Integration tests are intermittently failing. Initially thought to be timeouts, but there’s likely a real bug in the consensus logic - specifically in the consensus_3_nodes_with_failures::case_11_fail_on_proposal_committed test, Bob crashes and restarts ahead of others. Bob commits a block and advances, while Alice and Charlie lag behind. This causes consensus to stall due to inconsistent states:
running 1 test
Test artifacts will be stored in /tmp/consensus-integration-tests/.tmpTyiJGQ
...
Pathfinder instance Alice (pid: 188505) port 35693 decided h:r (2:0)
Pathfinder instance Alice (pid: 188505) port 35693 has block 2 < 6
Pathfinder instance Bob (pid: 188596) port 43529 decided h:r (3:0)
Pathfinder instance Charlie (pid: 188597) port 37891 decided h:r (3:0)
Pathfinder instance Charlie (pid: 188597) port 37891 has block 3 < 6
Pathfinder instance Bob (pid: 188596) port 43529 has block 3 < 6
Got SIGCHLD!
Respawning Bob...
Pathfinder instance Bob (pid: 188596) has already exited
Pathfinder instance Bob (pid: 188826) has been spawned
Pathfinder instance Alice (pid: 188505) port 35693 decided h:r (3:0)
Pathfinder instance Alice (pid: 188505) port 35693 has block 3 < 6
Pathfinder instance Charlie (pid: 188597) port 37891 decided h:r (3:0)
Pathfinder instance Charlie (pid: 188597) port 37891 has block 3 < 6
Pathfinder instance Bob (pid: 188826) ready after 1 s
Bob is ready again
Pathfinder instance Bob (pid: 188826) port 35765 decided h:r None
Pathfinder instance Bob (pid: 188826) port 35765 has block 4 < 6
Pathfinder instance Alice (pid: 188505) port 35693 decided h:r (3:0)
Pathfinder instance Alice (pid: 188505) port 35693 has block 3 < 6
Pathfinder instance Charlie (pid: 188597) port 37891 decided h:r (3:0)
Pathfinder instance Charlie (pid: 188597) port 37891 has block 3 < 6
...
Pathfinder instance Alice (pid: 188505) port 35693 decided h:r (3:0)
Pathfinder instance Alice (pid: 188505) port 35693 has block 3 < 6
Pathfinder instance Bob (pid: 188826) port 35765 decided h:r None
Pathfinder instance Bob (pid: 188826) port 35765 has block 4 < 6
Test timed out after 240s
thread 'test::consensus_3_nodes_with_failures::case_11_fail_on_proposal_committed' (188426) panicked at crates/pathfinder/tests/consensus.rs:201:10:
called `Result::unwrap()` on an `Err` value: Test timed out after 240s
Integration tests are intermittently failing. Initially thought to be timeouts, but there’s likely a real bug in the consensus logic - specifically in the
consensus_3_nodes_with_failures::case_11_fail_on_proposal_committedtest, Bob crashes and restarts ahead of others. Bob commits a block and advances, while Alice and Charlie lag behind. This causes consensus to stall due to inconsistent states: