Upgraded to 0.5.1, now container won't start

Oh yeah, that /etc/subuid and /etc/subgid is a mess and likely the issue here.
Incus 0.5 fixed a long standing bug where only the first entry in those files would be considered for use by containers.

But then on systems like yours that have multiple conflicting entries, it’s going to cause some issues. I’d recommend that you wipe both files and then only put a single line in each of those:

root:1000000:1000000000

Once that’s done, restart Incus with systemctl restart incus, then confirm that new containers work properly and after that, try restarting your existing containers as they will need to go through another automatic remap stage to get back to working order.

2 Likes