Non-confirmed (expired/started) nodes were spamming the network every 60 seconds by asking 7 random peers to check availability — all of which reject immediately since a367ec0 added confirmed-list validation to isFluxAvailable
Added a guard in checkDeterministicNodesCollisions to check local daemon status before calling checkMyFluxAvailability — if the node isn't CONFIRMED, skip the remote probe entirely
Fixed existing test that was missing status: 'CONFIRMED' in its stub data, and added a new test for the expired-node path
Context
Discovered on a node (38.247.82.127:16147) that expired due to a collateral collision. The node knew it was expired (daemon RPC returns status: "expired") but kept asking remote nodes to verify its availability. Every remote node rejected at the socketAddressInFluxList check without even probing ports, making the requests pointless.
Test plan
fluxNetworkHelper.test.js — all 103 tests passing, 0 failing
New test verifies expired node skips availability check
Existing "should not change dosMessage" test fixed to properly stub CONFIRMED status