<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<p>Yes, but<br>
If you increase your pool by adding a new vdev, your current data
are not auto-rebalanced. This will only happen over time with new
or modified data. <br>
<br>
If you want the best performance then, you must copy over current
data ex by renaming a filesystem, replicate it to the former name
and delete it then.<br>
<br>
Gea<br>
</p>
<br>
<div class="moz-cite-prefix">Am 23.06.2017 um 17:19 schrieb Artyom
Zhandarovsky:<br>
</div>
<blockquote type="cite"
cite="mid:CAOdMvyi=4p7k-4PPDe9ufN9Y=Z-KdHyvdzNq-z3xk4MrL+8jZA@mail.gmail.com">
<meta http-equiv="Context-Type" content="text/html; charset=UTF-8">
<div dir="ltr">So basically i need to add just more drives... ?</div>
<div class="gmail_extra"><br>
<div class="gmail_quote">2017-06-23 18:09 GMT+03:00 Guenther
Alka <span dir="ltr"><<a href="mailto:alka@hfg-gmuend.de"
target="_blank" moz-do-not-send="true">alka@hfg-gmuend.de</a>></span>:<br>
<blockquote class="gmail_quote">The fragmentation info does
not describe the fragmentation of the data on pool but the
fragmentation of the free space. A high fragmentation value
will result in high data fragmentation only when you write
or modify data.<br>
<br>
<a
href="https://utcc.utoronto.ca/%7Ecks/space/blog/solaris/ZFSZpoolFragmentationMeaning"
rel="noreferrer" target="_blank" moz-do-not-send="true">https://utcc.utoronto.ca/~cks/<wbr>space/blog/solaris/ZFSZpoolFra<wbr>gmentationMeaning</a><br>
So the best and only way to reduce data fragmentation is not
to fill up a pool say over 70-80%.<br>
<br>
You should also know that CopyOnWrite filesystems where a
complete datablock ex 128k is written newly even if you
change a "house" to a "mouse" in a textfile are more
vulnerable to fragmentation than older filesystems. This is
the price for the crash resitency where a power outage
during a write cannot lead to a corrupted filesystem like
with older filesystems where it can happen that the data is
modified "infile" while the according metadata update is not
happening. ZFS over-compensates this with its advanced
rambased read and write caches. A "defrag tool" is not
available for ZFS.<br>
<br>
Gea<span class=""><br>
<br>
<br>
Am 23.06.2017 um 16:13 schrieb Artyom Zhandarovsky:<br>
<blockquote class="gmail_quote">
there any way to decrease fragmentation of dr_tank ?<br>
</blockquote>
<br>
-- <br>
<br>
</span>
______________________________<wbr>_________________<br>
OmniOS-discuss mailing list<br>
<a href="mailto:OmniOS-discuss@lists.omniti.com"
target="_blank" moz-do-not-send="true">OmniOS-discuss@lists.omniti.co<wbr>m</a><br>
<a
href="http://lists.omniti.com/mailman/listinfo/omnios-discuss"
rel="noreferrer" target="_blank" moz-do-not-send="true">http://lists.omniti.com/mailma<wbr>n/listinfo/omnios-discuss</a><br>
</blockquote>
</div>
<br>
</div>
</blockquote>
<pre class="moz-signature" cols="72">-- </pre>
</body>
</html>