Date: Wed, 23 Dec 2009 14:56:18 -0600 From: Barry Pederson <bp@barryp.org> To: Bob Friesenhahn <bfriesen@simple.dallas.tx.us> Cc: freebsd-fs@freebsd.org, Patrick Proniewski <patpro@patpro.net> Subject: Re: snapshot implementation Message-ID: <4B3283F2.7060804@barryp.org> In-Reply-To: <alpine.GSO.2.01.0912231031450.1586@freddy.simplesystems.org> References: <32CA2B73-3412-49DD-9401-4773CC73BED0@patpro.net> <alpine.GSO.2.01.0912231031450.1586@freddy.simplesystems.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 12/23/09 10:41 AM, Bob Friesenhahn wrote: > On Wed, 23 Dec 2009, Patrick Proniewski wrote: > >> By the way, I'm also interested in ZFS: is the snapshot technology >> available in ZFS the same as the one available in UFS? > > I don't know anything about snapshots in UFS, but snapshots in ZFS are > certainly remarkably different. ZFS uses copy-on-write (COW) whenever a > data block is updated and snapshot creation simply adds a new reference > to existing blocks. The snapshot is made available as a (usually) hidden > directory (/filesystem/.zfs/snapshot/snapname) which contains the > complete filesystem content at the time the snapshot was taken. In my > experience, ZFS snapshots usually take less than a second to complete. > They are so efficient that some systems have snapshots scheduled to be > taken every five minutes as a defense against user/application error. I always liked this quote from this writeup on ZFS: http://www.sun.com/bigadmin/features/articles/zfs_part2_ease.jsp "...there's virtually no overhead at all due to the copy-on-write architecture. In fact, sometimes it is faster to take a snapshot rather than free the blocks containing the old data!" That's certainly not the case with UFS snapshots, which can take a long time to complete (we're talking freezing your machine's disk activity for many minutes), and are limited to 20 total. Barry
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4B3283F2.7060804>