PowerHA for AIX

 View Only

WAS Takeover with PowerHA not working

  • 1.  WAS Takeover with PowerHA not working

    Posted Wed April 22, 2015 11:47 AM

    Originally posted by: Sean Franklin


    Hello,

    The setup for the issue I have is  PowerHA 1.7.3 service pack 0 is running on an AIX v7.1 system with WebSphere 8.5.5.3 installed on a shared disk.  I did not use smart assist in PowerHA to setup WebSphere.
     
    The scenario for the issue is when one host is shut down unexpectedly (A "halt -q" command is used) a takeover host takes over and continues the WebSphere processes as if it never failed (what PowerHA is doing).  The issue is that is seems once the takeover occurs and switched to the takeover host then Batch jobs (submitted through the Compute Grid piece built into WebSphere)  continue to run, but then another takeover (Another "halt -q" command is used) occurs going back to the main host.  This breaks some of the communication between WebSphere and the system.
     

    The connection issue is that S1 (example server name for this situation) sends info to S2 (the second example server name).  S2 receives the info and processes it.  When S2 tries  sends information back to S1 it is send but S2 waits indefinitely for S1 to recieve it.  S1 has no sign that it was received in the trace, but S2 shows it was sent in the trace.  Later on the same process occurs where S1 sends info to S2 and S2 processes the info and then sends it back to S1, but this time S1 receives it and is able to process the info.

    Is this situation possibly due to the setup with PowerHA and WebSphere?  Is there a lag time where WebSphere is not properly up and running fully so the info send from S2 to S1 is lost?  Is there a config issue missing here? 

    Thanks a lot in advance,

    Sean Franklin