1. Get rid of all advertisements and get unlimited access to documents by upgrading to Premium Membership. Upgrade to Premium Now and also get a Premium Badge!

ORA-07445: core dump in kkogbro()

Discussion in 'General' started by dblabla, Oct 14, 2009.

  1. dblabla

    dblabla Active Member

    Messages:
    5
    Likes Received:
    0
    Trophy Points:
    55
    Dear all,

    when issuing a view query through a web browser (oracle forms) the remote client receives ora-3113, at the same time alert.log shows:

    ORA-07445: exception encountered: core dump kkogbro()+192 SIGSEGV Address not mapped to object 0x000000000.

    These are the details of the environment:
    Oracle Database 10g Release 10.2.0.1.0 - 64bit Production
    With the Real Application Clusters option
    System name: Linux
    Release: 2.6.9-55.0.0.0.2.ELsmp
    Version: #1 SMP Wed May 2 15:06:32 PDT 2007
    Machine: x86_64

    Is this an oracle bug or can it be solved?

    Thank you very much for any help!
    Kind regards!
     
  2. zargon

    zargon Community Moderator Forum Guru

    Messages:
    2,347
    Likes Received:
    348
    Trophy Points:
    1,430
    Location:
    Aurora, CO
    You should be running 10.2.0.4 as 10.2.0.1 is the base release and has no CPU updates available. And, most likely you'll find such errors are addressed by fixes in the later patchsets.
     
  3. dblabla

    dblabla Active Member

    Messages:
    5
    Likes Received:
    0
    Trophy Points:
    55
    Thanks for the reply. This is most probably a bug.

    Saw somewhere though that
    SQL> alter system flush shared_pool;

    can delay the this occurrence...what is your opinion?

    Kind regards.
     
  4. zargon

    zargon Community Moderator Forum Guru

    Messages:
    2,347
    Likes Received:
    348
    Trophy Points:
    1,430
    Location:
    Aurora, CO
    Why kill the shared pool on a regular basis when you can patch to a supported release of Oracle and most likely fix the problem at the source?
     
  5. dblabla

    dblabla Active Member

    Messages:
    5
    Likes Received:
    0
    Trophy Points:
    55
    The problem is solved.
    There was nothing wrong with the db, and we didnt have to apply patch to supported version. The problem was in syntax of plsql code in oracle form that used this query! DB should catch such problems in a better way (without core dump in alert.log).
     
  6. zargon

    zargon Community Moderator Forum Guru

    Messages:
    2,347
    Likes Received:
    348
    Trophy Points:
    1,430
    Location:
    Aurora, CO
    Really? That's interesting since ORA-07445 errors signal a problem at the operating system level and I can't understand how a syntax error (which you did not report) caused the ORA-07445 (which you did report).

    Provide proof of this, please, as your explanation makes no sense.
     
  7. dblabla

    dblabla Active Member

    Messages:
    5
    Likes Received:
    0
    Trophy Points:
    55
    Exactly, ora-7445 signals problem at OS level, that's why I spent 2 weeks in researching for problems there. This was misleading and wasted time. The smartest thing everybody were saying was: apply the patch. The fundamental thing is that application's syntax error made some prohibited memory references (% in pls/sql caused it) and the error was caught only at OS level without pointing to application! Seems as though the db is too "stupid" to see errors on application level. Misleading thing was also that application with its error was successfully compiled. After two weeks when all seemed to be fine at OS and DB, I checked the application and very soon found the error using the stack trace for troubleshooting in alert.log, correction of it solved the problem. I cannot provide proof because there are no problems anymore!
     
  8. zargon

    zargon Community Moderator Forum Guru

    Messages:
    2,347
    Likes Received:
    348
    Trophy Points:
    1,430
    Location:
    Aurora, CO
    Yet you didn't consult Metalink or any of the related patch documents to discover what functionality was affected by patchsets 10.2.0.2, 10.2.0.3 or 10.2.0.4? You failed to research what bugs in the base release of 10gR2 affect your application code? To call the database stupid when, in actuality, the 'research' performed was incomplete, at best, is foolish. Metalink provides an ORA-00600/ORA-07445 lookup tool to assist in troubleshooting such events; you chose to not use this facility and, instead, use your own limited knowledge of Oracle to 'solve' your problem. Investigate this issue properly (you should still have the original ORA-07445 error text in the alert log) and I expect you'll find you do need to patch your database, and after you do you'll also find that the original application code will likely function as intended.
     
  9. dblabla

    dblabla Active Member

    Messages:
    5
    Likes Received:
    0
    Trophy Points:
    55

    Its not fun to write here and be judged by you on a personal level. I wanted to share experience, it might help somebody experiencing the same problem, thought this was this forum's mission, but apparently admins just want to play smart on somebody elses expense. It is not professional! Calling db "stupid" wasn't an attack on you personally it means only that db is on a lower abstraction level than then application, every (even with limited knowledge) engineer should know this. No wonder people (me included) find the forum unusable when they get attacked if they share their experiences. Just for your notice I did consult Metalink, and lookup tool but my common sense worked better which I also proved by solving the problem.
    Enjoy!
     
  10. zargon

    zargon Community Moderator Forum Guru

    Messages:
    2,347
    Likes Received:
    348
    Trophy Points:
    1,430
    Location:
    Aurora, CO
    Refering to someone or something as 'stupid' does NOT 'means only that ... is on a lower abstraction level'; the definition of stupid is as follows:

    lacking or marked by lack of intellectual acuity

    What intellectual acuity does a program possess? None I am aware of so calling the program 'stupid' mystifies me.

    As to your complaint on 'personal attacks' there are none here; until this last post you provided no insight on what you did research in finding your 'solution' (which I still think would have been corrected had you decided to patch the Oracle software since 10.2.0.1 contains numerous bugs fixed in those subsequent patchsets) leaving those here to wonder what, exactly, you researched and how you arrived at your conclusions. No one here is 'playing smart on somebody elses expense'; comments made reflect on the information, or lack thereof, in prior posts in a topic.

    The assistance provided is only as good as the information given for the problem presented.
     
  11. Sadik

    Sadik Community Moderator Forum Guru

    Messages:
    1,906
    Likes Received:
    252
    Trophy Points:
    1,455
    Hi guys, relax... no need to take anything personally. We all help each other out here, sometimes people may make some inadvertent comments, but it's nothing personal.