From owner-freebsd-questions Wed Aug 23 8:36:46 2000 Delivered-To: freebsd-questions@freebsd.org Received: from one.net-noise.com (ansible.nwark.net [208.136.254.190]) by hub.freebsd.org (Postfix) with ESMTP id CCC3737B422 for ; Wed, 23 Aug 2000 08:36:22 -0700 (PDT) Received: from guinevere (guinevere.net-noise.com [192.168.1.14]) by one.net-noise.com (8.9.3/8.9.3) with SMTP id KAA11280 for ; Wed, 23 Aug 2000 10:36:17 -0500 Message-ID: <002801c00d18$206e4e50$0e01a8c0@guinevere> From: "J. Seth Henry" To: Subject: re: vinum preparation Date: Wed, 23 Aug 2000 10:38:02 -0500 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_0025_01C00CEE.366CE5B0" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.50.4133.2400 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG This is a multi-part message in MIME format. ------=_NextPart_000_0025_01C00CEE.366CE5B0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Sorry for the delay, I've been playing with the kernel configuration as well. I was able to successfully create a stripe set using the following: "/etc/vinum.conf" > drive subdisk1 device /dev/da1s1e drive subdisk2 device /dev/da2s1e drive subdisk3 device /dev/da3s1e drive subdisk4 device /dev/da4s1e drive subdisk5 device /dev/da5s1e drive subdisk6 device /dev/da6s1e volume mmvolc plex org striped 1024b sd length 2033m drive subdisk1 sd length 2033m drive subdisk2 sd length 2033m drive subdisk3 sd length 2033m drive subdisk4 sd length 2033m drive subdisk5 sd length 2033m drive subdisk6 < However, when I try to create a RAID5 array, I get an error that one of = the plexes doesn't have enough subdisks and there are two plexes, mmvolc.p0 = and mmvolc.p1, rather than just the one plex, mmvolc.p0. I simply changed the plex line to read "plex org raid5 1024b" The volumes are partitioned to fill the entire space (dangerously = dedicated) and I created equal slices of 4165120 blocks on each disk. The leftover space was left unallocated. I would prefer the safety factor of a RAID array, given that these are = old disks, so this is a bit problematic. Any help would be appreciated. I am still migrating from Linux, which is why I've been playing with the kernel. I need EXT2 support for a little while. Once I get the data on = those volumes moved around and backed up again, I will supply the create log from vinum. Thanks, Seth Henry jshenry@net-noise.com ----- Original Message ----- From: "Greg Lehey" To: "J. Seth Henry" Cc: Sent: Thursday, August 17, 2000 7:41 PM Subject: Re: vinum preparation > [Format recovered--see http://www.lemis.com/email/email-format.html] > > Please don't write one line per paragraph. > > On Thursday, 17 August 2000 at 10:21:39 -0500, J. Seth Henry wrote: > > I have a pile of 2Gb (more or less) SCSI hard disks laying around > > that I would like to bolt together into a RAID5 array using > > vinum. The trick is, all of the drives have slightly different = sizes, > > varying from 2033Mb to 2048Mb. I know that the subvolumes have to be > > exactly the same size, but is that the partition, or the slice? > > No. > > > I currently have them all dangerously dedicated to the full size of > > the drive, and I set the slices to be the same size. > > > > I have read all of the documentation I could get my hands on > > regarding vinum, but so far, I've found nothing that covers > > this. When I try to start the array, three of the disks come online, > > but the other three fail due to a size mismatch. From this, I would > > assume the partitions must be equally sized and the slices set to > > fill the partitions. > > No, this isn't correct. Your subdisks need to be the same size, but > Vinum should automatically trim the ones which are too large. Please > show (without mutilating the text!) the output from 'vinum create'. > > Greg > -- > When replying to this message, please copy the original recipients. > For more information, see http://www.lemis.com/questions.html > Finger grog@lemis.com for PGP public key > See complete headers for address and phone numbers ------=_NextPart_000_0025_01C00CEE.366CE5B0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
Sorry for the delay, I've been playing with the kernel = configuration=20 as
well.

I was able to successfully create a stripe set using = the=20 following:

"/etc/vinum.conf"
>
drive subdisk1 device=20 /dev/da1s1e
drive subdisk2 device /dev/da2s1e
drive subdisk3 = device=20 /dev/da3s1e
drive subdisk4 device /dev/da4s1e
drive subdisk5 = device=20 /dev/da5s1e
drive subdisk6 device /dev/da6s1e

volume = mmvolc
plex=20 org striped 1024b
sd length 2033m drive subdisk1
sd length 2033m = drive=20 subdisk2
sd length 2033m drive subdisk3
sd length 2033m drive=20 subdisk4
sd length 2033m drive subdisk5
sd length 2033m drive=20 subdisk6
<

However, when I try to create a RAID5 array, I = get an=20 error that one of the
plexes doesn't have enough subdisks and there = are two=20 plexes, mmvolc.p0 and
mmvolc.p1, rather than just the one plex,=20 mmvolc.p0.

I simply changed the plex line to read "plex org raid5 = 1024b"

The volumes are partitioned to fill the entire space = (dangerously=20 dedicated)
and I created equal slices of 4165120 blocks on each disk. = The=20 leftover
space was left unallocated.

I would prefer the safety = factor=20 of a RAID array, given that these are old
disks, so this is a bit=20 problematic. Any help would be appreciated.

I am still migrating = from=20 Linux, which is why I've been playing with the
kernel. I need EXT2 = support=20 for a little while. Once I get the data on those
volumes moved around = and=20 backed up again, I will supply the create
log from=20 vinum.

Thanks,
Seth Henry
jshenry@net-noise.com



-----=20 Original Message -----
From: "Greg Lehey" <grog@lemis.com>
To: "J. Seth = Henry"=20 <jshenry@net-noise.com>
Cc= :=20 <questions@FreeBSD.ORG>
Se= nt:=20 Thursday, August 17, 2000 7:41 PM
Subject: Re: vinum=20 preparation


> [Format recovered--see http://www.lemis.co= m/email/email-format.html]
>
>=20 Please don't write one line per paragraph.
>
> On Thursday, = 17=20 August 2000 at 10:21:39 -0500, J. Seth Henry wrote:
> > I have = a pile=20 of 2Gb (more or less) SCSI hard disks laying around
> > that I = would=20 like to bolt together into a RAID5 array using
> > vinum. The = trick is,=20 all of the drives have slightly different sizes,
> > varying = from=20 2033Mb to 2048Mb. I know that the subvolumes have to be
> > = exactly the=20 same size, but is that the partition, or the slice?
>
>=20 No.
>
> > I currently have them all dangerously dedicated = to the=20 full size of
> > the drive, and I set the slices to be the same = size.
> >
> > I have read all of the documentation I = could get=20 my hands on
> > regarding vinum, but so far, I've found nothing = that=20 covers
> > this. When I try to start the array, three of the = disks come=20 online,
> > but the other three fail due to a size mismatch. = From this,=20 I would
> > assume the partitions must be equally sized and the = slices=20 set to
> > fill the partitions.
>
> No, this isn't=20 correct.  Your subdisks need to be the same size, but
> Vinum = should=20 automatically trim the ones which are too large.  Please
> = show=20 (without mutilating the text!) the output from 'vinum = create'.
>
>=20 Greg
> --
> When replying to this message, please copy the = original=20 recipients.
> For more information, see http://www.lemis.com/questio= ns.html
>=20 Finger grog@lemis.com for PGP = public=20 key
> See complete headers for address and phone=20 numbers

------=_NextPart_000_0025_01C00CEE.366CE5B0-- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message