Topic
  • 3 replies
  • Latest Post - ‏2012-07-18T14:23:50Z by MichaelMcBride
kagorp
kagorp
2 Posts

Pinned topic ABEND S04E-00E4D5D2 on REORG for ORGANIZE BY HASH TABLESPACE

‏2012-02-14T11:38:30Z |
World,

A REORG SHRLEVEL CHANGE for an ORGANIZE BY HASH TABLESPACE takes the "emergency exit" ABEND S04E-00E4D5D2.

As of now, I haven't been able to figure out what is going wrong. The object can be empty or contain data, it can be small, large, or huge - you name it, but the REORG always (!) fails with the same ABEND. I'm pretty sure that I'm making some stupid mistake, and that the solution is a trivial little detail I'm overlooking.

The APAR database does not have too much in store, mainly PM49317 (see item 5. of problem description section), PM31000, and PM48325. The fix for PM49317, UK73284, recently went PE (AM57322, 12.039). Besides, in my case I don't see any indication of a SORT problem of any kind.

A quick look into the DUMP easily lets you track back the "sequence of events", at least to some extent/depth.

CSECT DSNUGSTA (09/16/10; UK60512)

At offset X' 00000992', DSNUGTRC (09/26/11; UK71458) is called. At offset X'000009C2', an internal subroutine (at offset X'0000265C') is called, which right away stumbles over GPR2 contents not being positive (in my case, always some value like X'F2xxxxxx’), and then issues the ABEND (offset X'000027E8').

Any help or idea is highly appreciated.

Thanks a lot in advance!

Karl
Updated on 2012-07-18T14:23:50Z at 2012-07-18T14:23:50Z by MichaelMcBride
  • ToddBurchDB2
    ToddBurchDB2
    74 Posts

    Re: ABEND S04E-00E4D5D2 on REORG for ORGANIZE BY HASH TABLESPACE

    ‏2012-02-14T11:47:03Z  
    From the description of the issue:

    Explanation
    This abend is placed at logic points that will occur only through a programming error.

    There is nothing you can do to avoid the error. You'll have to open a ticket with support.

    Todd
  • kagorp
    kagorp
    2 Posts

    Re: ABEND S04E-00E4D5D2 on REORG for ORGANIZE BY HASH TABLESPACE

    ‏2012-02-14T14:03:56Z  
    From the description of the issue:

    Explanation
    This abend is placed at logic points that will occur only through a programming error.

    There is nothing you can do to avoid the error. You'll have to open a ticket with support.

    Todd
    Thanks, Todd!

    I did read the description of the reason code, of course. However, experience has taught me not to be "scared away" by the really bad, bad reason codes. Agreed, the X'00E4D5D2' is definitely an APAR-able problem, but the expertise in this forum might still provide hints, tips, or a possible workaround technique.

    Am I really the very first person to REORG a HASH TS and hit this particular problem? If so, a PMR is the only way, of course.

    Yet, if there is something - anything - I can do to get around this problem, I'd happily try and test it.

    Thanks again.

    Karl
  • MichaelMcBride
    MichaelMcBride
    1 Post

    Re: ABEND S04E-00E4D5D2 on REORG for ORGANIZE BY HASH TABLESPACE

    ‏2012-07-18T14:23:50Z  
    • kagorp
    • ‏2012-02-14T14:03:56Z
    Thanks, Todd!

    I did read the description of the reason code, of course. However, experience has taught me not to be "scared away" by the really bad, bad reason codes. Agreed, the X'00E4D5D2' is definitely an APAR-able problem, but the expertise in this forum might still provide hints, tips, or a possible workaround technique.

    Am I really the very first person to REORG a HASH TS and hit this particular problem? If so, a PMR is the only way, of course.

    Yet, if there is something - anything - I can do to get around this problem, I'd happily try and test it.

    Thanks again.

    Karl
    We are encountering the same issue in a test environment that we replicate using DSN1COPY from a Production object that is REORG'd weekly using a LISTDEF REORG. Any similarities?

    How is the data populated into your offending table? What version of DB2 are you on? Did you ever open a ticket with IBM?

    Thanks, MCMcB