For the last several months LFS has been completely unusable on 
-current. Just creating a file or directory on a newly created LFS file 
system causes a kernel panic. There's been quite a bit of patching by 
Konrad a few weeks ago but I've pretty much given up on it in favor of 
ffs with soft dependencies. I'll try a new kernel later today and see 
if my PR on the kernel panic has been fixed by the latest patches. If 
it works and can take some torture testing maybe it can get pulled up 
into the 3.0 branch. Unfortunately there have been a lot of changes to 
the ufs code so I don't know if that will fly or is already ticketed.

On Mar 23, 2005, at 7:22 AM, Bruce O'Neel wrote:

> Hi,
>
> Has anyone had any success with lfs on 2.0?
>
> I ask because a few nights ago I set the new (to me) 7300 up with 2.0 
> and it's
> been lightly loaded, but running fine, until last night.
>
> Yesterday morning I started building some things in a copy of pkgsrc
> whos tree lives on a lfs formatted disk.  Everthing was fine until 
> yesterday
> evening when I noticed that the make seemed not to be proceeding.  
> Everything
> was still running, just nothing was happening.  I could log in and 
> everything
> was fine until I wrote to the lfs disk then the process would hang.
>
> A reboot later and now make continued, for a while, then it hung again.
>
> Another reboot and we get a assert panic somewhere (sorry, it was 
> midnight)
> in the lfs code.
>
> Ok, fine, I can take a hint.
>
> disklabel
> newfs_ffs
>
> and off we go again, this time with ffs.  It ran all night with no 
> problem.
>
> lfs seems cool and seems to have a very high write performance, but,
> does it work?
>
> Thanks in advance.
>
> cheers
>
> bruce
>
> -- 
> edoneel@sdf.lonestar.org
> SDF Public Access UNIX System - http://sdf.lonestar.org