Too many dialogs get stuck on CPIC stopped,SAP hangs intermittently on prod

[Former Member]

Hi All,

I'm having a problem with a production system in which users started complaining about getting the hourglass cursor intermittently, and this is happening from multiple different locations / networks. Around the time the problem occurs, SM66 shows many dialog processes in status "Stopped CPIC". The processes are from different users and processes.

I've checked the workprocess trace files, made sure there was no memory bottleneck with the gateway profile params, checked System log and oracle alert log, but cannot find any clues.

Could anyone please suggest some other places to look for clues or possible suggestions?

Thanks!

Ian

System info

Netweaver 7.0, ECC 6.0

Oracle 10.2.0.4.0

AIX 5.3

[manu_susankar · Accepted solution]

Hi Ian Comyns,

CPIC status is due to the problem with communication (some times user locked in the externel system may cause) check here for the diffarent [CPIC|https://www.sdn.sap.com/irj/scn/advancedsearch?query=cpic&cat=sdn_all] related issues.

also check SMQ1 and SMQ2 and SM58

Regards,

S.Manu.

[Former Member]

Hi Ian,

You see many processes running for quite some time with Reason = CPIC then you know that there is probably a problem with external interfaces that the system tries to call.

You could check the system logs for problems, STAD and so on - but you'd wish you could have taken a look at SM66 exactly at the time of the problem.

[Former Member]

Thanks for all your comments (points given).

The problem and solution was very interesting.

1. We found that an external AS400 system that communicates with SAP was somehow sending the IP address in EBCDIC format, which looks like strange characters. We saw the strange characters in the Gateway NI/Hostname buffer. During each hang situation, we saw these special characters in the Gateway trace file. The Gateway was attempting to resolve the IP address for the strange EBCDIC characters, but the EBCDIC characters were actually the IP address itself when converted from EBCDIC to ASCII.

2. As per SAP, we added the strange character hostname/IP to the /etc/hosts file with the correct IP, and the hangs stopped happening.

3. After these 60 second hangs stopped happening, we were still getting 10 second hangs every 3 hours.

4. SAP found that the Gateway logging option was causing the 10 second hangs.

5. SAP is working on a kernel fix so that the Gateway doesn't hang in these situation. We have the workaround in place for the time being.

Thanks,

Ian

[Former Member]

"CPIC Stopped" is generally associated with Remote Function Call (RFC). Check SM50 for any RFCs.

[Former Member]

We're having a major impact to production...if anybody can offer any advice or rough ideas I would be very greatful. I have a high priority case open with SAP, but its going a little slow so I'm trying anything I can

Many Thanks,

Ian