Basically the requirement would be to have two disks on your system. I am testing this on an Oracle VM Virtualbox environment, with a CentOS 7.7 (1908) as a VM guest.
The rest of the VM guest setup is pretty straightforward nothing out of the ordinary, with two disks of the same size (or virtual disks) attached. This requires further testing done on a bare metal x86 server, which I would hope to be able to do soon with EFI.
The first disk is /dev/sda, and the second disk is /dev/sdb.
The second disk is basically empty. And obviously the first one isn't. And this is how the partition table looks like:
# parted /dev/sda p
Model: ATA VBOX HARDDISK (scsi)
Disk /dev/sda: 42.9GB
Sector size (logical/physical): 512B/512B
Partition Table: msdos
Disk Flags:
Number Start End Size Type File system Flags
1 1049kB 1075MB 1074MB primary xfs boot
2 1075MB 34.4GB 33.3GB primary lvm
Partition 1 (or /dev/sda1) is a native partition, where as partition 2 (/dev/sda2) is under LVM. As you can see here:
# pvs
PV VG Fmt Attr PSize PFree
/dev/sda2 centos lvm2 a-- <31 .01g="" 8.00m="" p="">
# vgs
VG #PV #LV #SN Attr VSize VFree
centos 1 8 0 wz--n- <31 .01g="" 8.00m="" p="">
# lvs
LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert
home centos -wi-ao---- 5.00g
root centos -wi-ao---- 5.00g
swap centos -wi-ao---- 2.00g
tmp centos -wi-ao---- 3.00g
var centos -wi-ao---- 5.00g
var_log centos -wi-ao---- 5.00g
var_log_audit centos -wi-ao---- 3.00g
var_tmp centos -wi-ao---- 3.00g
And all filesystems are xfs, which is the default filesystem for all RHEL or CentOS 7.x install.
The steps
The fastest way to achieve this is to use the good ol' dd. I am lazy so, here's the command:
# dd if=/dev/sda of=/dev/sdb
You can add bs=512, or bs=1024 to make the dd cloning go faster (This is another topic altogether)
After cloning is complete, use parted to see if the exact same partition table is on the second disk (/dev/sdb). By right, you should be seeing:
# parted /dev/sdb p
Model: ATA VBOX HARDDISK (scsi)
Disk /dev/sdb: 42.9GB
Sector size (logical/physical): 512B/512B
Partition Table: msdos
Disk Flags:
Number Start End Size Type File system Flags
1 1049kB 1075MB 1074MB primary xfs boot
2 1075MB 34.4GB 33.3GB primary lvm
First things, first, test to see if /boot or /dev/sda1 is mountable. Before you do this you'll need to use xfs_repair (to zero the logs) in order to make it mountable:
# xfs_repair -L /dev/sda1
After that, the xfs file system on /dev/sda1 uuid needs to be regenerated because of the clone, where all uuid for the filesystems, PV and VG will be the same (duplicated) and only one can be activated at one time.
To do this:
# xfs_admin -U generate /dev/sda1
Once this is done, perform a test mount:
# mkdir -p /a/boot
# mount /dev/sda1 /a/boot
Examine it's contents, by running :
# ls -l /a/boot
total 124568
-rw-r--r--. 1 root root 152980 Sep 30 22:23 config-3.10.0-1062.1.2.el7.x86_64
-rw-r--r--. 1 root root 151923 Jun 19 00:39 config-3.10.0-957.21.3.el7.x86_64
drwxr-xr-x. 3 root root 17 Jun 27 23:18 efi
drwxr-xr-x. 2 root root 27 Jun 27 23:19 grub
drwx------. 5 root root 132 Oct 29 23:56 grub2
-rw-------. 1 root root 57160365 Jun 27 23:24 initramfs-0-rescue-7ff3c24ba1924fe1af8443c0ea440407.img
-rw-------. 1 root root 21143881 Oct 3 18:41 initramfs-3.10.0-1062.1.2.el7.x86_64.img
-rw-------. 1 root root 21125579 Oct 3 18:42 initramfs-3.10.0-957.21.3.el7.x86_64.img
-rw-r--r--. 1 root root 318717 Sep 30 22:23 symvers-3.10.0-1062.1.2.el7.x86_64.gz
-rw-r--r--. 1 root root 314128 Jun 19 00:39 symvers-3.10.0-957.21.3.el7.x86_64.gz
-rw-------. 1 root root 3595191 Sep 30 22:23 System.map-3.10.0-1062.1.2.el7.x86_64
-rw-------. 1 root root 3545891 Jun 19 00:39 System.map-3.10.0-957.21.3.el7.x86_64
-rwxr-xr-x. 1 root root 6639904 Jun 27 23:24 vmlinuz-0-rescue-7ff3c24ba1924fe1af8443c0ea440407
-rwxr-xr-x. 1 root root 6734016 Sep 30 22:23 vmlinuz-3.10.0-1062.1.2.el7.x86_64
-rwxr-xr-x. 1 root root 6643904 Jun 19 00:39 vmlinuz-3.10.0-957.21.3.el7.x86_64
Unmount /a/boot.
