Encountered an error with the robot communication.

[Dawid_Szalanski]

I'm getting an error while executing a process in PAD. 

Connection is directly to machine (VM) and process is in unattended mode. 

Has anyone had a similar problem and knows the solution?

[wnzn]

No, it's still reported to be "under investigation".

[FredTheRed]

Microsoft Access Database Engine 2010 was my problem. It had to be repaired. 

[JiriMraz]

Hello FredTheRed,

please how did you repair it? What had you to fix exactly?

Thank you 

[FredTheRed]

After I updated Power Automate Desktop on our production servers, the error message came. After long time, we found out that the problem was with the "Microsoft Access Database Engine 2010", so we tried to uninstall and install it again, and it worked.

[momlo]

I have a support case opened for this particular issue.

When the flow fails, you can see that robot.exe and other modules crashed in the Win event log.

After several tests, I learned from MS that they have discovered what is happening, but this is hard to fix.

They said this issue is caused by the way PAD communicates with parts of the OS/apps, and to fix that, they have worked with the Windows product team. They are releasing fix(es) for both Windows and PAD updated for some time, but it will take more months to fix permanently.

What I have noticed is the same with @FredTheRed - after reinstalling odbc, the issue seems to be gone for most of the flows. However, I have one flow where we added new functionality (not that complex). Unfortunately, it failed every time we enabled this new feature 😞 What is frustrating is that this is a very small sub-flow where we do not use any new or extra actions, just yet another page automation. Still, for some reason, this sub-flow is an instant trigger for that issue. Even after adding logging to each step, I was not able to pin the point to any particular action, and it just happens in random places of the workflow.

For those with the O365 office suite, even if you do not have a standalone Access Database Engine, you can try to go to ODBC and manually delete the drivers, then perform the online repair - that should help, hopefully. However, note that you need to delete the drivers first. Repairing without deleting did not do the trick.

I am so waiting for MS to fix it.

[Slavnyy_Avtomat]

What version of PAD are you running?  We haven't run into this issue for a while now.  Maybe back in 2.23?  2.24?  You may want to consider updating.

[momlo]

Latest 2.30, I keep updating our dev box as I work actively with MS on this and another issue.

[GSJACINTO]

This problem started to recur for me at the end of December. At that time I had only one production server to work with 20 scaled bots and today with a 2nd production server I could split 10 bots to each server, after that the errors disappeared.
Something I realized is that when the server usage is at 100% of CPU usage the executions take longer to finish and the RobinAgentFaulted errors are more common.

[momlo]

Hi @GSJACINTO 

I did not see a correlation between the issue and high CPU or memory usage on our side.

With the latest release of PAD, ver. 2.31 we no longer see the error affecting our flows.

As mentioned in one of the earlier posts, we had one flow where the issue was reproducible by enabling one of the sub-flows. We kept testing each release for improvement.

With 2.31, we have been running the flow without any issues for several days now; fingers crossed.

Out of curiosity - Did you update your PAD at the same time you split the workload?

[GSJACINTO]

Hi @momlo ,

I am using version 2.27.

Thanks for sharing that in the new version of PAD you are not getting any more errors, this will certainly help me in my evaluation to update the tool.