Results 1 to 9 of 9

Thread: Lru error

  1. #1
    Join Date
    Oct 2010
    Posts
    5

    Unanswered: Lru error

    What is the meaning of this in Oracle core dump trace file:
    LRU ERROR=(a 9 digit number)

    Once we got this, no users could do any work until database was bounced.

    Any idea? Is it some caching problem? Oracle Support say to change some undocumented init parameters but give no reasoning behind it. No patch either for 10.2.0.4 on Solaris.

    Any comment helpfull.

    Thankyou,
    Jayce

  2. #2
    Join Date
    Jun 2004
    Posts
    796
    Provided Answers: 1
    "Oracle told me something but I'm not telling you what & I want you to tell me why they said it"

    Maybe if you told us what parameters you were told to change we could help you...
    90% of users' problems can be resolved by punching them - the other 10% by switching off their PCs.

  3. #3
    Join Date
    Oct 2010
    Posts
    5
    The parameters are _kghdsidx_count ("reduce to <= 31") and _kgl_latch_count ("set to one").

  4. #4
    Join Date
    Jun 2004
    Posts
    796
    Provided Answers: 1
    What was the error (in full)?
    What are the current values of the parameters?
    90% of users' problems can be resolved by punching them - the other 10% by switching off their PCs.

  5. #5
    Join Date
    Oct 2010
    Posts
    5
    That would be many 5MB trace files, but here are some extracts:
    *** SERVICE NAMESYS$USERS)
    LRU ERROR: LINK = 380059620
    ...
    ksedmp: internal or fatal error
    ORA-07445: exception encountered: core dump [kgscDump()+232] [SIGSEGV] [Address not mapped to object] [0x000000010] [] []
    ORA-07445: exception encountered: core dump [kgscDump()+232] [SIGSEGV] [Address not mapped to object] [0x000000010] [] []
    ORA-00600: internal error code, arguments: [17126], [0x533E37560], [], [], [], [], [], []
    ...
    Any place I can dump the entire trace?

  6. #6
    Join Date
    Oct 2010
    Posts
    5
    From Tanel Poder,
    SQL> select n.ksppinm pd_name, c.ksppstvl pd_value, n.ksppdesc pd_descr
    from x$ksppi n, x$ksppcv c
    where n.indx=c.indx
    and (
    lower(n.ksppinm) like lower('%&1%')
    or lower(n.ksppdesc) like lower('%&1%')
    );

    _kghdsidx_count = 7, and _kgl_latch_count = 0.

    What's the effect of changing these params? _kgl_latch_count=0 is correct?

  7. #7
    Join Date
    Jun 2004
    Posts
    796
    Provided Answers: 1
    First off I would suggest you read up on the documentation for the 2 parameters, to understand what they do.

    Secondly, the parameters are undocumented, which means that you're not supposed to play with them unless Oracle tell you to. If Oracle have looked at your problem & told you to change them then it follows that the best people to ask 'why?' would be Oracle.

    Judging by the lack of response to your posting of this on several other forums I would go back to Oracle & ask why. We could guess some reasons for you, but who knows if they would be relevant?
    90% of users' problems can be resolved by punching them - the other 10% by switching off their PCs.

  8. #8
    Join Date
    Oct 2010
    Posts
    5
    Oracle have already been asked (repeatedly) for their justification in proposing these changes. No explanation, and is being escalated. That is why I'm going to forums for opinions. What are your guesses?

  9. #9
    Join Date
    Aug 2010
    Posts
    34

Posting Permissions

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