Gluster users - Jul 2011 - How can I user strpe + replicate mode in gluster 3.2.1?

Hey,

I can write the .vol in the client side to use the stripe + replicate mode to
store large files safely in the previous version. Now, I tested the latest
version gluster 3.2.1with the command line mode. But I find that I can't use
the stripe + replicate mode now. How can I use the stripe + replicate mode to
store data in the lastest version 3.2.1?

Wei Wu

Hi Wei Wu,

We currently (ie, all releases till 3.2.2, including 3.1.x), don't support
the stripe+replicate mode. The feature is already present in mainline
repository. (git's master branch). This feature will be officially available
with 3.3.0 version.

Regards,
Amar

On Mon, Jul 11, 2011 at 9:15 AM, ?? <wuwei.tsh at gmail.com> wrote:
> Hey,
>
> I can write the .vol in the client side to use the stripe + replicate mode
> to store large files safely in the previous version. Now, I tested the
> latest version gluster 3.2.1with the command line mode. But I find that I
> can't use the stripe + replicate mode now. How can I use the stripe +
> replicate mode to store data in the lastest version 3.2.1?
>
> Wei Wu
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://supercolony.gluster.org/pipermail/gluster-users/attachments/20110711/7bbddf59/attachment.html>

On Mon, Jul 11, 2011 at 11:45:28AM +0800, ?? wrote:
> I can write the .vol in the client side to use the stripe + replicate mode
to store large files safely in the previous version. Now, I tested the latest
version gluster 3.2.1with the command line mode. But I find that I can't use
the stripe + replicate mode now. How can I use the stripe + replicate mode to
store data in the lastest version 3.2.1?
I obtained it by using "gluster volume create" with "repliace
2"
and specifying 4 bricks. There is a trap to avoid: the order in
which you specify the bricks is important: the two first bricks
shall be on two different servers. gluster will not be smart if
you specify the two first bricks on the same server:  you will have
data replicated on the same machines, and therefore no failover.

-- 
Emmanuel Dreyfus
manu at netbsd.org