AllCoreDevs - Testing CL Breakout #093
Transcript
- Nico Flaig
I merged both prs
- Nico Flaig
yeah just nits before merging
- cayman
Reacted to "I merged both prs" with 🚢
- Eitan Seri-Levi
Jimmy has been testing and uncovered a few small bugs on our end
- Jason Vranek
commit-boost works with kurtosis with loadstar 🙂
- Nico Flaig
Reacted to "commit-boost works w..." with ❤️
- Nico Flaig
lodestar*
- pk910
We have public builders on the devnet: https://buildoor.plataberget.ethpandaops.io/ Pick any on the list (not the overview url itself)
- Dustin
we have a circuit breaker
- Nico Flaig
lodestar has global circuit breaker
- Caleb
Nimbus does
- Nico Flaig
not per builder yet
- Caleb
Reacted to "we have a circuit ..." with 👍
- pk910
lighthouse has one too afaik
- Nico Flaig
our works now, kinda, will be ~2 weeks probably to polish it
- Nico Flaig
I can deathstar do reorgs to test this
Call summary
Decisions
- •Beacon node retains final block selection authority; VC signs and submits whatever Beacon node returns, regardless of bid source - 00:09:44
- •Separate endpoint preferred over merged produce-block endpoint for direct-bid API (PR 627) - 00:04:38
Highlights
- Circuit Breaker Status:
- ·Circuit breaker discussed but no clean test completed; non-finality on devnet triggered informal discussion - 00:17:44
- ·Nimbus, Lodestar (global, not per-builder), and Lighthouse have circuit breakers; Lodestar needs ~2 weeks to polish - 00:18:10
- Trustless Builder Support:
- ·Lighthouse has closed-source trusted builder for devnet testing; trustless builder not yet implemented - 00:19:23
- ·Lodestar currently rejects non-HEAD-compatible bids via gossip rules; flood-publish approach proposed to cover reorg branches - 00:20:06
- ·Public builder (PK910) builds payloads for parent and grandparent, with empty and full payload variants; bids observed on devnet 8 - 00:21:20
- Builder Specs And Beacon Api:
- ·builder-specs PR 165 and Beacon API PR merged; Key Manager API merged last week - 00:01:43
- ·Lodestar tested direct bids with CommitBoost on Kurtosis; not yet on devnet - 00:02:16
- ·Beacon API PR 627 to be rebased today; key open question is separate vs. merged produce-block endpoint - 00:04:18
- ·Consensus: Beacon node should retain all decision-making (circuit breaker, validity); VC signs and submits whatever Beacon node returns - 00:09:44
- ·Beacon node should validate incoming bid before signing; can fall back to local block if bid is invalid - 00:12:56
Action Items
- •Luca (Vero) - Update Beacon API PR 627: rebase, clarify separate endpoint behavior, document VC vs. Beacon node decision responsibilities - 00:04:18
Key decisions
Beacon node retains final block selection authority; VC signs and submits whatever Beacon node returns, regardless of bid source
Consensus that the Beacon node should retain all decision-making (circuit breaker, validity) and the Validator Client (VC) signs and submits whatever the Beacon node returns.Separate endpoint preferred over merged produce-block endpoint for direct-bid API (PR 627)
Technical preference for the Beacon API design regarding the direct-bid API.
AI Disclaimer: Some content or metadata on EIPsInsight may be AI-inferred or automatically compiled. If you find any discrepancy, please contact us at dev@avarch.org.