Date: Tue, 10 Apr 2007 13:58:33 +0200 From: Pawel Jakub Dawidek <pjd@FreeBSD.org> To: Gergely CZUCZY <phoemix@harmless.hu> Cc: Ivan Voras <ivoras@fer.hr>, freebsd-geom@freebsd.org Subject: Re: volume management Message-ID: <20070410115833.GC85578@garage.freebsd.pl> In-Reply-To: <20070410113702.GA17344@harmless.hu> References: <20070408140215.GA54201@harmless.hu> <86k5wmu420.fsf@dwp.des.no> <20070408181916.GA59715@harmless.hu> <86bqhyu225.fsf@dwp.des.no> <461A4D93.3010200@freebsd.org> <20070409143818.GA86722@harmless.hu> <20070409152401.GG76673@garage.freebsd.pl> <20070409153203.GA88082@harmless.hu> <evfr82$n23$3@sea.gmane.org> <20070410113702.GA17344@harmless.hu>
next in thread | previous in thread | raw e-mail | index | archive | help
[-- Attachment #1 --]
On Tue, Apr 10, 2007 at 01:37:02PM +0200, Gergely CZUCZY wrote:
> The mot accurate description would be the functionality
> of linux's LVM in the scope of volume management, that's why
> i had given this subject to this email/topic.
>
> I have a storage pool (the last part of a disk), from which
> i want to chop little pieces for services, and i want to be
> able to increment these little pieces whenever needed.
>
> Like i have a pool of 100G. i have two services, i allocate
> 10G and 20G for them. 30G is used. Upon need, i allocate a bit
> more first to service1, then it will have 15G of space
> available. After this, the other one might need a bit more
> space, so i give it a bit more, let's say 8G, and it will
> have 28G of space.
>
> It's really what i have described here several times. Having
> a storage pool, from which volume management can be done. Dynamic
> allication of storage areas from the same pool, and having the
> abilit to increase the size of the previously allocated storage
> areas when it's needed. It's kinda like what's Linux's LVM's
> main feature. And i'm a bit surprised that FreeBSD doesn't
> have a similar thing.
Yeah... If you can afford running 7-CURRENT, you should really try ZFS.
Even if you don't want to use ZFS file systems, you can still use ZFS
ZVOLs and run UFS on top of them.
For example:
# zpool create tank mirror da0 da1
# zfs create -V 10g tank/foo
# zfs create -V 20g tank/bar
# diskinfo -v /dev/zvol/tank/{foo,bar}
/dev/zvol/tank/foo
512 # sectorsize
10737418240 # mediasize in bytes (10G)
20971520 # mediasize in sectors
/dev/zvol/tank/bar
512 # sectorsize
21474836480 # mediasize in bytes (20G)
41943040 # mediasize in sectors
# newfs /dev/zvol/tank/foo
# mount /dev/zvol/tank/foo /foo
# newfs /dev/zvol/tank/bar
# mount /dev/zvol/tank/bar /bar
Now, when you want to grow tank/foo:
# umount /foo
# zfs set volsize=30g tank/foo
# growfs /dev/zvol/tank/foo
# mount /dev/zvol/tank/foo /foo
Not to mention you have fast UFS snapshots, clones and other cool
stuff:)
The hardest part of ZFS was ZPL (ZFS POSIX Layer), which only comes into
play for ZFS file systems and is not used when using ZVOLs.
--
Pawel Jakub Dawidek http://www.wheel.pl
pjd@FreeBSD.org http://www.FreeBSD.org
FreeBSD committer Am I Evil? Yes, I Am!
[-- Attachment #2 --]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (FreeBSD)
iD8DBQFGG3vpForvXbEpPzQRAmJ1AKC79rCXXekqD5jmpcW7FqKJZWkIHwCgyxP4
TQBmw+4te/HxS0j/zHTjv3c=
=itv/
-----END PGP SIGNATURE-----
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070410115833.GC85578>