<?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: Disk Bottlenecks in Operating System - HP-UX</title>
    <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-bottlenecks/m-p/2421068#M98</link>
    <description>If have have sar running you can us the sar -d command to l list the disk drives. This will show you want disk drives are busy. Look at the %busy colum and the avwait and avserv colums.  %busy should never be over 50%.  And, avwait should not be greater then avserv.  &lt;BR /&gt;&lt;BR /&gt;The sar -d data is based on the disk drive, not the file system. You can use Glance to with the (v) option to isolate the problem to a logical volume If the disk having the problem has multiple logical volulmes, moving a busy lv to a new disk should help with disk bottlenecks.  &lt;BR /&gt;&lt;BR /&gt;Also, you could add a mirror to the volume group and mirror the offending lvol to the new disk. This will really help with reads, but if the disk bottleneck is mainly caused by writes, this wont really help much. &lt;BR /&gt;&lt;BR /&gt;If you only have one lvol on the busy disk, you could strip the lvol over several disks to help spread the load out. &lt;BR /&gt;&lt;BR /&gt;I hope this gives you some ideas.&lt;BR /&gt;&lt;BR /&gt;Norm Clary</description>
    <pubDate>Tue, 23 May 2000 20:36:43 GMT</pubDate>
    <dc:creator>Norm Clary</dc:creator>
    <dc:date>2000-05-23T20:36:43Z</dc:date>
    <item>
      <title>Disk Bottlenecks</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-bottlenecks/m-p/2421064#M94</link>
      <description>I was looking in glance plus at the disk performance.  It said our system was &lt;BR /&gt;receiving a bunch of disk bottlenecks.  Does anyone know what this means and &lt;BR /&gt;what I need to do to fix this.  Thanks in advance for you help.   -Tod</description>
      <pubDate>Thu, 06 Apr 2000 08:48:37 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/disk-bottlenecks/m-p/2421064#M94</guid>
      <dc:creator>Tod Isaacson</dc:creator>
      <dc:date>2000-04-06T08:48:37Z</dc:date>
    </item>
    <item>
      <title>Re: Disk Bottlenecks</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-bottlenecks/m-p/2421065#M95</link>
      <description>Disk bottlenecks happen when there are a lot of read/write requests to the same &lt;BR /&gt;physical disk.  One way to fix this is to identify the file systems or logical &lt;BR /&gt;volumes that are used most often and place them on separate physical disks, &lt;BR /&gt;thus spreading the number of reads/writes across multiple disks.  You can go &lt;BR /&gt;one step further and place them on separate IO channels if performance is still &lt;BR /&gt;suffering.&lt;BR /&gt;&lt;BR /&gt;Check IO by File Sys, IO by Disk, and IO by Logl Vol to get an idea of which &lt;BR /&gt;filesystems and/or disks are stressed.  You can also see which PIDs are the top &lt;BR /&gt;disk users.&lt;BR /&gt;&lt;BR /&gt;You may also want to check the Global Waits in glance to see just what &lt;BR /&gt;percentage of processes are waiting on Disk I/O.  All Glance is really saying &lt;BR /&gt;when it warns of a disk bottleneck is that there is a lot of activity on the &lt;BR /&gt;disks and you may run into a bottleneck.&lt;BR /&gt;&lt;BR /&gt;-Evan</description>
      <pubDate>Thu, 06 Apr 2000 09:23:06 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/disk-bottlenecks/m-p/2421065#M95</guid>
      <dc:creator>Evan Day_1</dc:creator>
      <dc:date>2000-04-06T09:23:06Z</dc:date>
    </item>
    <item>
      <title>Re: Disk Bottlenecks</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-bottlenecks/m-p/2421066#M96</link>
      <description>Tod,&lt;BR /&gt;&lt;BR /&gt;We are having the same issues here where I work. What I would like to know is your enviroment and what if any ERP systems or Databases. Maybe we could share information..</description>
      <pubDate>Tue, 23 May 2000 13:26:42 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/disk-bottlenecks/m-p/2421066#M96</guid>
      <dc:creator>Jeffrey Killian</dc:creator>
      <dc:date>2000-05-23T13:26:42Z</dc:date>
    </item>
    <item>
      <title>Re: Disk Bottlenecks</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-bottlenecks/m-p/2421067#M97</link>
      <description>Tod,&lt;BR /&gt;Sorry here is my email&lt;BR /&gt;&lt;BR /&gt;Dlabonte@ime.net</description>
      <pubDate>Tue, 23 May 2000 13:30:19 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/disk-bottlenecks/m-p/2421067#M97</guid>
      <dc:creator>Jeffrey Killian</dc:creator>
      <dc:date>2000-05-23T13:30:19Z</dc:date>
    </item>
    <item>
      <title>Re: Disk Bottlenecks</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-bottlenecks/m-p/2421068#M98</link>
      <description>If have have sar running you can us the sar -d command to l list the disk drives. This will show you want disk drives are busy. Look at the %busy colum and the avwait and avserv colums.  %busy should never be over 50%.  And, avwait should not be greater then avserv.  &lt;BR /&gt;&lt;BR /&gt;The sar -d data is based on the disk drive, not the file system. You can use Glance to with the (v) option to isolate the problem to a logical volume If the disk having the problem has multiple logical volulmes, moving a busy lv to a new disk should help with disk bottlenecks.  &lt;BR /&gt;&lt;BR /&gt;Also, you could add a mirror to the volume group and mirror the offending lvol to the new disk. This will really help with reads, but if the disk bottleneck is mainly caused by writes, this wont really help much. &lt;BR /&gt;&lt;BR /&gt;If you only have one lvol on the busy disk, you could strip the lvol over several disks to help spread the load out. &lt;BR /&gt;&lt;BR /&gt;I hope this gives you some ideas.&lt;BR /&gt;&lt;BR /&gt;Norm Clary</description>
      <pubDate>Tue, 23 May 2000 20:36:43 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/disk-bottlenecks/m-p/2421068#M98</guid>
      <dc:creator>Norm Clary</dc:creator>
      <dc:date>2000-05-23T20:36:43Z</dc:date>
    </item>
  </channel>
</rss>

