mantlenetworkio / mantle Goto Github PK

• multiple l2-geth nodes
• currently only support one sequencer, how to aceive ha?
• demo accounts should be replaced and protected
• any forgotten necessary service?

• Bridge pool contract
• Fast bridge contract

change native token from eth to bit

monitor

• batch-submitter
• dtl
• gas-oracle
• l2geth

log

• json-format

Currently It seems that @bitdaoio/sdk not worked for bridging custom erc20 token between L1 and L2.

We will change it by the pre-version of @optimism/sdk.

CommitId:

940cfacb 作者 Matthew Slipper [email protected]，2022/5/17 上午12:46

I used two sequencer to send the withdraw tx in the meantime, and the tx is in the pending state.

One of batch-submitter report the following errors:

15|10:06:04.465] Sequencer unable to craft batch tx       err="execution reverted: Attempted to append more elements than are available in the queue."
INFO [08-15|10:06:12.714] Sequencer fetching current block range 
INFO [08-15|10:06:12.718] Proposer fetching current block range 
INFO [08-15|10:06:12.982] Sequencer block range                    start=142 end=151
INFO [08-15|10:06:13.238] Proposer no updates                      start=142 end=142
INFO [08-15|10:06:13.258] Sequencer crafting batch tx              start=142 end=151 nonce=469 type=ZLIB
INFO [08-15|10:06:13.288] Sequencer testing batch size             calldata_size=12408 min_tx_size=32 max_tx_size=90000
INFO [08-15|10:06:13.288] Sequencer batch constructed              num_txs=9  final_size=12408 batch_type=ZLIB
ERROR[08-

When I restart the l2-geth in one cluster, it showed the following bug:

ERROR[08-15|10:18:29.206] Timestamp monotonicity violation         hash=0xaebc2f6e28717e3134acadd8452b10ac52c48a995a4e1196d88ca443c545491b latest=1660558709 tx=1660545605
ERROR[08-15|10:18:29.206] Blocknumber monotonicity violation       hash=0xaebc2f6e28717e3134acadd8452b10ac52c48a995a4e1196d88ca443c545491b new=7407764 old=7408448
ERROR[08-15|10:18:29.207] Monotonicity violation                   index=32    parent=1660555041 tx=1660545605
ERROR[08-15|10:18:29.207] Problem committing transaction           msg="nonce too high"
ERROR[08-15|10:18:29.207] Got error waiting for transaction to be added to chain msg="nonce too high"
CRIT [08-15|10:18:29.207] Sequencer cannot sync transactions to tip err="Verifier cannot sync transactions with backend l1: Cannot sync transactions with backend l1: Cannot apply transaction: nonce too high"

And it seemed that it can't restore from the pre-state, and we need to re-deploy the contracts on L1.

In production env, it is unbearable.

This is illegal operation. How should we avoid this when we also want the high availabity in the production env?

Summary

Implement decentralized sequencers, there are two main goals here:

• Anyone can participate in the sequencer set as a L2 block producer.
• Reduce MEV, each sequencer participating in the consensus process is able to determine transaction ordering.

Task List

Layer 1

• Sequencer set selection on L1 ,#139
  • Register Sequencer with mintAddress(l2geth/miner/miner.go#Miner.coinbase) and nodeID(l2geth/p2p/encode/node.go#Node.id).
  • Set Scheduler with nodeID.

Layer 2

L2geth

• Rotation mechanism
  • Weighted round-robin algorithm, #16
  • Get Sequencer set from L1's contract, and update mintpower each epoch #55
  • Scheduler broadcast producers #197
  • Switching sequencer for liveless issue
    • Heartbeat: Listen to the connection between Scheduler and Sequencer via l2/geth/p2p pingpong mechanism(Identity authentication by nodeID).
    • Timeout: Sequencer does not commit the block to the scheduler on time(Identity authentication by mintAddress).
• Convert l1->l2 transactions and submit these transactions to L2 : TBD
• Add scheduler consensus in l2geth/consensus:
  • Implement consensus interface to produce block, #59
  • Get the block producer from scheduler, #60
  • Get blocks from scheduler/l2geth (Identity authentication by nodeID), #61

• Websocket client

• Tss Core
  p2p communication
  tss algorithm

• Request Validation

Related sub issues:
#9
#35
#36

Get Sequencer set from L1's contract, and update mintpower each epoch.

Sequencer rotation mechanism on L2.

test point：newGasPrice = (EthPrice / BitPrice) * L1GasPrice

We change the ETH to Bit in bridge and l2geth, so we should support the same function in @bitdaoio/sdk

integrate with contracts
integrate with command(specially Config.)

Get Sequencer set from L1's contract, and update mintpower each epoch

Currently the fault-detector can't work as expected because of queryFilter will search from 0 to latest block.

We build a dev ENV to deploy mantle based on tss-enhancement branch for the integration test

Currently integratio_test doesn't work because some property are still use OVM.

Started the two same sequencer service and sent the withdraw tx in the meantime polluted the contracts in goerli testnet.

So need to redeploy new contracts and changed the @bitdaoio/sdk.

• design
• implement
• deploy and test

Implement consensus interface(l2geth/consensus) to produce block

• frontend
• backend

Sync block producers from scheduler, identified by peer.id.

• Erc20 token replacement
• Add token price service for GasPriceOracle #28
• Distribution of Tss rewards

• request signature data from tss manager
• apend signature to scc contract

• qa fix. replace the sequencer key
• add init-key.sh to generate privatekey secretly.

integrate with contracts
add interaction function with contract

fix github & circleci workflows

Sync blocks from scheduler, identified by peer.id.