> > Also, the time to recovery after crash is often more important than
> > performance - this thread is about large filesystems, so what will be the
> > fsck time for a 1TB FFS filesystem? I suspect it will be non-trivial.
> 
> It may not be too bad if the fs has a relatively small number of large files.

Are you sure? Doesn't fsck need to do a pass over all allocated blocks to
see if some aren't allocated to multiple files at once? This gives O(N)
time (N = used space) regardless of the distribution of space among files.
Or is O(N) acceptable? I don't know the slope of this dependency...

> oh - also:
> 
> # ls -sl big
> 128 -rw-r--r--  1 root  wheel  16492674416641 Jan 19 21:23 big
> 
> or about 15TB with FFSv1 and 1k fragments.

I thought FFSv1 used 32-bit fragment numbers. Apparently it doesn't?
Also http://www.netbsd.org/Misc/features.html#large-filesystems shows
something different.

Thanks and bye	Pavel