Having issues with PXE boot not working.
DHCP server and WDS server are on the same subnet 10.18.0.0/16 and the clients are on a separate one.
If I set the subnet of the client to the same as the servers it will PXE boot fine but the issue is seen only when traversing subnets.
I have set IP helpers on the layer 3 device for both DHCP servers and WDS server but still no luck.
The client does receive a DHCP but the PXE boot does get stuck post receiving the address.
Any advice of things to check?
What error are you getting on the client. You may need to do a wireshark capture of both cases(working and not-working) and compare to find which step is failing.
Do you have DHCP options 66 & 67 configured on the DHCP server? 66 should tell the client the IP address of the PXE boot server (TFTP server) and 67 should tell it the path and file to boot from.
I have both options set although option 66 is set with the name of the server not IP. Will this have any affect on communication?
Modified option 66 from the server name to server IP and still no different.
It stops at "Awaiting approval" and although it is not a network error as such like i say if i have the client on the same subnet it works fine.
Go with Johnd2310's suggestion of capturing the conversation with it working locally on the same subnet and not working across subnets and compare them.
I know I have been through this before and did manage to get it working, however it was ages ago on a Windows 2000 domain. I have searched through my documents and can't find any notes I made unfortunately...
It is not clear to me whether the clients in one vlan and servers in a different vlan are connected on the same switch or are connected to different switches. Can the original poster provide clarification about this?
Also would the original poster post the configuration at lest of the interfaces involved, and it might be better if we saw the entire switch config.
Clients are in a different VLAN from the Servers.
Vlan interface of layer 3 switch:
description Floor 3 DATA VLAN
ip address 10.20.30.1 255.255.254.0
ip helper-address 10.18.4.113
ip helper-address 10.9.14.113
ip helper-address 10.18.3.84
no ip redirects
no ip unreachables
no ip proxy-arp
Interface going to Switch where servers are connected:
description Link to PMR1-FLG-DC-SW1 Eth1/1
ip address 10.190.30.2 255.255.255.252
no ip redirects
no ip unreachables
no ip proxy-arp
ip ospf network point-to-point
ip ospf cost 100
Interface going to layer 2 switch where client exists:
description PO Floor 3 STACK
switchport trunk native vlan 999
switchport trunk allowed vlan 32-35,900,999
switchport mode trunk
spanning-tree guard root
Thanks for the additional information. Would you identify the IP address of the DHCP server and of the PXE server? Am I correct in understanding that the clients are in vlan 32? And in understanding that the clients are on a layer 2 switch which connects to a layer 3 switch which provides the gateway for the subnet the clients are in? And that the layer 3 switch has a routed port connecting to the switch where the server are?
I re-read the original post and am interested in this statement
The client does receive a DHCP but the PXE boot does get stuck
So this indicates that the clients are successful in communicating with DHCP server but perhaps not with the PXE server? Is there any indication in the logs of the PXE server that it is seeing requests from the vlan/subnet where the clients are? Are the clients able to ping the address of the PXE server? Is it possible that there is something in the DHCP scope that is not setting up the PXE environment correctly?
DHCP Server IP is 10.18.4.113
SCCM Distribution point Server - PXE - 10.18.3.84
The clients are in VLAN 32 and they connect as you describe.
The PXE server is reachable via ping from the client or any other IP in the client subnet for that matter. The DHCP scope looks to be set correctly with option 66 and option 67.
Check the configuration of the SCCM server. Could be that the client is "unknown" and is waiting for approval in SCCM.
I do not have enough expertise with SCCM or with PXE to know if there is something in their config that needs to change when clients are remote. It is interesting that it works fine if clients are local and does not work when clients are remote. The configured helper addresses are correct. And given that clients successfully receive DHCP assigned addresses and are successful in pinging the server I do not believe that there is any routing type issue impacting PXE. I wonder about the possibility that something along the path is impacting PXE traffic. And wondering about the possibility that PXE is using or referring to some resource that is not available when clients are remote.
Were you able to resolve this?
I have a similar issue; however, it appears to me that it is failing because the PXE "DHCP Offer" has a destination IP of 255.255.255.255 (not the dhcp relay agent's IP). The "DHCP Offer" from the DHCP server works fine because it responds back to the relay agent's IP Address directly.
It seems like it would work fine if the PXE Server's "DHCP Offer" was destined for the relay agent's IP (the switch) - the way the DHCP Offer works from the DHCP Server.
If the client requesting PXE is in a subnet remote from the PXE server, and if the PXE is sending its offer to the local broadcast address then this is certainly a problem. It would seem that there is some issue with the PXE and how it is set up. Am I correct in assuming that you have configured ip helper-address (or some equivalent command depending on platform) and that it is forwarding to both your DHCP server and your PXE?