Raid vs rsync -

[Chris Murphy]

On Tue, Mar 10, 2015 at 1:43 PM, Suvayu Ali <fatkasuvayu+linux at gmail.com> wrote:

> In my encouter with data corruption in the past, it has usually been
> entire drive failures rather than one particular filesystem failing on
> me, and the other keeps functioning.
>
> I would always encourage separate physical disks as backup partitions.
> If the OP has flaky power, maybe having them offline when not in use,
> would also be a good idea.

This is quite good advice.

It's one reason why I do use Btrfs. But I have multiple separate Btrfs
backups (one is raid0, one is raid1). One backup is on HFS+ just
because if I had to do a quick restore of the Mac, this is push a
single button type restore, but of course is subject to SDC concerns.
And before I was confident in both Btrfs and my ability to repair it
should that be necessary, I kept two additional backups on XFS on
separate drives and file systems. And then the 6th is select data in
an encrypted image stored in the cloud, which is yet again a separate
physical device and filesystem.

The purpose of Btrfs raid1 is not so much avoiding downtime for me,
but to take advantage of automatic detection of corruption and
autohealing. The raid0 is mainly for size, of course it totally dies
if anyone device dies. But in the meantime, I still get notifications
of any corrupt files (by full filename path) should that happen.


-- 
Chris Murphy