109 lines
6.8 KiB
HTML
109 lines
6.8 KiB
HTML
|
<?xml version="1.0" encoding="UTF-8"?>
|
||
|
|
<!DOCTYPE html
|
||
|
|
PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
|
||
|
|
<html lang="en-us" xml:lang="en-us">
|
||
|
|
<head>
|
||
|
|
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
|
||
|
|
<meta name="security" content="public" />
|
||
|
|
<meta name="Robots" content="index,follow" />
|
||
|
|
<meta http-equiv="PICS-Label" content='(PICS-1.1 "http://www.icra.org/ratingsv02.html" l gen true r (cz 1 lz 1 nz 1 oz 1 vz 1) "http://www.rsac.org/ratingsv01.html" l gen true r (n 0 s 0 v 0 l 0) "http://www.classify.org/safesurf/" l gen true r (SS~~000 1))' />
|
||
|
|
<meta name="DC.Type" content="concept" />
|
||
|
|
<meta name="DC.Title" content="CPYOPT Performance" />
|
||
|
|
<meta name="abstract" content="This topic discusses the fact that performance is a complex subject with many interdependent components." />
|
||
|
|
<meta name="description" content="This topic discusses the fact that performance is a complex subject with many interdependent components." />
|
||
|
|
<meta name="DC.Relation" scheme="URI" content="rzam4cpyopt.htm" />
|
||
|
|
<meta name="copyright" content="(C) Copyright IBM Corporation 2004, 2006" />
|
||
|
|
<meta name="DC.Rights.Owner" content="(C) Copyright IBM Corporation 2004, 2006" />
|
||
|
|
<meta name="DC.Format" content="XHTML" />
|
||
|
|
<meta name="DC.Identifier" content="rzam4cpyoptperformance" />
|
||
|
|
<meta name="DC.Language" content="en-us" />
|
||
|
|
<!-- All rights reserved. Licensed Materials Property of IBM -->
|
||
|
|
<!-- US Government Users Restricted Rights -->
|
||
|
|
<!-- Use, duplication or disclosure restricted by -->
|
||
|
|
<!-- GSA ADP Schedule Contract with IBM Corp. -->
|
||
|
|
<link rel="stylesheet" type="text/css" href="./ibmdita.css" />
|
||
|
|
<link rel="stylesheet" type="text/css" href="./ic.css" />
|
||
|
|
<title>CPYOPT Performance</title>
|
||
|
|
</head>
|
||
|
|
<body id="rzam4cpyoptperformance"><a name="rzam4cpyoptperformance"><!-- --></a>
|
||
|
|
<!-- Java sync-link --><script language="Javascript" src="../rzahg/synch.js" type="text/javascript"></script>
|
||
|
|
<h1 class="topictitle1">CPYOPT Performance</h1>
|
||
|
|
<div><p>This topic discusses the fact that performance is a complex subject
|
||
|
|
with many interdependent components.</p>
|
||
|
|
<p>By changing one component, you may adversely affect another. This interdependence
|
||
|
|
and the other factors that affect performance, prohibits providing a formula
|
||
|
|
for computing the time required to copy a given number of files. It is simply
|
||
|
|
a guideline to help you estimate how long your CPYOPT might take.</p>
|
||
|
|
<p>You may need to estimate how long it takes to copy an entire optical volume
|
||
|
|
when using the CPYOPT command. You can start by copying a single directory
|
||
|
|
that contains a known number of average sized files. Then take the difference
|
||
|
|
between the ending time and starting time to determine an elapsed time. Take
|
||
|
|
the elapsed time and divide by the number of files that are copied to figure
|
||
|
|
the average seconds per file. You can use this number as a basis to determine
|
||
|
|
the amount of time that is required to copy the entire volume of average size
|
||
|
|
files.</p>
|
||
|
|
<div class="p">To maximize copy performance, use the following set of guidelines as a
|
||
|
|
starting point:<ul><li> Having too few directories with too many files can affect performance.
|
||
|
|
Having too many directories with too few files can affect performance also.
