Thomas mentioned in the comment in IMPALA-7184,
- I create a Kudu table, insert some stuff into it, scan it back at READ_YOUR_WRITES. Everything works as expected. - I wait greater than 'tablet_history_max_age_sec' and attempt to scan it again at READ_YOUR_WRITES (using either the same KuduClient or a new one). This results in an error of the form 'Snapshot timestamp is earlier than the ancient history mark...'
Today, the snapshot used to scan in RYW mode is max(propagated timestamp + 1, clean time). If the tablet hasn't been written to recently, a tablet's clean time can fall well before the AHM. If there are no ops in flight, it might be worth considering some other timestamps, e.g. now, now minus the AHM, the current safe time, etc., or escalating a RYW scan to READ_LATEST if the propagated timestamp has passed.
- is related to
IMPALA-7184 Support Kudu's READ_YOUR_WRITES scan mode