Gluster users - Sep 2008 - Adding a new storage brick to cluster/afr

Hi,

I'm running gluster using cluster/afr.  This has been running fine
across a few nodes, but I need to add a couple more.  I've made the
necessary config changes, restarted glusterfsd and remounted the
shares.  However, whilst new files are stored on all bricks, the new
bricks are not healing with other files upon access, despite the
gluster client being told to read locally via "option read-subvolume".

Can anyone suggest what I should do to sync the missing files, and how
to avoid this pitfall in future?

Thanks,

Hi Rob,

Before adding the new server to the spec file do this:
(if you are using the latest source)
set trusted.glusterfs.version to "1" for the directories.
set trusted.glusterfs.afr.<newvolname> to 1 for the files.

You would probably write a script to do that recursively.

Next you can add the new vol to the spec file and remount
the FS and execute a script to open all the files in the FS
and see if it selfheals. Let us know if this procedure does
not work.

Thanks
Krishna

On Mon, Sep 15, 2008 at 8:03 PM, Rob S <rtshilston at gmail.com>
wrote:
> Hi,
>
> I'm running gluster using cluster/afr.  This has been running fine
> across a few nodes, but I need to add a couple more.  I've made the
> necessary config changes, restarted glusterfsd and remounted the
> shares.  However, whilst new files are stored on all bricks, the new
> bricks are not healing with other files upon access, despite the
> gluster client being told to read locally via "option
read-subvolume".
>
> Can anyone suggest what I should do to sync the missing files, and how
> to avoid this pitfall in future?
>
> Thanks,
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users
>