Chris Wedgwood:
> On Tue, Aug 18, 1998 at 08:19:26AM +0200, Matthias Urlichs wrote:
>
> Has anyone actually verified whether or not the dcache is indeed a
> win on large news servers with 50-100 GB of disk at all?
>
I have verified that it's a win with 10 GB.
Beyond that, you get to have so big directories with the standard storage
method that ext2 is _very_ bad at handling them; I'd strongly advise
anybody with any "real" amount of News server disk to use one of the
alternate storage methods.
INN's "timehash" method still stores one news article per file, but its
directory access patterns are such that some testing on said 10-GB system
indicated (albeit a few months ago, shortly after the thing became usable)
that the dcache is indeed a win.
> I would have though with seemingly `random' access to several million
> files/inodes any caching benefits for most files would be lost...
> (sure, some files are accessed very often, but in this case the
> buffer case would have them `hot' in 2.0.x anyhow).
>
Better ask anybody who's using a big squid system...
-- Matthias Urlichs noris network GmbH- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html