Cloning an Existing Node
NOTE: This method is no longer used. This page is kept for reference regarding cloning using g4l
The following method describes how to create a clone of an existing node.
This description makes the following assumptions:
- The node to be cloned is using the current configuration of Fedora Core 4 and Oracle 10g XE
The node to be cloned is one of the Dell PowerEdge 850 machines.
- SATA drive of 80 GB or larger for copy.
Ghost for Linux (G4L) bootable CD
- CD-ROM drive (should be included in standard node configuration)
The G4L program will copy an existing drive to a new drive that is the same size or larger.
Preparing to Copy
- Open the cover of the new node
Attach the Master Copy drive to the blue SATA cable on the right side. (The nodes from Dell should have an additional cable in place for a spare drive.)
Attach the power cable to the Master Copy.
Cloning the Drive
- Power on the machine.
- Press F2 to enter the setup screen.
- Select SATA Port 1 and press Enter.
- Change drive type from OFF to AUTO and press Enter.
- Press ESC to exit.
Select Save Changes and Exit
- Insert the Ghost for Linux (G4L) cd into the CD-ROM drive and reboot.
At the boot: prompt press Enter.
- Press Enter at the disclaimer page
- Press Enter on instruction page to continue.
- Press Enter on the release notes page to continue.
- You should now be at a command prompt. (Don't worry about an error message stating that eth0: link is not ready)
Enter the following command to start the process: g4l
Press Enter on the Yes option to continue.
At the Main Menu select RAW Mode
Under RAW Mode, select Click'n'Clone
Under Click'n'Clone, first select source. This is the master drive to be copied. In the physical configuration described above, this would be sdb. Use space bar to make selection. You should see an asterisk in the selected choice.
NOTE: Be sure you select the correct source and target or else you will end up copying the blank drive to the master and losing all data on the Master Copy!
Next Select Target, in the physical configuration described above this would be sda (The drive that came with the new node)
After selecting Source and Target, select Click'n'Clone
- Confirm that you have the right Source and Target, and then select Yes to continue.
- The cloning process will take about 1.5 hours.
- After the drive is cloned select Cancel on all the menus until you get back to the command prompt.
- Remove the g4l CD from the CD ROM drive.
- Power off the machine.
- Remove the Master Copy hard drive and replace the cover.
Power on the machine and select F2 to enter Setup. Set the SATA Port 1 drive back to OFF.
Configuring the Cloned Machine
Check the Node Master Copy Updates page for additional changes or updates that need to be applied.
- Change the hostname in the following two places
Check that the vwpoint and root passwords are set correctly.
Configure networking with the following information for testing in the GNW lab environment. See Configure Linux Networking
IP Address: 10.199.199.199 Mask: 255.255.255.0 Gateway: 10.199.199.1
- Reboot the machine to make the hostname and network changes take effect.
shutdown -r now
- At the time of this writing, changing the hostname causes the Oracle installation to become inoperable. Oracle will need to be reinstalled.
Un-install the primary oracle RPM. Run the following commands as root.
/etc/rc.d/init.d/oracle-xe stop cd /rpms rpm -e oracle-xe-10.2.0.1-0.1.i386
Follow the procedures for Installing Oracle XE on Nodes but only re-install the Oracle RPM. All other RPM's can be ignored.
Change the Linux run level to 3. See Configure Linux Run Level
Check firewall settings. See Secure the Node
Confirm that all the network settings are set correctly. See Configure Linux Networking
If there is a start-up warning for lm-sensors then as root run the command sensors-detect (the default answers should be fine).
- If there is a start-up error for auditd then as root run
chmod 600 /var/log/audit/audit.log
For FC9 and later: delete the file /etc/udev/rules.d/70-persistent-net.rules
rm -f /etc/udev/rules.d/70-persistent-net.rulesThis is done to clear out the saved mapping of the ethernet hardware MAC addresses. This file is created/modified automatically at start-up, but old ethernet MAC addresses are not removed and the actual MAC addresses are assigned to eth3 and eth4 instead. This prevents the actual ethernet adapters from working because of a conflict with the pre-existing MAC addresses from the master copy node.
After deleting the 70-persistent-net.rules file, reboot the node to allow the file to be re-created with the correct ethernet MAC addresses for that machine.