Description of problem:

Have following volume configuration:

Volume Name: vol1
Type: Distributed-Disperse
Volume ID: 44c0b7fa-62b6-4704-9819-57f1aac3c168
Status: Started
Number of Bricks: 2 x (4 + 2) = 12

Now, if I fail more than supported number of bricks  and recover them back, I see that the operations (like ls)from the clients hang. There is not way to get out of this state. 

Version-Release number of selected component (if applicable):

glusterfs-3.7dev-0.885.git0d36d4f.el6.x86_64

How reproducible:

1. Create a dist. disperse
2. Mount it to a client and start I/O.
3. Failed multiple bricks and bring is back online
4. Do "ls" on the mount 


Actual results:

Mount hangs on the client

Expected results:

No hang

Additional info: