All my containers have "boot.autostart" set to false.
Irrespective of this if they were running when I shut-down my workstation, they are running again after my workstation has started.
I assume this is by design. I can’t find any way of managing this. Any search I do is dominated by autostart.
Do I have to explicitly stop any containers I don’t want restarting?
After this had been happening for months. I confirmed that containers were restarting several times yesterday by rebooting my workstation.
I stopped all the containersfrom the command-line yesterday and rebooted. I can report there are no unsolicited boot.autostarts happening now.
I am both delighted and disappointed
Before reboot today
% lxc ls -c ns4tS pydev1
+--------+---------+-----------------------+-----------+-----------+
| NAME | STATE | IPV4 | TYPE | SNAPSHOTS |
+--------+---------+-----------------------+-----------+-----------+
| pydev1 | RUNNING | 10.232.221.142 (eth0) | CONTAINER | 3 |
+--------+---------+-----------------------+-----------+-----------+
a) Can I assume that if boot.autostart is not explicitly a container’s configuration it defaults to “true”?
I can see from my workstation’s shell history that yesterday after posting I ran:
lxc profile set default boot.autostart=false
b) All my containers have the default profile as a base. Does that explain why they all stooped rebooting by default? I’m used to other systems where state doesn’t mutate.
Can you check your logs to see if the containers are failing to start because a boot-time dependency has not been fulfilled (perhaps a parent network interface or something).