<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<style>
/* margin:(X):=T+B+R+L; (V H):V=T+B,H=R+L; (T H B):T,H=R+L, (T R B L) */
/*body { margin:.5cm .20cm;padding:.15cm .15cm;}*/
/* disable local font and span tags by default */
.l, body, .l p, .l table, .l tbody{
font: 14.3pt/100% Constantia, Verdana, "Lucida Bright",Serif;
font-size-adjust:.44;
background-color:#faf8ff; color:#105070;
-moz-tab-size:3; -o-tab-size:3; tab-size:3;
}
p { margin .5cm; text-indent:.5cm;}
p+p { margin-top.20cm;margin-bottom:.20cm}
.l small { font-size:85.18%;}
.l big { font-size:117.4%;}
.generic-font-families {
font-family:cursive,fantasy,sans-serif,serif,monospace;
}
.quoted { font-style:italic; font-size:90%;}
.l em { font-variant:small-caps }
.l h6 {font-size:78.615% }
.l h5 {font-size:100% }
.l h4 {font-size:127.202%}
.l h3 {font-size:161.803%}
.l h2 {font-size:205.817%}
.l h1 {font-size:261.803%}
.l h1,.l h2,.l h3,.l h4,.l h5 { font-size: inherit; font-weight:bold }
.l h5,.l h6 { font-size: inherit; font-variant:small-caps }
.l h5 {font: inherit; font-weight:800}
.l h6 {font: inherit; font-weight:700}
.l h1,.l h2,.l h3,.l h4,.l h5,.l h6 { margin:.5em }
.l blockquote {margin:.5cm .9cm;}
.l blockquote .l p{ margin: .20cm .25cm;}
.l blockquote .l p, blockquote .l p+blockquote .l p { text-indent:-.35cm; }
.l blockquote .l p+blockquote .l p { margin-top:.15cm;margin-bottom:.15cm}
.l pre, .l cite {
padding:1cm .9cm;
}
pre cite tt { font-style:oblique; background-color:#f0f4ff; color:#203040;
-moz-tab-size:3; -o-tab-size:3; tab-size:3;
font-family:lucida console,monospace;}
.l pre+pre {
font-inherit; font-style:oblique; background-color:#f0f4ff; color:#203040;
-moz-tab-size:3; -o-tab-size:3; tab-size:3;
margin:1ex .9cm}
address {
font inherit; font-style:oblique; font-family:Cambria;
}
.l address { font:inherit; margin:1em 2cm; background-color:#f8faff; }
.l address+address{ margin:0 1.5cm}
.l img { margin:.8cm }
.l q:before {content:open-quote }
.l q:after {content:close-quote }
.l a, a:link,a:focus,a:visited { text-decoration:underline }
a:link { color:#44BB33 }
a:focus { color: #22FF11 }
a:visited {color: #557722 }
.sig {
font oblique 3cm/1em "Lucida Handwriting",cursive;
text-indent:2cm;
margin: 3cm 8cm;
}
.sig:first-letter {
float:left;
font: italic 3in/1em "Lucida Calligraphy",cursive;
font-weight:200;
margin: -3cm -4mm inherit inherit;
}
#sig_fl {
float:left;
font:italic 3in/1em "Lucida Calligraphy";
font-weight:200;
padding:0;
margin-top: -3cm -4mm 0 0 ;
}
.MsoNormal, .MsoNormalTable{
padding:0; margin:0;
color:darkmagenta; background:honeydew;
font: oblique 14pt/95% Calibri,Verdana,Arial !important;
}
.MsoNormal span, .MsoNormalTable span{
font-family: inherit !important; font-size: inherit !important;
font-style: inherit !important; color: inherit !important;
}
span[font-family=Arial], span[font-family="Times New Roman"], font[face=Arial] ,font[face="Times New Roman"]{
font-family: inherit !important; font-size-adjust:inherit !important;
font-size: inherit !important; line-height: inherit !important;
color: inherit !important;
}
/* vim: sts=2 sw=2 ts=2 si sc ai
*/
</style>
</head>
<body>
<p> </p>
<br>
<br>
Mike Snitzer wrote:
<blockquote id="mid_20120725210434_GA20135_redhat_com"
cite="mid:20120725210434.GA20135@redhat.com" type="cite"><br>
<blockquote id="StationeryCiteGenerated_1" type="cite">
<pre wrap="">
I have 12 data disks in a RAID 50 (3 RAID5's in a RAID0) and use a suggested
stripesize of 64k, so a stripe-width of 768k.
</pre>
</blockquote>
</blockquote>
I must state this correction -- to the above -- I was thinking I had to<br>
use the least common multiplier -- but in reality, I <span class="moz-txt-underscore"><span class="moz-txt-tag">_</span>think<span class="moz-txt-tag">_</span></span> (subject to<br>
later revision, both -- reality and what I think, ;-)), I should be
able to<br>
use the greatest common denominator and that should work<br>
equally well.<br>
<br>
Nevertheless, this:<br>
<blockquote id="mid_20120725210434_GA20135_redhat_com"
cite="mid:20120725210434.GA20135@redhat.com" type="cite">
<pre wrap=""><!---->
The DM update for the 3.6 merge window adds non power of 2 support in
the kernel (for the stripe and thin-pool targets).
So the lvm2 constraints that require a power of 2 chunksize will be
relaxed very shortly.
</pre>
</blockquote>
is great news, as I'm sure that being able to synchronize stripes to<br>
chunksize AND extent size, would allow better performance for RAID aware<br>
file systems...<br>
<br>
I.e. As it is now, I need to have my extents line up, my chunks line up<br>
AND my file system block allocations line up....<b class="moz-txt-star"><span class="moz-txt-tag">*</span>ouch<span class="moz-txt-tag">*</span></b>....<br>
<br>
If any of those start at 'fixed offsets' from the beginnning of their<br>
space and that fixed size isn't a multiple of the extent size (for VG's)<br>
OR chunksize for LV's, then all of the VG's/LV's will won't line up --<br>
this is only an issue if you have multiple VG's per PV (i.e. the PV's<br>
will use the same physical disks, and the LV's would also use the same<br>
disks, but what you thought was an aligned database mirrored in<br>
1 partition might be completely non-aligned if anything under it is<br>
misaligned.<br>
<br>
I very well do NOT know, the full impact of such -- only that it would<br>
cause unpredictable or variable performance, which as an engineer,<br>
bothers me.<br>
<br>
<br>
<br>
</body>
</html>