Skip to main content
Unix & Linux

Return to Question

added 172 characters in body
Source Link
Adam Ryczkowski
Adam Ryczkowski
  • 5.9k
  • 9
  • 44
  • 62

I have decided to give btrfs raid capabilities a try. I set up a btrfs with

sudo mkfs.btrfs -m raid10 -d raid10 /dev/sda9 /dev/sdb9 /dev/sdc9 /dev/sdd9

Now I want to clone my existing btrfs partition, (which sits on top of linux-raid). can't use a simple cp -a, because there is over 40 snapshot-based backups (which I wish to preserve), and I would simply overfill all the storage I can spare multiple times.

So far I can see two options:

partclone.btrfs -s /path/to/original/fs -o /dev/sda9 -b

and I suppose I would also need to btrfs balance start /dev/sda9

and

do: incrementaly copy with cp -a as much as fits into the storage, and then use bedup to de-duplicate files, and loop.

What is the preferred (i.e. best-practice) method? I would much prefer the first one; it should take much less time. Or maybe there is some evil "gotcha" lurking in any of this procedures (beside the fact that btrfs is experimental, of course)

The first question is simply out of question; however wonderful tool the partclone.btrfs is, it obviously doesn't support multi-device file systems. :-(

I have decided to give btrfs raid capabilities a try. I set up a btrfs with

sudo mkfs.btrfs -m raid10 -d raid10 /dev/sda9 /dev/sdb9 /dev/sdc9 /dev/sdd9

Now I want to clone my existing btrfs partition, (which sits on top of linux-raid). can't use a simple cp -a, because there is over 40 snapshot-based backups (which I wish to preserve), and I would simply overfill all the storage I can spare multiple times.

So far I can see two options:

partclone.btrfs -s /path/to/original/fs -o /dev/sda9 -b

and I suppose I would also need to btrfs balance start /dev/sda9

and

do: incrementaly copy with cp -a as much as fits into the storage, and then use bedup to de-duplicate files, and loop.

What is the preferred (i.e. best-practice) method? I would much prefer the first one; it should take much less time. Or maybe there is some evil "gotcha" lurking in any of this procedures (beside the fact that btrfs is experimental, of course)

I have decided to give btrfs raid capabilities a try. I set up a btrfs with

sudo mkfs.btrfs -m raid10 -d raid10 /dev/sda9 /dev/sdb9 /dev/sdc9 /dev/sdd9

Now I want to clone my existing btrfs partition, (which sits on top of linux-raid). can't use a simple cp -a, because there is over 40 snapshot-based backups (which I wish to preserve), and I would simply overfill all the storage I can spare multiple times.

So far I can see two options:

partclone.btrfs -s /path/to/original/fs -o /dev/sda9 -b

and I suppose I would also need to btrfs balance start /dev/sda9

and

do: incrementaly copy with cp -a as much as fits into the storage, and then use bedup to de-duplicate files, and loop.

What is the preferred (i.e. best-practice) method? I would much prefer the first one; it should take much less time. Or maybe there is some evil "gotcha" lurking in any of this procedures (beside the fact that btrfs is experimental, of course)

The first question is simply out of question; however wonderful tool the partclone.btrfs is, it obviously doesn't support multi-device file systems. :-(

Source Link
Adam Ryczkowski
Adam Ryczkowski
  • 5.9k
  • 9
  • 44
  • 62

How to clone btrfs filesystem into different medium preserving snapshots' sharing data

I have decided to give btrfs raid capabilities a try. I set up a btrfs with

sudo mkfs.btrfs -m raid10 -d raid10 /dev/sda9 /dev/sdb9 /dev/sdc9 /dev/sdd9

Now I want to clone my existing btrfs partition, (which sits on top of linux-raid). can't use a simple cp -a, because there is over 40 snapshot-based backups (which I wish to preserve), and I would simply overfill all the storage I can spare multiple times.

So far I can see two options:

partclone.btrfs -s /path/to/original/fs -o /dev/sda9 -b

and I suppose I would also need to btrfs balance start /dev/sda9

and

do: incrementaly copy with cp -a as much as fits into the storage, and then use bedup to de-duplicate files, and loop.

What is the preferred (i.e. best-practice) method? I would much prefer the first one; it should take much less time. Or maybe there is some evil "gotcha" lurking in any of this procedures (beside the fact that btrfs is experimental, of course)