[K12OSN] Re: tuxpaint and others running slow

[Les Mikesell]

On Wed, 2005-10-26 at 08:12, gKw-X wrote:
> > This isn't making sense.  A machine like that as the only active
> > client should be nearly as fast as X on the console.  Is the
> > old switch still in the picture?
> 
> You're telling me. I don't know a whole lot about linux or ltsp.
> 
> Is it possible that resources are being limited for each client? This is 
> the only thing I can think of.

As far as the system is concerned it doesn't matter if you are
on the console or remote.  There could be limits on non-root users,
but if you log in as the same person, you should have the same
limits in both places.

>  Why would it only be taking 1.5% cpu for 
> the client when it takes 50-60% cpu running locally on the server?

I still think it is the network until proven otherwise.  If you
can find a crossover cable, try connecting the fast client
directly - or use a different switch with nothing else but
the server and one client.   My next guess would be the
server NIC or the slot it occupies in the server.

Also, did you look at the error counts from 'ifconfig'?  You
should not have any tx/rx errors, but you also need to know
if you have any on the switch side.  With a crossover cable
you'd be able to see both ends.

-- 
  Les Mikesell
    les at futuresource.com