Operations

Cluster status

The status subcommand performs read-only local checks and does not require sudo.

curl -fsSL https://github.com/MuNeNiCK/setup-k8s/raw/main/setup-k8s.sh | sh -s -- status

Use wide output for cluster-level information:

curl -fsSL https://github.com/MuNeNiCK/setup-k8s/raw/main/setup-k8s.sh | sh -s -- status --output wide

Text mode checks:

• Node role.
• kubelet, containerd, and CRI-O service status.
• Installed kubelet, kubeadm, and kubectl versions.
• Cluster nodes and kube-system pods when kubectl is configured.

Wide mode also checks:

• API server endpoint.
• Pod and service CIDRs.
• etcd health on control-plane nodes.

Logging

Persist logs with --log-dir:

setup-k8s.sh deploy \
  --log-dir /var/log/setup-k8s \
  --control-planes 10.0.0.1 \
  --ssh-key ~/.ssh/id_rsa

Log files use timestamped names and mode 0600.

Audit logging

Structured audit events are recorded for deploy, upgrade, remove, backup, restore, and renew operations. Events include timestamp, operation, outcome, and user.

When --log-dir is set, audit events are written to the log file. Add --audit-syslog to send events to syslog via logger -t setup-k8s.

Failure diagnostics

Enable automatic diagnostics collection with --collect-diagnostics.

Collected data includes:

• kubelet logs.
• containerd logs.
• recent cluster events.
• disk and memory status.

Diagnostics are saved under /tmp/setup-k8s-diag-* on the orchestrator.

Health checks

Post-operation health checks run after deploy, upgrade, and remove:

• API server readiness.
• Node readiness.
• etcd health and quorum.
• kube-system pod health.

Pre-operation health checks run before upgrade and remove.

Resume operations

Long-running deploy and upgrade operations can be resumed:

setup-k8s.sh deploy \
  --resume \
  --control-planes 10.0.0.1,10.0.0.2 \
  --ssh-key ~/.ssh/id_rsa

State is stored in /var/lib/setup-k8s/state/. Bundle generation and transfer are always re-run because remote temporary directories are cleaned on exit.