base-consensus fails with MissingL1InfoDeposit when starting from official pruned reth snapshot with empty SafeDB

[riyue121]

Hi team,

I upgraded my Base mainnet node this afternoon and downloaded the official snapshot from the Base documentation.

After restoring the snapshot, the execution layer seems to be working and the block height can continue updating. However, I keep seeing the following error in the execution logs:

{"timestamp":"2026-05-22T13:45:28.020319Z","level":"ERROR","fields":{"message":"Invalid JWT: Authorization header is missing or invalid"}}

At the same time, the consensus node keeps restarting. The main error from the consensus logs is:

{"timestamp":"2026-05-22T13:53:40.178737Z","level":"ERROR","fields":{"message":"Failed to drain engine tasks","err":"EngineReset(SyncStart(FromBlock(MissingL1InfoDeposit(0xc710d12366805dd28afb50869e76de6e49225ecbd1d7b131941f2ac204ed85f2))))"},"target":"engine"}

Current situation:

I upgraded Base today.
I used the official Base snapshot.
The execution layer can start and the block height is updating.
The consensus node keeps restarting.
Execution layer keeps logging Invalid JWT: Authorization header is missing or invalid.
Consensus fails with MissingL1InfoDeposit(...).

Could you please help confirm whether this is caused by a configuration issue, a snapshot compatibility issue, or something else required for the v1 upgrade?

Thanks.

[riyue121]

{"timestamp":"2026-05-22T13:53:35.292207Z","level":"INFO","fields":{"message":"Safe head unchanged","safe_head":0},"target":"engine"}
{"timestamp":"2026-05-22T13:53:37.241468Z","level":"INFO","fields":{"message":"Validator enqueuing external unsafe payload","block_number":46334335,"block_hash":"0xc3e31b4dada0e9f288f0dd503101b70a41cbd0346a5626ff4222bb40bf553b41","parent_hash":"0x8fce23a4e4513f44dd202b0980afeec53b6f0d1a4f6666559703098972ab75bc"},"target":"engine"}
{"timestamp":"2026-05-22T13:53:37.246252Z","level":"INFO","fields":{"message":"Inserted new payload","hash":"0xc3e31b4dada0e9f288f0dd503101b70a41cbd0346a5626ff4222bb40bf553b41","number":46334335,"payload_safety":"unsafe","total_duration":"4.730217ms","insert_duration":"3.35527ms"},"target":"engine"}
{"timestamp":"2026-05-22T13:53:37.246303Z","level":"INFO","fields":{"message":"Safe head unchanged","safe_head":0},"target":"engine"}
{"timestamp":"2026-05-22T13:53:37.664281Z","level":"INFO","fields":{"message":"Processing l1 head update","l1_head":"BlockInfo { hash: 0xeb52eb4407c037ed1a89544829e798f5c8a77422929e024b897207e5a0b5292e, number: 25151224, parent_hash: 0xd79d449a1e85e4baaa6504f8a178d3fc8172bc84dec1b1257829fe538383c0ad, timestamp: 1779458015 }"},"target":"derivation"}
{"timestamp":"2026-05-22T13:53:37.664306Z","level":"INFO","fields":{"message":"Skipping derivation.","derivation_state":"DerivationStateMachine { confirmed_safe_head: L2BlockInfo { block_info: BlockInfo { hash: 0x0000000000000000000000000000000000000000000000000000000000000000, number: 0, parent_hash: 0x0000000000000000000000000000000000000000000000000000000000000000, timestamp: 0 }, l1_origin: NumHash { number: 0, hash: 0x0000000000000000000000000000000000000000000000000000000000000000 }, seq_num: 0 }, state: AwaitingELSyncCompletion }"},"target":"derivation"}
{"timestamp":"2026-05-22T13:53:39.235505Z","level":"INFO","fields":{"message":"Validator enqueuing external unsafe payload","block_number":46334336,"block_hash":"0x02f541ada101c462c476fac0816bb1a0beb9656c7fb000306b74ccc52d872a21","parent_hash":"0xc3e31b4dada0e9f288f0dd503101b70a41cbd0346a5626ff4222bb40bf553b41"},"target":"engine"}
{"timestamp":"2026-05-22T13:53:40.177839Z","level":"INFO","fields":{"message":"Finished execution layer sync."},"target":"engine"}
{"timestamp":"2026-05-22T13:53:40.177863Z","level":"INFO","fields":{"message":"Inserted new payload","hash":"0x02f541ada101c462c476fac0816bb1a0beb9656c7fb000306b74ccc52d872a21","number":46334336,"payload_safety":"unsafe","total_duration":"942.330431ms","insert_duration":"939.738073ms"},"target":"engine"}
{"timestamp":"2026-05-22T13:53:40.178038Z","level":"INFO","fields":{"message":"Safe head unchanged","safe_head":0},"target":"engine"}
{"timestamp":"2026-05-22T13:53:40.178046Z","level":"INFO","fields":{"message":"Performing initial engine reset"},"target":"engine"}
{"timestamp":"2026-05-22T13:53:40.178737Z","level":"ERROR","fields":{"message":"Failed to drain engine tasks","err":"EngineReset(SyncStart(FromBlock(MissingL1InfoDeposit(0xc710d12366805dd28afb50869e76de6e49225ecbd1d7b131941f2ac204ed85f2))))"},"target":"engine"}
{"timestamp":"2026-05-22T13:53:41.217117Z","level":"ERROR","fields":{"message":"Engine processing channel closed unexpectedly"},"target":"engine"}
{"timestamp":"2026-05-22T13:53:41.217167Z","level":"WARN","fields":{"message":"EngineRpcProcessor received shutdown signal"},"target":"engine"}
{"timestamp":"2026-05-22T13:53:41.217183Z","level":"INFO","fields":{"message":"Received shutdown signal. Exiting network task."},"target":"network"}
{"timestamp":"2026-05-22T13:53:41.217204Z","level":"INFO","fields":{"message":"Received shutdown signal. Exiting derivation task."},"target":"derivation"}
{"timestamp":"2026-05-22T13:53:41.217218Z","level":"ERROR","fields":{"message":"Critical error in sub-routine","error":"ChannelClosed"},"target":"rollup_node"}
{"timestamp":"2026-05-22T13:53:41.217241Z","level":"INFO","fields":{"message":"Received shutdown signal. Exiting L1 watcher task."},"target":"l1_watcher"}
{"timestamp":"2026-05-22T13:53:41.217262Z","level":"ERROR","fields":{"message":"Failed to start rollup node service","error":"ChannelClosed"},"target":"rollup_node"}
{"timestamp":"2026-05-22T13:53:41.217395Z","level":"INFO","fields":{"message":"Pubsub service request channel closed. Shutting down."},"target":"alloy_pubsub::service"}
Error: ChannelClosed

