On Mon, Nov 12, 2007 at 09:57:21AM -0500, Justin Piszcz wrote:
> On Mon, 12 Nov 2007, David Chinner wrote:
> >A single spindle, regardless of it's size, will have similar
> >seek characteristics so scaling the number of AGs with size
> >is the wrong thing to do - you don't get better parallelism
> >out of a single spindle, just more seeks and lower performance.
> >hence keeping the number of AGs fixed up to the point where
> >the AG size tops out (i.e. 4TB) seems like a better scaling
> >factor to me. i.e. something like:
> >
> >
> > if (!multidisk) {
> > if (dblocks >= TERABYTES(4, blocklog)) {
> > blocks = XFS_AG_MAX_BLOCKS(blocklog);
> > goto done;
> > }
> > agcount = 4;
> > /* work out ag size here */
> > goto done;
> > }
> >
> >I'd also like to see some test results showing the mkfs output
> >for the different configurations to confirm it works correctly
> >(i.e. that the corner cases work correctly).
>
> Dave, when this is put into place do you recommend people re-format their
> XFS partitions for those with a 750GiB drive -or- with a < 2TB RAID5
> array,
No. If you are having performance problems, then changing the way the
filesystem is laid out *may* improve performance but if everything
is working fine then don't change it.
> would one see any increase in speed?
On a single disk, yes. On RAID5 - who knows. There are so many other
variables to raid5 performance (esp software raid) that such single disk
optimisations could degrade performance. On other RAID hardware, it
might improve - it really depends on the RAID implementation....
Cheers,
Dave.
--
Dave Chinner
Principal Engineer
SGI Australian Software Group
|