Questions regarding snapshots and backup to Amazon S3

[skybolt_1]

I'm fairly new to FreeNAS and have been using it for about 9 months in home environment. I have three separate machines which are backed up on an hourly interval to my FreeNAS server, which consists of an ASUS EEE PC running an Atom processor, 2 GBs of memory, and a 500 GB Western Digital Blue hard drive. The FreeNAS instance is running on an 8 GB SD card. Since the system is set up only as a backup, no files exist only on FreeNAS; hence the lack of a redundant hard drive solution. However, I am somewhat concerned that some catastrophic event (house fire, flood) could wipe out all copies of my data, and would like to create an automated cloud backup which runs from the FreeNAS system. The questions that I have right now are:

- Do ZFS snapshots contain all data required to recreate the ZFS filesystem in the event of a total hard disk failure?
- If so, what is the best means to back the snapshots up to an Amazon S3 bucket? Rsync?

- If ZFS snapshots are not sufficient to recreate the entire filesystem, what is the best method of backing up the data, compressing it, and sending it to an S3 bucket?

I've read through everything I can find on the forum / Google regarding this type of backup scenario, and it doesn't seem like many people have pursued it. Any suggestions or insight would be helpful.

 

[peterh]

you would not (normally) copy a snapshot to some device, copying your filesystem would be
a "normal thing" to do. thus rsync -a <filesystem> <backup filesystem> is what you suggest.

This contains the user data, but it does not contain the snapshots nor the zfs infrastructure.
Given less then 500 bytes of notes however is sufficient to build a new filesystem where
the above backup could be restored.

If you are interested in backuping the data AND all snapshots i would suggest
you send/receive to another zfs system. Note however that zfs is very picky about pool
versions, both systems must have the same version.