I’m running lxc 0.8.0~rc1-8+deb7u1 on Debian wheezy with a single container.
The issue is after master server reboots the container instance either fails to start up or starts without networking.
It’s just a timing thing, and restarting the lxc is normally all that is required.
I have just inherited this server and that’s all the details I have.
I can’t even test the behaviour as the master server is very important and runs many services.
So any reboots, downtime and experiments are restricted to absolute minimum.
Since it’s an old version of lxc it doesn’t seem to support certain tools and commands, e.g:
$ sudo lxc config set instance boot.autostart.delay 30
E: lxc-config - no such program
Even without the tool I should be able to set this parameter (30 seconds delay) by appending:
“lxc.start.delay = 30” to /vol/lxc/instance/config
But I doubt that is the answer as what it actually does is “the system should wait x seconds before starting the next container”.
So I would expect the first (and only) one to still boot normally.
It’s also based on assumption that lxc service itself is fully functional and doesn’t need a restart.
I’ve checked /etc/rc2.d (my default run level is 2) and lxc here is at the bottom of the list:
Are there any big risks associated with moving lxc to the very end of the startup list?
Is it a good idea at all?
I’m even more tempted to change:
# Required-Start: $syslog $remote_fs # Required-Stop: $syslog $remote_fs
# Required-Start: $network $syslog $remote_fs # Required-Stop: $network $syslog $remote_fs
Or maybe there an option in lxc config better suited for what I’m trying to achieve?