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
Our ClientPool rotation handling within runLoop pessimistically assumes that any client start is after a failure and always attempts to replayListenAddresses despite the original calls into ListenAddresses having just successfully completed. This causes our pool startup to broadcast listen addrs twice on first start.
In order to be more bandwidth efficient, we should find a way to detect whether we're recovering from a failure or not as part of our runLoop.
The text was updated successfully, but these errors were encountered:
Our ClientPool rotation handling within
runLoop
pessimistically assumes that any client start is after a failure and always attempts to replayListenAddresses despite the original calls into ListenAddresses having just successfully completed. This causes our pool startup to broadcast listen addrs twice on first start.In order to be more bandwidth efficient, we should find a way to detect whether we're recovering from a failure or not as part of our runLoop.
The text was updated successfully, but these errors were encountered: