FreeBSD Preferred RAID controllers
Gabe
nrml at att.net
Sat Feb 14 20:41:27 PST 2009
--- On Sat, 2/14/09, Matthew Seaman <m.seaman at infracaninophile.co.uk> wrote:
> From: Matthew Seaman <m.seaman at infracaninophile.co.uk>
> Subject: Re: FreeBSD Preferred RAID controllers
> To: nrml at att.net
> Cc: freebsd-questions at freebsd.org
> Date: Saturday, February 14, 2009, 8:30 AM
> Gabe wrote:
> > --- On Mon, 2/9/09, Matthew Seaman
> <m.seaman at infracaninophile.co.uk> wrote:
> >
> >> From: Matthew Seaman
> <m.seaman at infracaninophile.co.uk>
> >> Subject: Re: FreeBSD Preferred RAID controllers
> >> To: nrml at att.net
> >> Cc: freebsd-questions at freebsd.org
> >> Date: Monday, February 9, 2009, 3:08 PM
> >> Gabe wrote:
> >>
> >>> Now with a gstripe+gmirror setup, would it be
> possible
> >> to fail a
> >>> specific drive on purpose? I mean fail a
> (good) drive,
> >> pull it out,
> >>> replace it and rebuild(?) it. I know I know,
> but humor
> >> me.
> >>
> >> Yes.
> >>
> >> Cheers,
> >>
> >> Matthew
> >>
> >> Well, to 'fail' the drive, you'ld have
> to
> >> physically pull the drive
> >> from the chassis which will involve a power cycle
> unless
> >> you've got
> >> hot-swap drives. Of course, you should confirm
> that your
> >> system will
> >> boot with the RAID in a degraded state and that
> rebuilding
> >> the RAID will
> >> continue even if interrupted by a reboot.
> gmirror(8)
> >> passes those
> >> tests. You do have to type some commands to get a
> mirror
> >> to rebuild
> >> (examples are shown in the man page) unlike some
> hardware
> >> RAIDs where
> >> simply inserting an unused disk is sufficient.
> >>
> >> -- Dr Matthew J Seaman MA, D.Phil.
> 7
> >> Priory Courtyard
> >>
> Flat 3
> >> PGP: http://www.infracaninophile.co.uk/pgpkey
> Ramsgate
> >>
> Kent, CT11
> >> 9PW
> >
> > Hello again all,
> >
> > So I wanted to test out gmirror on software RAID so I
> installed a completely vanilla FBSD 7, as base an install as
> you can get, it hasn't even been on the network. Anyway,
> I did the following upon first boot to get gmirror going:
> >
> > # sysctl kern.geom.debugflags=17
> >
> > Then:
> > # gmirror label -vb round-robin gm0 /dev/ad0
> >
> > Then:
> > # gmirror load
> >
> > Then:
> > # echo 'geom_mirror_load="YES"'
> >> /boot/loader.conf
> >
> > Then I edited /etc/fstab to show:
> > /dev/mirror/gm0s1b none swap sw
> 0 0
> > /dev/mirror/gm0s1a / ufs rw
> 1 1
> > /dev/mirror/gm0s1e /tmp ufs rw 0 0
> > /dev/mirror/gm0s1f /usr ufs rw 2 2
> > /dev/mirror/gm0s1d /var ufs rw
> 2 2
> >
> > I then rebooted the system, once I setup the mirror:
> >
> > # gmirror insert gm0 /dev/ad1
> > # gmirror status
> >
> > and it shows as COMPLETE. Okay, here comes the
> annoying part, I've got hot-swappable bays and I went
> ahead and pulled the drive. I then tried to write to the
> disk so that it realizes the disk is no longer there:
> >
> > # touch file
> >
> > once I do that and execute: gmirror status it shows as
> degraded. All fine and dandy. However when it comes time to
> pop the drive back in the drive is not recognized at all. I
> mean, the green light on the bay comes on so it definitely
> makes a connection but then thats it, atacontrol list
> doesn't list it and gmirror status still shows the same,
> degraded.
> >
> > What gives? I wonder if this is hardware related? Bios
> related even? Any clues?
>
> See atacontrol(8)
>
> # atacontrol list
>
> shows what your system knows is there
>
> # atacontrol attach X
>
> where X is the channel number, probes and attaches any
> devices on that
> channel in exactly the same way it is done at system boot.
>
> See camcontrol(8) if you've got SCSI drives.
>
> Cheers,
>
> Matthew
>
> --
> Dr Matthew J Seaman MA, D.Phil. 7 Priory
> Courtyard
> Flat 3
> PGP: http://www.infracaninophile.co.uk/pgpkey Ramsgate
> Kent,
> CT11 9PW
They're SATA drives. The two drives are on the same channel when using atacontrol list. I'm unsure that atacontrol attach ata0 would work but I'll give it a shot, hopefully that works. I'll report back.
Thanks again
More information about the freebsd-questions
mailing list