Description of problem:
=======================
remove-brick commit is happening when rebalance failed due to one of volume brick down.


Version-Release number of selected component (if applicable):
=============================================================
glusterfs-3.7.9-2.el7


How reproducible:
=================
Always


Steps to Reproduce:
===================
1. Have two node cluster with distribute volume (3 bricks)
2. FUSE mount and write enough data 
3. bring down one of brick //kill -9 brick_pid
4. remove one active brick (remove-brick start)
5. Check remove-brick status // you will see rebalance failure
6. Try to do remove-brick commit //will get committed and you will lose the data.

Actual results:
===============
Remove-brick commit is happening without rebalance success


Expected results:
=================
remove-brick commit should not happen if rebalance failed.

Additional info: