IDoc Data Lost between ECC and MII Message Listener.

[Former Member]

We have recently experienced some network issues which caused the IDoc's to stop flowing outbound from ECC to MII Message Listener. This happened 3 times during a network switch reconfiguration. 2 of the three occasions the MII Message listener had to be restarted on the MII server in order to get data to flow again. Interestingly some of the IDocs that were generated during this outage time were processed by the message listener and some didn't.

We are running MII 12.0 sp11 and ECC 6.0. The ECC server and MII server are located in different geographic locations.

When we look at the ECC system in WE05 we see only sucessful status for the IDoc messages saying the Data passed to port OK.

When we check the Message listener on the MII end there it only shows processed successfully. There are no failed messages or messages without rules.

Where can I check further to see the IDocs that really didn't make it to be processed by the Message Listener on the MII server?

Also is it possible that these IDocs that got lost somewhere are still sitting on the ECC or MII servers somewhere where we can re-process them?

And the Last question is, Why didn't the Message Listener handle the network issue automatically instead of needing to restart it?

[Former Member · Accepted solution]

Hi Robert,

I know this is an old thread - but I was wondering if this ever got resolved? We face the same issue (MII 12.1) when there's some network issues or maintenance. I don't tend to see missing messages - but we see a single message being stuck in 'processing' and it doesn't complete. The only solution is to restart the MII services and this is obviously disrupts production. We've 6 MII instance in various locations and this causes a lot of monitoring when any network maintenance is planned.

Cillian

[Former Member · Accepted solution]

Cillian,  since the time of the original post we have upgraded to version 12.2.4 .   The issue does not seem to be happening now like it used to.  Because of the intermittent nature of this issue I can't say one way or the other that it is resolved or not.

One other change in our Wide Area Network (WAN) that has helped with the reliability of our SAP MII distributed installation was to place WAN traffic to and from the MII servers into a higher level of Quality Of Service (QOS).  This was overlooked when QOS was first implemented, resulting in SAP MII traffic being placed in the default lowest QOS queue which caused packets to be dropped on busy links.

[former_member185280 · Accepted solution]

Cillian,

This Blog may offer a possible solution:

Regards,
Christian

[Former Member · Accepted solution]

Hi Robert,

Did SAP ever respond/resolve the ticket you created for this issue? Someone told me that this was a known issue, but I have yet to have it verified by SAP. They asked me to simulate the issue - but our non-prod systems are on a VM server with other systems. So I can't exactly start pulling network cables out - I did drop it from the network a few time while sending IDOCs down - but this failed to hang the processing of the IDOCs.

We're hopefully going to MII 14/15 next year - but the business is keen to resolve this issue as it has impacted production a few times. The MII sites are in some remote regions and I don't think we can really improve the WAN links much further.

@Christian - thanks for the link, we don't really have an issue with IDOC's not being received, its just the message processing just hangs and we either need to restarted the services or sometimes the system.

[Former Member · Accepted solution]

looking back at the support ticket with SAP we were informed that this issue was fixed in v 12.0 patch 2 for SP11.  I believe that is when we decided to move forward with our upgrade to 12.2.4.  vs. applying the patch.

I also believe the upgrade did resolve this issue however due to the intermittent nature of the occurrence and the issues of duplicating the problem it is difficult to replicate it in a testing environment.

I hope this helps.

Bob

[Former Member · Accepted solution]

Thank you for contributing this possible solution.

Bob

[former_member185280 · Accepted solution]

An interim solution may be to put up some 14 boxes just to handle the IDOC integration in lue of a full port/upgrade.

Regards,
Christian

[Former Member · Accepted solution]

Hi Bob,

We're on 12.1 SP7 - SAP said it was fixed in release 12.2 SP11?

Regards,

Cillian

[Former Member · Accepted solution]

SAP said it was fixed in 12.0 SP11 patch 2.

I do not know about version 12.1

[former_member193328]

The IDoc handling between MII 12.1 and MII 12.0 is completely different at the technical level. So not sure if the fix for MII 12.0 would affect/fix the problem in MII 12.1. However MII IDoc Processing in 12.1 is heavily dependent on JMS. I had seen some instances where the JMS infrastructure of NW 711 on Windows Server. It may cause processing to get stuck. I think this was solved with NW 730. You may try it if this is a feasible solution.

Regards

Partha

[Former Member]

Hi Partha,

We're planning on going to NW 730 and MII 14/15 this time next year - so hopefully this will resolve the issue. But it would be helpful if there was something official from SAP outlining the issue and the releases that are affected. One year is still a long time to live with such an issue - if they recommend a patch we could quickly roll that out. I don't know how many times I tried to simulate the issue - it seems almost impossible to capture.

Regards,

Cillian

[jamie_cawley]

SM58 should show you any RFC errors that occured. You can resend them by selecting one, choosing the edit menu and selecting execute luw.

Jamie

[Former Member]

Thanks Jamie, I had my Basis guy check SM58 and there were no stuck IDOC's

[Former Member]

Robert,

Are there any clues in the NW log for the MII system? Also, have you entered a support ticket for this issue?

Regards,

Kevin

[Former Member]

Thanks Kevin for responding.

I did check the NW logs on the MII system, both SAP logs and Default Trace for the timeframe where the issue happened and did not find any errors or warnings relating to the issue.

The issue has been entered in SAP's support system.

Edited by: Robert SyWassink on Jan 26, 2011 9:06 AM

[agentry_src]

Robert,

Can you provide the ticket number?

And let me make sure I understand correctly. You are sometimes receiving the IDocs in the listener and sometimes they just disappear? Is there any pattern (time of day, type, etc.) to when they are successfully received?

Regards,

Mike

[Former Member]

Ticket Number is 0000073834.

So I beleive we receive the IDOCs 99 % of the time. We have been running in this configuration for over a year.

The issue was when we had maintanence outages on our main network switch between the ECC and MII sites. This doesn't happend very frequently.

ECC indicated the IDOCs were delivered to the port successfully in transaction WE05. Yet they did not show up in the MII Message listener. (message listener doesn't display and IDOC info. just by counting the number of enteries in the Message Listener). So dissappears is a good term.

recreating the problem is not really possible since it would interrupt our operations to entire site by taking the main network switch offline.