Richard Lohman
2016-05-19 20:08:27 UTC
Hey, all:
I'm using dnsmasq in conjunction with cobbler to provision our servers.
This process has been working nicely for some time, but just recently I ran
into an issue, and am at the end of my rope.
I have the following configured in my dnsmasq.conf file:
# Analytics
dhcp-option=net:analytics,6,10.52.69.2,10.53.132.1
dhcp-option=net:analytics,option:router,10.53.152.254
dhcp-range=set:analytics,10.53.152.1,10.53.152.250,255.255.255.0,1h,proxy
and further down:
dhcp-host=net:x86_64,00:23:7D:EA:22:40,10.53.152.71
The local PXE/Cobbler/dnsmasq host is 10.53.152.36.
While I have not yet provisioned hosts in the subnet specified above (I'm
somewhat new to the organization), I have used similar stanzas for other
subnets with the same format that work fine. So the symptoms are:
1. Boot the host, it reaches out for the PXE server. We see this in the
dnsmasq.log file with a DHCPREQUEST followed by a DHCPACK, and then I see
log entries for the configuration being handed out. I confirmed the same
through a packet capture of the exchange. The problem is, option 3 being
sent out to the requesting server is the IP of the host upon which dnsmasq
resides (10.53.152.36), rather than what is specified in the file
(10.53.152.254). Since I am confident that what's leaving the server is
incorrect, I'm struggling with trying to understand the decisioning that
determines which IP to use (I understand the local host is a valid option
by design - no concerns there; just don't know where I might look to figure
out why).
Any thoughts?
Thanks in advance,
Rich
I'm using dnsmasq in conjunction with cobbler to provision our servers.
This process has been working nicely for some time, but just recently I ran
into an issue, and am at the end of my rope.
I have the following configured in my dnsmasq.conf file:
# Analytics
dhcp-option=net:analytics,6,10.52.69.2,10.53.132.1
dhcp-option=net:analytics,option:router,10.53.152.254
dhcp-range=set:analytics,10.53.152.1,10.53.152.250,255.255.255.0,1h,proxy
and further down:
dhcp-host=net:x86_64,00:23:7D:EA:22:40,10.53.152.71
The local PXE/Cobbler/dnsmasq host is 10.53.152.36.
While I have not yet provisioned hosts in the subnet specified above (I'm
somewhat new to the organization), I have used similar stanzas for other
subnets with the same format that work fine. So the symptoms are:
1. Boot the host, it reaches out for the PXE server. We see this in the
dnsmasq.log file with a DHCPREQUEST followed by a DHCPACK, and then I see
log entries for the configuration being handed out. I confirmed the same
through a packet capture of the exchange. The problem is, option 3 being
sent out to the requesting server is the IP of the host upon which dnsmasq
resides (10.53.152.36), rather than what is specified in the file
(10.53.152.254). Since I am confident that what's leaving the server is
incorrect, I'm struggling with trying to understand the decisioning that
determines which IP to use (I understand the local host is a valid option
by design - no concerns there; just don't know where I might look to figure
out why).
Any thoughts?
Thanks in advance,
Rich
--
Richard J. Lohman
***@gmail.com
Richard J. Lohman
***@gmail.com