Content
@
https://warpcast.com/~/channel/stakediva
0 reply
0 recast
0 reaction
Diva Staking
@stakediva
The Pectra upgrade on Holesky exposed critical validator issues Per the postmortem (https://github.com/ethereum/pm/blob/master/Pectra/holesky-postmortem.md), 3 of 5 Execution Layer clients used mainnet’s deposit contract address instead of Holesky’s, splitting the chain’s perception across clients. Let’s see all the details, and what is Diva’s situation 🧵👇
1 reply
1 recast
1 reaction
Diva Staking
@stakediva
1/ Initial mitigation—shutting down misconfigured clients, clearing databases, and resyncing with correct peers—failed. The Consensus Layer issue compounded the problem: an invalid chain state was justified at an incorrect epoch, creating a cascading failure.
1 reply
0 recast
0 reaction
Diva Staking
@stakediva
2/ Validators attesting to the justified invalid checkpoint faced “surround” slashing conditions when pivoting to the valid chain. With consensus clients favouring the attestation-heavy invalid chain, the valid chain struggled with low block production and peer support, stalling recovery.
1 reply
0 recast
0 reaction
Diva Staking
@stakediva
3/ Natural recovery via inactivity leak (https://eth2book.info/capella/part2/incentives/inactivity/) is possible—validators lose balance until ejected below 16 ETH effective stake. However, this requires ~3 weeks and significant capital loss. Current efforts lean toward mass slashing to expedite finality.
1 reply
0 recast
0 reaction
Diva Staking
@stakediva
4/ In our case, Pectra lowers initial slashing penalties to 0.008 ETH from 1 ETH, but a correlation penalty after 18 days (https://eth2book.info/capella/part2/incentives/slashing/#the-correlation-penalty) claims full balances. Inactivity leak limits losses to 16 ETH over ~3 weeks, while mass slashing accelerates finality at a steeper cost. Client sync issues and incomplete slashing waves (https://dora-holesky.pk910.de/validators/slashings) signal needed upgrades.
1 reply
0 recast
0 reaction
Diva Staking
@stakediva
5/ On our end, Diva’s prior testnet wasn’t Pectra-compatible and was already nearing deprecation. We’ve opted to decommission it and align with the next public testnet, ensuring resources focus on a robust, future-ready deployment rather than patching legacy setups.
1 reply
0 recast
0 reaction
Diva Staking
@stakediva
6/ Progress update: a new Diva client image is expected early next week, targeting devnet validation first. Post-devnet, we’ll advance to testnet phases with a hardened implementation. Precision over haste—expect further details soon.
1 reply
0 recast
0 reaction