<br><br><div class="gmail_quote">On Fri, Jun 29, 2012 at 11:26 AM, Gary Buhrmaster <span dir="ltr"><<a href="mailto:gary.buhrmaster@gmail.com" target="_blank">gary.buhrmaster@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
On Fri, Jun 29, 2012 at 3:23 PM, Dave Cunningham <<a href="mailto:ml@upsilon.org.uk">ml@upsilon.org.uk</a>> wrote:<br>
,,,,<br>
<div class="im">> Thanks for that - this is something I had never even consider and after<br>
> googling a bit it seems you are absolute correct about RAID-5<br>
<br>
</div>That is one of the reasons that raid-6 (aka raidz2) became popular,<br>
along with background scrubbing, and hot spare fast rebuild,<br>
especially in the enterprise space. And, even two parity drives<br>
are not always enough as drives get bigger and bigger (and the<br>
reconstruct time goes up and up).<br>
<div class="HOEnZb"><div class="h5">_______________________________________________<br></div></div></blockquote><div><br></div><div>Also note that using the minimum number of drives to build an array (3 in RAID 5 or 4 in RAID 6) is by far the slowest. It is far better to use more smaller drives to achieve a given volume size with the caveat that the more drives you have increases the likely hood of loosing two drives. </div>
</div>