ACDC #153 summary
Action Items
- If you are a validator of the Hoodi testnet, please get in touch with ethPandaOps for handover if you haven鈥檛 already.
- Prepare for the Hoodi Pectra fork next week: Wednesday, March 26, 2025 2:37:12 PM
- Continue implementation work on PeerDAS
Summary
- Pectra
- The new Hoodi testnet is live!
- Hoodi Pectra fork is: Wednesday, March 26, 2025 2:37:12 PM
- Clients have or are working on releases with Hoodi support
- We touched on post-mortems for the Holesky and Sepolia Pectra upgrades
- Client teams are still in the process of following up on various bugs/issues observed during these testnet upgrades.
- Then we touched on milestones we want to see before setting a Pectra mainnet date. Client teams want to see the following before thinking about a mainnet date:
- Hoodi Pectra going smoothly
- Results from the Pectra bug bounty competition assessed
- Time for users/applications to test their Pectra updates sufficiently
- Next, we turned to an interaction with history expiry and deposit logs
- Expiring history could complicate how clients get deposit logs from the EL which are required until EIP-6110 in the Pectra upgrade.
- This implies we should see Pectra on mainnet before moving forward with the existing history expiry deployment plan to drop pre-Merge history on May 1 2025.
- Given the current timing with Pectra, the expected history expiry date may need to be moved forward until after Pectra.
- Client teams confirmed this issue, and agreed we should wait for Pectra on mainnet.
- And to be precise, there is a transition period that starts at Pectra and finishes when the new 6110 mechanism takes over and the history expiry deadline should be no sooner than the end of this transition period.
- We also touched on the rollout of history expiry so that we go through testnet(s) before mainnet deployment.
- The only relevant testnet is Sepolia, and everyone agreed to test on Sepolia.
- Given the caveat above, we agreed to ship history expiry on Sepolia by the May 1st date, and then once that has been verified can plan a mainnet date for the pre-Merge history expiry.
- To wrap up Pectra, we had an update around performance testing on
pectra-devnet-6with 60M gas- All together, things went well; some minor client interactions were surfaced that caused some issues at that gas limit.
- PeerDAS
- Touched on
peerdas-devnet-5; some ongoing client work to identify various issues.- Check these notes from the breakouts if you want more detail: PeerDAS Breakout Room notes - Google Docs
- On the way to
peerdas-devnet-6, we turned to the question of validator custody that opened an interesting conversation around some implementation questions.- The spec currently mandates validator custody, but leaves room for different conforming implementations.
- There are some implementation details around tracking validators by balance, and how frequently the CL updates its operations to custody the correct set of data, and how the user interfaces with their node to customize this behavior.
- We settled on a preference for automatic, periodic checks that piggyback on existing infrastructure in the client (e.g. the validator registrations used for block building). We may want to add implementation notes to the spec, and work here is ongoing.
- Had an update to wrap up PeerDAS today with ongoing work around the
getBlobsmechanism. Sunnyside Labs is doing some analysis and is seeking multiple client implementations to help ground data.
- Touched on
- Fusaka
- EIP-7688 for SSZ StableContainers was PFI鈥檇 for Fusaka. We agreed to focus on stable Fusaka devnets with the current SFI鈥檇 EIPs before visiting other features. Dmitry from the Lido team gave an overview of how this EIP would be helpful, and Etan from Nimbus highlighted some other benefits of this EIP for the protocol.