bhruska's picture

I'm a bit confused about the snapshots.  If I wanted to restore my server to a snapshot from a few hours ago, it looks like I can only 

Launch New Cloud Server

Create Volume

Delete Snapshot

I don't see an hour to revert my existing server to that snapshot.  So, do I have to launch a new server and reassign the elastic IP in order to revert to a snapshot?



Jeremy Davis's picture

TBH I haven't really played much with snapshots so I can't be 100% sure, but I would assume that you are correct.

I'm not really sure on the complete rationale for this setup (or whether it's a technical limitation or something...), but I assume that it is to prevent a situation where something goes wrong during the snapshot restore and you get left with a corrupted server...

I would suggest that it is probably poor practice to be restoring any backup (snapshot, TKLBAM backup or whatever) to a production server without first confirming that the backup is functioning as it should be...

So unless your current server is completely fried, I'd be wanting to restore your snapshot and check it is all ok before you adjust your ElasticIP. Although perhaps that's just me being overly cautious...?!

Alon Swartz's picture

You're correct in your assumption Brian, the way to acheive what your after is to launch a new server from a snapshot and then reassign the elastic ip.

Behind the scenes snapshots are point in time "snapshots" of the underlying EBS volume saved to S3 storage. The snapshot can be converted to an EBS volume for mounting or manipulation, and the resulting EBS volume can be used as the rootfs for a new instance (ie. cloning a server).

I'd recommend you read the blog post for more details.

Add new comment