Re: Bugs in Oracle 7.3.2

Stephen A. Rodgers wrote:
>
> We've encountered some rather nasty optimizer bugs in Oracle 7.3.2 on
> Solaris. One of the bugs causes Oracle to core dump with either a
> SIGBUS error or a SIGSEGV error, neither of which is true. Then, as a
> result of the bug, the SGA becomes corrupt and we sometimes even lose
> indexes. I've applied a patch which helps somewhat, but yesterday I
> still got the abend. We are considering dropping back to 7.2 before we
> go to production. There appear to be far too many bugs related to
> Oracle's cost optimizer for us to continue, and 7.3.2.3 is not ready for
> Solaris yet.
>
> Has anyone else encountered bugs with the 7.3.2 releases? I'm just
> curious.
>

We have had many problems with any of Oracle's production releases of 7.3. Currently we are still running the beta version we received 7.3.1 and have not been able to move to the production release because of the bugs we have encountered, and we can't go back to 7.2 because we need the ability to write threaded code with Pro*C which is a feature of 7.3.

The bugs we have encountered are:

Pro*C code running 10 times slower then the beta release. Troubles with the SQL*NET in 7.3.2.2 (which was to have fixed the slow Pro*C problem) in even being able to connect to an Oracle database.

Needless to say, we are hoping that Oracles 7.3.3 release has all of these bugs fixed. In my opinion, in Oracle's desire to come out with a Universal Server before Informex had one, Oracle sacrificed some of there normally very through testing they do on their database kernal. Hopefully they have solved these problems or Oracle 8 will become as reliable as some of Oracle other tools.

Just for your information, we are also running under Solaris.

Eric Mortensen
mortense_at_chirondiag.com Received on Wed Apr 02 1997 - 00:00:00 CST