|
||
|
|
Try to keep the number of files in a directory to less than 6000 files.</li>
|
||
|
|
<li>Consider performance when determining file size.</li>
|
||
|
|
<li>Avoid use of extended attributes on files. When a file has extended attributes,
|
||
|
|
they are storied separately from the data. When copying the data, the must
|
||
|
|
copy the attributes also. It is similar to copying a second file for each
|
||
|
|
user file copied.</li>
|
||
|
|
<li>Keep the source and target volumes in the same library.</li>
|
||
|
|
<li>Avoid copying to the opposite side of an optical cartridge.</li>
|
||
|
|
<li>If the copy processes can have dedicated use of the optical drives, use
|
||
|
|
the COPYTYPE *IOP parameter on the CPYOPT command.</li>
|
||
|
|
<li>Avoid optical drive contention from other optical processes.</li>
|
||
|
|
<li>Dedicate the use of two optical drives for copy activity.</li>
|
||
|
|
</ul>
|
||
|
|
</div>
|
||
|
|
</div>
|
||
|
|
<div>
|
||
|
|
<div class="familylinks">
|
||
|
|
<div class="parentlink"><strong>Parent topic:</strong> <a href="rzam4cpyopt.htm" title="This topic discusses the Copy Optical (CPYOPT) CL command and some of the optional parameters which you can specify.">Copy optical (CPYOPT)</a></div>
|
||
|
|
</div>
|
||
|
|
</div><div class="nested1" xml:lang="en-us" id="rzam4drivecontention"><a name="rzam4drivecontention"><!-- --></a><h2 class="topictitle2">Drive Contention</h2>
|
||
|
|
<div><div class="p">The following conditions can severely affect copy performance:<ul><li>Having only one drive available for use.</li>
|
||
|
|
<li>Copying from one side of an optical cartridge to the opposite side.</li>
|
||
|
|
<li>Having other optical processes that are running that attempt to use the
|
||
|
|
available drives.</li>
|
||
|
|
<li>A large number of files on source volume.</li>
|
||
|
|
</ul>
|
||
|
|
</div>
|
||
|
|
<p>To remove a volume, to store it in a slot, to retrieve a new volume, and
|
||
|
|
to mount it requires from 8-15 seconds. You should try to do your copy requests
|
||
|
|
when the process can have dedicated use of the optical drives. </p>
|
||
|
|
<div class="p">You should not try to copy a large number of files from one side of an
|
||
|
|
optical cartridge to the other side. Optical drives have only one read/write
|
||
|
|
head. The following conditions occur when copying from one side of an optical
|
||
|
|
cartridge to another:<ul><li>The system mounts the source volume.</li>
|
||
|
|
<li>A limited number of files that are to be copied are read and stored on i5/OS™ temporary
|
||
|
|
storage.</li>
|
||
|
|
<li>The source volume is removed, and the system mounts the target volume
|
||
|
|
by turning over the optical cartridge.</li>
|
||
|
|
<li>Files are read from i5/OS temporary storage and written to the target volume.</li>
|
||
|
|
<li>If there are more files to copy, the system removes target volume and
|
||
|
|
mounts the source volume again by turning over the optical cartridge.</li>
|
||
|
|
<li>The system repeats this process until it copies all the files. You may
|
||
|
|
need to turn the optical cartridge over many times to copy all the files.</li>
|
||
|
|
</ul>
|
||
|
|
Copy performance is always better when both the source and target volumes
|
||
|
|
are in the same library and when setting the COPYTYPE parameter for that library
|
||
|
|
controller to *IOP. Two conditions exist that require n extra processing step.
|
||
|
|
The first one is that the source and target volumes are in different libraries.
|
||
|
|
The second involves setting the COPYTYPE parameter to *SYSTEM and having the
|
||
|
|
volumes exist in the same library. This extra step requires moving the files
|
||
|
|
you want to copy to temporary storage on the iSeries™ server before writing them to
|
||
|
|
the target volume. You do not need to use temporary storage when the system
|
||
|
|
meets both of the following conditions:<ul><li>Both optical volumes are in the same library.</li>
|
||
|
|
<li>You set the COPYTYPE parameter on the CPYOPT command to *IOP.</li>
|
||
|
|
</ul>
|
||
|
|
</div>
|
||
|
|
<p>This allows for the direct transfer of data between the two optical drives.</p>
|
||
|
|
</div>
|
||
|
|
</div>
|
||
|
|
|
||
|
|
</body>
|
||
|
|
</html>
|