P2P track sync fails when moving from checkpoing to tracking sync

[CHr15F0x]

https://sepolia.starkscan.co/block/276582

The block number is insignificant, it looks like a problem with transitioning form checkpoint to track sync, or to be more precise: restarting tracking sync after a (not so graceful?) shutdown.

[CHr15F0x]

client.log
proxy.log

[CHr15F0x]

  2024-11-09T09:31:28 DEBUG p2p::main_loop: Received sync request, request: BlockHeadersRequest { iteration: Iteration { start: Number(276582), direction: Forward, limit: 1000, step: Step(1) } }, peer: 12D3KooWDJryKaxjwNCk6yTtZ4GbtbLrH7JrEUTngvStaDttLtid, request_id: 3
    at crates/p2p/src/main_loop.rs:478
    in pathfinder_lib::p2p_network::behaviour
    in pathfinder_lib::p2p_network::p2p

  2024-11-09T09:31:28 DEBUG p2p::main_loop: Received sync request, request: EventsRequest { iteration: Iteration { start: Number(276582), direction: Forward, limit: 1, step: Step(1) } }, peer: 12D3KooWDJryKaxjwNCk6yTtZ4GbtbLrH7JrEUTngvStaDttLtid, request_id: 3
    at crates/p2p/src/main_loop.rs:618
    in pathfinder_lib::p2p_network::behaviour
    in pathfinder_lib::p2p_network::p2p

  2024-11-09T09:31:28 DEBUG p2p::main_loop: Received sync request, request: StateDiffsRequest { iteration: Iteration { start: Number(276582), direction: Forward, limit: 1, step: Step(1) } }, peer: 12D3KooWDJryKaxjwNCk6yTtZ4GbtbLrH7JrEUTngvStaDttLtid, request_id: 3
    at crates/p2p/src/main_loop.rs:548
    in pathfinder_lib::p2p_network::behaviour
    in pathfinder_lib::p2p_network::p2p

  2024-11-09T09:31:28 DEBUG p2p::client::peer_agnostic: Propagating head, number: 298858, hash: 0x06726D91C996E8A764B5B4F891355D406BE968B6E812577C8F0BBB12ED46F78C, topic: blocks/0x534e5f5345504f4c4941
    at crates/p2p/src/client/peer_agnostic.rs:91

  2024-11-09T09:32:28 ERROR pathfinder_lib::p2p_network: Failed to handle P2P event: Sending item: send failed because receiver is gone
    at crates/pathfinder/src/p2p_network.rs:110
    in pathfinder_lib::p2p_network::p2p

Looks like a problem on the proxy side.

[CHr15F0x]

Todo: trace level logs.

[CHr15F0x]

This is an artificially triggered bug, which was caused by me bumping MAX_BLOCKS_COUNT to 1k on both. Reverting to 100 solves the issue.

Imo we should spend our investigation efforts on the original issue of libp2p swarm halting when handling many concurrent p2p_stream streams #2351.