Subject: experiments with lfs
To: None <current-users@netbsd.org>
From: Daniel Carosone <dan@geek.com.au>
List: current-users
Date: 04/06/2005 12:38:46
--cDtQGJ/EJIRf/Cpq
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

I did some preliminary experiments with LFS, which has seen some
considerable recent work.

I created a 10G vnd file, and an LFS in there.  I generated lots of IO
and files.  I filled it, and kept trying to fill it - it did
eventually hang the machine in this state, alas.

Until then, it seemed to behave flawlessly, and certainly worked
better than the last time I tried it.  Before filling it, I unmounted
it and ran a fsck_lfs, which came up clean (though it took a *very*
long time and seemed to be heavily CPU bound, I'll try and do some profiling
of it at some point later to figure out where its spinning).

Even after the hang, fsck_lfs came up clean too.  So it's clearly much
improved.

--
Dan.

--cDtQGJ/EJIRf/Cpq
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (NetBSD)

iD8DBQFCU0u2EAVxvV4N66cRApgeAKCvn2wHLIGhW2uv+vB9KQWEMazk+QCg7+2H
E7+JNFd1YUMM4OL75XFKoFs=
=Sr7R
-----END PGP SIGNATURE-----

--cDtQGJ/EJIRf/Cpq--