Skip to content

Verify the cluster

Run these after the last node has joined and storage is set up. Each line says what a healthy answer looks like.

Check Command Healthy
Ready stratumctl admin readiness Exit 0. No blocking finding. Warnings name only what you chose to skip, such as operator images.
Members stratumctl admin cluster members list Every node active, none joining or failed.
Nodes stratumctl nodes list Every node ready, with a recent last-seen.
Substrate stratumctl admin selfcheck Exit 0. Its cluster HA check covers Postgres, OpenBao, OVN and the service addresses on every node, and its PKI check that every node trusts one cluster CA.
Versions stratumctl admin doctor Every component supported, and the OVN central and ovn-controller versions match.
Reconcilers stratumctl admin reconcilers fleet converged, or disabled for services that wait for the object gateway.
Ceph curl -s "$API/v1/admin/storage/ceph/status" -H "Authorization: Bearer $TOKEN" HEALTH_OK, every OSD up and in, all placement groups active and clean.
Storage stratumctl storage-classes list A default class, and the ceph-rbd class.
Images stratumctl images list An active image.

The three Postgres, OpenBao and OVN clusters each show one leader and the other members following. The cluster at a glance shows how to read them.

The end-to-end check creates a VPC, subnet, security group, key pair and an instance from an existing image in the project, waits for it to run, and deletes everything:

Terminal window
stratumctl admin selfcheck --tenant --project p-system