I'm curious about the necessity of etcd backups in my setup. Since everything is deployed using ArgoCD, are etcd backups really required? If I do need them, is the best way to handle backups using a CronJob YAML? Also, should I locate the etcd leader node before proceeding with the backup?
4 Answers
It depends on your Recovery Time Objective (RTO) and how quickly you can get a new cluster up and running. Think about what failures you want to guard against and how you plan to respond. Sometimes a complete cluster reinstall can actually be a valid disaster recovery method.
It really comes down to your recovery strategy. If you want to recover PVCs, you might need the unique IDs stored in etcd. Using a backup solution specifically for PVCs would be best. Just keep that in mind!
We back up our Argo applications and app projects every hour, and we restore that when migrating to new clusters. But I'm not sure about the metadata for PVCs—do those get included?
If most of your applications are stateless and can be easily redeployed, plus you can quickly replace the clusters, you might not need to back up etcd at all. The 'cattle, not pets' philosophy applies here as well!
Related Questions
Can't Load PhpMyadmin On After Server Update
Redirect www to non-www in Apache Conf
How To Check If Your SSL Cert Is SHA 1
Windows TrackPad Gestures