[Gluster-users] master volume
sacharya at redhat.com
Thu Aug 15 03:42:43 UTC 2019
Consider that you have a volume, named vol1. Now you want to have a replica
of it, for disaster recovery. What do you do? You will create a new volume,
say, repica-vol1 in a different cluster.
To replicate data from vol1 to replica-vol1, you will set up a geo-rep
session between vol1 and replica-vol1. In geo-replication, we call your
primary volume vol1 as master. The replicated volume replica-vol1 is called
slave. We call it master and slave because you will see unidirectional sync
of data from master volume(vol1) to slave volume(replica-vol1).
Hope this helps.
On Thu, 15 Aug 2019, 3:35 am richard lucassen, <mailinglists at lucassen.org>
> On Wed, 14 Aug 2019 20:24:53 +0530
> Aravinda Vishwanathapura Krishna Murthy <avishwan at redhat.com> wrote:
> > > > If Master and Slave Volumes are in
> > > > the same cluster then it is not so helpful when disaster happens.
> > >
> > > So, do I need to do something like this?:
> > >
> > > cluster volume name: gvol0
> > > replication vol name: rep-gvol0
> > Volume names can be kept as your convenience.
> The doc is quite unclear for me as a gluster newbie. It's the old Linux
> paradox: the moment that you will understand the docs is the moment
> that you don't need them anymore.
> Why is there a MASTER_VOL and a SLAVE_VOL? If I have a volume called
> "gvol0" it seems logical to me that the geo-replication slave server
> also uses this volume name "gvol0". The geo-replication server knows it
> is not part of the active cluster because it was told to do so.
> OTOH, I think there must be an obvious reason for this MASTER_VOL and
> SLAVE_VOL, but I don't understand this from the docs.
> richard lucassen
> Gluster-users mailing list
> Gluster-users at gluster.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Gluster-users