Can I restart a task that failed from Probe?

[mgriffith]

Hi all,

I have a scenario where a I have a task in my workflow that fails and and I want to restart it using probe.  Is that possible? I don't see any way to restart it, all I see is the process id.

The task is a user task that also uses a spring delegateExpressionType on create to send a message to an e-Commerce system using a SOAP/RPC Web service call.  If the e-Commerce system is off line, the process stops at the task, which I could in theory recover from when the e-Commerce system is available again?

Any thoughts would be appreciated.  Thanks in advance!

[trademak]

Hi,

You can create a sequence flow back to the failing task when the request was not successful.
Then the request will be executed again.

Best regards,

[mgriffith]

Hi,

You can create a sequence flow back to the failing task when the request was not successful.
Then the request will be executed again.

Best regards,

Thanks for the reply. I'm not sure I fully understand what you mean.  However, I get the impression this can't be done without additional programming. Is there an example in the samples that would demonstrate this concept?

If not, can you provide a high level overview of the steps?

Thanks in advance,

MG

[trademak]

Hi,

You don't need addtional programming.
What you can do is add a form to the user task "Notify task" with a boolean field like requestCompleted. If the request fails you complete the form with false and in an exclusive gateway you test the value of the requestCompleted variable. If it's false you draw a sequence flow back to the "Notify task", if it's true you continue with the normal process flow.

Best regards,

[sudarshan]

Hi
Have similar query here and its found to be happening frequently

Scenario

1. A Process is getting triggered via WebService and that WebService for somehow triggering multiple instances for same caseid. Now there is a logic which is written to check if its duplicate case, while doing this Activiti engine  jar is throwing an exception "already taking a transition" .

This is happening (as I suspect) when a custom code is getting executed parallel to a Web Service triggering a new process for same case id
And due to this the case is not visible in ACT_RU…_TASK table

Request , if there is any solution to :-

1.  handle this scenarios
2.  To re-rerun the failed cases ( because of "already taking a transition" error, the case is not visible to re run in my exception activity)

[frederikherema1]

Can you elaborate a bit more on this "custom code" and where exactly you are seeing this exception?