Results 1 to 7 of 7
  1. #1
    Join Date
    Jul 2004
    Posts
    306

    Unanswered: Gaps/missing time in the db2diag.log

    db2 9.7 fp 4 on Linux

    Hey all,

    Has anyone experienced gaps or missing time in the db2diag.log at all?

    On two occasions (that I've noticed) I have a large gap in the log. This is a production system with 7 pretty active databases and I know there was activity during this gap periods as I'm investigating some problems that occur during that time based on websphere logs reporting database errors (lock timeouts mostly). I also have activity in the .nfy and other logs during the time period.

    Seems a bit alarming to me... anyone else seen this or have an explanation?

  2. #2
    Join Date
    Jan 2009
    Location
    Zoetermeer, Holland
    Posts
    746
    Sounds creepy indeed. Not sure, but I recall that FP4 was not the most 'robust' version in the 9.7 landscape. Might be a good start to apply the latest fixpack and then start collecting evidence. Every once-in-a-while just try to make a connection with a fake userid/passwd. That MUST show-up in your diaglog. When it does not I'd open a PMR if I were you.

    P.S. you did check your diag-level, didn't you?
    Somewhere between " too small" and " too large" lies the size that is just right.
    - Scott Hayes

  3. #3
    Join Date
    Jul 2004
    Posts
    306
    Yea we're in the process of going to v10...
    Diaglevel is never usually changed from 3, we get plenty of traffic even at 3 with archive logging and some STMM stuff etc. over the 7 DBs

    Oh well.. I might open a PMR I guess...

  4. #4
    Join Date
    Jul 2004
    Posts
    306
    Had it happen again yesterday!
    no diag entries for 4 hours... I forced an entry and it kicked in again... the nfy stopped writing too this time but the stmm logs kept going (all on the same LUN)

    Very weird.

  5. #5
    Join Date
    Aug 2008
    Location
    Toronto, Canada
    Posts
    2,367
    When it happens again, collect latches (db2pd with -latches) and, if possible, stacks (db2pd with -stacks) several times. Check which process is holding db2diag.log / notify using fuser command. Is there anything logged in Linux logs that is related to db2diag.log/notify log?

  6. #6
    Join Date
    Jan 2009
    Location
    Zoetermeer, Holland
    Posts
    746
    How about the size of the diag.log file? Do a "db2diag -A" to be sure. On windows I once had trouble when the logfile was 4Gb in size. It just stopped.... well db2 stopped en I could solve that with an archive of the log file. Try to alter the path as well, maybe there are trouble in the filesystem where db2 tries to write
    Somewhere between " too small" and " too large" lies the size that is just right.
    - Scott Hayes

  7. #7
    Join Date
    Jul 2004
    Posts
    306
    Quote Originally Posted by db2girl View Post
    When it happens again, collect latches (db2pd with -latches) and, if possible, stacks (db2pd with -stacks) several times. Check which process is holding db2diag.log / notify using fuser command. Is there anything logged in Linux logs that is related to db2diag.log/notify log?
    Good idea, I'll keep that one in the tool belt, thanks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •