0

We have a mongo instance which is hosted on Kubernetes as Statefulset. It's of around 3.5 TB size with persistent volumes attached.

We are looking for a way reduce backup time. What's the best way to backup and restore mongodb instance to and from AWS S3.

I've looked at Physical backup and Logical backup options using PBM. But not sure if they're suitable for instances deployed as statefulsets in k8s as they're of TBs size.

Underoos
  • 4,708
  • 8
  • 42
  • 85
  • Best way != fastest way. If you impose constraints on your system you can have a faster backup (e.g. if your backup doesn't handle sharded clusters). – D. SM Apr 15 '21 at 14:36
  • Best way is probably the MongoDB-provided tools. They are most likely to work correctly in complex scenarios (e.g. sharded transactions). – D. SM Apr 15 '21 at 14:37
  • So, what's the fastest way then? I wouldn't mind if we don't have option to restore specific collection. Just looking for a way to do backup faster. Looked at PBM's hot backup, also exploring other options. – Underoos Apr 15 '21 at 14:39
  • Run a standalone instance, use mongodump piped to your s3 uploader. – D. SM Apr 15 '21 at 14:59
  • You should also read https://docs.mongodb.com/manual/tutorial/backup-sharded-cluster-with-filesystem-snapshots/ and other MongoDB documentation pertaining to backups so that you understand what challenges are involved and how they might be addressed. – D. SM Apr 15 '21 at 15:03

0 Answers0