On Thu, 2005-06-23 at 09:58 +0200, Nicolas Mailhot wrote:
On Jeu 23 juin 2005 09:30, David Woodhouse wrote:
> On Wed, 2005-06-22 at 08:46 -0500, Chris Adams wrote:
>> Some type of "journalling RAID" would be a possible solution (and
>> would also allow for much faster re-syncs on unclean shutdown, as only
>> the last written blocks would need updating).
>
> This is why RAID is entirely the wrong answer, and redundancy should be
> implemented in the file system itself, instead of being hacked into the
> block layer.
Can't this be done today with something that takes lvm snapshots each night ?
lvm snapshots right now don't seem fully baked to me; too many times
I've had the kernel run out of memory while fiddling with them.
Maybe it's just me...
--
Peter