Last week I upgrade from 2019.3 to 2020.2 and we ran into a pretty major issue last night.
I scheduled about 8 tenants to deploy to live at 3am this morning, when I woke up nothing had happened, Octopus was stopped.
When I opened the server manager it told me that the Octopus installation wasn’t even installed yet, so I re-installed it and it started Octo like normal again.
The deployments were still scheduled, but it didn’t try to recover them, they all show this message:
The logs don’t show anything about the server shutting down or anything, what could be the issue here?
I do see quite a bit of events on the 15th at 11:01 PM, this is when we run the Windows Updates.
Oddly enough I do see logs in my log folder for the 16th, but there is almost nothing in that day’s logs. Compared the the 15th (which was Saturday so should have the same amount of no activity as the 16th Sunday).
The last event for that day is 00:55:11, just a general Info about a tentacle connecting. Which makes sense, because part of the Windows Update scripts, the server reboots itself at 1pm, so I do see shutdowns in the Windows logs.
So I do see this event at 1:00:57: The OctopusDeploy service entered the stopped state.
But then I don’t see any events anymore for when Octopus started again, probably because it was considered not installed.
This might have been a fluke because it’s the first maintenance window after the upgrade to 2020.3.2, so I’ll keep an eye on it.
It sounds like something happened to the Octopus service during the Windows Update process. I’m glad to hear that reinstalling seems to have rectified this issue.
