Mateusz Gozdek

Ah, one more thing which comes to my mind. Are you trying to reboot controller or worker node? I'm not sure if FLUO is currently able to reboot the node...

I've tested it again with Lokomotive on latest master on Packet and here are the results: - after deploying FLUO, all nodes (including the controller) has been sequentially rebooted and...

>weird, wonder why it works for you but not for me. I'll try another one, what did you do to trigger an update (reboot), simply wait? I did `locksmithctl send-need-reboot`...

Good to hear that you got it to work. 0.7.3 tag contains manifests, which deploys the same version of agent and operator: https://github.com/kinvolk/flatcar-linux-update-operator/blob/v0.7.3/examples/deploy/update-agent.yaml.

We should use https://github.com/coreos/go-systemd/blob/65cc0ad82617/dbus/methods.go#L153 to idempotently determine if system is shutting down and if no, trigger a reboot. Triggering a reboot must perhaps be done not via `go-systemd` until they...

>Triggering a reboot must perhaps be done not via go-systemd until they add an error handling for calling a reboot. But replacing a reboot should actually be simple, as it's...

I think a good solution would be to use before/after reboot annotations to simply wait some time before proceeding with node reboot. Perhaps looking at #37 will give you an...

>How hard would this be to implement? Dead simple to implement right now, but hard to maintain, this is why I suggested using hooks instead.

Reading the code, this is clearly a bug, as https://github.com/kinvolk/flatcar-linux-update-operator/blob/4036c440e59e886a6ee94af562ec3374d3292ba9/pkg/agent/agent.go#L303 assumes, that update should happen within 24 hours from starting the watch. However, if no updates are posted within 24...

Also between https://github.com/poseidon/fleetlock