Location:
crates/consensus/cli/src/node.rs:265:13

[riyue121]

I found another possible issue in consensus-entrypoint.

My environment has:

BASE_NODE_L2_ENGINE_RPC=ws://execution:8551

In consensus-entrypoint, there is the following readiness check:

until [ "$(curl -s --max-time 10 --connect-timeout 5 -w '%{http_code}' -o /dev/null "${BASE_NODE_L2_ENGINE_RPC/ws/http}")" -eq 401 ]; do
  echo "waiting for execution client to be ready"
  sleep 5
done

Because BASE_NODE_L2_ENGINE_RPC is set to ws://execution:8551, the expression:

${BASE_NODE_L2_ENGINE_RPC/ws/http}

expands to:

http://execution:8551

This means the entrypoint is sending an unauthenticated HTTP request to the execution client's Engine/Auth RPC endpoint.

As a result, the execution client logs the following error:

{"timestamp":"2026-05-22T13:45:28.020319Z","level":"ERROR","fields":{"message":"Invalid JWT: Authorization header is missing or invalid"}}

I confirmed that:

port 8551 is not exposed to the host
the JWT file exists in both the execution and node containers
both JWT files are identical
both JWT files are 64 bytes
the SHA256 checksum is the same in both containers
after stopping the node container, the Invalid JWT error stops appearing in the execution logs

So it looks like the error is caused by this readiness check in consensus-entrypoint, rather than an actual JWT mismatch between the consensus and execution clients.

Would it be better to check the regular execution RPC endpoint instead, for example http://execution:8545 with eth_chainId, or perform only a TCP port check for execution:8551 without sending an unauthenticated HTTP request to the Engine/Auth RPC endpoint?

[riyue121]

After fixing the JWT readiness check and adding a persistent SafeDB path, the node still fails.

I added:

BASE_NODE_SAFEDB_PATH=/data/safedb.db
BASE_NODE_P2P_PRIV_PATH=/data/p2p_priv.txt

The P2P key is now loaded successfully:

Successfully loaded P2P keypair from file path="/data/p2p_priv.txt"

However, base-consensus still starts with safe_head=0 and fails during the initial engine reset:

Safe head unchanged safe_head=0
Performing initial engine reset
Failed to drain engine tasks err="EngineReset(SyncStart(FromBlock(MissingL1InfoDeposit(...))))"

I am using the official Base mainnet pruned reth snapshot. Does the pruned snapshot require a pre-populated SafeDB, or is an archive snapshot required when SafeDB is empty?