Murali,

Additionally your VM's are likely attached to a physical port via a virtual bridge or vswitch of some sort.  Can you verify that they are connected to the correct physical interface? 

i.e. 
eth0 is connected to br0 (if the cobbler server is here) 
VMs are connected to br1 (and the vms are here) 

If the setup above is similar to what you have its not going to work without port forwarding.  I would recommend to check this provided that your firewalls are in good condition.

The setup that you should want would be something like:

eth0 → br0 → VMs

let us know how things have turned out.

Scott



On Thu, May 7, 2015 at 2:44 PM, Alan Evangelista <alanoe@linux.vnet.ibm.com> wrote:
On 04/02/2015 08:30 PM, Scott Mattan wrote:
In regards to this issue i know you have verified that this with a physical box successfully.  Is your there a router between your two servers by chance? However this seems like some kind of firewall issue.

In your environment is it possible for you to backup your firewall rules, clear your firewall to temporarily allow all access and try again?

Are your VMs using a different interface to connect to the cobbler server?  It might be best to check your firewall rules for filtering of DHCP option 93 on the port accepting DHCP requests from the Cobbler server.

Also make sure you didnt open those ports for just you VMs but for the entire subnet.


2015年4月2日木曜日、Alan Evangelista<alanoe@linux.vnet.ibm.com <mailto:alanoe@linux.vnet.ibm.com>> さんは書きました:

    On 04/01/2015 10:52 PM, Sethuraman, Murali wrote:

        Hi Alan,
        Thanks for your continued responses. When "dhclient" is
        executed without any options from the baremetal,
        /var/log/messages on the cobbler server does not get updated
        with any DHCP requests [similar to what happens when it tries
        to pxe boot]. It does not work.


    I know I have already asked again, but just confirming

    - Are you sure the DHCP server is in the same subnet of the client?
    - Are you sure there are no other DHCP servers in the same subnet?

    If both answers are "yes", there is something odd in your
    network/DHCP setup which prevents
    broadcasted DHCP requests to arrive in the DHCP server. I'd also
    check if there are any link layer firewalls
    which may be blocking the communication. If not, I'd use some
    network debugging tool such
    as tcpdump in another system of the same subnet and assert if it
    is receiving the broadcasted DHCP
    requests (UDP packages), then we know if the problem is specific
    to the DHCP server or not.


Murali, was this a firewall issue after all?



Regards,
Alan Evangelista

_______________________________________________
cobbler mailing list
cobbler@lists.fedorahosted.org
https://lists.fedorahosted.org/mailman/listinfo/cobbler