Monitoring for crashing programs? - IBM AS400

This is a discussion on Monitoring for crashing programs? - IBM AS400 ; Just curious off hand (before I go diggin in the archives/manuals) if there is an exit point or some other way to know when any program on the system crashes (like with a level check or anything else). Some way ...

+ Reply to Thread
Results 1 to 5 of 5

Thread: Monitoring for crashing programs?

  1. Monitoring for crashing programs?

    Just curious off hand (before I go diggin in the archives/manuals) if
    there is an exit point or some other way to know when any program on the
    system crashes (like with a level check or anything else).

    Some way to monitor for all of these. I assume they would all end up
    with MSGW status... but I bet there are cases where it won't and it just
    ends abnormally.

  2. Re: Monitoring for crashing programs?

    On Feb 22, 6:21 pm, "Bradley V. Stone"
    wrote:
    > Just curious off hand (before I go diggin in the archives/manuals) if
    > there is an exit point or some other way to know when any program on the
    > system crashes (like with a level check or anything else).
    >
    > Some way to monitor for all of these. I assume they would all end up
    > with MSGW status... but I bet there are cases where it won't and it just
    > ends abnormally.


    We have a home-grown program that checks the QSYSOPR message que for
    MSGW status every 10 minutes, and starts paging when it finds one.


  3. Re: Monitoring for crashing programs?


    "Bradley V. Stone" wrote in message
    news:ef9eb$45de257d$41a099a1$32094@myclearwave.net ...
    > Just curious off hand (before I go diggin in the archives/manuals) if
    > there is an exit point or some other way to know when any program on the
    > system crashes (like with a level check or anything else).
    >
    > Some way to monitor for all of these. I assume they would all end up with
    > MSGW status... but I bet there are cases where it won't and it just ends
    > abnormally.


    There's no guaranteed way of doing this that I know of.

    What you can do is specify the DFTPGM parameter for escape messages which
    are NOT monitored. You could modify the escape messages sent when a failing
    program escalates the exception to the caller, plus say CPF9999 to catch
    function checks, so that any unmonitored exceptions are logged by a program.

    You can also develop message handler programs to monitor QSYSOPR.

    But if the crashing program monitors for the errors, then unless there's
    some kind of snazzy API which I haven't heard of, I can't think of a generic
    way to capture the failure - other than by imposing an exception handling
    standard in software development and quality assuring it.

    Then again, you can have programs in some products (eg JBA) where errors are
    trapped in *PSSR's which after logging the error in a file (sometimes!!!,
    often the error logger fails!), the *PSSR subroutine then ends*normally* and
    the caller can carry on in blissful ignorance :-)
    (unless it checks the *LDA for a "return code")






  4. Re: Monitoring for crashing programs?

    Hi Brad,

    as far as it is ILE that should be possible. An ile Condition Handler (or
    some global monitor statement does one part and the other, like abnormal
    (or normal) ending programs could be handled with CEE4RAGE and an exit
    point for registration of CEE4RAGE.
    you will find an example at scott klements site, its not perfect (there is
    no need to export the exit handling procedure, because you pass a procedure
    pointer) but it works.

    Dieter Bender

    Bradley V. Stone wrote:

    > Just curious off hand (before I go diggin in the archives/manuals) if
    > there is an exit point or some other way to know when any program on the
    > system crashes (like with a level check or anything else).
    >
    > Some way to monitor for all of these. I assume they would all end up
    > with MSGW status... but I bet there are cases where it won't and it just
    > ends abnormally.



  5. Re: Monitoring for crashing programs?

    A message watch seems like your best candidate (available in V5R4).
    There are performance considerations and false positives, but it has the
    greatest change of working for conditions were error messages are issued
    regardless of whether they are monitored for or not.

    Sadly you still have the return code conditions for which you cannot
    account for unless the program/app issues a message.

    Bradley V. Stone wrote:
    > Just curious off hand (before I go diggin in the archives/manuals) if
    > there is an exit point or some other way to know when any program on the
    > system crashes (like with a level check or anything else).
    >
    > Some way to monitor for all of these. I assume they would all end up
    > with MSGW status... but I bet there are cases where it won't and it just
    > ends abnormally.



    --
    Rodney A Johnson
    Technical Team Lead for i5/OS (AS/400) Spool
    Dept GJC
    IBM Rochester, Minnesota

    The contents of this message express only the sender's opinion.
    This message does not necessarily reflect the policy or views of
    my employer, IBM. All responsibility for the statements
    made in this Usenet posting resides solely and completely with the
    sender.

+ Reply to Thread