Cloned DNS server issues

[luke_nexgen]

First let me preface this by saying I cloned our production DNS servers via clonezilla and restored that clone to a VM on our private network.

Here's what I have and what I'm trying to accomplish.

What I have:

DNS1 is currently a bare bones physical machine.  We are an ISP and it provides DNS to our customers.  It currently has two interfaces, one is a public IP and the other is a private (10.x.x.x) IP.

DNS2 is a bare bones slave to DNS1, we use it as a fail safe in case DNS1 fails.  It also has two interfaces, one public and the other private.

What I want to accomplish:

Virtualize both servers.  I have ESXi 5.1 running on a server in my office, on a 192 network.  I want to clone them both to my office server, get it operable, then do the same to the other two physical machines.  If one box fails, simply light up a VM on the other physical EXSi box.  Once I have them operable in my office, I will install the server in our data center and work on getting the other two configured the same.

I've restored the clone of DNS2 to my office ESXi box, but believe I have some networking issues.  I tried reconfiguring the network parameters in /etc/sysconfig/network-scripts/eth0 & eth1 to my local network, but get the following when I try restarting networking:  "Bringing up interface eth0:  Device eth0 does not seem to be present, delaying initialization."

Anybody have any advice?

PS-They are running CentOS 6.3

[KamilAzmer]

I've restored the clone of DNS2 to my office ESXi box, but believe I have some networking issues.  I tried reconfiguring the network parameters in /etc/sysconfig/network-scripts/eth0 & eth1 to my local network, but get the following when I try restarting networking:  "Bringing up interface eth0:  Device eth0 does not seem to be present, delaying initialization."

Anybody have any advice?

PS-They are running CentOS 6.3

There are reason of your error occur and faced because of your cloning before is generating new mac address, and the mac address was generated the OS will recognized as new hardware / new network adapter. your network adapter now probably is not eth0 and eth1 anymore, it's would be eth2 and eth3 cause of the eth0 and eth1 is belong to your old mac address network adapter. This is normal symptoms.

How to check your current network adapter name ? Login into your shell, type $ ifconfig -a (will shown all the interfaces)

You probably wanted to change the naming and still want to use eth0 and eth1, yes you can do that with changes the files inside here :-

$ vi /etc/udev/rules.d/70-persistent-net.rules

Then find the text eth2 and replace to be eth0 (vice-versa), for the eth3 replace to eth1.

Examples:

change eth2 to eth0,

change eth0 to eth2.

change eth1 to eth3.

change eth3 to eth1.

Just my one cents!

[markdjones82]

In addition to that in your ifcfg rules you can remove the HWADDR line in case your MAC address ever changes and you don't want to have an issue.

[luke_nexgen]

Thanks.  Think I got the private IP resolved.  If I want to add another eth interface, do I just add the interface in VMWare then go into the ifcfg CentOS and change the IP address and leave the MAC address off?  We are going to have one interface for MGMT, and another for public DNS.  The public DNS is on VLAN X.  Is there anything I have to do to configure the virtual interface for VLAN X?

[KamilAzmer]

Yes, that correct.

You just make sure that your network adapter is set to the portgroup of vlan X to ensure the connectivity is thru that vlan.

Sent by Maxis from my BlackBerry® smartphone

[luke_nexgen]

Thanks,

Think I pretty much got it.  One last question.  On a server running ESXi, does the physical NIC act more like a passthrough device passing all traffic to the virtual switch?  What if I need two or more physical NIC's connected?  I have two physical NIC's, each needing to be connected to two different switches.

[Josh26]

luke_nexgen wrote:

Thanks,

Think I pretty much got it.  One last question.  On a server running ESXi, does the physical NIC act more like a passthrough device passing all traffic to the virtual switch?  What if I need two or more physical NIC's connected?  I have two physical NIC's, each needing to be connected to two different switches.

That team will have a "routing policy", look in the vSwitch configuration.

If they are connected to different switches, the only (usually) valid option is "Route by port ID", which means each vNIC will be assigned to oen pNIC - not necessarily evenly distributed.

[luke_nexgen]

Not sure that's it.  I don't need to load balance via the physical NIC's.  One physical NIC will be for private management (VLAN X), and the other physical NIC will have a public IP address and serve as DNS to our customers (VLAN Y).

Any ideas?

[luke_nexgen]

I think what part of my problem is that my second physical NIC isn't showing up in Confiuration->Networking.

[KamilAzmer]

You just need to creating Snew vSwitch and set the UPLINK for NICs that you need and also key-in the VLAN required inside there for Virtual Machine Port Group.

[KTom]

Hmm.  I see this hasn't been answered yet.  There are a number of resolutions to this not just the obvious.  You may need to dig deeper into the actual service and what it is touching.  Try also ip -o link to see what the output is there and post it here:

Device eth0 does not seem to be present.

Cheers,

TK