RIS DHCP Router

[Ian Edwards]

We have just found an issue with RIS and are using a workaround that might
interest soem, of you. If you know how to overcome the real issue please
reply, thanks.

If you use RIS across a router that is setup with a Bootp Relay address
pointing to the DHCP/RIS server, RIS does not connect to the TFTP server.
An ip address is issues but the next step fails.

Placing the DHCP on a different server then all works fine.

For the workstations on the local subnet the dhcp server needs to be on the
same server as RIS. The RIS/DHCP server should only contain scopes for the
local subnet.

Also, ensure that the Bootp Relay setup in the router does not have the
RIS/DHCP server, only the address/es of DHCP only server.

 

[Johan Arwidmark]

One customer site had this problem and they solved it by using ISC
DHCP Server on solaris box instead of Microsoft DHCP. ISC DHCP has a
Next-Server option which actually helps the client locate a RIS Server
on a different VLAN.

I don't know if this option can be enabled on Microsoft DHCP

Here is some options you can enable through netsh

Using Dynamic Host Configuration Protocol Options 60, 66, 67 to Direct
PXE Clients to RIS Servers May Fail
http://support.microsoft.com/?kbid=259670

You can also solve this problem by adding the IP address of the RIS
servers to the IP helper tables.

regards
Johan Arwidmark

Windows User Group - Nordic
http://www.wug-nordic.net

 

[NIC Student]

As Johan mentions, you need the RIS server's ip address as a second dhcp
relay in your routers.

For example, we have the following scenario in our environment:

1 RIS server (Server A)
1 DHCP server (Server B)
Many different subnets, no clients on same subnet as RIS server or DHCP
server.
Routers are configured with relay helper addresses for both the RIS and DHCP
servers.
We do not use the DHCP options 60, 66 or 67.

Port 4011 is not blocked? Make sure it is open.

Make sure the ip address your client is getting is not in use somewhere -
usually as a static entry. Ping the ip address that the clients are being
given to see if they are already in use.

Spanning tree is turned on? Ask about that and you may need to adjust the
time out as the client will not wait long before it times out. An easy way
to test this is to begin the RIS and then when the client gets the address -
hit pause for 10 seconds. Then hit enter and see if the client gets the
files.

Are clients on the same subnet as the RIS server able to complete RIS?

Look on the RIS Server: Are there any error messages in the event log under
the System or Application logs specific to BINLSVC, TFTPD, DNS, or Active
Directory services?

We saw a post here one time with this issue: "Turns out this is relate to
either MetaIP DHCP or the switches and routers between the PXE clients and
the DHCP server." Whatever that means

Do you have ACLs in place on the client side of the router (VLAN's)? Disable
them for the tests.

Make sure you have the latest PXE Roms in your nics. Flash them if
necessary.