[K12OSN] Solving the bandwidth bottleneck - Solved

R. Scott Belford scott at hosef.org
Wed Aug 31 20:20:21 UTC 2005

Les Mikesell wrote:
> As I mentioned, I don't net boot on those interfaces - that
> was just an example for the syntax for other subnets.  To
> boot, you need to supply the rest of the values you find
> inside the other scope.  In particular you need
> next-server, filename and  root-path options for booting.
> I'm not sure if you can move that whole stack of
> if/else/ logic to the global section of the file
> or if you have to repeat it for each scope to handle
> the PXE and ppc variations (which is why I didn't give
> an example before). It might be best to repeat it and
> change the addresses to the appropriate server interface
> address for the subnet.  Anything within the {}' for
> one subnet isn't done on the others.

Excellent advice.  Almost elementary in retrospect, but such is the
stress of confusion.  I copied the WORKSTATION scope, renamed it
WORKSTATION2, edited the IP addresses to match the new subnet, and it
almost worked.  I recalled that little thing called a firewall, made
eth1 a trusted interface, and now both eth0 and eth1 boot clients while
eth2 firewalls the box.

In summary, should one want to have two NICS serve thin clients,

1. Edit settings for eth1
2. Edit dhcpd.conf to include eth1 IP settings
3. Edit /etc/exports to share on the eth1 subnet
4. Edit firewall settings to make eth1 a trusted device
5. Edit /etc/hosts to include settings for the eth1 subnet

Thanks, Les, for helping so much.


More information about the K12OSN mailing list