From: Andreas Dilger <adilger@turbolabs.com>
To: linux-lvm@sistina.com
Subject: Re: [linux-lvm] Re: =?unknown-8bit?B?W2xp?= =?unknown-8bit?Q?nux-lvm=5D_Verst=E4ndnis?= von Stripes bei LVM
Date: Fri, 14 Sep 2001 09:52:31 -0600 [thread overview]
Message-ID: <20010914095230.S1541@turbolinux.com> (raw)
In-Reply-To: <20010914121333.A11814@dragon.blacknet.de>
On Sep 14, 2001 12:13 +0200, Goetz Bock wrote:
> On Fri, Sep 14 '01 at 11:32, Pawlowski Julian wrote:
> > ich habe eine Frage zur Funktion von Stripes beim LVM. Wof?r genau
> > ben?tigt man diese Funktion? Wird damit das LV auf mehrere der reellen
> > Partitionen zwecks Performancegewinn verteilt?
>
> Yes, that's exactly the intended bahaviour. if you have n (e.g. 2 or
> more) PV than a striped LV is splitted between this n PV. This can
> increase performance (like RIAD0 does) but it may backfire, if your PV
> are on the same reald device.
> If you want to use striped LVs, only use one PV per real device.
>
> e.g.: 4 HDDs of 40GB each (sda, sdb, sdc, sdd)
>
> you boot from sda and decide to have a 512MB root/boot partition for
> this. Than make 3 swap partitiones on sdb, sdc and sdd.
>
> on each device make a seccond partition with the rest of the space.
> create a PV out of each of this 4 partitions.
>
> you now have:
>
> sda: sda1 512MB root/boot
> sda2 39.5GB PV1
> sdb: sdb1 512MB swap
> sdb2 39.5GB PV2
> sdc: sdc1 512MB swap
> sdc2 39.5GB PV3
> sdd: sdd1 512MB swap
> sdd2 39.5GB PV4
>
> VG1: 118.0GB PV1 PV2 PV3 PV4
>
> if you now make a striped LV of 20GB it will use 5GB on each PV and the
> stripes will be read from PV1 PV2 PV3 PV4 PV1 PV2 PV3 PV4 ...
A note here - very rarely can one really get the theoretical performance
from striped LVs. There are several problems with stipes:
- bus bandwidth is often limited, so you often do not get N*disk_bandwidth
- application bandwidth is often limited, so you do not NEED N*disk_bandwidth
- for EACH read from a striped LV, you need to seek all disks to that
spot, which means you often have to wait for the MAXIMUM seek time
of all disks. With non-striped LVs, you can do parallel read/write
from different parts of the each disk (e.g. /usr, /home, /var, swap)
- if you have problems with ANY disk, you have lost ALL of your data
Cheers, Andreas
--
Andreas Dilger \ "If a man ate a pound of pasta and a pound of antipasto,
\ would they cancel out, leaving him still hungry?"
http://www-mddsp.enel.ucalgary.ca/People/adilger/ -- Dogbert
next prev parent reply other threads:[~2001-09-14 15:52 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2001-09-14 9:32 [linux-lvm] Verständnis von Stripes bei LVM Pawlowski Julian
2001-09-14 10:03 ` [linux-lvm] Verst?ndnis " Jan-Benedict Glaw
2001-09-14 10:13 ` [linux-lvm] Re: [linux-lvm] Verständnis " Goetz Bock
2001-09-14 10:43 ` svetljo
2001-09-14 15:52 ` Andreas Dilger [this message]
2001-09-17 8:54 ` [linux-lvm] " Goetz Bock
2001-09-17 9:21 ` svetljo
2001-09-14 11:37 ` Heinz J . Mauelshagen
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20010914095230.S1541@turbolinux.com \
--to=adilger@turbolabs.com \
--cc=linux-lvm@sistina.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox