[Sheepdog] Few things

[MORITA Kazutaka]

At Sun, 22 May 2011 13:38:19 +0200,
krimson wrote:
> 
> Just ran into another issue, if I abort this operation:
> 
> qemu-img convert /dev/vmvg/ub01 sheepdog:ub01
> 
> this seems to result in corruption:
> 
> # collie vdi list
>    name        id    size    used  shared    creation time   vdi id
> ------------------------------------------------------------------
> failed to read a inode header 2701199, 0, 2
> failed to read a inode header 10955677, 0, 2
> failed to read a inode header 13864185, 0, 2
> 
> I have to collie cluster format before I can use the store again.

In my environment, this problem couldn't happen.  How many physical
machines do you use for Sheepdog?  What is the number of redundancy?
Does this problem happen always?  Does the sheep.log contain any error
messages?

Thanks,

Kazutaka