[Verse 1] When your server needs some care and love Time to take it offline from above But pods are running, workloads too Here's the gentle way to see it through [Chorus] Kubectl drain, move them all away Ignore daemon sets, they're here to stay Delete empty dir data, clean and bright Drain the node for maintenance night Kubectl drain, that's the way Keep your cluster running every day [Verse 2] First you tell Kubernetes the plan Mark the node as unschedulable man Then eviction notices go out Pods get time to shut down without doubt [Chorus] Kubectl drain, move them all away Ignore daemon sets, they're here to stay Delete empty dir data, clean and bright Drain the node for maintenance night Kubectl drain, that's the way Keep your cluster running every day [Bridge] System pods will stick around Daemon sets stay safe and sound But your apps will migrate free To other nodes gracefully [Verse 3] Type the command with the node name clear Add those flags so nothing you fear Ignore daemon sets is what you need Delete empty dir helps you succeed [Chorus] Kubectl drain, move them all away Ignore daemon sets, they're here to stay Delete empty dir data, clean and bright Drain the node for maintenance night Kubectl drain, that's the way Keep your cluster running every day [Outro] When maintenance is done and through Uncordon brings it back for you The node returns to active state Your cluster's healthy, running great
← List nodes with status | Cordon a node (prevent new scheduling) →