LinuxQuestions.org
Share your knowledge at the LQ Wiki.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Server
User Name
Password
Linux - Server This forum is for the discussion of Linux Software used in a server related context.

Notices

Reply
 
Search this Thread
Old 01-25-2012, 02:46 AM   #1
saavik
Member
 
Registered: Nov 2001
Location: NRW, Germany
Distribution: SLES11 / FC20/ OES / CentOS
Posts: 601

Rep: Reputation: 32
hdparm / hdd performancetesting


Hello!

I have a c7000 HP Blade Server with several ESX servers.

On that I attached a MDS600 with 70 HDD`s (SAS 15k).

As we have some SQL-Server timeouts now, I was tooled that this is because we have the HDDs are to slow.

Question:

How can I test the throughput of the HDDs (attached via iSCSI) without the cache.

I tested:
Code:
time dd if=/dev/zero of=/tmp/file bs=64k count=10000
time dd if=/tmp/file of=/dev/null bs=64k count=10000
time dd if=/dev/zero of=/tmp/file2 bs=128k count=5000
time dd if=/tmp/file2 of=/dev/null bs=128k count=5000
time dd if=/dev/zero of=/tmp/file3 bs=256k count=2500
time dd if=/tmp/file of=/dev/null bs=64k count=10000
time dd if=/tmp/file2 of=/dev/null bs=128k count=5000
time dd if=/tmp/file3 of=/dev/null bs=128k count=2500


hdparm -tT /dev/sda2
Would that be the appropriate performance test ?
 
Old 01-25-2012, 03:03 AM   #2
EricTRA
Guru
 
Registered: May 2009
Location: Gibraltar, Gibraltar
Distribution: Fedora 20 with Awesome WM
Posts: 6,805
Blog Entries: 1

Rep: Reputation: 1291Reputation: 1291Reputation: 1291Reputation: 1291Reputation: 1291Reputation: 1291Reputation: 1291Reputation: 1291Reputation: 1291
Hello,

You indicated 'without the cache', so reading from the man page of hdparm:
Quote:
-t Perform timings of device reads for benchmark and comparison purposes. For meaningful results, this operation should be repeated 2-3 times on an
otherwise inactive system (no other active processes) with at least a couple of megabytes of free memory. This displays the speed of reading
through the buffer cache to the disk without any prior caching of data. This measurement is an indication of how fast the drive can sustain sequen‐
tial data reads under Linux, without any filesystem overhead. To ensure accurate measurements, the buffer cache is flushed during the processing of
-t using the BLKFLSBUF ioctl.

-T Perform timings of cache reads for benchmark and comparison purposes. For meaningful results, this operation should be repeated 2-3 times on an
otherwise inactive system (no other active processes) with at least a couple of megabytes of free memory. This displays the speed of reading
directly from the Linux buffer cache without disk access. This measurement is essentially an indication of the throughput of the processor, cache,
and memory of the system under test.
Which indicates the use of -T as being used to test cache without disk access, I'd suppress that one and stick with -t for the device access.

Also, as indicated in the man page, run the command two or three times and calculate the average speed from the results.

Another tool you could use is iostat (man iostat, install sysstat package if you don't have the command) and/or bonnie++ as a benchmarking program.

Kind regards,

Eric
 
1 members found this post helpful.
Old 01-25-2012, 04:59 AM   #3
saavik
Member
 
Registered: Nov 2001
Location: NRW, Germany
Distribution: SLES11 / FC20/ OES / CentOS
Posts: 601

Original Poster
Rep: Reputation: 32
Thanks, that helped a lot!
 
Old 01-25-2012, 07:47 AM   #4
EricTRA
Guru
 
Registered: May 2009
Location: Gibraltar, Gibraltar
Distribution: Fedora 20 with Awesome WM
Posts: 6,805
Blog Entries: 1

Rep: Reputation: 1291Reputation: 1291Reputation: 1291Reputation: 1291Reputation: 1291Reputation: 1291Reputation: 1291Reputation: 1291Reputation: 1291
Hi,

You're welcome!

Kind regards,

Eric
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Computing HDD capacity from hdparm wagaboy Linux - Hardware 3 06-04-2011 02:10 PM
[SOLVED] prevent HDD from starting after hdparm -Y gujedan Linux - Hardware 1 02-08-2011 06:58 PM
[SOLVED] Issue with hdd standby (hdparm vs. udev) semiRocket Slackware 4 08-31-2009 09:18 AM
HDD setting with hdparm powadha Debian 1 04-20-2006 04:40 PM
help with hdd and hdparm xcaliber Linux - Software 6 04-06-2006 02:50 PM


All times are GMT -5. The time now is 05:52 PM.

Main Menu
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
identi.ca: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration