On Fri, 18 Jul 2014 05:34:22 -0700 Duncan <1i5t5.duncan@xxxxxxx> wrote: > Effectively, admins can choose NOCOW XOR frequent-snapshotting, altho > the fact that snapshots stop at subvolume borders can be used as a > partial workaround, by putting NOCOW files on a dedicated partition and > not snapshotting it, exactly as I mentioned. You can't backup running VM images and datafiles of an active database using "traditional" backup techniques such as file copy or rsync. The tail of a file you're copying for a backup will be long-inconsistent with the overall state or the head of the file when you started copying. Snapshots on the other hand are atomic, and can very much be used to create a static copy of the files for the purposes of compressing/copying away somewhere. And at worst, the "restored from backup" state of such a backed up VM or DB will be equivalent to it just having had a power-loss. Journalling FSes and databases can deal with that with no major problems. So just exercise moderation, snapshot e.g. once an hour or even a day, the result will still be better than not using NOCOW, and will deliver most of the benefits you get by snapshotting. Another option is to snapshot->backup->delete snapshot. -- With respect, Roman
Attachment:
signature.asc
Description: PGP signature
