Observation
Many clients will store caching layers on top of the database for state. Some of these caches will flush at certain intervals like after 128 blocks.
Suggestion
This is an issue to track testing these caching mechanisms. One thing in particular is that rpc methods should not use the MPT cache post flip.
I would recommend triggering cache invalidations etc, though this requires knowing the caching mechanism for each client to do effectively
Observation
Many clients will store caching layers on top of the database for state. Some of these caches will flush at certain intervals like after 128 blocks.
Suggestion
This is an issue to track testing these caching mechanisms. One thing in particular is that rpc methods should not use the MPT cache post flip.
I would recommend triggering cache invalidations etc, though this requires knowing the caching mechanism for each client to do effectively