# umount /a/boot
Perform the exact same thing with xfs_repair and xfs_admin on the LVs sitting on PV /dev/sdb2, which was cloned from /dev/sda2.
But before you can do that, you need to perform a VG import, but in this case, you will not be able to do so, because the uuid on the VG would have been duplicated (because of the dd clone).
Instead of using vgimport, use vgimportclone instead:
# vgimportclone --import -n centos1 /dev/sdb2
Where -n centos1 specifies the name of the new VG, because the original VG on /dev/sda2 (the source of the clone) is centos. /dev/sdb2 is it's PV.
Once you've done the above, you should be able to see both PVs and both VGs:
# pvs
PV VG Fmt Attr PSize PFree
/dev/sda2 centos lvm2 a-- <31 .01g="" 8.00m="" p=""> /dev/sdb2 centos1 lvm2 a-- <31 .01g="" 8.00m="" p="">
# vgs
VG #PV #LV #SN Attr VSize VFree
centos 1 8 0 wz--n- <31 .01g="" 8.00m="" p=""> centos1 1 8 0 wz--n- <31 .01g="" 8.00m="" p="">
Activate the VG centos1 if you are unable to access the LVs under it.
# vgchange -ay centos1
8 logical volume(s) in volume group "centos1" now active
Once activated, you should be able to see all the LVs on both centos and centos1 VGs.
# lvs
LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert
home centos -wi-ao---- 5.00g
root centos -wi-ao---- 5.00g
swap centos -wi-ao---- 2.00g
tmp centos -wi-ao---- 3.00g
var centos -wi-ao---- 5.00g
var_log centos -wi-ao---- 5.00g
var_log_audit centos -wi-ao---- 3.00g
var_tmp centos -wi-ao---- 3.00g
home centos1 -wi-a----- 5.00g
root centos1 -wi-a----- 5.00g
swap centos1 -wi-a----- 2.00g
tmp centos1 -wi-a----- 3.00g
var centos1 -wi-a----- 5.00g
var_log centos1 -wi-a----- 5.00g
var_log_audit centos1 -wi-a----- 3.00g
var_tmp centos1 -wi-a----- 3.00g
The next thing for us to do is to ensure the LVs under the VG centos1 is mountable:
# for lvname in home root tmp var var_log var_log_audit var_tmp; do
> xfs_repair -L /dev/mapper/centos1-${lvname}> xfs_admin -U generate /dev/mapper/centos1-${lvname}
> done
Once this is done, you should be able to mount these LVs. Remember the /a directory we created earlier. Mount all the relevant filesystems onto /a
# mount /dev/mapper/centos1-root /a
# mount /dev/sdb1 /a/boot
# mount /dev/mapper/centos1-home /a/home
# mount /dev/mapper/centos1-tmp /a/tmp
# mount /dev/mapper/centos1-var /a/var
# mount /dev/mapper/centos1-var_tmp /a/var/tmp
# mount /dev/mapper/centos1-var_log /a/var/log
# mount /dev/mapper/centos1-var_log_audit /a/var/log/audit
# mount -t proc none /a/proc
# mount -o bind /sys /a/sys
# mount -o bind /dev /a/dev
Chroot the /a by running
# chroot /a
Next find out what is the existing uuid for /dev/sda1 and /dev/sdb1.
# blkid | grep sd[ab]1
/dev/sda1: UUID="0f7eaa06-5707-45cc-99db-f3e507c44bcf" TYPE="xfs"
/dev/sdb1: UUID="9b6d885a-ea6d-4c28-997c-fb6f04dcec5e" TYPE="xfs"
Edit the contents of /a/etc/fstab, replace the UUID="<16 digit="" hexadecimal="">" from /dev/sda1 to /dev/sdb1 from the output above.
