pacer cluster-ops

estate dashboard docs

Starter box

checking…

Every other control needs this box. It is ephemeral and holds a lease — a running op keeps renewing it, so a long job is never reaped mid-flight. Start brings up the box and then has it build the cluster: from the barman archive if one holds a base backup, from scratch if none does. If a cluster is already ACTIVE it is left alone.

Cluster

not connected

Destructive

Type the cluster name to arm. These have no undo.

DR drill

Four phases. Each waits for the one before it, and phase 1 failing is a stop, not a retry.

1
backup
recoverability gate — CNPG, Velero, Longhorn
2
teardown
destroys; mints + verifies the recovery points
3
rebuild
recovers from the points phase 2 recorded
4
verify
every /healthz returns 200
Recovery points for phase 3 (optional — blank lets the rebuild list candidates)

Prefer the serverName whose last WAL is NEWEST — after a teardown that is the one phase 2's pre-teardown base backup wrote to.

Runs

runopstatusrcfinished
not connected

Log

select a run