On Sun, May 22, 2011 at 09:31:49AM +0000, ed wrote:
On Sun, May 22, 2011 at 10:25:54AM +0100, Hugo Mills
wrote:
Gaah. Ignore me. It was late, and I've
been spending a lot of time
on the btrfs mailing list recently. This looks just like the sort of
problem people get on there...
How do you rate btrfs compared with zfs, if you've got experience with
the two that is...
I've not used ZFS, so I can't compare the two directly. However,
btrfs is currently less mature: we recommend having good backups
regardless of your filesystem, but we suggest that you be prepared to
use them if you're running btrfs.
That said, the latest version is pretty stable in normal use. It
supports RAID-0, -1 and -10 independently for each of data and
metadata, has independently-moutable subvolumes and copy-on-write
snapshots, and supports online resize and disk insertion and removal.
It's fully checksummed, so if it encounters corrupt data it can tell
you, *and* attempt to go and read from the other copy in your RAID-1
(or -10) array -- MD RAID-1 won't do this for you, because it doesn't
have the checksums, so it can't tell which copy is good. There's a
mode for supporting SSDs.
What's not there yet, but planned: a working fsck(*),
deduplication, a cross-filesystem replication tool like ZFS's
send/receive, RAID-5/6 (in .41, probably), hierarchical storage,
per-subvolume RAID options, changing RAID levels, encryption.
Hugo.
(*) Rumour has it that we'll get the first cut of this at the end of
the month. But we've all heard that one before.
--
=== Hugo Mills: hugo@... carfax.org.uk |
darksatanic.net | lug.org.uk ===
PGP key: 515C238D from
wwwkeys.eu.pgp.net or
http://www.carfax.org.uk
--- Is a diversity twice as good as a university? ---