Application Development Discussions
Join the discussions or start your own on all things application development, including tools and APIs, programming models, and keeping your skills sharp.
cancel
Showing results for 
Search instead for 
Did you mean: 

CALL_FUNCTION_SEND_ERROR

0 Kudos
14,445

Hi Experts,

Getting CALL_FUNCTION_SEND_ERROR in production environment. Around 20,000 users are affected by this issue. This particular exception is raised when the user try to connect from third party tool to SAP. And to be specific this happens only when they try to connect to a particular custom function module.

I have gone thro' many threads and checked SM21 system logs. I have attached the same for reference.

In many threads they have mentioned "Connection to the partner '' broken (I/O) is specific to network issue. But our question is why it is occurring only when  it is trying to connect to this particular custom function module?

Actually this particular FM is a wrapper FM which is present in our portal system. The error occurred at the starting point of the FM (i.e., it has not executed the code in the FM).

If any one had faced similar issue before, kindly share your inputs. It will be helpful. Awaiting for your responses.

Thanks in Advance,

Fahmeetha M

1 ACCEPTED SOLUTION

Sandra_Rossi
Active Contributor
3,022

Okay, in fact the short dump is not interesting in that case 🙂

Anyway, CALL_FUNCTION_SEND_ERROR seems to be due to the client or the RFC session which terminates before the data of the function module can be sent back.

So, I think you should ask how the RFC is triggered from client side.

Moreover, could you make sure the attributes of this function module (in the Attributes tab in SE37) are like the other RFC function modules (as you say, these ones work).

After that, you should urgently contact the SAP support because you say it impacts 20.000 users.

23 REPLIES 23

Ashg1402
Contributor
0 Kudos
3,022

Hi,

Is the function module which all users are calling is checked as Remote-Enabled?

0 Kudos
3,022

Hi Ashish,

Yeah it is remote enabled. This is working fine in development and staging systems. Failing only in production .

0 Kudos
3,022

Hi,

According to log file it's just saying about connection broken (as you have mentioned) .

There was a similar issue back.

Runtime Error   CALL_FUNCTION_SEND_ERROR | SCN

Check it may be it can help. As there can be many issues for this error.

In meantime I'll also check for any solution.

Regards

Ashish

Sandra_Rossi
Active Contributor
0 Kudos
3,022

What is interesting is the short dump (ST22).

Sandra_Rossi
Active Contributor
0 Kudos
3,022

There are also SAP notes :

1421182 - Reasons for CALL_FUNCTION_SEND_ERROR -> http://service.sap.com/sap/support/notes/1421182

former_member185537
Participant
0 Kudos
3,022

Hi,

Have you been able to get it

0 Kudos
3,022

In ST22 it is mentioned :

CPIC-CALL: 'ThCMSEND' , communication rc : CM_PRODUCT_SPECIFIC_ERROR (cmRc=20) ,

taskhandler rc: NI_READ_FAILED (thRc=223)#CPIC program connection ended; read error

An error occurred while making a Remote Function Call.

0 Kudos
3,022

Hi,

As you say its occuring only in PRD, Can you check the SM59 setup with the destination.

0 Kudos
3,022


Could you please let me know what exactly needs to be checked in SM59.

0 Kudos
3,022

Try doing a connection test with the destination. It will let you know if there's a problem in the connection itself

Sandra_Rossi
Active Contributor
3,023

Okay, in fact the short dump is not interesting in that case 🙂

Anyway, CALL_FUNCTION_SEND_ERROR seems to be due to the client or the RFC session which terminates before the data of the function module can be sent back.

So, I think you should ask how the RFC is triggered from client side.

Moreover, could you make sure the attributes of this function module (in the Attributes tab in SE37) are like the other RFC function modules (as you say, these ones work).

After that, you should urgently contact the SAP support because you say it impacts 20.000 users.

0 Kudos
3,022

Hi Sandra,

Thanks for your reply.

