Mondo

[Ray Bell]

I have create a mondo image for RHEL5 and place it on a nfs filesystem. I want to clone another server that have a few different hardware devices. Anyway I stated mondo via nfs and accepted all the default and it ran successful. The server rebooted and came up as expected. From the console I login and try to activated the network card eth0 and I'm not successful. The ip address is valid, the gateway and subnet mask is correct. So why can't I activate the network card to put my server on the network.

[Patrick Wallek]

Any error messages anywhere (/var/log/messages)? Are you trying to activate it via the GUI or command line?

Can you use ifconfig from the command line to configure the card?

What about ethtool output?

[Ray Bell]

I went in /var/log/message and I did not see any errors. I went to /etc/sysconfig/network-scripts and add the following:

ONBOOT=yes
BOOTPROTO=static
IPADDR=209.82.230.162
NETMASK=255.255.255.0

The DEVICE=eth0 and HDWADDR=xx:xx:xx:xx
was already assign,

So I started it up with the following command:

service network restart

It restart the network and still I could not ping anything. Not sure why this is happening. Anyway suggestion would be great. Also if you have the syntax for the command to start it with the ifconf then please tell send me that as well. I am out of suggestion. I have done mondo 3 times and still it's faling.

[Ivan Ferreira]

Check the ouput of ifconfig -a and compare the MAC address of your eth0 card with the HWADDR of the /etc/sysconfig/network-scripts/ifcfg-eth0 file. They should match.

Also, after a hardware change, normally the ifcfg-eth0 file is renamed o ifcfg-eth0.bak. Check that your configuration file is valid for the eth0 card.

[Ivan Ferreira]

Please post the output of ifconfig -a and ensure that iptables is stopped.

[Michael Steele_2]

Hi

What happens when you ping your ip address? If this fails then your nic is a problem. If this succeeds then what happens when you ping your default gateway? If this fails then investigate.

Since your going over the network with mondo it appears all your cables, etc. are intact before clone.

[Michael Steele_2]

Hi

Did you clone the master's mac onto the client's?

[Cornec]

In additoin to the good hints given earlier, please check that no udev rule is creating an associatoi nbetween your MAC address and your ethx device (find /etc/udev -type f | xargs grep MAC-ADDR e.g.)

[Ray Bell]

I ran ifconfig -a and the MAC address math what for eth0.
I was not able to do a cut and paste so I type some of the information from eth0, eth1, eth2 and eth3:
{aspt5db1.national.aaa.com}/root-> ifconfig -a
eth0 Link encap:Ethernet HWaddr 00:22:64:99:52:8A
inet addr:209.82.230.121 Bcast:209.82.230.255 Mask:255.255.255.0
inet6 addr: fe80::222:64ff:fe99:528a/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 b) TX bytes:0 (0.0 b)


eth1 Link encap:Ethernet HWaddr 00:22:81:7c:29:B1
BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 b) TX bytes:0 (0.0 b)


eth2 Link encap:Ethernet HWaddr 00:23:7D:E2:BA:2E
BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:8296 errors:0 dropped:0 overruns:0 frame:0
TX packets:27 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:541701 (529.0 KiB) TX bytes:6773 (6.6 KiB b)
Interrupt:169 Memory:f8000000-f8012800

eth3 Link encap:Ethernet HWaddr 00:23:7D:E2:BA:2C
BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 b) TX bytes:0 (0.0 b)
Interrupt:177 Memory:fa000000-fa012800

"It look like when the server came online that eth2 try to start." Could that be the problem since it has no ip address associated with it?


- I can ping the ipaddress of the server but can't ping the gateway.

- I'm not sure what you mean about clone the master's mac onto the clinet's. I did a mondo and not usre if you can specify to not get the mac and if you do then how can I do it.

Anyway I have try everythong I know of and have done restore this server 4 time and still no success. Any other thoughts?

[John McNulty_2 · Accepted solution]

It could be that on the new hardware NICs are scanned by the kernel in a different order to the old server. This can happen even if the server type is the same but the cards are a different type and use a different driver. I experienced this recently when replacing NC360T (Intel) cards on a number of servers with NC382T (Broadcom) cards. The old cards scanned from right to left (looking at the back) and the new cards from left to right.

Login to the console as root and run: "dmesg -n 8" then pop the ethernet cables out and in one at a time and watch the link down/up messages on the console. This will confirm which NIC the system thinks is eth0, eth1, etc.

If it's different from what you expect, then change the network scripts accordingly. Use /etc/sysconfig/hwconf as a reference to see which hardware address the system has mapped to which eth device.

[John McNulty_2]

Oh, and you have change the IP address on the clone server from what it was before right? I mean, you're not running two servers with the same IP address at the same time?

[Tim Nelson]

I would guess that the image does not have a driver included for the new NIC hardware.

mkinitrd â f â v initrd-2.6.18-128.1.1.el5 2.6.18-128.1.1.el5

give it a reboot and see if that fixes your issue.

I was testing mondo for awhile but ran into too many issues. I am currently testing mkCDrec. Seems to be more stable, or at least easier to use. The biggest piece, as with mondo, if you are cloning you need to make sure all modules are included to ensure you have a driver after booting the cloned image.

Best of luck.

[Ray Bell]

I am trying to not doa fresh install but I just might have to do so. I'm still lerning Linux as I go and the mondo would be the best bet to get the system setup correctly. Anyway I have the following image:

HP-initrd-2.6.18-92.e15.img so how would I generate a new kernel that will include all new hardware? I added the porliant pack that the new kernel.

[Tim Nelson]

Let's take one step backwards first.

>>From the console I login and try to activated the network card eth0 and I'm not successful

Is there an error when you configure the NIC ? if so please post.

[Ray Bell]

The problem was that eth0 is now eth2. So thanks for the assistance from everyone. I did not know that the eth cards can change postion. I will take that in consideration when doing mondo for all futire system.

[Ray Bell]

The problem was the eth0 change postion after doing the restore.

[Michael Steele_2]

Right. Verify the master image's primary nic is also eth2. Else, how could eth2 overwrite eth0? You cloned the master over the client. Common imaging problem.