Clovertech

Thread Connections being lost – Processes Hung

› Clovertech Forums › Read Only Archives › Cloverleaf › Cloverleaf › Thread Connections being lost – Processes Hung

  • Creator
    Topic
  • June 13, 2007 at 9:10 pm #49342
    Femina Jaffer
    Participant

      Hello,

      Recently we have been experiencing losing connections from certain threads, especially after our logs have been archived.    Does anyone have any code that actually monitors network traffic on busy threads, before other clients call and tell us that they have not received any messages in the last so and so minutes or hours?  Surely, there has to be away of being alarmed.  Also, is there any code out there that monitors the recovery database to see the number of messages in the queue, and if the number is greater then something – notify someone or send an alert?

      We are currently on 3.8 and planning on migrating to platform 5.5 within the next couple of months.

      Any advise or tips on this would be very much appreciated.    Thanks in advance.

      Femina

    • Creator
      Topic
    Viewing 3 reply threads
    • Author
      Replies
      • June 13, 2007 at 9:51 pm #61575
        Jim Kosloskey
        Participant

          Femina,

          You should be able to almost all you want (at least to start) with the Alerts for the threads.

          For the Recovery and/or Error DB, there are posts on Clovertech of methods some have used to detect the content via the command line.

          Jim Kosloskey

          email: jim.kosloskey@jim-kosloskey.com 30+ years Cloverleaf, 60 years IT – old fart.

        • June 14, 2007 at 4:27 pm #61576
          Femina Jaffer
          Participant

            Thanks Jim.

            In the past we had alerts on and they had seemed to bring down our monitor daemons.  The monitor daemons would just hang.

            Thank you.

            fj

          • June 14, 2007 at 4:34 pm #61577
            Jim Kosloskey
            Participant

              Famina,

              If your Alerts went to a GUI (the default I think) then I recall there was something related to assuring the logon that started the NetMonitor had to be able to get the Alert or they would just queue up and consume memory until somethin froze. This is all very past recollection and may be completely wrong.

              The point is that many have their Alerts use email and/or paging to alert Operations (or whomever) rather than use the GUI Alert.

              Jim Kosloskey

              email: jim.kosloskey@jim-kosloskey.com 30+ years Cloverleaf, 60 years IT – old fart.

            • June 14, 2007 at 4:47 pm #61578
              Femina Jaffer
              Participant

                Thanks again Jim.  I will pursuit this avenue again.

                fj

            • Author
              Replies
            Viewing 3 reply threads
            • The forum ‘Cloverleaf’ is closed to new topics and replies.