Problems with large textarea in ITS transaction

[former_member498918]

We have an ITS template which contains a <textarea>. This is working ok for some text, however most of the time when the user enters their text and clicks save the web page hangs.

In SM51 I can see the report name and the process running, but it isn't doing anything. It is then using up a huge amount of memory and causing a bottleneck.

I have put an external breakpoint in the PAI of the report, but the system doesn't get that far. It hangs before running any of the code.

We have recently upgraded from 46c with an external ITS to ERP6 with the internal ITS. This same transaction was working fine before the upgrade.

We have checked all the memory parameter settings we can see. The em/global_area_MB is set to 3GB and the text is only a few paragraphs long. Is there another parameter we can change?

The textarea source code is as follows.

<textarea cols="95" rows="10" name="ASI:95[]">

The SM11 trace files produce the following errors.

-

---

X Tue Sep 23 17:42:18 2008

X *** ERROR => EmActiveData: Invalid Context Handle -1 [emxx.c 2214]

X

X Tue Sep 23 17:42:40 2008

X *** ERROR => EmActiveData: Invalid Context Handle -1 [emxx.c 2214]

X

X Tue Sep 23 17:43:25 2008

X *** ERROR => EgAlloc: MmxMalloc failed (9). see also file 'mm.dump' [egxx.c 521]

X *** ERROR => EmIAllocMmResourceEg: EgAlloc(1048576) failed 5 [emxx.c 3210]

X *** ERROR => EgAlloc: MmxMalloc failed (9). see also file 'mm.dump' [egxx.c 521]

X *** ERROR => EmIAllocMmResourceEg: EgAlloc(1048576) failed 5 [emxx.c 3210]

W *** ERROR => ipl_Malloc EmAlloc(512033) returns 7 [itsplmem.c 148]

W *** ERROR => diag: Memory of 512000 bytes allocation failed, increase em/global_area_MB ? [its_rtmem.cp 524]

M ***LOG W0B=> em/global_area_MB& [its_rtmem.cp 525]

W *** ERROR => Memory allocation failed, size=512000 [its_rtmem.cp 887]

W

W Tue Sep 23 17:43:27 2008

W *** ERROR => xgdk: CtxtAllocSpace, MmxMalloc failed [w3xxctxt.c 1481]

X *** ERROR => EgAlloc: MmxMalloc failed (9). see also file 'mm.dump' [egxx.c 521]

X *** ERROR => EmIAllocMmResourceEg: EgAlloc(1048576) failed 5 [emxx.c 3210]

X *** ERROR => EgAlloc: MmxMalloc failed (9). see also file 'mm.dump' [egxx.c 521]

X *** ERROR => EmIAllocMmResourceEg: EgAlloc(1048576) failed 5 [emxx.c 3210]

W *** ERROR => ipl_Malloc EmAlloc(512033) returns 7 [itsplmem.c 148]

W *** ERROR => diag: Memory of 512000 bytes allocation failed, increase em/global_area_MB ? [its_rtmem.cp 524]

M ***LOG W0B=> em/global_area_MB& [its_rtmem.cp 525]

W *** ERROR => Memory allocation failed, size=512000 [its_rtmem.cp 887]

W

-

---

It then repeats this section every couple of seconds until the session is terminated via SM51.

W Tue Sep 23 17:43:27 2008

W *** ERROR => xgdk: CtxtAllocSpace, MmxMalloc failed [w3xxctxt.c 1481]

X *** ERROR => EgAlloc: MmxMalloc failed (9). see also file 'mm.dump' [egxx.c 521]

X *** ERROR => EmIAllocMmResourceEg: EgAlloc(1048576) failed 5 [emxx.c 3210]

X *** ERROR => EgAlloc: MmxMalloc failed (9). see also file 'mm.dump' [egxx.c 521]

X *** ERROR => EmIAllocMmResourceEg: EgAlloc(1048576) failed 5 [emxx.c 3210]

W *** ERROR => ipl_Malloc EmAlloc(512033) returns 7 [itsplmem.c 148]

W *** ERROR => diag: Memory of 512000 bytes allocation failed, increase em/global_area_MB ? [its_rtmem.cp 524]

M ***LOG W0B=> em/global_area_MB& [its_rtmem.cp 525]

W *** ERROR => Memory allocation failed, size=512000 [its_rtmem.cp 887]

We have recently installed patch 178.

The problem is happening in IE and Firefox.

Is anyone able to help us resolve this issue?

Many thanks

Karen

[Former Member]

The simplest solution is either to reduce the system load by stopping unnecessary processes or to increase the scope of the swap

When you set the memory for the integrated ITS (1), you determine the maximum amount of memory that the integrated ITS may consume.

When you set the number of slots in the template cache (2), you determine how much of this memory is to be used for the preparsed templates.

Changing the slots in the template cache has the following effect:

If you increase the number of slots, the template cache becomes larger. The performance of the ITS applications improves. However, there is less memory for ITS sessions, and you can run fewer ITS sessions in parallel.

If you reduce the number of the slots, the template cache becomes smaller. The applications take more time to run because templates have to be parsed again and again. There is more memory for ITS sessions, and you can run more ITS sessions in parallel.

If there is sufficient memory, set the number of slots so that there are always empty slots in the template cache. There is then no displacement in the template cache. (Best performance)

[former_member316351]

Hello Karen,

The most likely explanation is that em/global_area_MB is not what you think it is for some reason. I would double check in RZ11 for the current value and note 742048.

If the value is 3gb and the issue persists then have a look in transaction SITSPMON and the "Memory statistics" tab when the issue is reproduced. Check the "Currently available to ITS:" and the Session Memory Used for this users session.

Edgar

[former_member498918]

Thanks for your response. I have been checking transaction sitspmon which is where I can see that the ITS system is running out of memory.

In the live system the basis team have told me that there is 3GB available and that they won't increase it becasue they say the transaction shouldn't be using that much.

I guess the problem is why is the transaction using up all the memory when all it is sending is a report with 4 paragraphs of text and a list of project information? Surely 3GB should be enough?

Is there another memory parameter or any other parameter that we can change to help stop this problem.

Do you know what the trace file is saying. Is it only complaining about the size of the em/global_area_MB or does the rest of the report mean anything to anyone?

Thank you

Karen

[former_member316351]

Hello Karen,

Based on your reply it is not clear what value is exceeding 3gb or if it is at all 3gb.

Please reproduce and state the following values from SITSPMON Memory statistics tab:

- Memory Consumption: Overview

Sessions:

Templates:

Sess. & Templ.:

Currently available to ITS:

- Session - Memory used

ITS Session

Total

The main issue here is if this one session is taking up a lot of memory or if it is the ITS in general. This will also show what the current em/global_area_MB value is. Please note that the em/global_area_MB is not only used by the ITS.

The trace shows that the solution is to increase the em/global_area_MB.

Edgar

[former_member498918]

I have checked the sitspmon ITS memeory stats.

The system is saying 0MB of 3000MB available, which I expected as the problem is a memory bottleneck.

The strange thing is that the user causing the memory to be used up is in the table with a huge negative value in the Peak and Current Memory, 1,746,409,275-

Does anyone know what is causing this or what this means?

Thank you

Karen

[former_member316351]

Hello Karen,

The Memory Used values for a workprocess frequently have a "-" at the end. It's being corrected or is corrected already with newer 700 systems.

What's more important is what the user is doing to use all the em/global_area_MB or why it's not being increased when the system has 0 em/global_area_MB left.

Edgar

[former_member498918]

I have reduced the column size of the textareas in the report from 95 columns to 90 and this seems to have fixed the problem with the memory being used up.

Karen