Summary - Disaster recovery

Combine snapshot, Velero and RPO to write a recovery plan.

Summary: disaster recovery

In this group:

Three layers to back up: etcd (objects), PersistentVolumes (data), Git (definitions); the 3-2-1 rule.
etcdctl snapshot save and etcdutl snapshot status/restore; the snapshot contains Secrets and is stored encrypted.
Velero: namespace-level object and volume backups; backup create, schedule create --ttl.
velero restore create --from-backup, opening for inspection with --namespace-mappings.
RPO is data loss, RTO is downtime; lowered with frequent backups, replication, warm standby and drills.

velero backup get
velero restore create --from-backup gece-20261103020000

Exercise A drill

Write three commands for a restore drill: switch to the staging-dr context (kubectl config use-context); create a restore named tatbikat from the backup gece-20261103020000; list the pods in magaza.

Show example solution

kubectl config use-context staging-dr
velero restore create tatbikat --from-backup gece-20261103020000
kubectl get pods -n magaza

Test yourself

How do you know a disaster plan really works?
- By Velero being installed
- By never having a disaster
- With regular restore drills measuring time and data
- By looking at the size of the backup files
Show answer

With regular restore drills measuring time and data — An untested backup may turn out broken when needed.

Restore from the backup.

velero restore create ____ gece-20261103020000

Show answer

velero restore create --from-backup gece-20261103020000

Match.

etcdctl

etcd snapshot
Velero

Object + volume backup
RPO

Accepted data loss
3-2-1

Copy rule

Next lesson: Challenge - CRD

We Are Enhancing Your Experience

Cookie Preferences

Strictly Necessary Cookies

Analytical Cookies