Bug 987418

Description of problem:
quotad process runs only on one node of the cluster, this is a major issue.
as if the node having quotad running goes down or quotad gets killed
we will have no one to take care about quota policy.

Version-Release number of selected component (if applicable):
[root@nfs1 ~]# rpm -qa | grep glusterfs
glusterfs-3.4.0.12rhs.beta5-2.el6rhs.x86_64
glusterfs-fuse-3.4.0.12rhs.beta5-2.el6rhs.x86_64
glusterfs-server-3.4.0.12rhs.beta5-2.el6rhs.x86_64


How reproducible:
always

Steps to Reproduce:
setup 4 rhs nodes[1, 2, 3, 4], 1 client
1. create a volume, start it
2. enable quota on root of the volume on rhs-node1
3. set limit of 1GB 
4. mount over nfs from node3
5. start creating data. create 1MB file tille quota limit gets exceeded.
6. kill quotad on node1.

let I/O keep happening.

Actual results:
the quota limit is surpassed 

[root@rhsauto030 nfs-test]# du -h .
2.8G	.
[root@rhsauto030 nfs-test]# pwd 
/mnt/nfs-test
[root@rhsauto030 nfs-test]# mount | grep nfs-test
10.70.37.196:/quota-dist-rep on /mnt/nfs-test type nfs (rw,addr=10.70.37.196)
[root@rhsauto030 nfs-test]# 


[root@nfs1 ~]# ps -eaf | grep quotad
root      6709  2366  0 05:43 pts/0    00:00:00 grep quotad
[root@nfs1 ~]# 


Expected results:
we should have mechanism to make it highly available or distributed.

Additional info: