Lodestar Standup 2026-10 (March 10, 2026)

[matthewkeil]

Lodestar Standup

Tuesday March 10, 2026, · 3:00 – 4:00pm UTC
Video call link: https://meet.google.com/fwq-bhrs-pey

Please post any topics/comments below and we will add it to the stand-up discussion. Examples are issues, discussions or proposals that require feedback, consensus from the team and/or community members.

Agenda

Announcemnets

N+1 Fork (Glamsterdam)

• Devnet-0 launched with no fanfare
• Following devnet and are healthy. (Teku and Nimbus are not following. LH cannot sync. Prysm is missing blocks)
• Nico is keeping up with interop issues
• Nico found bug spec computing the PTC, spec discussion ongoing
• Bigger PR's under review (state caches and payload input)
• Some other small PRs have been put up as well (unknown and range sync, reqresp, checkpoint sync)
• Goal for before end of devnet-0 is to get everything from devnet-0 scope onto unstable

FOCIL

• notified breakout that we need to wait for devnet-0 branch to get merged to unstable before we can rebase

Non-Fork Protocol Features

Fast Confirmation Rule

• performance regressions mostly resolved but still needs tuning

Partial Messages & Cell-Level DAS

• actively kurtosis debugging

Execution Proofs

• @spiral-ladder researched the spec and looked at the other client implementations
• analyzed lodekeeper implementation and thinks there might be some pieces of our impl that are missing
• first meeting is tomorrow and @spiral-ladder will attend from our team

6-Second Slots

• no updates. need to rebase on gloas

Other Features

BLST

• Found perf regression and got perf up to blst-ts speed!!! 🎉
• Bindings PR is up and under review
• PR to Lodestar consuming the new blst-z
• Unlocks all the other TS performance updates (pubkey cache, no TS worker pool)

Flat-Files for Data-Columns

• pushed updates responding to all updates and ready for another round of comments
• will soak for next release
• maybe look at migrations next week but no rush
• ready for review by @lodekeeper. will let soak and do a metrics analysis when we are ready

Quic

• Found the segfault!!!! Got it fixed
• Got "fix" actually fixed and has been soaking
• Perf looks good!!! 🚀
• ping @AgeManning

libp2pV3 Memory Leak

• regressions were found and lodekeeper has been working on a couple of avenues
• taking heap snapshots to try and get it isolated
• will holdup the release until fixed

Segfault Discussion

• its a red herring. root cause was on shutdown from panic of quic segfault

TSgo

• @nazarhussain is on the mission and worked through the details
• PR is up and under review

Zig Roadmap

• Experimented with struct storage optimization in tree view. Great perf improvement
• Worked on fuzzing for blst-z. No crashes
• Currently working on generating extra test data to seed more fuzz tests in the future
• Looking at approaches to refactor common code
• Can look at migrating to the zig-std fuzzer when we move to v0.16 but for now need the external lib for completeness
• Working on fork-choice-z. Built out plan of attack and setup milestones for development. Researched spec and current implementation. Started building the first pieces of the implementation. First estimate for timeline is 4 weeks.

Devops/Infrastructure

• Vault is no longer needed for deploying to dev servers
• Working on getting bots access to the hosts
• Project Butter is up for discussion and we are going to push it forward
• New repo created for dev servers lodestar-ansible
• PR for modernization and server lifecycle under review
• New account made for our VPS that we have full control over

[matthewkeil]

Meeting minutes:
https://docs.google.com/document/d/1rhYBmxGhodhbL7Aytr0eFiO9dzTthyG-r10Hb2NFVtA/edit?usp=sharing

Recording:
https://drive.google.com/file/d/1qE1WqDol5FMJTqilS-vGk6Wah-AO4UHg/view?usp=sharing

Full transcript:
https://docs.google.com/document/d/1rhYBmxGhodhbL7Aytr0eFiO9dzTthyG-r10Hb2NFVtA/edit?usp=sharing