maillog: 25/03/2004-12:01:32(+0100): Paul de Vrieze types
> On Thursday 25 March 2004 08:52, Georgi Georgiev wrote:
> > and have vim on my laptop. However, if I chroot to /remote/laptop, I'd
> > end up making all communication over NFS and using the hard disk of
> > the laptop for compiling as well, which is not very fast. I was doing
> > something with
> 
> Isn't this /remote/laptop partition local to the server? using chroot or 
> not does not in any significant way change the performance of access to 
> this location.

/remote/laptop is mounted over NFS. /remote/laptop is the / of my laptop and it
is physically on the laptop. The desktop is compiling and hopefully, using
"ROOT=/remote/laptop" is emerging the software remotely to the laptop.

> 
> > mount -o bind /var/tmp/portage/ /remote/laptop/var/tmp/portage/
> 
> Using a fast disc for /var/tmp/portage is allways advantageous.
> 
> > which speeds up things a bit, but it is still slower and too annoying
> > to do every time.
> 
> Then do it automatically from fstab

I don't have /remote/laptop always mounted (only when the laptop is powered on
and I want to emerge packages there), and I'd need to do the mount -o bind
commands after I mount /remote/laptop. I agree that this is not much of an
issue. A proper shell script would do the mounts right when I need them.

There is still the problem of using the slow disk of the laptop (for all
headers, libraries, etc. while compiling) over NFS, if I chroot to
/remote/laptop.

-- 
/\   Georgi Georgiev   /\ Cropp's Law: The amount of work done varies  /\
\/    chutz@...    \/ inversly with the time spent in the office.  \/
/\  +81(90)6266-1163   /\                                              /\

--
gentoo-dev@g.o mailing list