Sorry for the long story here but this is everything I did to trouble shoot and resolve this issue over the course of a long time and I wanted to share in case anybody else is stuck as I was.
Whenever I schedule scripts to run using Windows task Scheduler, sometimes they run and sometimes they do not. It seems intermittent and random for when they would choose to fail. 0xF6 is the “Last Run Result” we would receive in windows task scheduler.
Note that this problem only started after I converted my scripts from python 2.7 to 3.x and started using the pro interpreter. 2.7 never had any problems but I need to switch to 3.x
Note also that the scripts always run no matter what when I manually run them. This is only a scheduled task problem
We are using VM server with windows datacenter 2022 but we were experiencing the same problems on an older VM running an older OS, maybe Windows Server 2016.
We tried switching from python.exe to pythonw.exe did not resolve the issue.
We tried setting up a batch file (.bat) to execute the scripts and then have task scheduler execute the batch file and this had the same result.
I tried making lots of changes in task scheduler like Run with Highest Privileges, run whether user is logged on or not, run only when user is logged in, making sure user is logged in, etc., no combination of these had any effect.
Speaking with Esri support they recommended the following:
1. Import script into .tbx as script tool then schedule to run from within Pro, this will add it to task scheduler anyway but it creates its own task with error logging.
2. Use a .tbx rather than .atbx (since .atbx is new and might have kinks to iron out.
3. Open the task in task scheduler to find the path to the new task.py and open that to see the path to where the logs are created. Should be something like this: "C:\Users\me\AppData\Local\ESRI\ArcGISPro\Geoprocessing\ScheduledTools\NameOfYourProject\task.py"
4. Then you can open the logs to see meaningful errors not just the meaningless 0xf6 in Task Scheduler. With meaningful errors you can work to resolve the issue.
So I did that and it still failed with 0xF6 and I found that in fact the logs were not being created at all which suggested that the pythonw.exe was not even being allowed to start executing since not even an empty log was created.
So we asked IT to check the anti-virus and other security systems on the server that might be preventing it from executing. They found no issues.
Finally, IT checked some local security policies. There’s a local security policy on each server that they set to prevent bad guys from mining bitcoins I guess. They adjusted these to add the “logon as a batch job” permission for my user on that server. Finally the 0xf6 errors stopped and everything runs as it should for the past week.
So hopefully I’m not speaking too soon but this looks resolved for me. I hope this helps somebody else.
