[SOLVED] How the reboots are handled when run inside on the containers

[kristian.kirilov]

Hello guys,
I'm working on an playbook (Ansible) intended to perform so called "unattended upgrades" of the whole environment. Because the playbook includes actions like setting up scheduled downtime in Nagios, rebooting the server and wait till it came up again online I created a logic, all of these actions to be performed only on KVM virtual machines (based on Ansible facts).
So far so good, but today I got a thought that what will happen if I try to reboot the same way LXC container? I logged into one of it and simply issued reboot - and it worked.
On the hypervisor I see the uptime change of the process, but no event was generated, most probably because it not comes from the Proxmox hypervisor itself.

So regarding the tests everything is fine, but I'm not sure is it wise to extend the playbook with these actions for LXC containers as well.
What do you think?
Thanks in advance.

 

[oguz]

hi,

So far so good, but today I got a thought that what will happen if I try to reboot the same way LXC container? I logged into one of it and simply issued reboot - and it worked.
On the hypervisor I see the uptime change of the process, but no event was generated, most probably because it not comes from the Proxmox hypervisor itself.

when you do a reboot from inside the VM, also no event is generated.
if you want to see the reboot event in the task log you'll need to call the API to reboot that VM/CT instead.

So regarding the tests everything is fine, but I'm not sure is it wise to extend the playbook with these actions for LXC containers as well.
What do you think?

if everything works what seems to be the issue?

 

[kristian.kirilov]

Ah, sorry, no issue at all. Just wondering if this is the right way to do the things.
I'm 100% sure when using VM, because this is completely separated/emulated process from the host OS, but what about CT's - it seems to work.

Just a matter of discussion. If you want, I can close the topic in order to not flood the forum.
All good by far

 

[oguz]

Ah, sorry, no issue at all. Just wondering if this is the right way to do the things.
I'm 100% sure when using VM, because this is completely separated/emulated process from the host OS, but what about CT's - it seems to work.

Just a matter of discussion. If you want, I can close the topic in order to not flood the forum.
All good by far

no worries the forum is for asking questions ^^

you can mark the thread [SOLVED] if you don't have any other questions