SonicESB Process Fault Handler Not Being Invoked

Posted by cwiseman on 20-Apr-2011 10:08

Encountering an issue where a process which contains a custom service throws an uncaught runtime exception.  There is:

  • No RME message sent to the RME endpoint (dev.RME)
  • The exception stacktrace gets printed out into the dev_ESBTest container logfile
  • The fault handler process configured to the main process containing the custom service is not invoked

The main process is a mapper operation of a top down SonicConnect exposed web service.  The fault handler is necessary to respond to the SonicConnect client when this fault occurs (to form up a SOAP Fault basis).

I posted a bit more information with the container log stacktraces in this forum post:   http://communities.progress.com/pcom/message/125370#125370

I have added the dev_ESBTest container logfile exceptions and the application_logfile to this post.

The web service operation gets the incoming message routed to two separate processes, one is a process local and the other is a process response.  They are independent processes.  Local does some necessary database process and response is responsible for responding to the client requestor.  Both processes are fronted by a security check.  The security check disallowed the processing by throwing a SecurityException - seen in the application logfile and then in the container log file.  Rethrowing this exception is not leading to the fault handlers configured to the local and response processes being invoked.

I would appreciate any insights.  Thanks in advance.

application_logfile.txt.zip

container_logfile.txt.zip

All Replies

Posted by jerem on 04-May-2011 03:37

Hello Chris,

Your error sounds like a bug found with the Sonic Connect error management

see bug SNC00080194

A possible workaround (if the error is occuring in a process step is to disable intracontainer messaging)

Regards

Chan Jeremy

Posted by mnair on 04-May-2011 22:57

Hi Chris,

We recently identified an issue related to fault handling with processes exposed via Sonic Connect. You could also escalate your issue thru support to determine if this is the same issue. The issue we identified will be fixed in 8.1.

If intracontainer messaging was enabled then can you retry your scneario after disablilng it and let me know if it fixes the issue?

Regards

-Mahesh

Posted by cwiseman on 10-Nov-2011 13:06

Can anyone confirm if this issue was fixed in v8.1 or v8.5?  Thanks.

Posted by dbroberg on 10-Nov-2011 13:21

SNC00080194 is related to (or more or less a duplicate of) SNC00080534.  That's resolved in 8.5.

In the 8.5 release notes:

SNC00080534     
RESOLVED in 8.5.0 SonicESB: Fault handling not triggered for async dispatch with ICM enabled
---
If there an exception is thrown in a process that is invoked using the async
dispatch API and if ICM is enabled then the exception is ignored and fault
handling is not triggered.

This thread is closed