Four snappy DUTs are stuck in "Repairing":

$ atest host list -b board:snappy,pool:bvt -s Repairing | awk '{print $1}'
Host
chromeos6-row4-rack20-host2
chromeos6-row4-rack20-host6
chromeos6-row4-rack20-host14
chromeos6-row4-rack20-host16

They're being reported in the pool inventory e-mail:
Status for pool:bvt, by model:
Model                    Bad  Idle  Good Total
[ ... ]
snappy                     0     4     7    11
[ ... ]
Idle Host List:
Hostname                       Model                Pool
chromeos6-row4-rack20-host2    snappy               bvt
chromeos6-row4-rack20-host6    snappy               bvt
chromeos6-row4-rack20-host14   snappy               bvt
chromeos6-row4-rack20-host16   snappy               bvt

They've been like this for a while.