<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic Re: IO testing and tuning in Operating System - HP-UX</title>
    <link>https://community.hpe.com/t5/operating-system-hp-ux/io-testing-and-tuning/m-p/5086510#M441629</link>
    <description>Hi Chris (and Happy New Year to you too):&lt;BR /&gt;&lt;BR /&gt;Creating I/O on disks can be as simple as using 'dd'.  For example, to read a disk in block sizes of 1024 characters having skipped 2048 blocks before starting and quit after reading 100 blocks, do:&lt;BR /&gt;&lt;BR /&gt;# dd if=/dev/rdsk/c0t6d0 of=/dev/null bs=1024 iseek=2048 count=100&lt;BR /&gt;&lt;BR /&gt;Notice the use of the raw device file.  This means that you don't involve the Unix buffer cache in your reads.&lt;BR /&gt;&lt;BR /&gt;As for writing, you could use '/dev/zero' as your input ('if') file, BUT MAKE VERY SURE that your output disk is not one with any viable data, since you will be smearing zeros all over it!&lt;BR /&gt;&lt;BR /&gt;Clearly, a script with a number of 'dd' processes can be quickly hacked together to build up so I/O.  As always, see the manpages for 'dd(1)' for more information.&lt;BR /&gt;&lt;BR /&gt;Regards!&lt;BR /&gt;&lt;BR /&gt;...JRF...</description>
    <pubDate>Thu, 03 Jan 2008 17:40:59 GMT</pubDate>
    <dc:creator>James R. Ferguson</dc:creator>
    <dc:date>2008-01-03T17:40:59Z</dc:date>
    <item>
      <title>IO testing and tuning</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/io-testing-and-tuning/m-p/5086507#M441626</link>
      <description>Hi all,&lt;BR /&gt;&lt;BR /&gt;hope everyone had a good christmas and newyear .... I worked for most of it but will be getting lots of $$$$$ (happy days) ...&lt;BR /&gt;&lt;BR /&gt;Let me get back to the question in hand:&lt;BR /&gt;&lt;BR /&gt;I want to play around with system IO tuning and understand some of the parameters and the effects of changing.  I have a couple crash and burn servers running hpux and AIX and I was hoping someone had a program or could point me to some method where I can increae the IO on the disks to 100% and create bottlenecks etc so I can determine different tuning methods.&lt;BR /&gt;&lt;BR /&gt;does anyone have such a program?&lt;BR /&gt;&lt;BR /&gt;many thanks&lt;BR /&gt;&lt;BR /&gt;Chris</description>
      <pubDate>Thu, 03 Jan 2008 17:29:34 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/io-testing-and-tuning/m-p/5086507#M441626</guid>
      <dc:creator>lawrenzo_1</dc:creator>
      <dc:date>2008-01-03T17:29:34Z</dc:date>
    </item>
    <item>
      <title>Re: IO testing and tuning</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/io-testing-and-tuning/m-p/5086508#M441627</link>
      <description>Bonsoir Chris,&lt;BR /&gt;&lt;BR /&gt;Maybe IOZONE is your answer. Take a look at &lt;A href="http://www.iozone.org/" target="_blank"&gt;http://www.iozone.org/&lt;/A&gt;&lt;BR /&gt;&lt;BR /&gt;Regards&lt;BR /&gt;&lt;BR /&gt;Eric</description>
      <pubDate>Thu, 03 Jan 2008 17:38:27 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/io-testing-and-tuning/m-p/5086508#M441627</guid>
      <dc:creator>Eric SAUBIGNAC</dc:creator>
      <dc:date>2008-01-03T17:38:27Z</dc:date>
    </item>
    <item>
      <title>Re: IO testing and tuning</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/io-testing-and-tuning/m-p/5086509#M441628</link>
      <description>hi chris ;&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;a) Bonnie+&lt;BR /&gt;&lt;BR /&gt;&lt;A href="http://www.coker.com.au/bonnie++/" target="_blank"&gt;http://www.coker.com.au/bonnie++/&lt;/A&gt;&lt;BR /&gt;&lt;BR /&gt;b) IOzone&lt;BR /&gt;&lt;BR /&gt;&lt;A href="http://www.iozone.org/" target="_blank"&gt;http://www.iozone.org/&lt;/A&gt;&lt;BR /&gt;&lt;BR /&gt;c) IOstone&lt;BR /&gt;&lt;BR /&gt;&lt;A href="http://www.acnc.com/benchmarks/iostone.zip" target="_blank"&gt;http://www.acnc.com/benchmarks/iostone.zip&lt;/A&gt;&lt;BR /&gt;&lt;BR /&gt;d) IObench&lt;BR /&gt;&lt;BR /&gt;&lt;A