Re: kern/38717: sysinst shouldn't create LFS file systems

[Andrew Doran <ad%NetBSD.org@localhost>]

On Tue, May 27, 2008 at 09:55:18PM +0200, Manuel Bouyer wrote:

> On Mon, May 26, 2008 at 11:25:03PM +0000, Jeremy C. Reed wrote:
> > The following reply was made to PR kern/38717; it has been noted by GNATS.
> > 
> > From: "Jeremy C. Reed" <reed%reedmedia.net@localhost>
> > To: gnats-bugs%NetBSD.org@localhost
> > Cc: 
> > Subject: Re: kern/38717: sysinst shouldn't create LFS file systems
> > Date: Mon, 26 May 2008 18:16:33 -0500 (CDT)
> > 
> >  >  Both lfs and unionfs have come a long way, and are quite usable for 
> > some.
> >  
> >  I enabled LFS for the first time a few days ago. My system which had 
> >  around a 55+ day uptime, after a day became unstable and any process (it 
> >  seemed) that wrote to my other partition (not LFS) would hang. LFS 
> >  appeared to make my system unusable. I powercycled and newfs to FFS and 
> >  remounted and all is well.
> 
> /usr/src on my home box has been lfs for several years now. I've not had
> major problems with it.

I have done a lot of stress testing on the file system code recently. Using
a mixture of SGI fsstress, fsx, bonnie and postmark over different versions
of NetBSD, my experience is that:

- ffs, ext2fs, and tmpfs in -current /seem/ unbreakable
- lfs and ffs+softdep can be broken within minutes
- lfs in -current breaks within seconds

Andrew