You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
On a hive testnet with Deneb genesis, sending a valid /eth/v1/beacon/pool/bls_to_execution_changes POST request to the Beacon API results in error:
Oct-26 22:02:04.910[api] �[31merror�[39m: Error on submitPoolBlsToExecutionChange [0] validatorIndex=0 - PublishError.InsufficientPeers
Error: PublishError.InsufficientPeers
at Eth2Gossipsub.publish (file:///usr/app/node_modules/@chainsafe/libp2p-gossipsub/dist/src/index.js:1547:19)
at async NetworkCore.publishGossip (file:///usr/app/packages/beacon-node/lib/network/core/networkCore.js:290:32)
Even when there is a second lodestar node connected and blocks and attestations are being gossiped correctly.
The error occurs in regardless of the epoch in which the request is sent to the beacon node.
A similar test where the Deneb fork happens at epoch 1, and then the BLSToExecutionChanges are sent, does not exhibit the same issue, and the changes are broadcasted successfully.
The main difference in the lodestar logs between the two tests is that the successful test is subscribed to topics:
Subscribe to gossipsub topic topic=/eth2/<CAPELLA DIGEST>/bls_to_execution_change/ssz_snappy
Subscribe to gossipsub topic topic=/eth2/<DENEB DIGEST>/bls_to_execution_change/ssz_snappy
whereas the failing test is only subscribed to topic:
Subscribe to gossipsub topic topic=/eth2/<DENEB DIGEST>/bls_to_execution_change/ssz_snappy
And later in the logs of the successful test we can see that the message is indeed gossiped to topic /eth2/<CAPELLA DIGEST>/bls_to_execution_change/ssz_snappy, which leads me to believe that in fact none of the lodestar peers are listening to the topic where they both broadcast to.
Describe the bug
On a hive testnet with Deneb genesis, sending a valid
/eth/v1/beacon/pool/bls_to_execution_changes
POST request to the Beacon API results in error:Even when there is a second lodestar node connected and blocks and attestations are being gossiped correctly.
The error occurs in regardless of the epoch in which the request is sent to the beacon node.
A similar test where the Deneb fork happens at epoch 1, and then the BLSToExecutionChanges are sent, does not exhibit the same issue, and the changes are broadcasted successfully.
The main difference in the lodestar logs between the two tests is that the successful test is subscribed to topics:
whereas the failing test is only subscribed to topic:
And later in the logs of the successful test we can see that the message is indeed gossiped to topic
/eth2/<CAPELLA DIGEST>/bls_to_execution_change/ssz_snappy
, which leads me to believe that in fact none of the lodestar peers are listening to the topic where they both broadcast to.Expected behavior
bls_to_execution_changes
should work.Steps to reproduce
Using hive branch https://github.com/marioevz/hive/tree/eth2-dencun, run the following command:
Additional context
No response
Operating system
Linux
Lodestar version or commit hash
Branch deneb-builder
The text was updated successfully, but these errors were encountered: