Have a deployment hanging, which I cannot cancel. The task is marked as cancelled (black icon), however it seems that the state is not updated in the database, since new deployments are queued and waiting for the cancelled task to finished. Prior to v3, I would issue the tentacle.exe --reset-activities command on the particular machine, however this seems to have been deprecated in v3. Any ideas?
Sorry for the troubles. Can you try a restart of the Octopus Server service.
Ouch - kind of invasive fix…
I’ll try to schedule a time slot where this is possible.
Not a whole server restart, just the service? I’m surprised an activity reset in 2.6 on the tentacle would fix a queue on the server. And when the activity reset was run on the server you needed to stop and start the service then too.
Hi Vanessa -
Got ops to do a service restart last night and it seemed to solve the particular problem - Got another team today with similar problems and I’ll be looking into these tomorrow. Very odd - Have you had any other cases of mismatch of task state?
On a side note - didn’t say that the reset-acitivities would have solved the problem, but it would have been my next move, if it still existed.
I’ve had a couple of cases, I need to gather them up and find if there is some kind of pattern.
Are you able to grab deployment logs from the ones that hung or had mismatched states?