href="http://www.acnc.com/benchmarks/iobench.tar.Z" target="_blank"&gt;http://www.acnc.com/benchmarks/iobench.tar.Z&lt;/A&gt;&lt;BR /&gt;&lt;BR /&gt;e) Bonnie&lt;BR /&gt;&lt;BR /&gt;&lt;A href="http://www.acnc.com/benchmarks/bonnie.tar.gz" target="_blank"&gt;http://www.acnc.com/benchmarks/bonnie.tar.gz&lt;/A&gt;&lt;BR /&gt;&lt;BR /&gt;f) nbench&lt;BR /&gt;&lt;BR /&gt;&lt;A href="http://www.tux.org/~mayer/linux/bmark.html" target="_blank"&gt;http://www.tux.org/~mayer/linux/bmark.html&lt;/A&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;from the url &lt;BR /&gt;&lt;BR /&gt;&lt;A href="http://forums12.itrc.hp.com/service/forums/questionanswer.do?admit=109447627+1199381950289+28353475&amp;amp;threadId=1187415" target="_blank"&gt;http://forums12.itrc.hp.com/service/forums/questionanswer.do?admit=109447627+1199381950289+28353475&amp;amp;threadId=1187415&lt;/A&gt;&lt;BR /&gt;&lt;BR /&gt;Hasan</description>
      <pubDate>Thu, 03 Jan 2008 17:40:19 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/io-testing-and-tuning/m-p/5086509#M441628</guid>
      <dc:creator>Hasan  Atasoy</dc:creator>
      <dc:date>2008-01-03T17:40:19Z</dc:date>
    </item>
    <item>
      <title>Re: IO testing and tuning</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/io-testing-and-tuning/m-p/5086510#M441629</link>
      <description>Hi Chris (and Happy New Year to you too):&lt;BR /&gt;&lt;BR /&gt;Creating I/O on disks can be as simple as using 'dd'.  For example, to read a disk in block sizes of 1024 characters having skipped 2048 blocks before starting and quit after reading 100 blocks, do:&lt;BR /&gt;&lt;BR /&gt;# dd if=/dev/rdsk/c0t6d0 of=/dev/null bs=1024 iseek=2048 count=100&lt;BR /&gt;&lt;BR /&gt;Notice the use of the raw device file.  This means that you don't involve the Unix buffer cache in your reads.&lt;BR /&gt;&lt;BR /&gt;As for writing, you could use '/dev/zero' as your input ('if') file, BUT MAKE VERY SURE that your output disk is not one with any viable data, since you will be smearing zeros all over it!&lt;BR /&gt;&lt;BR /&gt;Clearly, a script with a number of 'dd' processes can be quickly hacked together to build up so I/O.  As always, see the manpages for 'dd(1)' for more information.&lt;BR /&gt;&lt;BR /&gt;Regards!&lt;BR /&gt;&lt;BR /&gt;...JRF...</description>
      <pubDate>Thu, 03 Jan 2008 17:40:59 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/io-testing-and-tuning/m-p/5086510#M441629</guid>
      <dc:creator>James R. Ferguson</dc:creator>
      <dc:date>2008-01-03T17:40:59Z</dc:date>
    </item>
    <item>
      <title>Re: IO testing and tuning</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/io-testing-and-tuning/m-p/5086511#M441630</link>
      <description>I have been playing round with IO test as well. I have used IOZONE, which is better than dd because you can create a lot o threads. And the measuring is build in.&lt;BR /&gt;&lt;BR /&gt;Remember to play around with Block Size and the filesystem cache (dbc_min_pct and dbc_max_pct in 11i until v3 then it is fs something).</description>
      <pubDate>Thu, 03 Jan 2008 21:27:47 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/io-testing-and-tuning/m-p/5086511#M441630</guid>
      <dc:creator>Jannik</dc:creator>
      <dc:date>2008-01-03T21:27:47Z</dc:date>
    </item>
    <item>
      <title>Re: IO testing and tuning</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/io-testing-and-tuning/m-p/5086512#M441631</link>
      <description>Thats great,&lt;BR /&gt;&lt;BR /&gt;I have been running the dd command to test times and activity so thanks James.&lt;BR /&gt;&lt;BR /&gt;and also thanks for the suggested programs.&lt;BR /&gt;&lt;BR /&gt;Chris.</description>
      <pubDate>Mon, 07 Jan 2008 13:11:08 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/io-testing-and-tuning/m-p/5086512#M441631</guid>
      <dc:creator>lawrenzo_1</dc:creator>
      <dc:date>2008-01-07T13:11:08Z</dc:date>
    </item>
  </channel>
</rss>

