Using an Amazon AWS S3 Bucket as backup storage

I currently have a cronjob set up to copy the entire datastore directory from PBS to an S3 compatible storage platform. Just tested restoring by installing PBS on a new machine and downloading the datastore from the bucket. Works almost flawlessly and backups are verified and restorable.

The main downsides is that it's another manual thing to maintain, doesn't provide notifications, and PBS has no idea it wasn't the one to create the backups in the first place, so the verification checkmarks will show as good, even if they aren't. Just have to run a verify job again after the download
 
May I add another option here? using https://juicefs.com you get a fully posix compatible FS backed by S3 and (to name an example) redis as metadata server. This is NOT s3fs or alike.

I've been using it in a lot of projects even with databases and it works flawlessly and it's safe enough. I had it running for years now (with other things that aren't PBS, so **be advised to test in non-prod environments and absolutely DO NOT take this post as a direct advice on using JFS or OFS as storage backend**).
There is another S3 FS like JuiceFS but with a commercial license called ObjectiveFS. As good as JuiceFS, maybe even better, and it does not need separate metadata server so it it incredibly convenient.

Anyway, I will try JFS by myself.
 
Last edited:
You must log in or register to reply here.
Top Bottom