[sheepdog] [PATCH v2] zookeeper: exit program on unrecoverable error

MORITA Kazutaka morita.kazutaka at gmail.com
Wed Jun 12 13:35:37 CEST 2013


At Wed, 12 Jun 2013 10:04:44 +0800,
Kai Zhang wrote:
> 
> Hi
> 
> Could you wait for my patch about improving zookeeper driver?
> 
> Actually, my patch is to fix some fatal errors about current zookeeper, which including:
> - change the way of deciding master to avoid error when concurrent start up
> - wait for master transfer other than exit
> - rejoin cluster when session timeout
> 
> My patch is almost done and under testing. I will submit for reviewing in this week.
> 
> I think it would be better that we fix fatal errors before this patch.

Well, why is it better?  Is it easier to rebase my patch onto yours?
My patch fixes a critical problem that epoch information will be
corrupted after session timeout.  Currently, there is no way to fix
the broken epoch other than re-formatting the cluster.

If you could review my patch and add your Reviewed-by tag, I would
apply this patch asap.  Yuan is offline this week, so reviews from
other developers are highly appreciated.  But if there is a reason
your patch should be applied first, I'll wait for your patch. :)

Thanks,

Kazutaka



More information about the sheepdog mailing list