Power and shutdown

Goal	Command
Reboot or power off one node	systemctl reboot or systemctl poweroff on it
Power off the whole cloud	stratumctl admin cluster shutdown
Stop everything, power off by hand	stratumctl admin cluster shutdown --prepare-only
See where a shutdown is	stratumctl admin cluster shutdown status
Undo it before the power-off	stratumctl admin cluster shutdown cancel

One node

On a cluster, use stratumctl nodes reboot and stratumctl nodes shutdown: they hold the node out of scheduling and check that the cluster keeps its majority. See Shut down, reboot and cold start.

A plain systemctl reboot or poweroff is safe. On the way down the node asks every instance to shut down (ACPI), asks again every 20 seconds, and powers off the ones still running after 90 seconds; only then do the platform’s services stop. On the way back, the instances it stopped start again, unless their owner set them to stopped.

A crash or a power cut skips this: instances lose what they had not written.

The whole cloud

stratumctl admin cluster shutdown --reason "power work"

Confirm by typing the cluster ID, or pass --confirm <cluster-id>. The platform then stops every running instance (each gets 60 seconds to shut down), keeps each one’s desired state, and powers the nodes off. Instances with stop protection shut down with their node instead.

On a single node there is nothing else to coordinate: it stops the instances and powers the node off.

Coming back: power the nodes on. Nothing else is needed: the platform waits for every node, then starts the instances that were running.

status shows each phase; readiness carries ClusterShutdown while a shutdown or its recovery is in progress. Console: Admin › Substrate › Cluster › Shutdown.