AWS S3/Google Cloud Storage remote file system
MiBa last edited by
Hi all, anyone know a way to set this up or is this a task for you @olivierlambert? (If feasible and others think it's a good idea )
@olivierlambert, If not possible already then an integration for S3 only would work as Google can emulate S3 api. As long as it is possible to define a custom target endpoint.
This is intended for offsite backups of the VMs for when s*** hits the fan.
It's planned for this year, at least for XOA with our Cloud Backup feature
Did this make any progress?
With the new extremely cheap S3 glacier tier, I could totally see value in making a full backup to S3 once a month or so.
The downside of S3 is that's complicated (but we'll see soon) to edit files in place. So it's good for full backup, but delta seems more complicated. Stay tuned.
S3 has write only permission tiers. So I could see a situation where the VM is backed up to S3 with no ability for the backup to be accessed if XOA is compromised. This would make perfect sense against crypto ransomware attacks.
Delta backups really don't interest me for any S3 related scenarios.
So a kind of "write once" for "full" (eg XVAs) backup, right?
Yep, you got it.
S3 automatic storage tiers can do the rest, including object expiration.
Well, it's always a tradeoff, how much "standard" code of XO we'll use to do that VS doing spaghetti code for every different kind of remote
Anyway, it will be in the pipes!
Perhaps there are some middleware solutions to save you from writing too much spaghetti code, ie https://github.com/gaul/s3proxy
olivierlambert last edited by olivierlambert
Adding existing spaghetti is not better than creating your own spaghetti, especially when those pasta are written in Java
We'll see what we chose in time