Yes, it is quite shocking - and doesn't make much sense to me, either.
Instead, you can have the servers be responsible for AFR, and let the
clients be unaware of each other. This solves the n-fold bandwidth
issue entirely (and seems more "normal").
Are you saying there is a way to configure gluster to do this? How?
You can load all the translators either on client or on serverside by that
"programming" how the Gluster works...
For example:
On the server (and you do the same on the other storage server) you export
one local volume (local directory) and grab one remote volume and make AFR
volume over both and export that to clients..
.
That way the afr-copying is done on the "server" level..
Only drawback in this is your clients can mount only one server so if one
goes down you have to remount or implement some other HA method (like dns or
heartbeats)).
rr
