Back when I set up the server I wasn't sure how I'd handle backups. I'm a little more sure, now. Most of what is on the server is replicas of some other media: my notebooks' backups, my CD and DVD collection, and so forth. Loss of the server does not mean loss of data, so backups in this case are a matter of convenience: it would take less time to clone a dataset than it would to re-rip all of my CDs.
A refresher: I have a single RAID-Z pool with 4 x 2TB disks and I run a scrub on a weekly basis because that is recommended for consumer grade disks. Right now I have only one backup disk, which I would technically call a replica rather than a proper backup. It's a 2TB disk in a USB enclosure. The disk is partitioned with its own flat zpool. The backup process is manual. When I want to run off replicas I first import the external disk's pool. For each dataset that I want to back up I make a snapshot, use zfs send and zfs receive to copy it to the external disk, and then do some cleanup on that external disk. For now there is only one snapshot for each dataset on the external disk, but I keep the snapshots on the server. Like I said, the replica is for convenience rather than durability. Once finished I export the replica pool and unplug the enclosure. I have a 2-disk enclosure that I intend to press into service. My plan is to have each disk be its own pool. I intend to alternate disks so that I will have a current and a previous replica available. --Rich P. _______________________________________________ Discuss mailing list [email protected] http://lists.blu.org/mailman/listinfo/discuss
