How to determine who is sending SIGKILL - Aix

This is a discussion on How to determine who is sending SIGKILL - Aix ; Is there a way to set up something in AIX, version 5.3.0.50, to track who would be sending a SIGTERM to a process? We have a process that runs and accesses DB2 and after a while of running, a few ...

+ Reply to Thread
Results 1 to 3 of 3

Thread: How to determine who is sending SIGKILL

  1. How to determine who is sending SIGKILL

    Is there a way to set up something in AIX, version 5.3.0.50, to track
    who would be sending a SIGTERM to a process?

    We have a process that runs and accesses DB2 and after a while of
    running, a few times we have hit a problem where an EDU process of DB2
    is killed with signal 9. Nobody did it from the command line, our code
    does not have SIGKILL for DB2 processes, and DB2 does not kill itself
    like that according to DB2 suport. So, we must somehow track down this
    rogue process and I am not sure how to go about it.

    I posted this in the DB2 group but I don't think that was the right
    place since DB2 itself is not the problem.

    Any help or suggestions would be appreciated.


  2. Re: How to determine who is sending SIGKILL

    On 21 Nov 2006 13:10:26 -0800, "soccertl" wrote:

    >Is there a way to set up something in AIX, version 5.3.0.50, to track
    >who would be sending a SIGTERM to a process?


    A system trace for syscalls would tell you. Check out the AIX trace
    facility.

    To rule out the obvious, is the system running out of paging space?
    Anything in the error log (errpt)?

    -- David

  3. Re: How to determine who is sending SIGKILL


    David Luner wrote:
    > A system trace for syscalls would tell you. Check out the AIX trace
    > facility.
    >


    Thanks. I will check into using the trace stuff, I just wish it would
    occur on the same machine more than once. The occurrences have all been
    on different machines.

    > To rule out the obvious, is the system running out of paging space?
    > Anything in the error log (errpt)?
    >


    The error log had absolutely nothing around the time the process died.
    Other entries way before and soon after that time but not near the time
    of the failure.

    The system did not run out of paging space from what another developer
    told me. I didn't personally look at this myself.


+ Reply to Thread