[Gluster-users] Force resync between two nodes with 3.0.2 version

Stephan von Krawczynski skraw at ithnet.com
Wed Mar 10 07:26:10 UTC 2010


On Tue, 9 Mar 2010 13:19:58 -0500
Benjamin Long <benjamin.long at longbros.com> wrote:

> On Tuesday 09 March 2010 12:24:29 pm carlopmart wrote:
> > Hi all,
> > 
> >   I have two Rhel5.4 servers with glusterfs version 3.0.2 installed. Due to
> >  hardware failure, node2 have remained shutdown over 2 days. During this
> >  time, node1 have incremented his data over 70GB. I have startup node2 but
> >  info doesn't resyncs (node2 have created all new files with 0 bytes). Is
> >  this normal?? How can I resync all info between nodes??
> > 
> >   I have tried to launch a "ls -ltR" from one client without luck.
> > 
> > Many thanks.
> > 
> 
> I'd assume it should sync that data, but If it's that far gone I would use 
> rsync to bring the nodes in sync. I'm sure rsync would be a lot faster since 
> it doesn't have the overhead that gluster does.

My understanding is that you cannot rsync the data on the backend side to
speed up replication because this will confuse glusterfs further on.


-- 
Regards,
Stephan




More information about the Gluster-users mailing list