RIS clients cannot boot via PXE

[Jerry]

I am setting up a single W2KSP3 RIS server on an AD
network. RIS server is not running DHCP, but has been
authorized already and the verify button under Remote
Install tab returns no problems. 1045 event does appear
in event log and no errors are found there relating to
BINL or TFTP.

Clients on same subnet as server return "no boot filename
received" error, even though they've had their BIOS
updated and ON w/PXE is enabled via BIOS.

Server has been rebooted several times and problem is
duplicated on several PXE compliant systems on same and
different subnets.

Is it possible for something to be disabling PXE traffic
on a network? What gives? I get no DHCP address when
booting via F12.

 

[NIC Student]

Welcome to the group, Jerry.

So you hit F12 and you never get a dhcp address? I ask again because the
exact moment of failure is important.

You can RIS on other subnets other than the one the RIS server lives on?

Is the dhcp server on the same subnet as the RIS server?

Can operational clients on the RIS server subnet get dhcp addresses?

Do you have VLANs on your subnets with ACLs in place? Try turning them off.

 

[Jerry]

Correct, I hit F12 to network boot and it never SHOWS me
the IP address of a DHCP request.

HOWEVER....I had one of our network engineers put a
sniffer on the client port and did see it requesting a
DHCP lease, the DHCP server responded, and an IP address
was leased out to that client's MAC address.

When I go into the DHCP server and view the MMC, it
doesn't show that client's MAC address or that IP address
anywhere in the scope as leasing anything.

DHCP and RIS server are on same subnet. VLANs are being
used, but no ACLs.

Nothing has been able to contact the RIS server so far.
RIS server is pingable and I can terminal service into it
from a different subnet, if this helps.

Jerry

 

[Jerry]

Update to my issue:

I built out another RIS server on a separate subnet and
was able to successfully boot via PXE from a client on
that same subnet (plugged this server and a client into
the same 5-port switch).

I then unhooked my 1st RIS server, moved the known working
RIS server to its connections, changed the IP of the known
working system to match the new subnet static address of
the 1st server, and duplicated the problem again.

At this point I'm quite sure it is a network/switch
related issue and not a problem with the way the RIS
server was built/authorized/used.

Thanks for your help!

Jerry

 

[Oli Restorick]

Glad you got it working. It was sounding like you had a switch/router
blocking traffic in some way or another.

Thanks for updating us on the issue.

Regards

Oli