<?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 utilization in Operating System - HP-UX</title>
    <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936842#M634799</link>
    <description>There's nothing you can do to really improve anything in the kernel. There are several distinct categories of files in Oracle and each category should have a separate hpysical disk (separate lvols on the same disk will do nothing). The Oracle data areas are randomly accessed and should be on a separate disk (or disks). Archive/rollback logs are intense serial activities and must be on a separate disk.&lt;BR /&gt;&lt;BR /&gt;Striping might help a bit but you just don't have enough disks to make much of a difference.</description>
    <pubDate>Sun, 30 Mar 2003 01:11:48 GMT</pubDate>
    <dc:creator>Bill Hassell</dc:creator>
    <dc:date>2003-03-30T01:11:48Z</dc:date>
    <item>
      <title>Disk utilization</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936836#M634793</link>
      <description>Hi,&lt;BR /&gt;&lt;BR /&gt;We are running an application on HP-UX L3000 server. The disk layout is explained below.&lt;BR /&gt;Three 36.4GB hardidsks.&lt;BR /&gt;&lt;BR /&gt;First disk used for the oracle indexes and root partition (18.2 GB each)&lt;BR /&gt;&lt;BR /&gt;Second disk is used for oracle tables and Application logs. (18.2 GB each)&lt;BR /&gt;&lt;BR /&gt;Third disk is used for Oracle logs (18.2 GB) and remaining 18.2 GB is free in the third disk.&lt;BR /&gt;&lt;BR /&gt;On the second disk almost the disk IO rate is 100% always, is anybody have any idea why this is always 100% do we have to tune any kernal parameters. &lt;BR /&gt;Presently we are trying to test with following options&lt;BR /&gt;&lt;BR /&gt;1. Moving the Application logs from  second disk to 3 disk. (i.e. in third disk we have 18.2 GB free space, we are planning to create another lvol in the volume group of third disk and move all the logs from disk 2 lvol to newly created lvol)&lt;BR /&gt;&lt;BR /&gt;2. Adding separate disk for application logs and moving all the logs data from second disk to new disk.&lt;BR /&gt;&lt;BR /&gt;Is any options mentioned above will solve the disk I/O problem? or is there any body have any ideas please..&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;</description>
      <pubDate>Thu, 27 Mar 2003 06:44:16 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936836#M634793</guid>
      <dc:creator>Kourla</dc:creator>
      <dc:date>2003-03-27T06:44:16Z</dc:date>
    </item>
    <item>
      <title>Re: Disk utilization</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936837#M634794</link>
      <description>yes,you are right.&lt;BR /&gt;&lt;BR /&gt;The Disk is going 100% as the Oracle processes go on writing with the logfiles and the datafiles.&lt;BR /&gt;&lt;BR /&gt;Distribution the datafiles and logfiles should improve the situation.&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;I do not think any kernel parameter can help you on this.&lt;BR /&gt;&lt;BR /&gt;If you have Glance then you can get a clear picture on the process which has a high I/O rate on the disk.&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;Thanks</description>
      <pubDate>Thu, 27 Mar 2003 06:53:27 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936837#M634794</guid>
      <dc:creator>T G Manikandan</dc:creator>
      <dc:date>2003-03-27T06:53:27Z</dc:date>
    </item>
    <item>
      <title>Re: Disk utilization</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936838#M634795</link>
      <description>Whilst you only have three disks you will have stacks of Disk IO. Your best bet is to add as many disks as possible to even out the load, particularly for logging. Moving some of the load from disk 2 to 3 will only be a temporary solution.</description>
      <pubDate>Thu, 27 Mar 2003 06:56:58 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936838#M634795</guid>
      <dc:creator>Michael Tully</dc:creator>
      <dc:date>2003-03-27T06:56:58Z</dc:date>
    </item>
    <item>
      <title>Re: Disk utilization</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936839#M634796</link>
      <description>Adding more disks and STRIPING will take care of your problem.  When you stripe you evenly distribute the I/O to multiple disks hence multiple mechanical devices and you can see signigicant performance increase.&lt;BR /&gt;George</description>
