Cloning a psoft DB 1

[nohandlenah]

I am a new oracle DBA who will be working on this PeopleSoft database running on PS financial 8.1(AP, AM, AR, Project ..). I need your advice on a couple of things:

1) If I want to create a duplicate test oracle database(using another SID name) from production, what should I do? Presently, I can clone an identical oracle db with another oracle SID easily, but should I change just the psdbowner table like 7.5? Anything else I need to change? Also, should I use oracle imp/exp, or just cerate a blank db and you use datamover or import manager?

2) What are the biggest tables on your system? What are the most trouble making batch? At least I have a feel of what hits me when a production problem happens. Does the application engine provide any tracing tool aside from turning on an instance wide oracle trace for the batches?

3) Has anyone tried using table partioning? Does it give any performance improvement? I don't find it in PeopleTools although I know it's there.

Any of your advice would be much appreciated.

Thanks,
Jonathan.

 

[skhowell]

1) Once you create the test db via either exp/imp or through datamover you need to run the last portion of the GFDMOORA.dms. Cut out the import statements in the top portion. You have to run the grant.sql as well.

2) Your table sizes are basically dependent on what application modules within the Fins application you have. In 8.1 there is limited tracing of Application Engine. I have found two useful ways to trace. Via the process scheduler trace settings in the psprcs.cfg and running the application engine from the command line. There is documentation in PeopleBooks regarding each method.

3) Partitioning the tables works well in PS as it does in other applications. There is a definite performance improvement.