cancel
Showing results for 
Search instead for 
Did you mean: 

SOP to Restart PI System

suwandi_cahyadi
Contributor
0 Kudos

Dear Expert,

What is the right Standard of Procedure to restart PI system?

Do we need to first close all of the CCs before restarting the server?

Thank you,

Suwandi C.

Accepted Solutions (1)

Accepted Solutions (1)

Dimitri
Active Contributor
0 Kudos

Hi Suwandi,

You can restart the system without stopping the CC's.

But if the system is restarted, all traffic will be triggered again. To avoid that, you can stop the sender CC's and restart the system. After the restart you can start the CC's one by one to keep it under control.

Kind regards

Dimitri

Answers (4)

Answers (4)

marksmyth
Product and Topic Expert
Product and Topic Expert
0 Kudos

Hello,

In addition, refer to Controlled restart of a J2EE engine in a PI system with multiple application servers if this

matches how your PI system is set up.

Regards

Mark

Former Member
0 Kudos

Hi

Refer below blog on   XI/PI Maintenance activities.

Regards

Mastan

former_member184720
Active Contributor
0 Kudos

Hi Suwandi -

Can you go through the below link? SAP has provided SOP to restart a PI system.

All the stop and start steps explained very cearly...

http://help.sap.com/saphelp_nwpi71/helpdata/en/12/5dd23aaad7b3408dde3c1577fda5b3/frameset.htm

Former Member
0 Kudos

Hello Suwandi,

the "right" procedure depends on your particular installation. But generally, yes, close all CCs and second, de-register all queues to avoid IDOCs or proxy requests coming in (SMQR) in case of double-stack system. Wait till all running processes (mappings, open queues etc.) are finished. Be careful when you restart the system and don't open queues in ABAP before the Java part is up and running. Big backlog may also cause problems, so better check that before and make a controlled one-by-one start before you open the gates completely.

This is not a must, the system should be able to handle a stop in the middle of business, but you never know. Depending on the processes running and on custom implementation you may run into trouble. So this is just a recommendation from experience.

Regards,

Jörg

Former Member
0 Kudos

Hi Suwandi,

In case you are using a clustered environment do also check if there are any messages stuck/in peocessing in any of the instances, status like To be delivered status, waiting. Wait for them to be processed before restarting the server.

Thanks,

Sudhansu