<div class="gmail_quote">On Thu, Mar 3, 2011 at 2:33 AM, Alex Butcher <span dir="ltr"><<a href="mailto:mythlist@assursys.co.uk" target="_blank">mythlist@assursys.co.uk</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<div>On Thu, 3 Mar 2011, Fedor Pikus wrote:<br>
<br>
</div><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"><div>
On Thu, Mar 3, 2011 at 12:07 AM, Alex Butcher <<a href="mailto:mythlist@assursys.co.uk" target="_blank">mythlist@assursys.co.uk</a>><br>
wrote:<br>
</div><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"><div>
I don't like parity RAID, but that's because I've seen poor write<br>
performance be a botttleneck in real-world performance. Yes,<br></div>
capacity/Ł is<br>
</blockquote><div>
<br>
<br>
I have a RAID5 made of old 750G WD drives (which were actually some of the<br>
faster drives when I bought them), write speed is about 130-140MB/s. A pair<br>
of Seagate 1.5TB Barracudas in RAID-0 give 205MB/s. How much faster do you<br>
want in a desktop?<br>
</div></blockquote>
<br>
That's pretty good; my experience with parity RAID has been with hardware<br>
RAID controllers where the write performance has been truly miserable (a few<br>
10s of MB/s). Yet another example of how the embedded processors used in<br>
most hardware RAID controllers have fallen behind software RAID running on<br>
the main CPU (providing you have a core to spare).<br><br></blockquote><div><br></div><div>Yes. About 6 years ago the company I work for bought large rack-mounted file servers (16 250G disks). We bought them from a vendor specialized in Linux clusters, whole racks with 1U servers and local file storage. The vendor tested both hardware and software RAID. High-end Adaptec RAID controllers were saturated at 90MB/s, while soft-RAID was going over 130MB/s easily. The reason was insufficient processor power of the RAID controller.<br>
</div></div><br clear="all">
<br>-- <br>Fedor G Pikus (<a href="mailto:fpikus@gmail.com" target="_blank">fpikus@gmail.com</a>)<br><a href="http://www.pikus.net" target="_blank">http://www.pikus.net</a><br><a href="http://wild-light.com" target="_blank">http://wild-light.com</a><br>