Vijay Bellur wrote:
On 11/08/2012 03:55 PM, Joop wrote:
> Continuing my quest for a system consisting of oVirt en gluster I came
> across the following.
> I'm using the latest nightlies and had tried something with gluster but
> it didn't work out. So I tried starting over en did a stop on my gluster
> volumes, that went OK, removed them, that went OK too.
> Then I tried to create a nieuw Volume using the same name and same
> location and got an unexpected exception. Look at the folder which had
> contained the data I still see a .glusterfs folder but all other data is
> gone. Ok.
> Googling I found that the following two commands should give me the
> ability to reuse that folder for glusterfs.
> setfattr -x trusted.glusterfs.volume-id /gluster-data
> setfattr -x trusted.gfid /gluster-data
> Tried again from the webui and the volume was created and I could
> start it.
> Stopped it, removed it and tried to recreate, same error. See the
> attached logs.
>
This is expected behavior as you need to be careful while re-using
bricks across different volume types. You can actually have a striped
volume and use a brick from there to create a replicated volume after
the striped volume has been deleted. If the brick contains data from
the striped volume, then the results of accessing such a volume from a
client mount point would not be desirable. xattr removals are
necessary so that the admin is aware of the consequences of re-using
bricks across different volume types.
Thanks,
Vijay
Sounds sensible but I haven't found it documented anywhere, even on
gluster.org I couldn't find it.
Is it possible to catch this error and present it to the admin telling
him what todo and why. Unexpected error is a bit errr unexpected ;-)
Joop