Re: [Ext2-devel] disk throughput

Andrew Morton (akpm@zip.com.au)
Tue, 06 Nov 2001 00:48:29 -0800


Alexander Viro wrote:
>
> On Tue, 6 Nov 2001, Alan Cox wrote:
>
> > Surely the answer if you want short term write speed and long term balancing
> > is to use ext3 not ext2 and simply ensure that the relevant stuff goes to
> > the journal (which will be nicely ordered) first. That will give you some
> > buffering at least.
>
> Alan, the problem is present in ext3 as well as in all other FFS derivatives
> (well, FreeBSD had tried to deal that stuff this Spring).
>

Yep. Once we're seek-bound on metadata and data, the occasional
seek-and-squirt into the journal won't make much difference, and
the other write patterns will be the same.

Interestingly, current ext3 can do a 600 meg write in fifty
seconds, whereas ext2 takes seventy. This will be related to the
fact that ext3 just pumps all the buffers into submit_bh(),
whereas ext2 fiddles around with all the write_locked_buffers()
stuff. I think. Or the intermingling of indirects with data
is tripping ext2 up. The additional seeking is audible.

-
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/