How to Tune / Configure SAP XI to Process a bigger size file ( > 50 MB) ?

[Former Member]

Hi Gurus,

I am facing problem while processing a file in XI whose size is greater than 20 MB.When ever a file size more than 20 MB is given for processing in XI it is giving error.

So to deal this situation we split the file into smaller portions and process in XI after that we again join the files and some time this is triggering some other issues.

So,I want to know if any Tuning ,Configuration or Extra Hardware can increase the processing limit of XI from 20 MB.

Thanks in Advance.

Satya

[Former Member · Accepted solution]

Hi Satya,

I am not sure, but I guess the java-parameter HEAP-SIZE could solve your problem.

Regards Mario

[Former Member · Accepted solution]

Will HEAP_SIZE solve the problem or do i have to consider any other parameters in XI box?

Could you please guide me provide me any links or document for this?

Thanks,

Satya Jethy

[Former Member · Accepted solution]

Hello.

In my current project I have a similar scenario.

We handle large XML files.

My set design, thanks to recommendations from SAP, which files to be sent to be equal to or less than 5 megabytes.

If the file at its source is greater than 5 megabytes is recommended particionalos No files in less than or equal to 5 megabytes each.

This ensures a good performance in the system issuer, XI / IP and receiver.

Sincerely,

Claudio Enrique León Hernández

[Former Member]

Satya,

File size depends on your Java Heap (Java heap depends on your Hardware Config).

What we can do as a developer is try to tune and use optimum resources. If your Message gets errored out in Integration engine then use Large Queues (XBTM-Message size is specific to requirements) and also carefully design the number of parallel Inbound and Outbound Integration engine Queues(Refer to Tuning Guide).

Also try to optimize the mapping program to maximum. Eg if you are using Java Map, go for SAX Processing instead of DOM. Try to avoid Queues in case of Graphical Mapping, etc.

Regards,

Sudharshan N A

[Former Member]

This may seem like repetative requests for information on this but for one - is it possible to break the file into smaller pieces or is it necessary to process the file as a whole? Additionally, what is your current HEAP setting set to and your version of XI/PI?

Thanks for providing more details so we can help to resolve your issue.

[Former Member]

Hi ,

Currently we are processing the file in after split and after processing i am again joining it.

Currently we are using XI 3.0.

Thanks,

Satya Jethy

[Former Member]

Jethy,

Is there a specific error that you are getting while processing the files? With what you are doing, you should be able to process this with standard settings. There are a number of ways to increase performance in PI/XI but each is suited to a specific set of issues of which there really isn't one specifically for file sizes. Of course if your J2EE Engine is crashing with an OutOfMemory Error, that is Heap - otherwise it is going to be something else.

Thanks and if you can be as specific as possible, everyone can assist to get you by your current issues.

[Former Member]

Hi,

What type mapping type r u using. Try to use ABAP XSLT mapping.

this can elimanate the switching time to AE while performing Pipeline Steps.

[Former Member]

Hi Pradeep,

Thanks for your answer.We are using graphical mapping.

Could you please describe bit little of your answer.

Thanks,

Satya

[bhavesh_kantilal]

Hi pradeep,

The problem with this huge amount of data comes not in the mapping steps but even before the message can enter the pipeline step from the AE to IE.

I have tried something of this and everytime the thresholds where reached the message gets stuck in the AE itself and hence even though you have a valid point that will help therotically if the Message enters the IE, i wonder if in real time it would help as the message might not even enter the pipeline steps.

Would like to hear your thoughts as well and if having tried this option has helped you in realtime.

Thanks,

Bhavesh

[Former Member]

Hi Bhavesh,

Increase Current Value of the Parameter Nmame : rdisp/max_wprun_time

using Tcode RZ11 and try to process.

we had processed files Sizes up to 50 Mb.

[Former Member]

Hi Sathaya,

In one of my scenarios we have been processing files of Size 300 Mb but here we have been splitting the files using script of size 10 mb and then merge the files after receiving all the files in the receiver.

we have tried using JAVA XSLT mapping it was throwing time out errors and then we used ABAP XSLT it worked fine. ABAP XSLT mapping can process files up to 50 MB( for a single receiver).The diffrence is that when we use java xslt mapping the heap size is 40 times the size of the message but whe we use ABAP XSLT the Heap Size was only 4-5 times.

[Former Member]

Can you ensure that the PROCTIMEOUT parameter is set to at least

900 on the system? Refer to section 9.3 of the attached XI

Troubleshooting Guide for details on this. Then try the scenario

again.

Note 807000 usually resolves issues like this.

[Former Member]

HI Pradeep ,

Is the ABAP XSLT mapping is the only possible solution b'coz already we have used graphyical mapping.

For your information as per business requirement we have also done few sorting on records.

Thanks,

Satya Jethy