Testing call summary
Pectra
pectra-devnet-4/Mekong
Grandine is now fixed and network is at >98% participation
Missing rate is from ethereumJS perf issues
pectra-devnet-5
We discussed existing open PRs:
- Pass max_blobs_per_block into Engine API methods: Decided to include it in a future devnet to give us more time for consensus, so it would not be included in the latest release
- Rename PartialPendingWithdrawal field index to validator_index: Included in devnet-5 since its a minor renaming
- eip7251: Limit consolidating balance by validator.effective_balance: Included in devnet-5 since its a minor renaming
- changes required for EIP-7742: Waiting to be merged in
EIP-7762: Increase MIN_BASE_FEE_PER_BLOB_GAS was brought up in the context of finding out the increased testing load. The testing team had an initial look and the tests that need updating indeed overlap with EIP-7691
, So the topic of inclusion of this EIP would be brought up on ACD with this new information.
Update EIP-7762: Move to Draft was brought up for discussion as well, the decision was to close the EIP as it introduces a decent amount of testing delays.
Update BLS precompile costs was also discussed. EL devs please have a look before ACD this week, Marek might reach out for an update.
An up to date spec-sheet can be found here
Gas limit change
- The community has recently been pushing for a gas limit increase, some issues were found with the increase and have been outlined here.
- 3/5 CL clients on the call agreed that they were going to increase the
GOSSIP_MAX_SIZE
value in a future release (before pectra). The other 2 CLs were not present to state their viewpoint, but potentially one client is against and one client is for the change. - The worst case of a staggered rollout were then discussed. There should be no issues as long as the block size is not bigger than the lowest
GOSSIP_MAX_SIZE
set on any client on the network. Beyond that size, we might see the client with smallestGOSSIP_MAX_SIZE
missing slots. - Such a change would introduce a non-equal p2p network and we usually strive to avoid that as well as strike to batch such changes at hard forks.
- Since we lacked all parties to make a decision, we would do so on the PR here or in existing async channels.
- In the meantime, the testing team will carry out experiments with a staggered rollout and other attempts at breaking the chain
TLDR: Please read the ethresearch post here. As long as the recommendations are followed we should continue to see a safe network.
peerdas & eof
No change, waiting on pectra unblocking