Automic Workload Automation

Looking for an Event-driven method to force an EXQUERY after an agent is stopped.

  • 1.  Looking for an Event-driven method to force an EXQUERY after an agent is stopped.

    Posted Apr 03, 2017 01:04 PM
    I have created an OS patching infrastructure using Automation Engine v11 and Operating System Agents on each server that is to be patched.  My biggest requirement is to have no patch operations in a "paused" or looping state.  Instead, i want each action to process and complete which then causes the next server action.  I have made this patch process modular so that each application group can control when their server(s) are affected.  The application groups also can arrange the order in which their respective servers come up and down (database instances come down after app and web servers etc).  

    With all the above as background information, my patch system works well but could be more efficient and event driven if i had a way to force Automation Engine to do an EXQUERY function.  EXQUERY is what the Automation Engine uses to query an agent to see if its up or down.  I guess its similar to an agent ping in the world of Automic.  This EXQUERY is automatically issued at an interval that is specific to the KEEP_ALIVE value with up +/-60 seconds of variance.  My plan is to shut down the HOST using the agent, i am currently using a 'shutdown -h now' and waiting up to 5 minutes for output.  if more than 5 minutes occurs, then i will force the server down using the hypervisor.  Because i know the server should be coming down imminently, i would like to issue a "forced" EXQUERY instead of waiting the KEEP_ALIVE timeout.  During our monthly patch window, around 1000 servers will be shutdown and rebooted; by waiting the KEEP_ALIVE value, i am wasting unnecessary time.  My work around is to lower the KEEP_ALIVE time to its lowest value of 60 seconds, but this is unnecessary network traffic when i would rather leave it to 600 seconds during all other normal processing.  Also, performance of the Automation Engine is compromised with such an aggressive KEEP_ALIVE value.  

    I apologize for being wordy or for not explaining my intentions clearly.  Please reply with any questions or solutions that i can use.

    Terry
    Clemson University
    Linux/Unix Systems Administrator