Brick sizes after rebalance aren't as uniform as expected in RHS

Issue

After adding two bricks to an existing replica 2 volume and performing a rebalance, the distribution of data as shown by the df command doesn't appear to be as uniform as expected.

# for i in {01..12}; do ssh node${i} 'df -h /rhs/brick1'; done
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/RHS_vg1-RHS_lv1
                       27T   18T  9.0T  67% /rhs/brick1
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/RHS_vg1-RHS_lv1
                       27T   20T  7.1T  74% /rhs/brick1
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/RHS_vg1-RHS_lv1
                       27T   18T  9.2T  66% /rhs/brick1
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/RHS_vg1-RHS_lv1
                       27T   20T  7.7T  72% /rhs/brick1
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/RHS_vg1-RHS_lv1
                       27T   21T  6.3T  77% /rhs/brick1
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/RHS_vg1-RHS_lv1
                       27T   22T  5.6T  80% /rhs/brick1
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/RHS_vg1-RHS_lv1
                       27T   17T   10T  63% /rhs/brick1
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/RHS_vg1-RHS_lv1
                       27T   17T   10T  63% /rhs/brick1
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/RHS_vg1-RHS_lv1
                       27T   14T   14T  50% /rhs/brick1
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/RHS_vg1-RHS_lv1
                       27T   14T   14T  50% /rhs/brick1