Re: Corrupt table problem

From: bs <shatzman_at_netcom.com>
Date: Mon, 2 May 1994 23:52:18 GMT
Message-ID: <shatzmanCp78z7.BLK_at_netcom.com>


Anders - this may not be your problem but it's something that can cause table corruption so i thought i'd mention it. If you have multiple free lists and do a TRUNCATE TABLE, your table can become corrupted. Oracle's known about this bug only since March, so i don't think it's fixed in 7.0.15. The workaround is simple - don't TRUNCATE a table if you have multiple free lists. Better to drop and re-create it.

Barry

anos_at_elmrd6.ineab.ikea.se wrote:
: Hi
 

: I'm not really sure where to begin, but I have run into a problem that I
: not sure how to handle.
 

: Two weeks ago I installed Perfect Cache on our DEC 7000-610 (VMS 1.5). It
: gave us a real performance boost and everyone was happy. A couple of days
: later we also upgrades Oracle to 7.0.15 on the same system. Again, no probs
: for almost two weeks.
 

: Sunday night, at about 22.00, one batch job went berzerk. Instead of running
: for 15 min as ususal, it was RUNNING for more than 2 hours, doing +13.000.000
: i/o (avg 1040 i/o per sec). We decided to abort the job and did so. During
: the rollback, Oracle detected that one table was *completely* corrupt and
: therefore that database was unable to continue. We had to restore the entire
: database (+8 GB) on monday morning (I don't have the exact error number here,
: but can provide it if somebody needs it to help me out).
 

: Monday night, the same thing happended again (another batch, another table).
: We have spoken to Oracle and Raxco. Both parties says "not us!". Surprise !
 

: Anyway, my question is; can a product like Perfect Cache really screw up a
: database table ? Until we know for sure, I have been told to disable the
: caching (with a significant performance loss as a result). As I see it, it
: is almost impossible to prove (???) what's the source of the problem. What
: to do next ? Hope for a new crash with PC disabled ? Will it happen ? Who's
: next ? There is at least a dozen other products that we can put the blame on.
 

: Sigh
 

: Anders
 

: +-------------------------------------------------------------------------+
: | Internet anos_at_ineab.ikea.se |
: | Phone +46-42-25 73 08, Anders Ostling |
: | IKEA Northern Europe AB, Sweden |
: +-------------------------------------------------------------------------+
: --
: +-------------------------------------------------------------------------+
: | Internet anos_at_ineab.ikea.se |
: | Phone +46-42-25 73 08, Anders Ostling |
: | IKEA Northern Europe AB, Sweden |

 +-------------------------------------------------------------------------+
Received on Tue May 03 1994 - 01:52:18 CEST

Original text of this message