<?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: Block size in Operating System - HP-UX</title>
    <link>https://community.hpe.com/t5/operating-system-hp-ux/block-size/m-p/2644683#M646567</link>
    <description>vxfs/JFS is an *extent*-based filesystem, so the term "fragment" is somewhat incorrect, because that is an *HFS* term.&lt;BR /&gt;&lt;BR /&gt;With vxfs/JFS you can only specify the *block* size. A block size of 1KB is normally fine, because, for larger files, vxfs/JFS will use an extent, which is normally a, large, multiple of the block size.&lt;BR /&gt;&lt;BR /&gt;So unless the filesyetm is (too) fragmented because (at one time or another) it was too full, and you can not unfragment it, vxfs/JFS block size may be an use.&lt;BR /&gt;&lt;BR /&gt;Very, very likely the 8KB recommendation is, implicitly, an *HFS* recommendation.</description>
    <pubDate>Tue, 15 Jan 2002 10:11:08 GMT</pubDate>
    <dc:creator>Frank Slootweg</dc:creator>
    <dc:date>2002-01-15T10:11:08Z</dc:date>
    <item>
      <title>Block size</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/block-size/m-p/2644680#M646564</link>
      <description>We have a hp-ux 11.11 system. The filesystems were all created with a 1k fragment size(vxfs).&lt;BR /&gt;   I want to change the size to 8k so that my oracle database does not have any performance issues.My sysadmin says that there will not be any performance improvement as the bottleneck would be still there at the fiber optic level(ours is a SAN environment) HP L3000 connected to a XP512 .He says the fragment size there is 1K and that would still be the bottleneck.&lt;BR /&gt;   &lt;BR /&gt;Any thoughts?&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;Murali</description>
      <pubDate>Mon, 14 Jan 2002 20:58:10 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/block-size/m-p/2644680#M646564</guid>
      <dc:creator>muralip</dc:creator>
      <dc:date>2002-01-14T20:58:10Z</dc:date>
    </item>
    <item>
      <title>Re: Block size</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/block-size/m-p/2644681#M646565</link>
      <description>Murali,&lt;BR /&gt;&lt;BR /&gt;I agree with you that the block size should be 8k for oracle (AT LEAST 8k, if not LARGER!!!!) It's a tough call. What are you blcoked on, IO calls or IO waits?&lt;BR /&gt;&lt;BR /&gt;live free or die&lt;BR /&gt;harry</description>
      <pubDate>Mon, 14 Jan 2002 21:08:20 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/block-size/m-p/2644681#M646565</guid>
      <dc:creator>harry d brown jr</dc:creator>
      <dc:date>2002-01-14T21:08:20Z</dc:date>
    </item>
    <item>
      <title>Re: Block size</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/block-size/m-p/2644682#M646566</link>
      <description>&lt;A href="http://www.docs.hp.com//hpux/onlinedocs/os/11.0/tuningwp.html#diskperf" target="_blank"&gt;http://www.docs.hp.com//hpux/onlinedocs/os/11.0/tuningwp.html#diskperf&lt;/A&gt;&lt;BR /&gt;&lt;BR /&gt;live free or die&lt;BR /&gt;harry</description>
      <pubDate>Mon, 14 Jan 2002 21:16:33 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/block-size/m-p/2644682#M646566</guid>
      <dc:creator>harry d brown jr</dc:creator>
      <dc:date>2002-01-14T21:16:33Z</dc:date>
    </item>
    <item>
      <title>Re: Block size</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/block-size/m-p/2644683#M646567</link>
      <description>vxfs/JFS is an *extent*-based filesystem, so the term "fragment" is somewhat incorrect, because that is an *HFS* term.&lt;BR /&gt;&lt;BR /&gt;With vxfs/JFS you can only specify the *block* size. A block size of 1KB is normally fine, because, for larger files, vxfs/JFS will use an extent, which is normally a, large, multiple of the block size.&lt;BR /&gt;&lt;BR /&gt;So unless the filesyetm is (too) fragmented because (at one time or another) it was too full, and you can not unfragment it, vxfs/JFS block size may be an use.&lt;BR /&gt;&lt;BR /&gt;Very, very likely the 8KB recommendation is, implicitly, an *HFS* recommendation.</description>
      <pubDate>Tue, 15 Jan 2002 10:11:08 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/block-size/m-p/2644683#M646567</guid>
      <dc:creator>Frank Slootweg</dc:creator>
      <dc:date>2002-01-15T10:11:08Z</dc:date>
    </item>
    <item>
      <title>Re: Block size</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/block-size/m-p/2644684#M646568</link>
      <description>Hi&lt;BR /&gt;&lt;BR /&gt;vxfs is 1k fragment &amp;amp; 8k block!&lt;BR /&gt;I believe these are set in stone.&lt;BR /&gt;&lt;BR /&gt;However, if you use raw disk you can do what you like (within the limits of the technology).&lt;BR /&gt;&lt;BR /&gt;We have an fc60 &amp;amp; have LUNS set up with a 64k stripe ...  There are multiple ways of achieving this stripe&lt;BR /&gt; o Using LVM striping (lvcreate -i &lt;N_DISK&gt; -I &lt;STRIPE size=""&gt;&lt;BR /&gt; o Using the HW striping.&lt;BR /&gt;It is all accessed as RAW LVM logical volumes.  &lt;BR /&gt;&lt;BR /&gt;A major difference here though is that we use Informix as the database but, I'm sure Oracle can do the same.&lt;BR /&gt;&lt;BR /&gt;Tim&lt;BR /&gt;&lt;/STRIPE&gt;&lt;/N_DISK&gt;</description>
      <pubDate>Tue, 15 Jan 2002 10:32:36 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/block-size/m-p/2644684#M646568</guid>
      <dc:creator>Tim D Fulford</dc:creator>
      <dc:date>2002-01-15T10:32:36Z</dc:date>
    </item>
    <item>
      <title>Re: Block size</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/block-size/m-p/2644685#M646569</link>
      <description>See also the current "block size default" thread, talking about the same issues:&lt;BR /&gt;&lt;BR /&gt;&lt;A href="http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x59d828e43106d6118ff40090279cd0f9,00.html" target="_blank"&gt;http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x59d828e43106d6118ff40090279cd0f9,00.html&lt;/A&gt;&lt;BR /&gt;&lt;BR /&gt;</description>
      <pubDate>Tue, 15 Jan 2002 10:41:21 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/block-size/m-p/2644685#M646569</guid>
      <dc:creator>Frank Slootweg</dc:creator>
      <dc:date>2002-01-15T10:41:21Z</dc:date>
    </item>
  </channel>
</rss>

