Re: Arrghh!!

"Billy Verreynne" <vslabs_at_onwe.co.za> wrote in message news:bd6icb$5kr$1_at_ctb-nnrp2.saix.net...
> Fix one problem and then you run smack into another one.
>
> Oracle 8.1.7.4 (32bit) on HP-UX B11.
>
> Been getting this:
> ORA-00600: internal error code, arguments: [711], [1], [1074596568],
[kdiulk
> - prefix context], [], [], [], []
>
> Traced it to a potential problem with compressed indexes. Rebuild indexes
> without compression. This error disappears...
>
> ...only to be replace with this shortly after:
> ORA-600 [25012]
>
> Metalink reports this as an index problem (bugs 1228658, 1138239, 1312233,
> 1297674, etc) - but all fixed with 8.1.6.2 latest.
>
> I know that this should be TAR'ed.. but I'm pushing for a 9i upgrade on
all
> our platforms anyway. And we're likely to see that before we see this
> problem patched (given that 8.1 is being desupported end of this year).
>
> In the meantime I have production jobs that fail. Ones that never have
> failed before with an ORA-00600. The instance itself is unchanged. Been
> patched to 8.1.7.4 last year - no other changes (except dealing with
larger
> volumes of data than before).
>
> Bigger indexes on bigger tables, and using compression btw, on the _same_
> instance work fine. Not a single ORA-00600 except for two specific tables
> on that instance.
>

I'd look at Note 100073.1 which I found through the ora-0600 lookup Utility which metalink supplies if you try to log an 0600 tar - It is *sometimes* pretty useful this.
I'd hazard a guess that the root cause of the compression problems is related to this and not to compression per se.

cheers (er.. i think)

-- 
Niall Litchfield
Oracle DBA
Audit Commission UK
*****************************************
Please include version and platform
and SQL where applicable
It makes life easier and increases the
likelihood of a good answer
******************************************