SMQ2 - Queue monitoring & reprocessing

[former_member185846]

Dear Experts,

We need to automate qRFC/tRFC queue monitoring and reprocessing. Which is the best way to do that out of the 3 mentioned below or any better options?

1. Write custom ABAP programs to monitor and trigger alerts incase of queue issues  - I've seen the code somewhere in SCN but unable to track it now. Any help please?

2. Use the standard program as described in http://scn.sap.com/people/community.user/blog/2005/11/29/xi-how-to-re-process-failed-xi-messages-aut....

But, not sure whether this will park failed messages in SMQ3?

3. Use CCMS templates to monitor and alert queue issues - We already have a set up to monitor PI mapping/adapter engine failures. http://help.sap.com/saphelp_nwpi711/helpdata/en/f0/02a63b9bb3e035e10000000a114084/content.htm - is this the way to configure this? Or any other document available?

Thanks,

[Former Member]

Hi,

have a look at the CCMS setup blogs:

http://scn.sap.com/people/aravindh.prasanna/blog/2005/12/23/configuring-scenario-specific-e-mail-ale...

http://scn.sap.com/people/aravindh.prasanna/blog/2005/12/24/configuring-scenario-specific-e-mail-ale...

http://scn.sap.com/people/aravindh.prasanna/blog/2006/02/20/configuring-scenario-specific-e-mail-ale...

CCMS setup is easy to setup and you can receive alerts by setting up the sapconnect.

Apart from queue monitoring you can check the health of adapter engine.

Some can use the feature to design some use cases like:

1) More than 200 are there at a time for a interface.

2) Message processing time is more than the threshold value set.

3) messages suck in queue

4) Error in BPM

Thanks,

Sudhansu

[udo_martens]

Hi Joe,

SAP XI/PI: Alerts for tRFC(SM58) Errors

SAP XI/PI: Alerts for Queue Errors

The screenshots are gone, but the ABAP source code is still there - i love the new scn design

Regards,

Udo

Message was edited by: Udo Martens

[former_member185846]

Thanks for the prompt response guys. But, I think I need to go with the custom program option as the standard approach may throw an alert but won't save off the failed messages from SMQ2 (for the other messages to get processed from the queue). Right?

@Udo, which one would you suggest? Or any other alternates available?

Thanks again.

[Former Member]

That is correct. Is there some special reason in your environment that makes you want to do that? If the queue is in error status because of connection problems, all subsequent messages would also fail. If it is just one scenario that is causing problems and blocks all others, why not try other measures like assigning this scenario different queues or solving the problem that causes the error? How do you deal with the possible inconsistencies because of overtaking messages? In the end you will have automated queue monitoring, but you'll have to deal manually with SMQ3.

Regards,

Jörg

[udo_martens]

Hi Joe,

>Udo, which one would you suggest?

Usually there is more risk that message could hang in SMQ2 of PI. Sometimes messages stuck in SM58 of the ERP system. Theoretically they can stuck in SMQ2 and SM58 of both systems.

I think the reports are very useful, because you might  not realize locked queues.

Regards,

Udo

[Former Member]

you should go for mix of 2 and 3.

for restarting the failed one you need 2nd approach, ans if you want immidiate alert CCMS is the best one.

Regards,

[Former Member]

Hi Joe,

I recommend you schedule the standard jobs in any case. This is easily done and might solve a lot of errors automatically. It won't take entries from SMQ3, since it is supposed that you would park entries there manually and for a special reason only. So it wouldn't make sense to pick them up automatically. If you park entries there and reprocess them later there is always a risk of inconsistency - a risk that no machine rule could calculate. So this will be left to the administrator.

You can use CCMS to avoid active monitoring, it is the standard way and should work reasonably fine. Unless you go for the more recent versions of PI this is the way I would recommend.

Regards,

Jörg

[Former Member]

Hi Joe,

The third option provided by you is the best option and standard for monitoring of queues.

The advantage of using CCMs alerts is that it gives you a consolidated email based on multiple queue failures at the same time so it works properly and as well as do not flood your mail Box.

You can search for more document available or should contact your BASIS team to configure CCMS alerts.

Thanks.