[Gluster-users] geo-replication maintenance

[Venky Shankar]

On Apr 23, 2014, at 4:18 PM, John Ewing <johnewing1 at gmail.com> wrote:

> Hi,
> 
> I have a geo-replicated volume using 3.4.3 and I was wondering about what happens when I need to perform maintenance on the master and the slave.
> 
> If I want to reboot the slave , should I stop geo-replication on the master first?

You can reboot the slave without stopping geo-replication. If you do this, geo-rep session status would turn faulty and geo-rep would try to connect to the slave periodically. After the slave node is back online, things would be fine thereafter.

> 
> When the slave is rebooted is a full re-crawl / re-sync needed or will it just pick up where it left off?

Geo-rep would pick up where it left off.

> 
> Similarly if I need to do work on the master what's the correct procedure and will it need a full recrawl ?

In this case it’s good to stop the geo-rep session. After it’s back up, it will continue from where it left off. No full crawl is required. OTOH, do you plan to use slave (for writes, etc.) when master is offline? In that case geo-rep supports failover-failback, where a reverse sync (from slave to master) can be done efficiently.

> 
> Thanks
> 
> J.
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://supercolony.gluster.org/mailman/listinfo/gluster-users