Description of problem:
-------------------------

When remove-brick fails on a volume, because of brick process being killed, the user is still required to perform either commit or stop for that failed remove-brick task, before another task can be started on the volume.

After the failure of a task, the user should not be expected to perform further operations on that task before being able to start other tasks.

Version-Release number of selected component (if applicable):
glusterfs-3.4.0.49rhs

How reproducible:
Always

Steps to Reproduce:
1. Start remove-brick operation on a volume.
2. While remove-brick is in progress, kill a brick process.

Actual results:
The remove-brick task fails. But the user is expected to perform either stop or commit, before another task can be started on the volume.

Expected results:
User should not be expected to perform further operations on a failed task.

Additional info: