Problems with performance of PI PRD

[Former Member]

Hi experts:

I have a problem with PI server in PRD, the File System and the Database haven't been debugged (delete of messages) for about 2 years and the system has become slow and we have also had problems transporting objects from QA environment to Production because these aren not being updated (or refreshed) and therefore we had problems with the version of the objects.

The solution for Basis Team was to reboot the server and they suggest to do this task every 2 months but I don't think this should be done because it's Production environment what we are talking about and it should always be available to do the integration processes.

I would like to know what do you think about the Basis team solution, is it a good idea to reboot the production server every couple of months or every so often?

Thanks for your answers.

Regards,

Julio

[Former Member]

hi,

no, prod should not be restarted as often.

If PI is restarted and if AE is linked to your PI, then all your communication channel will run at the same time (could have an overload) and perhaps these CC will run at a time which is not wanted by business (could have conflict between employee's changes and interface flow, especially if interfaces work normally in batch nigth).

Before a restart of a PI prod, you should analyze the impact for the Business. And normally this action should be approved by the Business itself. And after a PI server restart, you should control that nothing is "waiting / holding / in error" and/or stuck in ECC / PI queue. Do you want to do these controls every 2 months ?

It's better to analyze the reason of a slow PI and why not to:

- Archive old messages (or eventually delete them, that depend on your business needs and legal laws).

- Archive or delete exchanged files, if these ones are stored directly on PI server (itself)...

- Increase the allocated tablespace of your hard disk to PI database

- Increase physical / virtual memory.

regards.

Mickael

[Former Member]

Dear Julio,

Please do all the checks as mentioned in the earlier response.

Along with that, the BASIS team should set up routine jobs for the Health check and clean up of the PI PRD server.

This is in general called as the House Keeping activities.

This is a normal problem for PRD servers that have last 2+years OR if the message volume is very high.

There are some standard jobs that have to be set up which Archive/Deleted the old processed data which is not needed any more. The retention period needs to be decided as per the business need.

Also refer the Note 872388 - Troubleshooting Archiving and Deletion in PI.pdf

Thank you

Younus Mohammed

[Former Member]

Hi Julio,

Before the System Restart is performed, in my opinion the System's Analysis is needed.

The overall performance problem for me is to know if SAP response time is acceptable... If not, do this :

- at AIX level : utilities are topas, to catch what's wrong : Disk I/O ? CPU bottleneck ? Too much swap ?

- at Oracle level : tkprof but don't know much about this.

- at SAP level : you can manage SAP memory, Oracle and OS too.

If SAP response time is acceptable, then try this :

- Try to have a look at OS response time : ST06 : look at CPU, Memory, Swap, Disk and Lan response time.

- Try to have a look at buffer quality : ST02, if many fields are red, investigate each fields dependant parameters.

- Try to have a look at DB response time : ST04N, databuffer quality (SGA zone), how much physical reads / logical reads Ratio, wait times, number of user calls, Shared Pool cache hit ratio should be >96%, Sort usage, Redo logging.

Take a look also to this 3d: