Oracle 7.2.3 lock problem on AIX 4.1.5

We have been running Oracle 7.2.3 on AIX 4.1.5 in a stable state for some time. We are running an HA system on concurrent SSA disk. There are two G40 machines in the cluster. Last Friday we upgraded both machines from 512Mb memory to 1Gb. At the same time the Oracle attribute db_block_buffers was increased from 9000 to 27000. The number of Oracle locks remained the same but the distribution of the locks was changed. In the last 3 days Oracle has been crashing regularily (but not predictably) with the following error:

> Dump file /home/oracle/admin/TRNG/bdump/lck0_24316.trc
> Oracle7 Server Release 7.2.3.0.0 - Production Release
> With the Parallel Server option
> PL/SQL Release 2.2.3.0.0 - Production
> ORACLE_HOME = /home/oracle/product/live
> ORACLE_SID = TRNG1
> Oracle process number: 8 Unix process id: 24316
> System name: AIX
> Node name:
> Release: 1
> Version: 4
> Machine:
>
> Tue Mar 10 08:18:12 1998
> *** SESSION ID:(8.1) 1998.03.10.08.18.12.000
> error 50 detected in background process
> OPIRIP: Uncaught error 447. Error stack:
> ORA-00447: fatal error in background process
> ORA-00050: O/S error occurred while obtaining an enqueue. See o/s
error.

We have placed calls with IBM and Oracle support. Oracle suggested to us that the problem is a "file table overflow" with the fix being to increase the number of locks, resources or processes for DLM. Your assistance in determining the contentious "resource" is greatly appreciated.

thankyou,
Randal

--
Randal Kornelsen                email: rkornelsen_at_burnt-sand.com
UNIX System Administrator       phone: (403) 231-6264
Burnt Sand Solutions Inc.       FAX: (403) 290-0060
                                web: www.burnt-sand.com