Recent versions of ZFS have solved the problem of drives ignoring the synchronize cache command. Now ZFS always keeps around the latest 3 transaction groups, r

Recent versions of ZFS have solved the problem of drives ignoring the synchronize cache command. Now ZFS always keeps around the latest 3 transaction groups, regardless if any data on any one of those transaction groups has been freed/updated already. So if the last transaction group gets corrupted during an abrupt power down, it can always go back to the latest consistent one.