Topic
3 replies Latest Post - ‏2012-03-30T19:08:34Z by SystemAdmin
gilgen
gilgen
3 Posts
ACCEPTED ANSWER

Pinned topic CWPZC8136E: The application status is currently UNKNOWN - how to avoid?

‏2012-03-28T20:42:50Z |
We have a customer PMR where after a reboot of their OS, we try to start two zero servers. One starts cleanly and the other gets the following message:

2012-03-05 00:27:06 zero.management.tasks.Start::isAppStarted Thread-0
WARNING

So now I have the following questions about the -force parameter:

1) is the -force parameter the message refers to be passed on the zero start call, like "./zero start -force"
2) what does the force parameter really do?
3) is there a downside to always specifying -force when we start our processes?
4) if there is no downside, why isn't it the default behaviour
5) if we add the -force parameter and it does not help, how do we debug this? Enable more trace, and if so, what trace spec would you suggest?

Thanks,
Dave Gilgen, IBM Workload Deployer
Updated on 2012-03-30T19:08:34Z at 2012-03-30T19:08:34Z by SystemAdmin
  • gilgen
    gilgen
    3 Posts
    ACCEPTED ANSWER

    Re: CWPZC8136E: The application status is currently UNKNOWN - how to avoid?

    ‏2012-03-28T20:45:09Z  in response to gilgen
    Sorry, looks like the pasted error message did not come accross cleanly. The messages I see are:

    2012-03-05 00:27:06 zero.management.tasks.Start::isAppStarted Thread-0
    WARNING
    2012-03-05 00:27:06 zero.cli.tasks.commands.ZeroBaseTask::onCliTask Thread-0
    SEVERE CWPZT0601E: Error: Command start failed
    • gilgen
      gilgen
      3 Posts
      ACCEPTED ANSWER

      Re: CWPZC8136E: The application status is currently UNKNOWN - how to avoid?

      ‏2012-03-28T20:48:46Z  in response to gilgen
      Well, looks like a problem with the message box I am pasting into. The message appears in the box but when I submit it doesn't end up in the post. Loks like it is choking on the brackets in the message. I'll remove them manually and try one more time:

      2012-03-05 00:27:06 zero.management.tasks.Start::isAppStarted Thread-0
      WARNING CWPZC8136E: The application status is currently UNKNOWN, perhaps as the result of system crash or other error. To force application to start, use "-force" parameter.
      2012-03-05 00:27:06 zero.cli.tasks.commands.ZeroBaseTask::onCliTask Thread-0
      SEVERE CWPZT0601E: Error: Command start failed
  • SystemAdmin
    SystemAdmin
    9224 Posts
    ACCEPTED ANSWER

    Re: CWPZC8136E: The application status is currently UNKNOWN - how to avoid?

    ‏2012-03-30T19:08:34Z  in response to gilgen
    I believe you are seeing the effects documented in the last two posts in this thread: https://www.ibm.com/developerworks/forums/thread.jspa?messageID=14652592&#14652592

    confirm by deleting the zso.pid file to see if 'zero start' without the -force parm works.

    1) is the -force parameter the message refers to be passed on the zero start call, like "./zero start -force"
    yes

    2) what does the force parameter really do?
    It ignores the logical checks for a valid current state and goes ahead with trying to start zso. If it can be started it should start without complaint. If it is already running, the port will be in use and this start attempt will fail.

    3) is there a downside to always specifying -force when we start our processes?
    Only to user awareness and a tiny bit of engineering efficiency. Specifying -force bypasses the easy checks for current status and allows a real resource failure (port in use) tell you that your node is already started. That's a lot more code to get to the same result in the most common failure case. Impolite to the CPU, but not critically important. If it really cannot start, the -force parm will not help or hurt.

    4) if there is no downside, why isn't it the default behaviour
    Engineering efficiency - we don't try to start if we already think we're started or we don't know. Its also the way we can expose the fact that we have arrived at an UNKNOWN status.

    5) if we add the -force parameter and it does not help, how do we debug this? Enable more trace, and if so, what trace spec would you suggest?
    zero.management I believe, but lets get there first