[sheepdog-users] re-start a cluster using md

Valerio Pachera sirio81 at gmail.com
Fri May 3 17:10:21 CEST 2013


2013/5/3 Liu Yuan <namei.unix at gmail.com>

> On 05/03/2013 10:45 PM, Valerio Pachera wrote:
> > May you add one or more check on this:
> >   check if the first directory (/meta-store) actualy contains the
> > subfolder config
> >   check if other object dirs contains only sheepdog object files.
> >
> > This to reduce the risk of messing up the cluster because of a simple
> > typing error.
>
> Ah, good suggestions, I'll cook a patch based on your suggestions and
> abort when find something unexpected for simplicity. Won't try to be
> smart to do auto-path-parse because it deals with the data, one wrong
> step might compromise the data integrity.
>


Right, just exit with an error so the administrator can check what's wrong.

(It may be also good checking that not folder get repeated: sheep
/mnt/sheepdog/,/mnt/sheepdog/,/mnt/sdb1).
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.wpkg.org/pipermail/sheepdog-users/attachments/20130503/bc47cad2/attachment-0005.html>


More information about the sheepdog-users mailing list