcimonitord:11/22/2011 15:33:15] msiRegionDelete: Can’t remove region semaphore: No such file or directory

      [prod:prod:ERR /0:  hcimonitord:11/22/2011 15:33:15] Initialization of global shared memory failed

      Any ideas on how to correct this?

    Viewing 2 reply threads
    • Author
      Replies
      • #75558
        Paul Johnston
        Participant

          Hi Jim ,

          We had this problem recently with our Test server. We recently upgraded to 5.8 running Linux (have not changed my signature yet) .  In order to correct your error you will need to increase the size of the semaphores in the Kernel. This does not require a reboot.

          To do this use this command to change the kernal values.

                               sysctl -w =””

          To View the values for semaphores ;

                              sysctl -a|grep -i sem

          To Change the values  ;

                              sysctl -w kernel.sem=”250 32000 100 512″

          We have our Production server set as ;

                               250     62000   1200    1600

          Hopefully this works for you on HPUX. But at least you know what needs to be corrected.

        • #75559
          Josh Deere
          Participant

            On Linux (not sure on HPUX) you’ll also have a /etc/sysctl.conf where you can store the kernel.sem=x x x x value.

          • #75560
            Jim Beall
            Participant

              We shut down another non-prod site and, sure enough, we were able to start up monitord and all the processes/threads on the new non-prod site.  We think we’ve figured out which kernel parm we need to increase as well.  Thanks to both of you for your help.

          Viewing 2 reply threads
          • The forum ‘Cloverleaf’ is closed to new topics and replies.