public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Christian <christiand59@web.de>
To: linux-kernel@vger.kernel.org
Subject: Device mapper/LVM throughput problem
Date: Fri, 20 Oct 2006 18:41:31 +0200	[thread overview]
Message-ID: <200610201841.31435.christiand59@web.de> (raw)

Hello lkml!

I've just tested the disk throughput on my machine using hdparm and found some 
puzzling results. I have  a two disk nvidia fakeraid 0 configuration using 
dmraid and lvm.

When testing the single disks, I get good results just like when testing the 
dmraid mapping. When I test the throughput on a raid partition or the LVM 
Volumes I get very bad performance. Has anybody seen something like this 
before?


single disks: (good results)

user@ubuntu:~$ sudo hdparm -Tt /dev/sd[ab]

/dev/sda:
 Timing cached reads:   3384 MB in  2.00 seconds = 1693.17 MB/sec
 Timing buffered disk reads:  210 MB in  3.02 seconds =  69.56 MB/sec

/dev/sdb:
 Timing cached reads:   4096 MB in  2.00 seconds = 2049.10 MB/sec
 Timing buffered disk reads:  208 MB in  3.01 seconds =  69.01 MB/sec


dmraid mapping: (good results)

user@ubuntu:~$ sudo hdparm -Tt /dev/mapper/nvidia_cbcaiacc

/dev/mapper/nvidia_cbcaiacc:
 Timing cached reads:   3380 MB in  2.00 seconds = 1690.50 MB/sec
HDIO_DRIVE_CMD(null) (wait for flush complete) failed: Inappropriate ioctl for 
device
 Timing buffered disk reads:  392 MB in  3.01 seconds = 130.05 MB/sec
HDIO_DRIVE_CMD(null) (wait for flush complete) failed: Inappropriate ioctl for 
device


third (or any other) partition on raid (bad throughput):

user@ubuntu:~$ sudo hdparm -Tt /dev/mapper/nvidia_cbcaiacc3

/dev/mapper/nvidia_cbcaiacc3:
 Timing cached reads:   3272 MB in  2.00 seconds = 1636.72 MB/sec
HDIO_DRIVE_CMD(null) (wait for flush complete) failed: Inappropriate ioctl for 
device
 Timing buffered disk reads:  158 MB in  3.00 seconds =  52.63 MB/sec
HDIO_DRIVE_CMD(null) (wait for flush complete) failed: Inappropriate ioctl for 
device


LVM Volume (even worse)

user@ubuntu:~$ sudo hdparm -Tt /dev/VolGroup00/LogVol00

/dev/VolGroup00/LogVol00:
 Timing cached reads:   3636 MB in  2.00 seconds = 1818.96 MB/sec
HDIO_DRIVE_CMD(null) (wait for flush complete) failed: Inappropriate ioctl for 
device
 Timing buffered disk reads:  114 MB in  3.01 seconds =  37.82 MB/sec
HDIO_DRIVE_CMD(null) (wait for flush complete) failed: Inappropriate ioctl for 
device

So on LogVol00 residing on a two disk striped set I get only 38MB/s 
throughput! I would anticipate something in the range of 130-140 MB/s


user@ubuntu:~$ uname -a
Linux ubuntu.localnet 2.6.18-rc7 #2 SMP Wed Sep 13 11:28:41 CEST 2006 x86_64 
GNU/Linux

-Christian

                 reply	other threads:[~2006-10-20 16:41 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=200610201841.31435.christiand59@web.de \
    --to=christiand59@web.de \
    --cc=linux-kernel@vger.kernel.org \
    /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