Proper way to reboot nodes that need reboot

[Stephan Kristyn]

My nodes are signaling

container-linux-update.v1.coreos.com/status=UPDATE_STATUS_UPDATED_NEED_REBOOT

after auto-update.

How can I reboot them without breaking the scheduled pods?

[alex.s...@coreos.com]

Hi Steve,

I'm assuming this node is part of a Tectonic cluster. Is that right?

That status label is specific to Container Linux Update Operator, a Tectonic component that coordinates node updates. The CLUO should schedule the node reboot automatically and take care of draining the pods that might be running on it. Is that not actually happening for you?

How many nodes do you have in your cluster in total?

Thanks,

Alex

[Stephan Kristyn]

Hi,

exactly, it’s not rebooting automatically. I read that this feature will only come up from Mid-September tectonic releases, we are stilll on 1.7.3

We have 3 Master Nodes and 1 Worker and 1 ECD, but I tolerated the pods to run on Masters, so rebooting should be done with care.

If nothing helps I could reschedule the deployments, but its 50 already.

See here:

https://imgur.com/a/1E1CK

Regards,

Steve

--
You received this message because you are subscribed to the Google Groups "CoreOS User" group.
To unsubscribe from this group and stop receiving emails from it, send an email to coreos-user...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Alex Crawford]

On 10/17, Stephan Kristyn wrote:
> exactly, it’s not rebooting automatically. I read that this feature
> will only come up from Mid-September tectonic releases, we are stilll
> on 1.7.3
>
> We have 3 Master Nodes and 1 Worker and 1 ECD, but I tolerated the
> pods to run on Masters, so rebooting should be done with care.
>
> If nothing helps I could reschedule the deployments, but its 50
> already.
>
> See here:
>

> https://imgur.com/a/1E1CK <https://imgur.com/a/1E1CK>

What is the status of the pod scheduled by the
container-linux-update-operator deployment? Is there anything
interesting in the logs?

-Alex