Gb NICs they all have but server running at Gb, clients at 100mb. I ask you, who can afford gb to the desktop for 500+ clients!!!!
Gb NICs they all have but server running at Gb, clients at 100mb. I ask you, who can afford gb to the desktop for 500+ clients!!!!
we pretty much have gb switches through out the network with fibre backbones.
not all our pc's have gb nic's though, id say about 70% have
750ish machines here
We haven't the money or the space for gb switches - such luxury!
Having troubles doing this chaining stuff myself.
We have some of these SIS cards and they're driving me loopy.
On clients that we don't have such cards we have FOG working like a *dream* (multicast and all).
However the SIS card machines.....
We've managed to get gPXE to boot on them, but gPXE doesn't find the dhcp server for round 2 of the chain (it just sits there going DHCP............ Connection Timed Out)
Clearly its getting hold of DHCP for the first round, and these clients are booting off a different config via DHCP to end up with gPXE instead of pxelinux.0 (fog) but I'm just not getting the hand-over.
Servers = TFTP/Fog on one machine and DHCP is windows based. Followed the guide on the gPXE website and all seems in order...
Any ideas?
To answer my own question:
You have to make bin/sis900.pxe in order for the sis cards to properly do the chaining. For some reason the tutorial mentioned before uses undionly.kpxe which SHOULD work but doesn't (according to the chaps at etherboot).
This overcomes the chain problem, but will introduce a new problem. FOG sends an interrupt to the BIOS to choose the next available device, but gPXE picks it up and attempts a boot but fails! so you end up going gPXE->FOG->gPXE. This is a totally different issue to the DHCP thing though
However, my problem seems to have been rectified by eliminating gPXE from the chain altogether but I am not entirely sure how...
There are currently 1 users browsing this thread. (0 members and 1 guests)