Recycling issue

Product: PowerShell Universal
Version: 2.5.3


I’m using PowerShell Universal with IIS and for some time now I’m experiencing issues when application pool of PowerShell Universal is recycled. Afterwards no scheduled jobs are running. I configured the settings as described in IIS - PowerShell Universal:

Help or an information if someone is experiencing the same issue would be much appreciated.

Best regards,

Hi @zweailltienrger,

In the event log, you should see a reason for the recycling. I would also suggest setting Regular Time Interval to 0 to avoid it recycling every 29 hours.

Hello adam,

thank you, I know why the recycling happens, and from my point of view it’s necessary since I’m backing up the database as well. Why is the application failing to start or why won’t the jobs run afterwards since start mode is set to always running?

Best regards,

Our typical guidance has been to avoid recycling because of this issue. I’m not sure why the app pool wouldn’t restart (or schedule jobs as recycle) when the app is set to AlwaysRunning. Jobs are scheduled when the PSU service starts up so it seems like IIS isn’t restarting the process completely or not starting it at all.

If you navigate to the website after a recycle, do the schedules set correctly?

I understand. Is there some other way, expect recycling, for the database to be in a consistent state? When I recycle the application and start the website, jobs are working fine until the next recycle.

So, what I just found is if you implement something like a healthcheck you could go with recycling, since access to the healthcheck endpoint seems to start the instance fully?! Also when using the endpoint with a ServerStarted trigger seems to work around this issue.

That’s very bizarre. We are seeing strange issue with Azure Web Apps as well (uses IIS) where if you restart the web app, it stops working but redploying the web app works. It almost seems like Azure is doing a recycle and causing problems similar to IIS on premises.

So I tried with a trigger on ServerStarted event, but no luck so far. I created a scheduled task, which will call the unauthenticated healthcheck endpoint and will let you know if this is an applicable workaround.