lightninglabs
add itests covering transport-connection interruption and reconnection logic
Description
When Pool's auctioneer or various subservices that interact with Pool client are restarted, trading clients often times have a don't reconnect well to the Auctioneer
Pool's main state machine should [] properly wrap all errors [how] [] Inspection if it was a connection issue/interruption [] Pass connection interruption events to the reconnection logic
Action Items: [] Add itest to validate all connection termination events are properly handled by Pool.
Shaping
Check what kind of error is returned when Aperture is killed? - Is it a wrapped error? Cannot be decoded because it’s just a string Only works if it’s an error struct Prob easy to test with setup appt kill and observe the error
https://github.com/lightninglabs/pool/blob/master/auctioneer/client.go#L1049
UX expectations
As long as the Auctioneer and the Client are running and connected to the network, they will eventually reconnect regardless of previous connection fluctuations.
