NW Business Client - timeout in Sales - premature EOS error in SMICM logs

[Former Member]

Hi

We are running Netweaver Business Client and when we click on the Sales menu in the left hand pane the system hangs with the timeout icon.

In the dev_icm trace we see the following message:

• ERROR => PlugInHandleNetData: server: premature EOS (0/-1) in request [http_plg_mt. 2004]

Note that we found:

SAP Note 1585064 - ERROR: Premature EOS in request

but the symptoms are NOT exactly the same - we do not get the second line of the error (IcmPlCheckRetVal) nor do we get the blank times above this error. In fact, even with the trace level set to full, there is no other detail/error messages -just the premature EOS error.

Any ideas?

Already tried changing XSRFCHECK back from 1 to 0 and deactivating/activating the webgui service as per several forums on here; it made no difference.

Regards

Ross

[former_member128291 · Accepted solution]

Dear Ross,

Kindly note the PlugInHandleNetData is a function which analyses the input stream from the network. This error you get means that some error occured during receiving the data stream from the related server. This can be caused by a temporary network error or some synchronization fault.

This message appears in the icm trace when you are executing stateful applications and the user session (in SM04) is already deleted. This session is deleted after the timeout in rdisp/plugin_auto_logout.

After this time the user session in SM04 will be deleted. This parameter is only used for stateful sessions.

Because the HTTP protocol is a stateless protocol we cannot determine when the browser is closed or the user navigates to another page. So, we had to introduce this timeout. If the user does not connect to the server in a defined timeframe again then the session is deleted and if now a request comes in which want to use this session you get this error.

Additionally please check note

1. 824554 - ICM and dispatcher Web timeout parameter

The parameter gw/so_keepalive should be set to 1 on all servers to prevent any orphaned rfc connections from causing problems. Note 743888:

1. 743888 - GW: gw/so_keepalive profile parameter

Regards,

Abhishek

[Former Member]

Thanks but the timeout values are already optimally set, besides, it times out IMMEDIATELY.

gw/so_keepalive is already set to 1.

[Former Member]

It seems some other timeout values weren't set correctly - all ok now.

[sebastian_sebaraj]

Ross

We have same problem. Can you list what other timeout you changed.

[Former Member]

Ross

we are facing the same error and we cannot find out what the root course is

we are on CRM 7.0 EHP 3 Kernel 32, and we are about to update to Kernel 42.

Perhaps this helps

Thank you

Frank

[Former Member]

Hi Sebastian & Frank

I'm sorry but this was two years ago and at another company, I can't even find any notes on it...

From what I recall though, it wasn't something I changed anyway, someone else made an adjustment whilst I was away a few days and I don't think we ever established clearly what is was! All I can suggest is that you review all related NWBC time parameters, as well as standard time parameters, and ensure that they are set appropriately.

Good luck!

Regards

Ross

[Former Member]

Hello,

I suggest please change Keepalive  timeout parameter of ICM .HTTP according to your requirement.

for example increase the TIMEOUT parameter value in the below RZ10 Profile parameter. (It is just for example)

icm/server_port_x  = PROT=HTTP,PORT=80xx,TIMEOUT=60,PROCTIMEOUT=600

Cheers,

Karthick