[ZK Stack] - Withdrawal Finalizer not picking up withdrawals properly

[therealdannzor]

Module Affected

• System contracts (era-system-contracts)
• Smart contracts (era-contracts)
• zkEVM (era-zk_evm)
• Server (zksync-era)
• Prover (era-boojum)
• Fee withdrawer (era-fee-withdrawer)
• Block Explorer (clock-explorer)
• Portal - Wallet + Bridge (dapp-portal)
• Other (please specify below)

Rust Version

1.78.0-nightly (d44e3b95c 2024-02-09)

Operating System

• Ubuntu
• MacOS
• Windows
• Other (Please specify below)

Issue Description

Setup:

• Already deployed zkSync hyperchain (our testnet)
• Local running instance of dapp-portal
• Running the binary of zksync-withdrawal-finalizer

I have:

• deployed the WithdrawalFinalizer contract in zksync-withdrawal-finalizer successfully
• made sure that all the env variables are set up properly
• properly created and migrated to local Postgres database

Despite that, it doesn't perform automatic withdrawals on our zkSync hyperchain to L1 Sepolia.

Expected Behavior

Claims are automatically performed on L1

Current Behavior

Repository Link (if applicable)

https://github.com/matter-labs/zksync-withdrawal-finalizer

Additional Details

• our hyperchain is not EIP-4844 compliant yet
• enabled FINALIZE_ETH_TOKEN as this only concerns ETH
• hardcoded contract constants before deployment
• WITHDRAWAL_FINALIZER_ACCOUNT_PRIVATE_KEY is funded to make txs
• DATABASE_URL indexes events in my local database
• no attempts to send any txs from the backend service at all
• no observed calls to CONTRACTS_WITHDRAWAL_FINALIZER_CONTRACT
• GAS_LIMIT=30000000 and BATCH_FINALIZATION_GAS_LIMIT=10000000000000
• tried TX_RETRY_TIMEOUT_SECS=10..60
• set correct L1 chain id: wallet = wallet.clone().with_chain_id(11155111u64);
• verified withdrawal finalizer contract on Sepolia etherscan

Tables in my database:

l2_to_l1_events and tokens are empty.

Prior Research

• Yes, I searched for similar issues and did not find any matching ones.

[bxpana]

Thanks for reporting, will check with the team on this and get back to you.

[therealdannzor]

Much appreciated.

Are you able to provide the env variables GAS_LIMIT and BATCH_FINALIZATION_GAS_LIMIT that your team is running in the deployed service? In your docs the amount 2^32 is mentioned but I've set it lower to avoid the error, encountered in my initial post (see edit). Based on experience, a lot of issues comes from gas configuration.

[therealdannzor]

To give an update, I don't think it has to do with the gas since the dry-run with eth_call works fine without reverting and I even added (superfluous) functionality by calling estimate_gas() to update the config dynamically. Despite that, all the withdrawals I observe are predicted to fail and the ones that I initiate myself are discovered by the watcher module but never acted on by the finalizer module.

[therealdannzor]

I also asked the dev for the project (PR-498) why the table l2_to_l1_events are no longer needed as it might be an explanation for why it doesn't work in our non-4844 compliant hyperchain.

[bxpana]

I'm still waiting to hear back from the team on this and will let you know.

[montekki]

Hi, @therealdannzor , thanks for reporting this. First we need to look at the transaction that initiate withdrawals and what logs they are generating. Can you provide such an example? Idk, if you have the block explorer for your chain deployed. But maybe you can just get transaction and logs for it from the RPC?

[therealdannzor]

x-post matter-labs/zksync-withdrawal-finalizer#435