Forum Discussion

Sulaiman_85782's avatar
Sulaiman_85782
Icon for Nimbostratus rankNimbostratus
Aug 12, 2014

GUI Connection lost to device

Hi

 

I have a scenario where some of my devices that is setup in our DMZ environment, periodically looses connection to the GUI but not the CLI. The connection is lost on the GUI for about 2mins and then it re-establishes connection on its own. The errors that gets logged at the same time are: Aug 12 11:20:33 bigip notice mcpd[6301]: 01070430:5: end_transaction message timeout on connection 0x817786c8 (user admin) Aug 12 11:20:33 bigip notice mcpd[6301]: 01070418:5: connection 0x817786c8 (user admin) was closed with active requests

 

Any ideas as to what my cause this or how I can trace what is causing it.

 

Thanx

 

18 Replies

  • What is the value of System -> Preferences -> Idle Time Before Automatic Logout ? As I recall, the original setting is quite low.

     

    • Sulaiman_85782's avatar
      Sulaiman_85782
      Icon for Nimbostratus rankNimbostratus
      The Idle timeout is set to 1200 seconds. It happens while working on the GUI, not necessarily when in idle mode.
  • Check if http or tomcat daemon have restarted, once you observe the trouble again. Could be a memory issue.

     

    If the two daemons are up and running for long time, try restarting them.

     

    • Sulaiman_85782's avatar
      Sulaiman_85782
      Icon for Nimbostratus rankNimbostratus
      Hi Sudheer I have checked and the tomcat service does not restart after experiencing the issue, it has been running since the device has restarted. It also happens on multiple devices, not only one.
    • Sulaiman_85782's avatar
      Sulaiman_85782
      Icon for Nimbostratus rankNimbostratus
      Hi nitass There are no errors in /var/log/httpd. We have a limited number of admins, so we not reaching the allocation listed in the Maximum HTTP Connections. I am seeing the error being logged in the early hours of the morning when no one is online logged onto the GUI as well.
    • Sulaiman_85782's avatar
      Sulaiman_85782
      Icon for Nimbostratus rankNimbostratus
      Hi nitass There are no errors in /var/log/httpd. We have a limited number of admins, so we not reaching the allocation listed in the Maximum HTTP Connections. I am seeing the error being logged in the early hours of the morning when no one is online logged onto the GUI as well.
  • We have a limited number of admins, so we not reaching the allocation listed in the Maximum HTTP Connections.

     

    i think some browser creates number of connections.

     

    I am seeing the error being logged in the early hours of the morning when no one is online logged onto the GUI as well.

     

    if the log also happens when no one logs into the system, could the log relate to the issue (GUI Connection lost to device)?

     

    • Sulaiman_85782's avatar
      Sulaiman_85782
      Icon for Nimbostratus rankNimbostratus
      When the issue happens and I have the CLI open then that log is always generated, thus the reason I am thinking that the 2 are linked. It also is generated on the other devices that I am experiencing the same issue. I am thinking that it may be network related but the network teams, response is that it wont work and not work a few minutes later, which I understand, so I dont know what to do.
    • boneyard's avatar
      boneyard
      Icon for MVP rankMVP
      if you can open a support ticket with F5, specially when you have some errors to share they might be able to find a cause / solution.
  • We have a limited number of admins, so we not reaching the allocation listed in the Maximum HTTP Connections.

     

    i think some browser creates number of connections.

     

    I am seeing the error being logged in the early hours of the morning when no one is online logged onto the GUI as well.

     

    if the log also happens when no one logs into the system, could the log relate to the issue (GUI Connection lost to device)?

     

    • Sulaiman_85782's avatar
      Sulaiman_85782
      Icon for Nimbostratus rankNimbostratus
      When the issue happens and I have the CLI open then that log is always generated, thus the reason I am thinking that the 2 are linked. It also is generated on the other devices that I am experiencing the same issue. I am thinking that it may be network related but the network teams, response is that it wont work and not work a few minutes later, which I understand, so I dont know what to do.
    • if you can open a support ticket with F5, specially when you have some errors to share they might be able to find a cause / solution.
  • I am experiencing the exact same issue with 8 devices in our DMZ environment. 11.4.0 HF3. Was there ever any resolution that you can share? We have opened a support case and already tried restarting tomcat and httpd services with no luck.

     

  • We are also playing around with the EM central manager and these devices. Any correlation?

     

  • We have opened a support case and already tried restarting tomcat and httpd services with no luck.

     

    have you tried to capture traffic (i.e. tcpdump) when the problem is happening? just wondering whether traffic hits bigip (since you mentioned restarting httpd and tomcat do not help).

     

  • Funny thing is that once I removed these devices from Enterprise Manager, the problem went away. Not sure if EM requires a dedicated IP or vlan for discovery instead of using the management IP?