[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 13 Dec 2017 16:58:30 -0700
From: Scott Bauer <scott.bauer@...el.com>
To: Randy Dunlap <rdunlap@...radead.org>
Cc: dm-devel@...hat.com, snitzer@...hat.com, agk@...hat.com,
linux-kernel@...r.kernel.org, keith.busch@...el.com,
jonathan.derrick@...el.com
Subject: Re: [PATCH v3 2/2] dm unstripe: Add documentation for unstripe target
On Wed, Dec 13, 2017 at 04:18:06PM -0800, Randy Dunlap wrote:
> Use ASCII characters ___ or ---, not whatever those bottom block characters are.
Argh, sorry, I removed those for an internal review but forgot to remove them here.
>
> > +
> > + The purpose of this unstriping is to provide better QoS in noisy
> > + neighbor environments. When two partitions are created on the
> > + aggregate drive without this unstriping, reads on one partition
> > + can affect writes on another partition. This is because the partitions
> > + are striped across the two cores. When we unstripe this hardware RAID 0
> > + and make partitions on each new exposed device the two partitions are now
> > + physically separated.
> > +
> > + With the module we were able to segregate a fio script that has read and
> > + write jobs that are independent of each other. Compared to when we run
> > + the test on a combined drive with partitions, we were able to get a 92%
> > + reduction in five-9ths read latency using this device mapper target.
>
> 5/9ths
> although I can't quite parse that sentence.
I'll reword it, thanks.
Powered by blists - more mailing lists