longhorn-backup-dr

name	longhorn-backup-dr
description	Configure Longhorn backup target (S3 or NFS), author RecurringJobs for snapshot/backup/filesystem-trim, and restore a volume from a backup. Use when wiring up backups for the first time, editing recurring schedules, or performing a DR restore. Not for PVC create (longhorn-volume-ops) or disk work (longhorn-node-ops).

longhorn-backup-dr

Backup and disaster-recovery operations for Longhorn on anton.

Status

Not yet configured. Plan 0001 installs Longhorn; backup target wiring is a follow-up (expect an ADR + plan 0002+). The recipes below are forward-looking — validate against the canonical 1.11.1 docs at setup time.

Platform notes

Nothing Talos-specific for backup target wiring itself; see longhorn-node-ops/references/talos.md for the platform-level gotchas that already apply.

Recipes

Set backup target (S3)

S3 requires a Kubernetes Secret with AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, and optional AWS_ENDPOINTS (for non-AWS S3). Prefer an ExternalSecret sourced from the 1Password anton vault — see add-flux-app for the template.

Configure the target via HelmRelease values:

defaultSettings:
  backupTarget: s3://<bucket>@<region>/
  backupTargetCredentialSecret: longhorn-backup-credentials

Author a RecurringJob

apiVersion: longhorn.io/v1beta2
kind: RecurringJob
metadata:
  name: daily-backup
  namespace: storage
spec:
  cron: "0 3 * * *"
  task: backup            # snapshot | backup | filesystem-trim
  groups: [default]
  retain: 7
  concurrency: 2

Volumes with the label recurringjob.longhorn.io/default: enabled are picked up automatically. Prefer the default group so new PVCs opt in via label rather than per-volume job references.

Restore-as-new-PVC

Use a CSI VolumeSnapshot as dataSource — the Kubernetes-idiomatic path. See restore-flows for the exact shape and when to reach for Longhorn's own restore CRs.

DR volume

Continuously-syncing cross-cluster replication. Not an anton requirement today.

References

• Backup target config — S3/NFS target shapes, credentials, anton guidance
• Recurring jobs — task types, groups, retention, scheduling suggestions
• Restore flows — CSI snapshot restore vs Longhorn restore vs DR volume
• Backup discipline — snapshots ≠ backups, test restores, RPO tiering