Yes, thats good stuff. Anyone planning to implement Lustre for the first time must surely read it.<br>It helps you visualize your storage requirements very nicely.<br><br>And yes, the examples are similar. <br>The configuration options are quite common when use Dell Storage.<br>
<br>It is also mentioned in the 'Lustre 2.0 Operations Manual' (821-2076-10.pdf), <br>10.1.1 Selecting Storage for the MDS or OSTs, Page 175.<br><i>(May have been moved to some other page in the newer version of the doc)</i><br>
<br>Had implemented a similar configuration (with lesser disks) for one my clients.<br><br>Regards,<br><br><br>Indivar Nair<br><br><br><br><div class="gmail_quote">On Mon, Mar 11, 2013 at 9:20 PM, Christopher J. Walker <span dir="ltr"><<a href="mailto:C.J.Walker@qmul.ac.uk" target="_blank">C.J.Walker@qmul.ac.uk</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">On 11/03/13 15:30, Indivar Nair wrote:<br>
> The best configuration to go for should be based on your file sizes,<br>
> file count and read / write patterns.<br>
> But as such you are right, Jerome.<br>
><br>
</div>> *In general -*<br>
<div class="im">> -----------------<br>
><br>
> The best RAID Configuration would be to create one that aligns with the<br>
> 1MB I/O size of Lustre.<br>
><br>
> Say you have 1 x MD3200 and 4 x MD1200 expansion arrays. That would give<br>
> you 60 Disks.<br>
> So the best option here would be to create 6 RAID6 Arrays of 10 Disks each.<br>
> In this case, you would end up having - 6 RAID6 Arrays = 6 LUNs = 6 OSTs.<br>
><br>
> In each of the 10 Disk RAID6 Array, you would get 8 Data Disks and 2<br>
> Parity Disks.<br>
</div>> Now if you divide 1MB (I/O Size) by 8 Data Disks, it gives you - *128KB*<br>
<div class="im">> - the segment / chunk size you should go for.<br>
> This config will align the I/O reads / writes, giving you the best<br>
> performance possible with the disk set.<br>
><br>
> Divide the 6 OSTs among 2 OSS Nodes, and configure the nodes to act as<br>
> failover to each other.<br>
> Just ensure that each node has enough RAM to support 6 OSTs, in case one<br>
> of them fails.<br>
><br>
<br>
<br>
</div><a href="http://content.dell.com/uk/en/enterprise/d/hpcc/cambridge-hpc-solution-centre" target="_blank">http://content.dell.com/uk/en/enterprise/d/hpcc/cambridge-hpc-solution-centre</a><br>
<br>
Has links to a couple of white papers with Dell MD3200s and MD1200s in a<br>
failover configuration. They use 2*(MD3200 + 4*MD1200 +server) to give a<br>
failover solution - which looks like exactly the setup you describe.<br>
<br>
Chris<br>
<div class="im"><br>
<br>
> Hope this helps.<br>
><br>
> Regards,<br>
><br>
><br>
> Indivar Nair<br>
><br>
><br>
> On Mon, Mar 11, 2013 at 6:31 PM, Jerome, Ron <<a href="mailto:Ron.Jerome@ssc-spc.gc.ca">Ron.Jerome@ssc-spc.gc.ca</a><br>
</div><div class="im">> <mailto:<a href="mailto:Ron.Jerome@ssc-spc.gc.ca">Ron.Jerome@ssc-spc.gc.ca</a>>> wrote:<br>
><br>
> I am currently having a debate about the best way to carve up Dell<br>
> MD3200's to be used as OST's in a Lustre file system and I invite<br>
> this community to weigh in...<br>
><br>
> I am of the opinion that it should be setup as multiple raid groups<br>
> each having a single LUN, with each raid group representing an OST,<br>
> while my colleague feels that it should be setup as a single raid<br>
> group across the whole array with multiple LUNS, with each LUN<br>
> representing an OST.<br>
><br>
> Does anyone in this group have an opinion (one way or another)?<br>
><br>
> Regards,<br>
><br>
> Ron Jerome<br>
> _______________________________________________<br>
> Lustre-discuss mailing list<br>
</div>> <a href="mailto:Lustre-discuss@lists.lustre.org">Lustre-discuss@lists.lustre.org</a> <mailto:<a href="mailto:Lustre-discuss@lists.lustre.org">Lustre-discuss@lists.lustre.org</a>><br>
> <a href="http://lists.lustre.org/mailman/listinfo/lustre-discuss" target="_blank">http://lists.lustre.org/mailman/listinfo/lustre-discuss</a><br>
<div class="HOEnZb"><div class="h5">><br>
><br>
><br>
><br>
> _______________________________________________<br>
> Lustre-discuss mailing list<br>
> <a href="mailto:Lustre-discuss@lists.lustre.org">Lustre-discuss@lists.lustre.org</a><br>
> <a href="http://lists.lustre.org/mailman/listinfo/lustre-discuss" target="_blank">http://lists.lustre.org/mailman/listinfo/lustre-discuss</a><br>
<br>
_______________________________________________<br>
Lustre-discuss mailing list<br>
<a href="mailto:Lustre-discuss@lists.lustre.org">Lustre-discuss@lists.lustre.org</a><br>
<a href="http://lists.lustre.org/mailman/listinfo/lustre-discuss" target="_blank">http://lists.lustre.org/mailman/listinfo/lustre-discuss</a><br>
</div></div></blockquote></div><br>