Change this:
UUID=0f7eaa06-5707-45cc-99db-f3e507c44bcf /boot xfs defaults 0 0
To this:
UUID=9b6d885a-ea6d-4c28-997c-fb6f04dcec5e /boot xfs defaults 0 0
Edit the contents of /a/boot/grub2/grub.cfg and locate the line containing:
### BEGIN /etc/grub.d/10_linux ###
menuentry 'CentOS Linux (3.10.0-1062.1.2.el7.x86_64) 7 (Core)' --class centos --class gnu-linux --class gnu --class os --unrestricted $menuentry_id_option 'gnulinux-3.10.0-1062.1.2.el7.x86_64-advanced-18586e25-cb00-4ef5-b913-e0613ede485a' {
load_video
set gfxpayload=keep
insmod gzio
insmod part_msdos
insmod xfs
set root='hd0,msdos1'
if [ x$feature_platform_search_hint = xy ]; then
search --no-floppy --fs-uuid --set=root --hint-bios=hd0,msdos1 --hint-efi=hd0,msdos1 --hint-baremetal=ahci0,msdos1 --hint='hd0,msdos1' 0f7eaa06-5707-45cc-99db-f3e507c44bcf
else
search --no-floppy --fs-uuid --set=root 0f7eaa06-5707-45cc-99db-f3e507c44bcf
fi
linux16 /vmlinuz-3.10.0-1062.1.2.el7.x86_64 root=/dev/mapper/centos1-root ro crashkernel=auto rd.lvm.lv=centos1/root rd.lvm.lv=centos1/swap rhgb quiet
initrd16 /initramfs-3.10.0-1062.1.2.el7.x86_64.img
}
Replace the uuid marked in yellow to 9b6d885a-ea6d-4c28-997c-fb6f04dcec5e
Amend the vg name from centos to centos1 in the file /a/etc/default/grub. Ensure these lines containing:
GRUB_CMDLINE_LINUX="crashkernel=auto
rd.lvm.lv=centos/root rd.lvm.lv=centos/swap rhgb quiet"
is changed to :
GRUB_CMDLINE_LINUX="crashkernel=auto
rd.lvm.lv=centos1/root rd.lvm.lv=centos/swap rhgb quiet"
Update the grub.cfg file under /a/etc/grub2/grub.cfg
# grub2-mkconfig -o /boot/grub2/grub.cfg
Perform a touch relabel to relabel all selinux context.
# touch /.relable
Exit from the chroot /a environment:
# exit
And now you should be back to the environment on the first disk.
Remember the /a/boot/grub2/grub.cfg entry you modified under the chroot /a environment ?
Use the same entries and edit the first disk's /boot/grub2/grub.cfg
And add the following entry, this entry is similar to what we have on the second disk (which we chroot /a into earlier)
menuentry 'CentOS Linux (3.10.0-1062.1.2.el7.x86_64) 7 (Core) on Disk 2' --class centos --cla ss gnu-linux --class gnu --class os --unrestricted $menuentry_id_option 'gnulinux-3.10.0-1062 .1.2.el7.x86_64-advanced-18586e25-cb00-4ef5-b913-e0613ede485a' {
load_video
set gfxpayload=keep
insmod gzio
insmod part_msdos
insmod xfs
set root='hd1,msdos1'
if [ x$feature_platform_search_hint = xy ]; then
search --no-floppy --fs-uuid --set=root --hint-bios=hd0,msdos1 --hint-efi=hd0,msdos 1 --hint-baremetal=ahci0,msdos1 --hint='hd0,msdos1' 9b6d885a-ea6d-4c28-997c-fb6f04dcec5e
else
search --no-floppy --fs-uuid --set=root 9b6d885a-ea6d-4c28-997c-fb6f04dcec5e
fi
linux16 /vmlinuz-3.10.0-1062.1.2.el7.x86_64 root=/dev/mapper/centos1-root ro crashker nel=auto rd.lvm.lv=centos1/root rd.lvm.lv=centos1/swap rhgb quiet
initrd16 /initramfs-3.10.0-1062.1.2.el7.x86_64.img
}
Save the file /boot/grub2/grub.cfg. And finally reboot to test.
Once reboot, your grub menu should look like this:
And all the above steps can be scripted. No problem.
