Oracle Database Discussions

I'm posting this, not in the hope of a solution, but in the hope that you will confirm my outrage at the stupidity of the response from Oracle Support. The following is cut and pasted from my SR.

--------------
UPDATE
=======

Hi Matthew,

This is caused by the following unpublished bug.
Bug 6818207 - SIGSEGV AT QESHIHINSERTPROWINBKT()+36 DURING SELECT GROUP BY

You should use the workaround and wait until a patch is available.
o Set "_gby_hash_aggregation_enabled" = FALSE.
o Wait until a patch is available.

Thanks,
REMOVED

27-JAN-09 11:30:01 GMT
Email Update button has been pressed: Sending email to REMOVED

27-JAN-09 16:54:03 GMT
New info : REMOVED
This is not a 'new' SQL, it has been run several times on Production and non-Production environments. Nothing
has been changed recently.

I am a bit worried about using this hidden parameter. Are there any side effects? (Like could it generate an ORA-600 if
used in conjunction with another parameter for example).

Thankyou.

28-JAN-09 11:03:47 GMT
UPDATE
=======

Hi Matthew,

Your developers should know whether this parameter should have an effect on your system or not, please try to reproduce the pro
blem on a test environment first and try to see if the parameter should stop and
whether it will affect anything else.

Oracle 10g release 10.2 Introduces a new feature called Hash Group by aggregation which allows a hash algorithm to proce
ss group by statements. This functionality can be disabled using the following d
ynamic database parameter:

SQL> alter session set "_gby_hash_aggregation_enabled" = FALSE;

Thanks,

28-JAN-09 11:03:52 GMT
Email Update button has been pressed: Sending email to REMOVED

28-JAN-09 05:26:03 : CHANGES MADE VIA MetaLink
NOT YET FORWARDED TO OUR INTERNAL SYSTEMS :

New info :
Hello REMOVED,

"Your developers should know whether this parameter should have an effect on your system or not, please try to reproduce the problem on a test environment first and try to see if the parameter should stop and whether it will affect anything else."

Quite frankly I'm shocked. The developers here didn't develop the code for the Oracle Server. The bug, as you said, is an Oracle bug, how would developers who develop a product that uses Oracle know if a bug fix sanctioned by Oracle - a hidden parameter for an unpulished bug - impacts their product?

Please ensure that this SR is dealt with by one of your seniors.

I shall be posting the details of the SR on OTN.

------------------------------

Was I being harsh? Should the developers of a product that uses the Oracle Server know about the impact of a hidden parameter?

Would love to hear your opinions.