Skip to content

Recreate

Backup

  1. Run backup-pvcs in cluster's directory.
  2. Perform application specific backup for the applications considered critical
  3. baikal
  4. home-assistant
  5. linkace

Destroy

Run buvisctl destroy in cluster's directory. It may fail if providers versions were updated since the cluster creation. In that case, run terraform init -upgrade in infrastructure/terraform directory and try again.

Bootstrap

  1. Set workstation's DNS to 1.1.1.1, because Blocky isn't running
  2. Make sure that GitHub PAT (stored in GITHUB_TOKEN environment variable) is still valid, and update it eventually
  3. Run buvisctl bootstrap in cluster's directory.
  4. Wait for the bootstrap to finish. It may take a while, because it will download all images and install all applications.
  5. Really, be patient, wait for no new workloads being created.
  6. Check status of kustomizations with flux get ks -A. If there are any errors, fix them.
  7. Check status of helmreleases with flux get hr -A. If there are any errors, fix them.
  8. if helmrelease's resources started after it reached timeout, you can run flux reconcile hr -n <NAMESPACE> <RELEASE> to retry
  9. if the helmrelease is still not ready, restart: flux suspend hr -n <NAMESPACE> <RELEASE> and then flux resume hr -n <NAMESPACE> <RELEASE>
  10. as the last resort, you can delete the helmrelease and let Flux recreate it: flux delete hr -n <NAMESPACE> <RELEASE>

Restore storage

  1. Run recreate-backup-jobs in cluster's directory.
  2. Run restore-pvcs in cluster's directory.

Recreate Authentik outpost

Run trigger-outpost-creation in cluster's directory.