[Gluster-users] read access to replicate copies

[Rosario Esposito]

Hello,
first of all I would like to thank the gluster developers for the great 
job they are doing.
I hope someone can give me more details about how replicate copies of 
the same file in a distributed/replicated gluster volumes are accessed 
by clients.
My specific question is: if I have a distributed/replicated volume and a 
gluster native client needs to read a file, which server will be chosen ?

Let's say I have a 2-nodes cluster running the following gluster 
configuration:

---
Volume Name: myvolume
Type: Replicate
Status: Started
Number of Bricks: 2
Transport-type: tcp
Bricks:
Brick1: host1:/brick
Brick2: host2:/brick
---

host1 and host2 are also gluster native clients, mounting "myvolume" in 
/gluster

e.g.

[root at host1 ~]# mount | egrep "brick|gluster"
/dev/sda1 on /brick type ext3 (rw)
glusterfs#host1:/myvolume on /gluster type fuse 
(rw,allow_other,default_permissions,max_read=131072)

[root at host2 ~]# mount | egrep "brick|gluster"
/dev/sda1 on /brick type ext3 (rw)
glusterfs#host2:/myvolume on /gluster type fuse 
(rw,allow_other,default_permissions,max_read=131072)


If host1 needs to read the file /gluster/myfile will it use the local 
copy from host1:/brick or the other copy from host2:/brick over the 
network ?
Is there a way to force the client to read the local copy ?

Cheers, Rosario