All uses of Epoched past epoch tracking and handling of last_update may not be consistent

[brentstone]

A storage dump from epoch 8 at height 35201 had the following storage ket structure for a validator:

"#tnam1qgqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqc8j2fp/validator/#tnam1q96tsxckvm4n937efh7t78w4j40m5783wsqcz3ye/state/last_update" = "0900000000000000"
"#tnam1qgqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqc8j2fp/validator/#tnam1q96tsxckvm4n937efh7t78w4j40m5783wsqcz3ye/state/lazy_map/data/000000000000E" = "00"
"#tnam1qgqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqc8j2fp/validator/#tnam1q96tsxckvm4n937efh7t78w4j40m5783wsqcz3ye/state/lazy_map/data/000000000000G" = "01"
"#tnam1qgqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqc8j2fp/validator/#tnam1q96tsxckvm4n937efh7t78w4j40m5783wsqcz3ye/state/lazy_map/data/000000000000I" = "00"
"#tnam1qgqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqc8j2fp/validator/#tnam1q96tsxckvm4n937efh7t78w4j40m5783wsqcz3ye/state/oldest_epoch" = "0700000000000000"

I would not have expected last_update = 9 and would've expected at least one more epoch in the past stored, since pipeline_len epochs should be kept in the past for the ValidatorStates. It is possible that function calls that manipulate this data may not be performed consistently.

Check that all these operations are consistent. Furthermore, recheck that the NestedEpoched objects like the validator sets and positions and addresses are storing what is expected.