cancel
Showing results for 
Search instead for 
Did you mean: 

ICM - Emergency Mode - Not Running

Former Member
0 Kudos

Hello,

This seems to happen for the third time in the last two months. All of a sudden we get a call from storage space monitoring team that this particular drive is outgrown by the 'DEV_ICM.txt' log file This happens in one of the three application servers in our WAS 6.20 environment where ICM server is running. Within two hours, the log file grows to 8 GB. The only log error message I see all through the 8 GB log file is this:

[Thr 2952] *** ERROR => HttpLogParseItem: premature EOS in format

[http_log.c 2370]

When this happens, when I try to execute 'SMICM', I would notice that ICM is in emergency mode and status 'not running'.

All of a sudden in a day or two, this would get fixed by itself and ICM would get back to normal. We created a message with SAP and the response we got from them was 'as this got fixed by itself, we cannot tell exactly what could have caused the problem'. Have any one of you came across such a situation and can one of you throw some light on what's going on?

I would be more than happy to reward anybody and everybody who can help me fix this problem.

Thank you.

Accepted Solutions (1)

Accepted Solutions (1)

Former Member
0 Kudos

Hi,

I also use ICM on WAS 6.20 (6.40 kernel) and did not experience this issue even with a high web services and BSP load.

Did you check the ICM trace level ? maybe it has been increased to the maximum and in that case the ICM is VERY verbose.

Regards,

Olivier

Former Member
0 Kudos

Hi again,

This may be an ICM bug.

If it's possible, you should try to upgrade your kernel (and so the ICM executable).

Regards,

Olivier

Former Member
0 Kudos

Olivier,

Thank you for both the suggestions. Here are my responses / questions:

1. As I had mentioned earlier, the ICM is in emergency mode with status 'not running'. In this mode, I tried to change the trace level but, I always get the message that this action cannot be performed in the emergency mode. As you said, let us assume that the trace level had caused this problem and caused the ICM to shutdown / emergency mode. How do I bring the ICM back up and running, so that I could change the trace level to 1 if it is set to 3. Right now, the trace level is 0 because, that's what happens whenever the ICM is shutdown / emergency mode. It won't show the actual trace level.

2. With regards to Kernel upgrade, here is some background information. As I had mentioned in my previous message, we have three application servers for WAS 6.20 (kernel 6.40). When we had ICM emergency mode / shutdown last time, we noticed that the kernel patch level was different to this app server compared to the other two. So, we brought the patch level of this app server equal to the other two. That didn't fix the problem. Also, another piece of information - we don't see ICM in emergency mode / shutdown when we log into the other two app servers and execute SMICM. They seem to run fine.

Your response is highly appreciated.

Thank you.

Former Member
0 Kudos

SAP007,

Your problem is really weird.

As it happens on one only of your 3 app servers, you should check every difference of this app server compared with the 2 others.

Do you have a webdispatcher to load balance the HTTP requests between the 3 servers ?

Or do you have different kind of http requests for each server ?

Regards,

Olivier

Answers (1)

Answers (1)

Former Member
0 Kudos

I think I got this resolved by updating kernel. Don't remember exactly. Thank you for all your support. Closing the message now.