Rohit, you might need to send ED logs of the execution.

cd /var/log/Avaya/services/EngagementDesigner
ce dlogon EngagementDesigner
run your scenario
copy EngagementDesigner.log and send it to us.

Hi Shikha,

Here you go.

Regards,
Rohit

The workflow looks ok, dont see the issue with drop call you mentioned in the logs you attached, can you attach logs from time when you are able to see the issue

Also, you are on ED 3.3 which is bit old, would it be possible to upgrade to later ED 3.4

Hi Anurag,

I would love to but it is customer who wants to know the significance difference between 3.3 and 3.4.

Also, The issue I have highlighted is the one stuck on call drop and and exclusive gateway.


Would you be able to look in to this as per the instance Id?

Regards,
Rohit

3.3 is two years old, it had no known issue which i can relate to here but getting new code is generally preferable, but staying at 3.3 should be ok

the logs you provided dont have the instance id you showed in word document, if you are able to reproduce the issue - can you please provide the logs from the time you are able to reproduce

Anurag, I was able to replicate this issue. I have emailed you logs and WFD.
What I observed is that if gateway is being executed when it received call_ended event, it will error out with NPE, and if its on any other task, say PNC, it gracefully exits via interrupt event. Instance also gets completed.

Hi Guys,

Any update on the issue?

Regards,
Rohit

Rohit, we understand the reason a bit, PA is completed when wither announcement is done or call is ended, we listen for both things

in parallel, you have interrupt event for call ending

so when call ends, we start at both places and PA happens first, next there is drop call - but call is already dropped - so this task throws error

one way is that you check output of PA - if you are coming out of PA because of call ending then you don't do drop call (the output has fields which allow to differentiate between announcement done and call ending)

other way would be to do some code change so that somehow interrupt handling is done first and we terminate the process after that and drop call is never run. This is bit more difficult and we would try to do it but not in 3.3 or 3.4, this would be in 3.5 or later

Hi Anurag,

Are you saying I should remove the drop call that is one part of the issue. What about issue on the gateway that goes in hung state?

Also what do you mean by PA here?

Sorry for knowledge

Regards,
Rohit

in gateway you have some condition like if interaction id > 4 and <12, but if call was dropped interaction id is null so can you please modify expression to interaction id != null && (interaction id between 4 and 12)

Hi Anurag,

I have made changes but it is getting stuck on other parts of play announcement. I have attached instance Id and logs.

Please let me know what else can be done.

Regards,
Rohit

forgot to attach the instance.

I am increasing the priority that interrupt event shold take precedence over anything else, but this would still not be available before 3.5,. In mean time you would have to do error handling in required places (or ignore errors if that does not harm your use cases)

So I have to include error handling on every block to release the instances. What about when instance is showing complete but not clearing?

what does clearing mean ?