From: Noons on
On Jun 2, 11:07 am, zigzagdna <zigzag...(a)yahoo.com> wrote:

> I changed default 800 seconds undo_retention to 1 day. I got past that
> error, but now another error came on a different table:
> ORA-00600: internal error code, arguments: [b], [23786], [121640375],
> [121640408],
> Could not find anything on  kdliSyncRead on google or metalink.
> Looked at trace file, seems like some core dump.
>
> Oracle 11.1.0.7.1 has all kind sof bugs; problem is now I am in
> production so cannot back out of this release.



Narh! Not possible!
So let me see: those of us who Oracle called "dinossaurs" because we
refused to upgrade - knowing the problems with new releases of Oracle
- were right after all?
Wonders will never cease...
From: Steve Howard on
On Jun 2, 12:45 am, Noons <wizofo...(a)gmail.com> wrote:
> On Jun 2, 11:07 am, zigzagdna <zigzag...(a)yahoo.com> wrote:
>
> > I changed default 800 seconds undo_retention to 1 day. I got past that
> > error, but now another error came on a different table:
> > ORA-00600: internal error code, arguments: [b], [23786], [121640375],
> > [121640408],
> > Could not find anything on  kdliSyncRead on google or metalink.
> > Looked at trace file, seems like some core dump.
>
> > Oracle 11.1.0.7.1 has all kind sof bugs; problem is now I am in
> > production so cannot back out of this release.
>
> Narh!  Not possible!
> So let me see: those of us who Oracle called "dinossaurs" because we
> refused to upgrade - knowing the problems with new releases of Oracle
> - were right after all?
> Wonders will never cease...

This is really scary. The problem is, everyone will just say, "well,
you should be on 11.2.0.1, anyway". In a year or so when another post
comes through talking about how buggy that release is, the cycle will
repeat itself.
From: John Hurley on
Zigzag:

# I changed default 800 seconds undo_retention to 1 day. I got past
that error, but now another error came on a different table:
> ORA-00600: internal error code, arguments: [b], [23786], [121640375], [121640408], Could not find anything on  kdliSyncRead on google or metalink. Looked at trace file, seems like some core dump.

.... Oracle 11.1.0.7.1 has all kind sof bugs; problem is now I am in
production so cannot back out of this release.

Pretty darn stable for me so far all things considered.

Seems like your plan for testing out your planned release level before
moving into production was missing some really big pieces.

Why on earth would one put in a very large undo tablespace with a very
small value set for undo_retention unless one had missed many of the
important concepts involved here in this very important piece of
system architecture?

From: John Hurley on
Noons:

> > Oracle 11.1.0.7.1 has all kind sof bugs; problem is now I am in
> > production so cannot back out of this release.
>
> Narh!  Not possible!
> So let me see: those of us who Oracle called "dinossaurs" because we
> refused to upgrade - knowing the problems with new releases of Oracle
> - were right after all?
> Wonders will never cease...

Well actually some of the dinosaurs are running that release and
finding it pretty good. Some of the dinosaurs though actually do a
systematic job of planning and testing things.

You have to remember the OP in this thread did not seem to know where
to find resources and information on snapshot too older ... so take
some of the complaints from the OP with a large degree of skepticism.

There is a large difference between refusing to upgrade because one
has tested and found things in their environment and and their
applications which cause serious ( or critical ) errors and refusing
to upgrade just because something is new. Ya gotta remember that 11.1
has been out now for a pretty long time eh kimosabe?

We will be starting the phase of seriously checking out 11.2 just as
soon as the first patchset is available although it might be sooner.
From: zigzagdna on
On Jun 2, 7:31 am, John Hurley <hurleyjo...(a)yahoo.com> wrote:
> Zigzag:
>
> # I changed default 800 seconds undo_retention to 1 day. I got past
> that error, but now another error came on a different table:
>
> > ORA-00600: internal error code, arguments: [b], [23786], [121640375],  [121640408],  Could not find anything on  kdliSyncRead on google or metalink. Looked at trace file, seems like some core dump.
>
> ... Oracle 11.1.0.7.1 has all kind sof bugs; problem is now I am in
> production so cannot back out of this release.
>
> Pretty darn stable for me so far all things considered.
>
> Seems like your plan for testing out your planned release level before
> moving into production was missing some really big pieces.
>
> Why on earth would one put in a very large undo tablespace with a very
> small value set for undo_retention unless one had missed many of the
> important concepts involved here in this very important piece of
> system architecture?

Every release has soem issues. I tested Oracle 11g for 6 months; yet
many of the problems found in production never showd in testing. I
ahev following major issues with 11.1.0.7.1 on HP UNIX 11i which I
used:
1. db control does not work. I had problem in creating db control in
QA, but in production I did not get any errors during database
instance creaion. BUt then I found in laert.log, it is generating core
dmp every minute. I removed db control from databas eand problem went
away.

2. Export dump has errors as in this thread.

3. rman has bugs sometimes it ends up in an infinite loop during
archived log backup and my lof file keeps growing and fills my file
system.

4. Once I restarted the database and it showd it mounte dbut never
showd db opened, eventhough db was opened.


Prem