What’s the most efficient way to clone a Linux hard drive?
I have about 20 hard disks to clone for a new student lab with identical hardware. I’d like not to have to purchase any additional software. I do have a hard drive dock which holds one hard disk.
Below is the procedure I just used and it worked, but I’m sure there’s a better way.
Starting with a “master” workstation running Ubuntu (Machine 1)
1. Install linux on a different machine (Machine 2). Just the initial installation, no configuration needs to be done. (Hardware is identical to the 1st machine.) I suppose all I’d actually need to do is format this hard disk. But it’s a small format case and there isn’t a power connector for a second drive.
2. Remove the hard disk (Hard disk 2) and place it in the hard drive dock. Connect the hard drive dock to Machine 1 via USB.
3. Boot Machine 1; it sees Hard Disk 2 as /dev/sdb
4. (as root) dd if=/dev/sda of=/dev/sdb
5. Place Hard Disk 2 back in Machine 2 and boot. It comes up in maintenance mode.
6. fsck /dev/sda
7. Now Machine 2 boots as a clone of Machine 1.
8. Edit /etc/hostname and /etc/hosts to reflect the Machine 2’s correct name.
Below is the procedure I just used and it worked, but I’m sure there’s a better way.
Starting with a “master” workstation running Ubuntu (Machine 1)
1. Install linux on a different machine (Machine 2). Just the initial installation, no configuration needs to be done. (Hardware is identical to the 1st machine.) I suppose all I’d actually need to do is format this hard disk. But it’s a small format case and there isn’t a power connector for a second drive.
2. Remove the hard disk (Hard disk 2) and place it in the hard drive dock. Connect the hard drive dock to Machine 1 via USB.
3. Boot Machine 1; it sees Hard Disk 2 as /dev/sdb
4. (as root) dd if=/dev/sda of=/dev/sdb
5. Place Hard Disk 2 back in Machine 2 and boot. It comes up in maintenance mode.
6. fsck /dev/sda
7. Now Machine 2 boots as a clone of Machine 1.
8. Edit /etc/hostname and /etc/hosts to reflect the Machine 2’s correct name.
I think you are looking for Clonezilla server : https://www.youtube.com/watch?v=cRXNipWrg4U
Dd is a low level media copy.
Dumo/restore might be faster.
Using Clonezilla .. To image unconfigured system
If you have 20 to setup per session. Use of a deployment, or disk less network based ... But that takes time to setup, test. Are these lab systems use LDAP to be in a position for central management of user accounts?
Virtualized?
An option? I,e. Users access a VM rather than physical host.
Dumo/restore might be faster.
Using Clonezilla .. To image unconfigured system
If you have 20 to setup per session. Use of a deployment, or disk less network based ... But that takes time to setup, test. Are these lab systems use LDAP to be in a position for central management of user accounts?
Virtualized?
An option? I,e. Users access a VM rather than physical host.
ASKER CERTIFIED SOLUTION
membership
Create a free account to see this answer
Signing up is free and takes 30 seconds. No credit card required.
ASKER
Thanks John, arnold and Dr. Klahn.
Yes, I agree I should probably use Clonezilla. But if dd works then I'll stick with it. I'll try "bs=16M"
dump/restore seems like a very heavy-duty tool. I'll consider it in the future maybe. I installed it and looked at the man page.
I agree, we should use more virtualization. But I'm not the decision-maker in this regard.
Thanks again. I love Experts-Exchange.
Yes, I agree I should probably use Clonezilla. But if dd works then I'll stick with it. I'll try "bs=16M"
dump/restore seems like a very heavy-duty tool. I'll consider it in the future maybe. I installed it and looked at the man page.
I agree, we should use more virtualization. But I'm not the decision-maker in this regard.
Thanks again. I love Experts-Exchange.
ASKER
cloning-a-linux-hd.pdf