Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Oracle PL/SQL & Solaris Weirdness.... Help Please.

Oracle PL/SQL & Solaris Weirdness.... Help Please.

From: XPipe <XPipe_at_spamproof.net>
Date: Mon, 1 Jun 1998 22:14:59 -0500
Message-ID: <6kvqp0$r4$1@excalibur.flash.net>


Ok, I have no idea if I will be providing enough info on this, but on the off chance that someone in the group has seen this...

I have a team that has developed a generic multi-threaded server architecture that has recently been ported from AIX to Solaris. We have multiple threads running, each with its own database context, banging up against Oracle by calling a PL/SQL package and
then returning the results, if any, from that call. We are experiencing the following intermittent bug. Trace files show the following:

  1. Our software receives a transaction. Pass to package.
  2. Package recieves the request and passes SQL to oracle for processing.
  3. Oracle goes hay-wire and tries to execute that same code 30 or 40 times. Ultimately it returns a success code, with no data, even when the same query had received results before.
  4. PL/SQL reflects success code from DBMS and returns NULL to caller.
  5. Our routine returns NULL to client who sent the transaction.

Now, the weird thing about this is it doesn't happen every call. I can send the
same transaction to the server with the same data and it will only happen every
fourth or fifth time. The rest of the time I get valid results back. We are pretty
sure this is an issue with the DBMS and not our software since we can see this
strange behavior in the oracle trace logs. Oracle of course, insists that it must
be us since it can never be them. We have so far talked to them in Florida, London,
and Australia.

Oracle version is 7.31
Soloris: ???? Not sure. Most recent I beleive.

This bug happens on both a Ultra 2 and an Ultra Enterprise 5000.

Does anyone have any remote ideas on what might be causing this? I am rapidly
coming to the conclusion that Oracle is going to be little or no help in this matter
unless we can debug the problem for them.

All input welcomed,

Grego Received on Mon Jun 01 1998 - 22:14:59 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US