xMII 11.5 Slowness

[Former Member]

Hi Folks-

We are running an instance of 11.5.3 xMII and have seen a continual performace slowdown, particularly when the application is first started/called. We have mitigated some of this via query caching but I wonder if there is something else going on. We are running on a Windows Server class machine with 4GB of RAM. I was wondering what if anything I can check to insure performance is AOK or needs to be adjusted. In addtion I was wondering if I can update the Java Plug-in as I am running as shown below:

Java(TM) Plug-in: Version 1.4.2_12

Using JRE version 1.4.2_12 Java HotSpot(TM) Client VM

Any inisghts would be appreciated. Thanks and have a good weekend.

Kevin Fitzgerald

[Former Member]

Thanks for the suggestion but I remain a bit confused. Are you saying the JRE should be at 5.0 or 1.4.2.x?? Currently the server machine is set up as follows:

Java(TM) Plug-in: Version 1.4.2_12

Using JRE version 1.4.2_12 Java HotSpot(TM) Client VM

User home directory = C:\Documents and Settings\Administrator

Proxy Configuration: No proxy

Is this OK?? Thanks.

Kevin

[jcgood25]

Kevin,

Server-side symptoms of an application induced memory leak would be a continual and progressive slow down culminating with a potential out of memory error or complete unresponsiveness. I would also expect that increased user activity would accelerate this issue.

Client-side symptoms would seemingly be remedied by just closing your browser instance and starting fresh.

You mention sluggishness when it first starts or when it is called - this sounds like it might be going to 'sleep', which looks more like some of the Windows IIS application pool tuning options giving you fits. There are 2 checkbox settings on a Windows 2003 environment that are related to this (which I typically recommend unchecking):

1. Recycle worker process (in minutes)

2. Shutdown worker process after being idle for (time in minutes)

The cms logs would be the first place to look for any signs into what might be causing any application related issues, and the system stats log would be a good way to see if you are experiencing any memory leak issues over time.

Because it's 11.5 you have to stay with Java 1.4.2_xx on the server (5.0 is just a different name for 1.5 which is not allowed by the server platform until MII 12.1 with NetWeaver CE).

If the issue is related to your application then it would obviously require a bit deeper study than my subjective guesswork here on the forum

Regards,

Jeremy

[Former Member]

I hope you got in depth clue of slowness

Well, just one more thought on this way, hope you have not tried to disable "file Caching" in ServletExec for the Web application( xMII).

In ServletExec "File Caching" enables/ disables the caching of the Web Applications static contents (HTML pages, GIF images, etc) in memory. Enabling caching improves performance but uses more memory. So always it should remain checked (Enabled).

[Former Member]

Even though your server has 4 GB of memory, ServletExec can only use 1.5 GB of max. Check your General Logs to see if you are getting any out of memory errors. See if there are any queries / BAPI calls that might be consuming large resources.

I used microsofts process explorer to watch Java.exe when I was having a problem like this.

In our case, the error would come from a BAPI call "BAPI_PRODORD_GET_DETAIL" if the user sent an invalid workorder, xMII would either lock up, spike memory, or slow WAY down. I ended up reengineering the BLS transaction to prevent this from happening.

[Former Member]

Even though your server has 4 GB of memory, ServletExec can only use 1.5 GB of max. Check your General Logs to see if you are getting any out of memory errors. If so, check what your memory is set in your StartServleExec.bat file.

I used microsofts process explorer to watch Java.exe ; when I was having a problem like this.

In our case, the error would come from a BAPI call "BAPI_PRODORD_GET_DETAIL" if the user sent an invalid workorder, xMII would either lock up, spike memory, or slow WAY down. I ended up reengineering the BLS transaction to prevent this from happening. See if there are any queries / BAPI calls that might be consuming large resources.

[agentry_src]

You may also want to check the size of your transactions. If you have any that are much larger than a few KB, you may have too much sample data stored in the transaction. If you find any up in the MB range, you definitely have too much sample data. Check your Local and Transaction properties for large xml property contents. It is nice to have sample data, but just enough to provide links.

Good luck,

Mike

[Former Member]

Check this pdf for JDK versions / system configuration / RAM size (as per concurrent users)

https://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/b7c79bd4-0a01-0010-4cafbfb68ab98...

Although, Java applications & Applets should load & run faster, and consume less memory when used with JRE 5.0 so it should improve the client machines performance (the latest release of JRE 1.4.2 was supported for the client machine, I am not sure about JRE 5.0)

[Former Member]

typo error in previous URL

https://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/b7c79bd4-0a01-0010-4caf-bfb68ab9...