Hi, this issue might be similar to this one: volume attached to destroyed machine
But I had the following situation:
- machine immediately shuts down due to bad config in volume
- can’t SSH into the machine (due to it rebooting immediately)
- I know what is wrong in the config
I was hoping to be able to do something like specifying
--entrypoint for a docker image. I tried things like the following:
fly m run -a forgejo-instance --volume forgejo:/data .
but all I got in return was
Error: No unattached volumes in region ‘ams’
this made me think that the app had a machine running, so i did do
fly m stop as well, but I suspect it instantly tried to restart a machine attaching the volume as well.
This is all sadly in the past tense since I ended up with the idea that destroying an
app would for some reason not destroy the app’s
volume. It’s gone now, granted nothing important was on it, but I wanted to learn how to do some maintenance for the future. Is there any newbie guide on how to do some maintenance on volumes? I wanted to be able to do any of these:
- specify an entrypoint for the docker image/container. I tried the
--shellflag for the
fly machine run, but maybe there’s something similar that could start a machine but set the docker --entrypoint to a bash shell? (preferred solution), this would prevent the machine from running whatever the normal entrypoint was that would cause it to crash instantly.
- attach a volume to a different machine that i can use to fix some config inside it. It kept telling me it was attached already… so I didn’t succeed with this.
I kinda hope to be able to recreate this issue in the future, and if I do, i’ll post here if I find a working solution.