Date: Wed, 02 Sep 2009 10:27:40 +0200 From: Mark Stapper <stark@mapper.nl> To: Emil Mikulic <emikulic@gmail.com> Cc: Maciej Jan Broniarz <gausus@gausus.net>, freebsd-stable@freebsd.org, Miroslav Lachman <000.fbsd@quip.cz>, Thomas Backman <serenity@exscape.org> Subject: Re: zfs on gmirror slice Message-ID: <4A9E2C7C.6030904@mapper.nl> In-Reply-To: <20090902074445.GA13588@dmr.ath.cx> References: <a7454e2f0909010704g2fb27216hacb3ffd2cae5594c@mail.gmail.com> <061541E3-F301-46C4-8ECB-5B05854F0EAA@exscape.org> <a7454e2f0909010904s224be2ectdd18eb887f4c1311@mail.gmail.com> <4A9D558A.9070609@quip.cz> <4A9E1CB5.6030906@mapper.nl> <20090902074445.GA13588@dmr.ath.cx>
next in thread | previous in thread | raw e-mail | index | archive | help
[-- Attachment #1 --] Emil Mikulic wrote: > On Wed, Sep 02, 2009 at 09:20:21AM +0200, Mark Stapper wrote: > >> updating a zfs filesystem which you are running from is next to >> impossible. >> > > [citation needed] :) > Well, to update your zfs filesystem version, the filesystem is first unmounted, then updated, and then mounted again. citation coming up! # umount / umount: unmount of / failed: Invalid argument > >> So, i would recommend setting up gmirror to mirror your whole disks, >> install the base system(boot and "world") on a small UFS slice, and use >> the rest of the disc as zfs slice. >> > > As Thomas Backman pointed out, this means you won't get self-healing. > self-healing sounds very nice, but with mirrorring you have data on two discs, so in that case there no "healing" involved, it's just checksumming and reading the non-corrupted copy. From the gmirror manpage: "All operations like failure detection, stale component detection, rebuild of stale components, etc. are also done automatically." This would indicate the same functionality, with a much less fancy name. However, i have not tested it the way they demonstrate zfs's "self-healing" property. I might, if I get the time to run it in a virtual machine one of these days.. > I don't know if a ZFS mirror performs smarter disk access scheduling > than gmirror. Someone oughta measure. ;) > NCQ should help here, but still, very interesting. gmirror is fast though. I've even heard it's faster than software RAID1/ataraid. (not confirmed). Even if this is true, running zfs on top of gmirror probably isn't faster then running zfs with a mirrored pool. [-- Attachment #2 --] -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkqeLH4ACgkQN9xNqOOVnWB/RgCfd29VJcnGXFCP0pe3MUoG0fJm U6oAn3y+NPSeNeDAQzoEhJlupr9gKHTN =Vzij -----END PGP SIGNATURE-----
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4A9E2C7C.6030904>