Description of problem: ======================= We make it mandate to stop the geo-rep session before performing remove-brick commit with the successful migration of data with start. But in a case we remove the brick from each subvolume, in other words if we decrease the replica count from 3 to 2 for example, both remove-brick start and commit will fail with usage to use "force". By using remove-brick force we bypass the validation of geo-rep and remove the bricks instantly which is expected with force, but this wont kill the gsyncd for all the removed-bricks. Hence, a note section in document is needed to cover this corner case. Version-Release number of selected component (if applicable): ============================================================= glusterfs-3.7.1-1.el6rhs.x86_64