SOAP Request not reaching PI

[former_member235717]

Hello Experts,

We have a SOAP to RFC interface. The interface is working fine until recently we observed when the vendor is sending the web request for the first time in the day, he is getting exception ( Exception has been thrown by the target of an invocation ) but we see no logs in PI. After few minutes when he try to send again, it works fine. This happened every first sign in after couple of hours.

Since no logs are in PI, I assume the request itself is not reaching PI, but later it is working. Could someone throw some light on why this is happening.

Thanks

[former_member204100 · Accepted solution]

Hi Swetha,

I think it is important to see in the first place what is the endpoint the sender system is calling.

It is directly pointing to the PI server, or is there any web dispatcher between the sender system and the PI? Maybe you should not search for the logs in PI but in the web dispatcher first.

If there are no such dispatcher in the picture, than its worth to analyze how the followings look like exactly:

- the endpoint they are calling

- the whole SOAP request - including with the HTTP header

- the error message what the sender system will receive during the first call

Also you might setup ICM trace to check the traffic, or there are also other ways to monitor the communication - like TCPGateway tool. Note: #856597 - FAQ: XI 3.0 / PI 7.0/7.1/7.3 SOAP Adapter

--> "How can I trace the whole message"

If it turns out that the call reaches the SOAP adapter than in the default trace file - on the server node which processed the request - must be some error in the PI system.

Best Regards,

Viktor

[Ryan-Crosby]

Hi Swetha,

What is the functional nature of the RFC that is executed when the vendor makes the SOAP request?  If I had to guess just based on what you are describing I would suspect some type of database buffering issue where things are getting swapped out and every so often have to be fully loaded again.

Regards,

Ryan Crosby

[former_member235717]

the RFC was a simple FM which takes in the input as the user name and sends out the response with few fields.

I would like to mention that this is happening only in Development.

"would suspect some type of database buffering issue where things are getting swapped out and every so often have to be fully loaded again." - do you mean the issue is lying with the web service side?

[Ryan-Crosby]

Hi Swetha,

Ok, so it's not any sort of time out related to buffering but did you check the ST22 logs in the back end system to see what might be happening?

Regards,

Ryan Crosby