      <pubDate>Fri, 28 Mar 2003 08:02:13 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936839#M634796</guid>
      <dc:creator>George Petrides_1</dc:creator>
      <dc:date>2003-03-28T08:02:13Z</dc:date>
    </item>
    <item>
      <title>Re: Disk utilization</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936840#M634797</link>
      <description>You are limited in what you can do with only 3 disks.  As others have already said, your first step should be to get more disks and more evenly distribute the load.  I always have my operating system and only the operating system in volume group 0 (vg00).  Then create additional volume groups on other disks for the applications and data.  And, use striping.&lt;BR /&gt;Having said all that, if you can not get additional disks for this system and keep everything in 1 volume group, I would use Ignite and rebuild the system.  Create a make_tape_recovery tape and rebuild from that striping all of your logical volumes across the 3 disks.  &lt;BR /&gt;In an Oracle environment it can be difficult to pinpoint where the majority of your activity will be.  If it is for the most part static data and mostly reads happening, there will not be much activity in your logs.  So putting the logs on a seperate disk means that disk will not be busy but, your data disk(s) may be overworked.&lt;BR /&gt;If a lot of data is being added and or deleted, the data and log volumes will be active.  &lt;BR /&gt;If you have inadequate memory you may be swapping so the swap devices can be very active.  &lt;BR /&gt;With only 3 disks I would stripe across all of them and hope it all balances out somewhat.  More disks is really the answer.&lt;BR /&gt;</description>
      <pubDate>Fri, 28 Mar 2003 15:24:54 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936840#M634797</guid>
      <dc:creator>Dave Wherry</dc:creator>
      <dc:date>2003-03-28T15:24:54Z</dc:date>
    </item>
    <item>
      <title>Re: Disk utilization</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936841#M634798</link>
      <description>Hi&lt;BR /&gt;&lt;BR /&gt;  for this issue ,if you have free memory u can change some kernel parmeter to increase the i/o performance &lt;BR /&gt;nbuff 5&lt;BR /&gt;dbcmax 25&lt;BR /&gt;dbcmin 5&lt;BR /&gt;&lt;BR /&gt;this will give some performanc enchanment ,as it copies more data to memory ,so it doesnt aceess disk quite frequently&lt;BR /&gt;&lt;BR /&gt;Rizwan&lt;BR /&gt;</description>
      <pubDate>Sat, 29 Mar 2003 16:18:21 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936841#M634798</guid>
      <dc:creator>Rizwan Mohammed</dc:creator>
      <dc:date>2003-03-29T16:18:21Z</dc:date>
    </item>
    <item>
      <title>Re: Disk utilization</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936842#M634799</link>
      <description>There's nothing you can do to really improve anything in the kernel. There are several distinct categories of files in Oracle and each category should have a separate hpysical disk (separate lvols on the same disk will do nothing). The Oracle data areas are randomly accessed and should be on a separate disk (or disks). Archive/rollback logs are intense serial activities and must be on a separate disk.&lt;BR /&gt;&lt;BR /&gt;Striping might help a bit but you just don't have enough disks to make much of a difference.</description>
      <pubDate>Sun, 30 Mar 2003 01:11:48 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936842#M634799</guid>
      <dc:creator>Bill Hassell</dc:creator>
      <dc:date>2003-03-30T01:11:48Z</dc:date>
    </item>
    <item>
      <title>Re: Disk utilization</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936843#M634800</link>
      <description>Oracle does not recommend striping for optimal performance.&lt;BR /&gt;&lt;BR /&gt;Though it can help you out in your currenet situation, you'll get better performance using Raid 1/0 mirroring.&lt;BR /&gt;&lt;BR /&gt;lvextend -m 1 /dev/vg01/oradata /dev/dsk/c#t#d#&lt;BR /&gt;&lt;BR /&gt;The archive logs and such don't benefit as much from this setup.&lt;BR /&gt;&lt;BR /&gt;Striping will improve reliability, especially if you strip across more disks, but there is a disk performance penalty.&lt;BR /&gt;&lt;BR /&gt;SEP</description>
      <pubDate>Mon, 31 Mar 2003 04:09:31 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/disk-utilization/m-p/2936843#M634800</guid>
      <dc:creator>Steven E. Protter</dc:creator>
      <dc:date>2003-03-31T04:09:31Z</dc:date>
    </item>
  </channel>
</rss>

