Huge file cause lot's message in "To be delivered" status

[Former Member]

hi, I have a IB senario use input file 500M output file size 300MB.

It is a file-file transformation.

If we run it one time it can output file, but when I repeat to call this senario per 60 seconds, there will be no output.

The RWB message mornitor reports all message in "To be deliverred " status.

Sometime there may create one output file if I wait for a very long time --- a couple hours?

Are there any parameters can be set to resolve this issue?

I already restart the server many times and no output.

BTW, I increated the icm/HTTP/max_request_size_KB to make this senaro worked.

[abhay_rajhans2 · Accepted solution]

Hi shiwudao

Notice my issue is not a real case, I just want to test how to let SAP work on a huge input .

Consider as mentioned above you can set following parameters to process bulky files in system.

In RZ11 transaction Please set the Profile Parameters gw/reg_keepalive and

rdisp/max_wprun_time both to 0.

Please use the J2EE Visual Administrator to check that the following

parameters are set as recommended:

Go to -> J2EE Visual Administrator -> Cluster -> Server -> Kernel

ApplicationThreadManager

MaxThreadCount

should be set to >= 350

If you have more than one server node, please check this parameter for

all the server nodes.

Set the Instance Profile parameter, icm/server_port_0 as below:

PROT=HTTP,PORT=<icmhttpport>,TIMEOUT=30,PROCTIMEOUT=900

Please increase the parameters CPIC_MAX_CONV, gw/max_conn &

gw/max_sys

to 1000

Edited by: Rajhans Abhay on Dec 22, 2009 1:13 PM

[Former Member]

There was a java.lang.OutOfMemoryError on developer trace of J2EE, so I increased 2 VM parameters

maxHeapSize = 10240

maxPermSize = 1024

And now the continue BIG FILE tests works.

[former_member200962]

hi, I have a IB senario use input file 500M output file size 300MB.
Are there any parameters can be set to resolve this issue?

Check this tuning guide: http://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/2016a0b1-1780-2b10-97bd-be3ac6221...

See section 3.2.9.2 .....alternately you can refer this blog which deals with the same method:

/people/sravya.talanki2/blog/2005/11/29/night-mare-processing-huge-files-in-sap-xi

Regards,

Abhishek.

[Former Member]

thanks, I did read the guide, it just suggested to use 1Mb-5Mb message size.

It did not help to resolve my issue.

Notice my issue is not a real case, I just want to test how to let SAP work on a huge input .

[Former Member]

thanks, I did read the guide, it just suggested to use 1Mb-5Mb message size.

It did not help to resolve my issue.

Notice my issue is not a real case, I just want to test how to let SAP work on a huge input .

[former_member187339]

Hi,

Try to configure a priority queue (in SXMB_ADMN) for this interface and see whether you get any performance benefit..

Regards

Suraj

[former_member200962]

Did you see the accompanied blog? Instead of thinking of the file in MB split the file in terms of records....suppose you have 100 records then split the file at every 20th record.

[Former Member]

thanks.

I want to do a performance test for big message hence can not split my input file to some pieces.

[Former Member]

Hi,

What he means is to split the operational messages at the XI level for handling the huge volume data with recordset per message option... and not to split the file at source system itself

HTH

Rajesh

[former_member200962]

As mentioned by Rajesh you dont do any changes at source side....the settings are effective for XI/ PI

Are you trying load testing of your server? If yes then do not follow the method.....in such a case heap size is what you should be looking for to get optimum performance

[Former Member]

are you refering a flat text file?

My input file is an XML structure, how do I specify the content conversion?