Pending job fails on robot disconnect

Help Orchestrator
,
1

Hi,
environment:
orchestrator and only one robot

imagine following scenario:

  • JOB1 starts another job using Start Job sctivity.
  • It results in one running (JOB1) and one pending job (JOB2).
  • Before JOB1 finishes the robot machine disconnects. It results in JOB1 failure with error
    “Job stopped with an unexpected exit code: 0x40010004”
  • Orchestrator starts pending JOB2 despite the robot is disconnected
  • JOB2 immediately fails with error
    “Could not start executor. Rdp connection failed: Message: The disconnection was initiated by the user logging off their session on the server., Last error: 65548”

Desired behaviour:
Orchestrator will detect robot disconnect.
JOB2 will stay in pending state until robot reconnects.

Any idea how to achieve this?

Cheers

2

Hello @J0ska

Have you checked the reason why the bots are getting disconnected?

I hope inside the workflow you are calling the second job?

Thanks

3

Hi @J0ska

Firstly its weird to see that job 1 disconnects before the execution is complete which is unusual

But you can try the below thread this might help those disconnects

Cheers

4

Hi @J0ska, did you solve it ?

5

HI @J0ska ,

At what step in Job1, the job is getting failed? Is it failing at Start Job activity step?

6

Hi @gediminas,
No, I did not. I found this behaviour while “playing” with Orchestrator and jobs scheduling. But as it was not important for real operation I did not follow-up this case.

I still believe that Orchestrator should behave as described in “desired behaviour” and any pending job should wait untill robot is available.
The truth is that I did not test it recently and maybe it is fixed.

Cheers

7

Ok, fine. It is in pending state in current Orchestrator as your described desired behavior. I am more interested in cases where jobs stop with an unexpected exit code: 0x40010004. But it’s more related question to sys admins I think.

8

Right. In my case it was caused by robot machine restart - I do restart it in case of an unrecoverable error to “clean the table”.

Cheers