Process randomly getting stuck on pending status and never running

Help Orchestrator
,
1

Hello there, I have this issue where a process created by a trigger gets “randomly” stuck on pending status, the problem is that the trigger works perfectly fine and has only stuck on a never-starting pending job two times so far: first time 2 weeks ago on Saturday and the second last Saturday arround the same hour as the first. The previous job was executed succesfully and in a few seconds as usual for both cases.

The trigger activates every 2 minutes and when this happens it is unable to create new jobs which eventually disables the trigger as it’s to be expected. Then the job has to be manually killed before the trigger can be enabled again. Also the job has no logs and the only information it contains is that the start and end time is the same it was killed. No changes have been done to neither the settings on the orchestrator or the robot.

My question is then: what is the cause for the jobs getting stuck on pending and how can be fixed?
I’m aware it is possible to set a “timeout” for jobs and stop or kill them or both, but let’s keep that as a last resort for it would interfere with the robot execution under certain circumstances.

Thank you in advance.

2

Hey @procesorpa

When this issue comes, did you kill and restart the job directly or you go to the robot machine login and do some stuff & then re-trigger ?

Thanks
#nK

3

I do everything from the orchestrator page, I just go to the jobs tab and kill the process then enable the trigger, nothing else.

4

Post which it is running fine right ?

Okay little clueless. Did you install the bot in service mode in the robot machine ?

5

Hello @procesorpa

Have you checked the logs of the failures? Any other installations or patches happened during that time?

Please have a check of the memory utilization as well.

Thanks

6

Robot is installed in service mode.

7

The job doesn’t have any logs, nor when is pending state nor after being stopped/killed. Also the were no updates/installations during that time to my knowledge. However the first time this happened was when the queue policy update dropped I belive, but I don’t think that has to do anything with it since we don’t use queues, or so I hope. Although looking throught windows events to see if there were any anomalies I found there are 2 recurring errors regarding the UiPath.Service.Host. However these errors seem to have occurred since April and there have been no problems until now…
Anyway, I’m attaching them in case they give some insight.
image
image

8

Okay, is this intermittent or regular issue ?

9

Until now we’ve never had this or a similar issue, and this one in particular has only happened twice: on the same day of the week arround the same hour in a 2 week interval… I see a pattern there however I can’t say for sure if it will happen again, I’m asuming it will unless proven otherwise. As for if it is intermittent or regular, I’d say it’s intermittent or rather sporadic until further events occur.

I’m sorry I can’t provide more useful information but I’m unable to replicate the issue, also it happening under normal, work-usual circumstances doesn’t help the slightest.

10

Okay @procesorpa

If possible trace the Event log at the same time to see if there is anything logged which may give us some clue.

Thanks
#nK

11

No anomalies on the Event log either, last robot’s execution before the faulty job was completed correctly and the next recorded log is from Monday, when the trigger was renabled. I also checked the logs of the previous and following days but I didn’t find anything unusual.

12

Okay bit tougher to find actually but will see if that occurs again.

13

our team has encountered a similar. in our case, we had one server with two licenses, two jobs were running and a third one queued up. The third job kept showing up as stopped after some time. We found that the third job’s trigger had Kill after 10 minutes, which has making the job show as Stopped with no logs or execution.

In other cases, we have seen the error: “Could not start executor. A specified logon session does not exist. It may already have been terminated.” We are still looking on our end as to what causes these instead of leaving the job in Pending until a license frees up.