| 1 |
On Tue, May 27, 2014 at 7:25 AM, Mick <michaelkintzios@×××××.com> wrote: |
| 2 |
> |
| 3 |
> Also how big is each snapshot of / and why are these necessary on an hourly |
| 4 |
> basiszfs ? |
| 5 |
|
| 6 |
Btrfs is COW, so snapshots only consume space as files change. If you |
| 7 |
have a read-only filesystem and snapshot it hourly the only space |
| 8 |
consumed by a snapshot will be a few metadata records. |
| 9 |
|
| 10 |
Snapshotting hourly would mostly be a convenience - in theory it |
| 11 |
should get you time-machine-like functionality just like hourly |
| 12 |
backups would, but with far less overhead and space use. |
| 13 |
|
| 14 |
In practice I stopped doing this, as btrfs can misbehave when you |
| 15 |
start getting a lot of snapshots accumulated (we're talking |
| 16 |
thousands). It probably doesn't help that I have VM images |
| 17 |
snapshotted (though these images have fairly low write volumes - the |
| 18 |
most active one does most of its writing to an nfs volume so only OS |
| 19 |
updates, logs, etc change the VM). When snapper would go to cleanup |
| 20 |
snapshots I'd get panics. I ended up having to write a script that |
| 21 |
deleted one snapshot every 30min over the course of days to clean up |
| 22 |
from that. Now I only manually snapshot periodically and I haven't |
| 23 |
had a problem with it. |
| 24 |
|
| 25 |
I suspect that as with many things btrfs-related that it will be |
| 26 |
worked out in time, though snapshots will always cause fragmentation |
| 27 |
as long as the filesystem does partial diffs. |
| 28 |
|
| 29 |
Rich |
Archives