Bring back the old behavior
Signed-off-by: Sasha Bogicevic <[email protected]>
Signed-off-by: Sasha Bogicevic <[email protected]>
> because we have no recent view of the chain since we are out of sync
> NodeState now tracks currentChainPoint :: ChainPointType tx, instead of currentSlot :: ChainSlot + currentChainTime :: Maybe UTCTime > NodeSynced and NodeUnsynced now carry ChainPointType tx, instead of chainTime :: UTCTime > extended IsChainState class with chainPointTime :: ChainPointType tx -> UTCTime > initialChainTime = posixSecondsToUTCTime 0
This reverts commit 037aea1d9f15f65db00113f0606f9e09d0b03548.
> when calculating the slot drift
* extended NodeSynced and NodeUnsynced with time and slot drifts * added Haddocks to both currentSlot and currentChainTime * made NodeState.currentChainTime non optional * defined initialChainTime = posixSecondsToUTCTime 0
This reverts commit d82108a1af55afbbad813711b4ca0ae4eab13cf0.
> NodeState now tracks currentChainPoint :: ChainPointType tx, instead of currentSlot :: ChainSlot + currentChainTime :: Maybe UTCTime > NodeSynced and NodeUnsynced now carry ChainPointType tx, instead of chainTime :: UTCTime > extended IsChainState class with chainPointTime :: ChainPointType tx -> UTCTime > initialChainTime = posixSecondsToUTCTime 0
We encountered a limit of maximum messages we can submit through the etcd-backed network layer (2MB) when we had scenarios with expensive to validate or just lots of transactions queuing up. This led to very big snapshots which eventually hit that limit. Limiting the number of transactions in a snapshot effectively results in a bounded block size for the Hydra L2. It must not be configurable or it becomes a protocol parameter.