Topic
  • 3 replies
  • Latest Post - ‏2012-08-14T19:35:33Z by SystemAdmin
xf00657
xf00657
20 Posts

Pinned topic Recycle Agentless Adapter Resources

‏2012-08-13T13:27:45Z |
Hello,
we want to use the "Restart.." button für some AL-Resources but the restart ill ever been interrupted. Why SAAM waits only 1 second??:

8/13/12 15:20:58:659 CEST 0007134d EEZMDBMsg I EEZRestartStateMachine logMessage EEZJ1015I Restart of resource "IBM.RemoteApplication:travic_retail_zone_nzitrv04_ITU2:nsitrv01" starts as requested by "xf00657".
Explanation: The restart request is validated successfully. The stopping phase of the restart cycle begins.
User action: No action required.
System action: The automation manager sends a stop request to the resource.
8/13/12 15:20:59:868 CEST 0007134f EEZMDBMsg E EEZRestartStateMachine logMessage EEZJ0053E A restart request by "xf00657" failed for resource "IBM.RemoteApplication:travic_retail_zone_nzitrv04_ITU2:nsitrv01" after 1 seconds. The state of the restart cycle is "Stopping". The reason code is: "The observed state is Online".
Explanation: The restart cycle was interrupted by an event.
User action: Check the status of the affected resource. If needed issue a new request.
System action: Terminates the restart cycle of the resource.

So we must then use "Bring online..." button.
Regards
Rüdiger
Updated on 2012-08-14T19:35:33Z at 2012-08-14T19:35:33Z by SystemAdmin
  • SystemAdmin
    SystemAdmin
    46 Posts

    Re: Recycle Agentless Adapter Resources

    ‏2012-08-13T20:44:52Z  
    Hi Rüdiger,

    this is not the way it should work.... The restart-cycle which is implemented in the SA AppMan layer depends on a correct sequence of states the referenced resource is traversing-through in order to work correctly. So, if the referenced resource was first stopping, after the offline request has been sent and then suddenly it is online again - then the SA AppMan will stop the restart cycle with an error, as you posted.

    I have one idea, which could be the reason for this behavior. Is the start-script you specified for this ALA resource blocking? Or is it asynchron? I mean, when the start command finishes does the monitor command first return offline and only after a while finds that the resource is now started?
    This behavior is not good for ALA resources, since the AL Adapter does not have a notion of "RunCommandSync" as you might know from SAMP resource instrumentation.

    Lemme know, if this hint helps you to solve the problem for now - and then lets discuss, if the behavior is acceptable.

    Greetings,

    Josi.
  • xf00657
    xf00657
    20 Posts

    Re: Recycle Agentless Adapter Resources

    ‏2012-08-14T07:18:11Z  
    Hi Rüdiger,

    this is not the way it should work.... The restart-cycle which is implemented in the SA AppMan layer depends on a correct sequence of states the referenced resource is traversing-through in order to work correctly. So, if the referenced resource was first stopping, after the offline request has been sent and then suddenly it is online again - then the SA AppMan will stop the restart cycle with an error, as you posted.

    I have one idea, which could be the reason for this behavior. Is the start-script you specified for this ALA resource blocking? Or is it asynchron? I mean, when the start command finishes does the monitor command first return offline and only after a while finds that the resource is now started?
    This behavior is not good for ALA resources, since the AL Adapter does not have a notion of "RunCommandSync" as you might know from SAMP resource instrumentation.

    Lemme know, if this hint helps you to solve the problem for now - and then lets discuss, if the behavior is acceptable.

    Greetings,

    Josi.
    Hi Josi,

    thanks for your quick answer. In our case the AL-Resources are SUN zones. The monitor scripts was written by our SUN collegues. They wouldn't use the provided script from your Lab. Their script only returnes "online" and "offline" but nothing between. So when shutting down the zone the monitor still returns "online" until the zone is complete down.
    Regards
    Ruediger
  • SystemAdmin
    SystemAdmin
    46 Posts

    Re: Recycle Agentless Adapter Resources

    ‏2012-08-14T19:35:33Z  
    • xf00657
    • ‏2012-08-14T07:18:11Z
    Hi Josi,

    thanks for your quick answer. In our case the AL-Resources are SUN zones. The monitor scripts was written by our SUN collegues. They wouldn't use the provided script from your Lab. Their script only returnes "online" and "offline" but nothing between. So when shutting down the zone the monitor still returns "online" until the zone is complete down.
    Regards
    Ruediger
    Hi Rüdiger,

    The question is not about your monitor script. I think you need to have a look at the start script. Once the ALA gets a start request it will execute the start script. As long as this script is running the remote resource is in state observed "starting". The next assumed observed state would be "online". As said - as long the start script executes the resource will be in starting and the monitor script is not called. After the start script finishes the monitor script is called. If at this moment the monitor script returns "offline" we have an unexpected state transition - from starting to offline. This causes the restart logic to stop continuing the restart cycle.

    So the start script must not finish as long as the resource is not actually online. Maybe you need to call the monitor command in a loop inside the start script and only end, when thus the returned state is actually really online?

    Same logic as explained here also applies to the stop command.

    Greetings,

    Josi.