ConsenSys developer: Ethereum mainnet merger does not need to be perfect, Goerli testnet issues are not a big deal

Note: The original author is ConsenSys developer Ben Edgington. In this article, he mentioned the problems encountered during the Goerli testnet merge. He explained that the problem was not a big deal and would not affect the subsequent Ethereum mainnet merge plan. In his opinion, the Ethereum mainnet merge does not need to be perfect.

Compiled by: Overnight Porridge, The Way of Metaverse

The TTD for the Ethereum Mainnet merge is set! The number 58750000000000000000000000 joins the list of the most important integers of our time. Once Ethereum’s Proof-of-Work chain accumulates this much difficulty, the entire network will switch to Proof-of-Stake (PoS).

This number was determined during a consensus developer call following the successful Goerli testnet merge. The goal is to have the mainnet merge around September 15th, although the exact time the merge occurs may vary by a few days due to hashrate fluctuations. You can follow the latest merge predictions as usual on bordel.wtf‌ or on Matthew Rabinowitz’s spreadsheet‌.

TTD's selection is based on Mario Havel's analysis‌, please note that this is not yet fully confirmed by the developers and may be revised before the final client version is released next week.

Therefore, the merged path looks like this:

1. August 18: Approved merging TTD numbers during all core developers call;

2. August 22: The client version including the final TTD is completed;

3. August 23: The Ethereum Foundation published an official blog announcement;

4. September 6th, 11:34:47 UTC: The beacon chain Bellatrix was upgraded at epoch 144896, slot 4636672;

5. Sometime around September 15: Champagne celebration;

Goerli testnet merge review

Ethereum’s Goerli testnet reached its scheduled TTD around 01:50 UTC on August 11 and merged with the Prater beacon chain testnet.

The merge didn't go perfectly, and various issues meant it took 40 minutes to complete, rather than the 15 minutes we had ideally hoped for. But it did go through, and at the end of the day it was a success.

We discussed the related issues in the consensus dev call, which I won't rehash here. There's also an issue tracker with links to some more detailed analysis of individual issues. Some of the issues stemmed from having two candidate blocks that met the TTD criteria. Some Besu clients mistakenly rejected one of them, which caused some of our Teku nodes to end up on the wrong fork for a period of time. It's possible that we'll have multiple terminal candidates on mainnet, but the chances of that are much smaller than on these testnets. In any case, I'm hopeful that this bug will be fixed. Also, if one of the client teams hadn't misconfigured their node (again), we'd probably be done a long time ago.

A few hours after the merger, Goerli–Prater returned to near-normal performance.

So, all in all, nothing major has gone wrong and we are confident enough to proceed to the next step. Honestly, I would be very happy if the mainnet merge went like that. The merge doesn't need to be perfect!

The Goerli merge livestream by Bankless and EthStaker has received over 21,000 views.

In addition, CryptoGucci wrote a summary article on the Goerli merger process, and media such as The Block, Decrypt , The Defiant , Blockworks , and CNBC also reported on it.

Goerli Shadow Fork (6)

It seems like a long time ago, but GSF6 took place on August 4th and went very well, and here is a summary based on Pari's notes.

1. There were no client-side issues, all combinations were in sync, and transitions were handled correctly;

2. Two nodes were running low on disk space, which explains the drop in participation (from 97% to 94%);

3. About 30% of the network was running MEV-Boost during the transition, and apart from latency-related issues on one node in India (which was likely a defective machine), we didn’t notice any other issues.

Preparing for the merger

Somer Esat updated his staking guide to include the Goerli merged configuration. Somer's guide has pretty much become the standard for beacon chain staking (along with CoinCashew's), and I used his Teku guide when setting up my own configuration. These only cover the merged Goerli testnet, but I'm sure it will be updated with mainnet setups: including Teku, Lighthouse, Nimbus, and Prysm.

Check out EthStaker’s latest validator workshop, which also covers setting up MEV-Boost on your node. It was a great session with lots of helpful advice and Q&A. The next one is scheduled for the 24th.

Infura's excellent merge workshops continue, here are the latest ones:

1. August 2: Execution Client and Consensus Client Workshop, hosted by my esteemed colleague Matt Nelson.

2. August 9: Running multiple validators and staking services (SaaS)‌;

Here is a list of merged configurations‌ .

Research and Development

1. If we are going to do MEV-Boost, we need to consider that it requires trust. In particular, in the current design, stakers have to trust that the relayers work properly. That is, relayers report their block values ​​correctly, they provide valid blocks, and they do not withhold blocks after validators sign them. One way to keep relayers honest is to monitor them, so Alex Stokes proposed a relayer monitoring design that allows failure information to be collected and reported as a score for each relayer. Stakers can then make informed decisions about which relayers to visit.

2. As I write this, ethresear.ch is not accessible yet, but I’m aware of this breakthrough research into building KZG multi-proofs, which will help improve Danksharding performance.