Subject: Re: LFS [was: Re: CVS commit: src]
To: Hubert Feyrer <firstname.lastname@example.org>
From: Konrad Schroder <email@example.com>
Date: 03/14/1999 17:54:57
On Wed, 10 Mar 1999, Hubert Feyrer wrote:
> Thank you for doing this work, I'm really glad that there's someone
> working on LFS now. Great! :-)
Hey, my pleasure.
> I'm looking forward to try this out soon - can you let me/us know in what
> state LFS is these days?
I've just committed a CHANGES and updated TODO in src/sys/ufs/lfs, that
should explain in some detail what I've done to it. In terms of how
usable it is:
- I've been using it as /tmp on my home system for a few months,
running newlfs at startup time, but not rebooting but once every few
weeks. My home system has been wonderfully stable.
- My test system, on which I really hammer on things, still crashes
sometimes, but when it comes back up the fs is clean. It generally
crashes during on-demand cleaning, which is pretty torturous for LFS, but
which it really should be able to withstand without trouble if it's going
to be useful to anyone.
Note that this is with the patch I just posted to tech-kern, not yet
committed to -current because I need to clean it up; otherwise it
will be clean except for a bunch of unreferenced inodes ;^).
- Filesystem performance, for the few tests I've done of it, is comparable
to the results in Seltzer, Smith, et al. (no year? available off of
Seltzer's web page), i.e., they seem pretty good unless you sync a lot or
the cleaner is running, then they're comparable to but worse than FFS. (I
don't have numbers to back this up, the tests were rather informal.)