Actually the dump is occurring in the portal system.  Is it like the connection is getting terminated before it could get the data from R/3?  Then we should get time_out dump on R/3 system side right..?

0 Kudos
3,022

Is it like the connection is getting terminated before it could get the data from R/3?  Then we should get time_out dump on R/3 system side right..?

I'm not sure of the exact "direction" of the error. I thought it could be the RFC server which attempts to send the answer back to the RFC client, but there's no more connection, maybe closed by the RFC client (asynchronous RFC call without waiting for the answer, or error).

ST22 dumps only happen in ABAP systems, so do you mean your portal is an ABAP system?

R/3 being also an ABAP system, that would mean there's an RFC issue between those two ABAP systems.

Could you please explain more about what is exactly done?

You should urgently contact the SAP support because you say it impacts 20.000 users.

0 Kudos
3,022

In Portal system we have wrapper RFC which is called by the external third party tool. Then this wrapper RFC will call our RFC in R/3 system. It will retrieve the data and send it back to our portal system.

If it is a connection issue we need to know what is specific about this custom RFC alone.

We have raised an incident with SAP also. SAP replied saying it is a custom RFC. Involve the developer.

0 Kudos
3,022

So:

Third party tool ---> "wrapper RFC" in portal (in fact an ABAP system?) ---> function module in R/3 system

The problem is that you don't give us a lot of information. Is it a wrapper RFC in ABAP, java? Show the code. Which RFC destination does it call in R/3 system? Is there any short dump in R/3?

Please paste the discussion between SAP and you, so that we have more information to help you.

0 Kudos
3,022

Hi Fahmeetha,

There are two batch jobs for this? how are the functions getting executed. Can you give us the results for each of the call and let us know which one fails. As indicated let us know the type of systems. And did you test the connection between SAP and src system

0 Kudos
3,022

Hi Venkat,

It is not a batch job. From the dot net application , they call the custom RFC . This custom wrapper  RFC is present in our portal system (ABAP) . From there it will connect to the R/3 client , retrieves data and send back the details.

One clue which I got by debugging in staging system is the data is getting retrieved from the R/3 system. This RFC is returning huge data. So checking in a perspective that the dot net application is unable to handle this large amount of data. In the dot net application they are getting memory exception. However we are not getting any dump in our SAP system (staging). Dump occurs in production system.

0 Kudos
3,022

The below shown screenshot shows the point at which the dump occurs:


Earlier I assumed that when the control is coming in , at that point itself it was failing. But it is not the case. The data is retrieved successfully from the R/3 system. But I am not sure why it is pointing to the start of the RFC.

0 Kudos
3,022

Can you also post the text and reasons that they have given in the system level  from the DUMP

0 Kudos
3,022

I think you should not focus why it's pointing to the start of the RFC, probably there's a general error. What you indicate in your post at 5:47am is very interesting. Maybe you can ask SAP support again with this extra information.

Moreover, you must be aware that RFC is subject to timeout because it runs in dialog workprocess. If you retrieve a big amount of data, it may take a while, so the timeout may happen anywhere in the chain of RFC (both ABAP systems). The only solution is to reduce the amount of data, or you may temporarily increase the time limit (profile parameter rdisp/max_wprun_time , ask the admin) in both ABAP systems to test whether it solves the issue.

0 Kudos
3,022

Thanks Sandra..

Is there any possibility for the presence of huge data to be a cause for the connection reset?

0 Kudos
3,022

Why not, there could be bugs, features, ..., but I don't know how to detect it. But what means "huge" for you? (some people may say "huge" for something not so "huge"; from my point of view, huge would mean hours of transmission...)

0 Kudos
3,022


Thanks all for your inputs.

I changed the amount of data getting transferred in debugging mode in staging system. It is working fine for less amount of data. Exactly we didn't get the reason for the dump but this must be related to the data volume .

Now we have proposed for a new change for restricting the data being fetched.