performance bottleneck in Linux MD RAID-1

Andrew Sharp andy.sharp at lsi.com
Thu Jul 15 16:32:52 CDT 2010


On Thu, 15 Jul 2010 12:10:08 -0600 Bond Masuda <bond.masuda at jlbond.com>
wrote:

> 
> 
> > Call me crazy, but I'm guessing you never let it finish syncing in
> > the first place.  When syncing, it caps its own throughput at some
> > arbitrary amount the code deems sensible.  Sounds like it never
> > un-capped, suggesting you never let it complete the sync process
> > before benching it.  Further evidence that both numbers are so
> > remarkably the same also suggests that.  Depending on the size of
> > the dataset, it could take a while to finish.  I could swear that I
> > recall there being some parameter which allows you to create a
> > mirror but skip the sync step.  If you do that, then basically
> > you're swearing on the storage bible that both volumes were zeroed
> > before the mirror creation, and the first thing you did after
> > creating the volume was mkfs. Personally I've never tried it, and
> > searching the documentation, I can't find it now, so perhaps that
> > was just a bad dream I had.
> 
> Hi Andrew,
> 
> Thanks for the response. Indeed, we did let it finish syncing.. we
> had to let it run overnight to finish syncing. But, we also noticed
> that while it was syncing /dev/sda to /dev/sdc, the sync rate as
> observed in /proc/mdstat and iostat was limited to 200MB/sec.

Drat, I was hoping I caught an easy one ~:^)  The similarity in those
two numbers does seem a bit coincidental.  Is it possible that
something with involved with updating the sync information slows it
down?  That seems a bit far fetched, nm.

a



More information about the Linux-PowerEdge mailing list