more efficient forkchoiceUpdated usage #4055
Conversation
| @@ -47,6 +48,10 @@ type | |||
| # ---------------------------------------------------------------- | |||
| eth1Monitor*: Eth1Monitor | |||
|
|
|||
| # Allow determination of whether there's an upcoming proposal | |||
| # ---------------------------------------------------------------- | |||
| attachedValidators*: ref ValidatorPool | |||
There was a problem hiding this comment.
we need to consider proposers that come in via rest api as well
There was a problem hiding this comment.
ActionTracker could be an alternative that takes care of VCs, no?
func getNextProposalSlot*(tracker: ActionTracker, slot: Slot): Slot
There was a problem hiding this comment.
2809f41 uses getNextProposalSlot
There was a problem hiding this comment.
looking over the code, in addition to the action tracker we should also be checking https://github.com/status-im/nimbus-eth2/blob/7cac6f46db950ed7bdf758f82d3533526a1c1f4f/beacon_chain/spec/eth2_apis/dynamic_fee_recipients.nim which collects calls from https://ethereum.github.io/beacon-APIs/#/Validator/prepareBeaconProposer which is the offiical way of preparing - actiontracker derives "active" proposers from attestations which is a good enough proxy in many cases but not all
There was a problem hiding this comment.
Has risk of false positives, but if people prefer that, sure.
tersec
force-pushed
the
OXg
branch
2 times, most recently
from
88e16cf
to
e93c995
Compare
…slot proposal; use wall slot + 1 rather than state slot + 1 to correctly check when missing blocks
| # Some attached validator is next proposer, so prepare payload. As | ||
| # updateHead() updated the DAG head, runProposalForkchoiceUpdated, | ||
| # which needs the state corresponding to that head block, can run. | ||
| asyncSpawn self.consensusManager.runProposalForkchoiceUpdated( |
There was a problem hiding this comment.
is the spawn here guaranteed to complete before the next fcu call?
There was a problem hiding this comment.
As far as Nimbus is concerned, yes
nimbus-eth2/beacon_chain/consensus_object_pools/consensus_manager.nim
Lines 304 to 319 in e367f61
FORKCHOICEUPDATED_TIMEOUT is 8 seconds.
| # TODO after things stabilize with this, check for upcoming proposal and | ||
| # don't bother sending first fcU, but initially, keep both in place | ||
| asyncSpawn self.runProposalForkchoiceUpdated() | ||
| if self.checkNextProposer(wallSlot).isSome: |
There was a problem hiding this comment.
this is called inside rpfu as well?
No description provided.