# zpool status
pool: zfs
state: ONLINE
scan: scrub in progress since Mon May 11 04:27:02 2015
7.98T scanned out of 7.98T at 42.1M/s, 115970327h32m to go
0 repaired, 100.01% done
When data is written to the pool after scrub begins, more than 100% of the original data will get scrubbed. This confuses the time remaining calculation causing it to seemingly underflow.