Currently, my desktop computer has two storage devices attached: one 1TB NVME SSD, which has both Windows 10 and Linux Mint 21.2 installed on it (Each OS getting ~ 500 GB), and a 1TB SATA hard drive mostly used for Timeshift backups of the Linux Mint partition (Including my Home folder, for the record).
Later today I’m expecting to receive two more 1TB SSDs. When I’ve finished the upgrade process, I’d like to have my Linux Mint installation transferred to a RAID 1 array comprised of the two new drives and expand the Windows 10 partition to take up the whole existing SSD.
My current plan for doing this is to use my existing installation USB drive to install a fresh Linux Mint 21.1 installation on the two new drives, then use Timeshift to ‘restore’ my most recent backup from the existing installation. Is there a better way of going about this that I’m not already aware of?
why not?
sudo cp -ax
foots the bill.I assume people prefer rsync because you may need to run it twice, but unless you tick all the boxes rsync won’t copy capabilities (see
getcap /usr/bin/rsh
)sudo cp -ax is short and sweet and does everything right.
Why not DD? Dd is agnostic to anything, just copy over the entire partition and you’re golden
Man I always forget about dd and jump to gparted.
deleted by creator
dd is good if the destination disk is equal or bigger, unless you are brave enough to shrink the source partition.
if you are moving to a smaller disk for whatever reason (hdd to sdd) then you need to fallback to a different method, which takes us back to cp/rsync.
Because
cat
andcp
are faster if you don’t pick a good block size fordd
.How are you supposed to use
cat
to copy files?cat
ting binary files does, interesting things.You cat the device files of the partitions directly into each other.
cat /dev/sda/ > /dev/sdb
?I’m always hesitant to use the “disk destroyer”, even as a regular user.
rsync
does a good job and it’s maybe even more agnostic thandd
since it doesn’t really care about the partition size, as long as all data fits.rsync
is also more reliable: in cases the transfer is interrupted it only transfers what’s missing and it can run the checksums making sure there were no transfer errors. I don’t see a good reason to usecp
.