[BBLISA] ZFS Anyone?

[Peter Galvin]

Hi Dewey, 

Yes, I meant that ZFS scales linearly with the number of spindles,
essentially being able to drive them at their max throughput.  I'm not sure
on your RAID 6 calculation. For reads certainly you'd scale with number of
spindles (basically a striped read).  For writes, without cache, you would
certainly see a large decrease from that in general. ZFS, for asynchronous
I/O, essentially always tries to do sequential full-stripe writes even of
random-write data, greatly improving over random performance. If it's
synchronous writes it has to get them out to disk fast, and that's when the
logzilla really improves performance.

--Peter

Blogs: http://ctistrategy.com http://www.galvin.info


On 4/3/09 3:59 PM, "Dewey Sasser" <dewey at sasser.com> wrote:

> Peter Galvin wrote:
> 
>>> ZFS will go at spindle speed for sequential I/O. For random I/O it's best to
>>> have flash memory for the ZIL (intent log). I assume you'd use NFS if you
>>> went with an external device? If so that would be mostly random I/O and
>>> you'd want the ZIL to be on flash.
>>   
> 
> "Spindle speed" is an interesting concept.  I know I need to go way
> faster than a single spindle will give me.  I have an idea of how
> various RAIDs scale with number of spindles, but how does ZFS do?  If I
> buy a 7210 with 46 drives, do I effectively see 46 * (1 drive's worth of
> IOps)?  With RAID 6 I would expect 1/6*46*(1 drive's worth of IOps).
> 
> thanks,
> 
> --
> Dewey
> 
> 
> _______________________________________________
> bblisa mailing list
> bblisa at bblisa.org
> http://www.bblisa.org/mailman/listinfo/bblisa