Trace Entries Missing.

[bhavesh_kantilal]

All,

We are on a production system and I just observed something strange.

In my mapping ( java mapping) we have added a few Trace Entries. What I found was that the Trace was missing for a few runs on the scenario and the trace was found in the Trace of another Scenario.

In other words, one of the Runs , had the multiple trace entries belonging to different runs of the same scenario. We are using the Idoc number in the trace and so we can confirm that the Trace Entries are getting merged / lost in some runs.

Any ideas as to why this is happenning? or how to go about finding the root cause of this issue?

We are on XI 3.0 Sp 14 .

Regards

Bhavesh

[Former Member]

Hi Bhavesh,

What about the trace entries put in by XI itself ?

Are they also getting mixed up/missed ?

Thanks.

[bhavesh_kantilal]

Deepu,

The Default trace Entries are correct.

Rajeev~

The Prod Server has been quicksized and is working perfect without any issues. Also, 30 K idoc's is a small number as we have earlier had a peak load of 120 K idoc's hitting XI without any issues. This seems ti be a one time problem, that happened like you told correctly due to some server issues.

Am just trying to find out why so that this doesnt happen again.

Thanks for the ideas.

Regards

Bhavesh

[Former Member]

Hi Bhavesh,

i am not sure why this problem is occuring on your productin XI system...but i guess that this may happen that your XI system is sometimes running out of memory because of some background process of either your OS or some other application, so what happpens is that your trace is written in memory buffer, but due to slight lack of CPU thread or memory, this buffer is not able to be written in the correct message log.....but when the scenario is again run, then if CPU threads are slightly free or some more memory is there, then the trace of current run is written in buffer and then system writes the completer buffer of previous run and current run in the current message trace.......- i just guess this may be happening in your XI production system.

Well today i am surprised to see that you also ask question on this sdn forurm

Thanks,

Rajeev Gupta

Message was edited by:

RAJEEV GUPTA

[bhavesh_kantilal]

Rajeev,

Well yes, this happens when the data load is high on our server. Around 30 K Idoc's hitting XI at the same time.

Though it does not affect our scenarios , they run well without any errors , the trace entries is something that we have added for monitoring purposes.

You answer does help though.

Regards

Bhavesh

PS : There is lots on XI which I still have not figured out.

[Former Member]

Bhavesh,

If so much data as 30K hits your XI system at same time - if this is frequent, then i will advise you to please ask your basis guys to increase your XI prod memory, XI virtual memory(cache memory) and if possible to add some more CPUs for parallel processing of your XI messages......so that your this problem of trace getting mixed up gets resolved.

Well it was nice to know that my guess in your this problem - helped you

Thanks,

Rajeev Gupta