<br><br><b><i>Lincoln Peters <anfrind@gmail.com></i></b> wrote:<blockquote class="replbq" style="border-left: 2px solid rgb(16, 16, 255); margin-left: 5px; padding-left: 5px;"> On Dec 31, 2007 10:57 AM, Scrappy Laptop <scrappylaptop@yahoo.com> wrote:<br>> This may be simplistic, but just make the partitions on each physical disk<br>> the same size? It'll leave extra space on some of the drives, but you can<br>> lump those into another RAID volume if you so desire and don't mind the<br>> performance hit of two volumes' stripes on the same physical drive(s).<br><br>Unless I'm misunderstanding what you're suggesting, that would mean<br>two or more RAID partitions for the same array on a single drive.<br>Which means that if one drive fails, you could lose more than one<br>part, and RAID-5 is NOT guaranteed to recover fully when more than one<br>part of the array fails.<br><br>><br>> Hey! What if you created another abstraction layer that would join the
two<br>> RAIDs into a single volume? Totally Rube-Goldberg-esque, but an interesting<br>> way to use all available disk space...<br><br>I've thought about writing a program that would present itself as a<br>file server (maybe Samba, most likely WebDAV for the first version),<br>and manage a set of files distributed over multiple hard disks. Kind<br>of like RAID-5, but running at a filesystem level rather than a block<br>level. It would give me the flexibility I have in mind, but I'm not<br>sure whether or not it's already been done (although I know Google has<br>something similar).<br><br><br>-- <br>Lincoln Peters<br><anfrind@gmail.com><br><br>_______________________________________________<br>talk mailing list<br>talk@nblug.org<br>http://nblug.org/cgi-bin/mailman/listinfo/talk<br></anfrind@gmail.com></scrappylaptop@yahoo.com></blockquote>Nope; I was actually talking about two separate arrays (and therefore volumes) that happen to be on the same set of physical
disks. Lose a drive and two arrays would need to be rebuilt. I would not really advise it since the scattered I/O would kill your response times; it just led to the further extrapolation of the extra abstraction layer that would ultimately present a single volume that actually consisted of two separate RAID arrays on the same set of drives for the sole purpose of using every last available block of storage. Just a silly mind game, that's all.<br><p> 
<hr size=1>Looking for last minute shopping deals? <a href="http://us.rd.yahoo.com/evt=51734/*http://tools.search.yahoo.com/newsearch/category.php?category=shopping">
Find them fast with Yahoo! Search.</a>