Ah, interesting, can you edit /var/log/lxd/databases/lxc.conf and modify that first line to read lxc.log.file instead of lxc.logfile?
Assuming you haven’t rebooted yet.
root@abc ~ # cat /var/log/lxd/databases/lxc.log
root@abc ~ #I don’t see what I should change… You asked me to change something to itself?
Run sed -i s/lxc.logfile/lxc.log.file/g /var/log/lxd/*/lxc.conf and see if your containers behave better after that.
I tried replacing the first line on just databases conf, and databases container now seems to behave much better… I can exec bash to it
Now I’ll do the sed trick
Then I’ll reboot, ok?
Well, if the sed trick works, you won’t have to reboot at all, but sure, you can still do it for peace of mind 
sed edited correctly all the files. Now I rebooted. Will let you know how that went.
Worked fine, thanks…!
Sent a patch that will do a one time re-gen of all those configs which should catch any case where the configs still contain invalid keys.
1 Like