Migration from the KQ Implementation

Gordan Bobic gordan.bobic at gmail.com
Wed May 4 20:50:28 EDT 2011


On 04/05/2011 23:41, Uncle Stoatwarbler wrote:
> On 04/05/11 03:34, Brian Behlendorf wrote:
>> We don't have any plans for an offline fsck-like tool.
>>
>
> I really wish you'd reconsider that.
>
> It's hard to sell ZFS as a possibility in my $orkplace if there's no 
> last resort option for data recovery (several hundred Tb involved)
>
> ($ork is a university space lab)
>

I'm also pretty sure that whenever any FS was pushed for kernel 
inclusion, fsck was one of the mandatory requirements. Even for log 
structured file systems like nilfs2 that should in theory always be 
usable even if the most recent writes got corrupted.

The simple fact is that data going missing in crashes isn't the only 
thing that can lead to requiring a fsck tool.
Disks (both mechanical and flash) can lose and/or corrupt data, and no 
degree of crash-proof safety will make up for that.

Gordan



More information about the zfs-discuss mailing list