The failing tests showed big variations in startup times, i.e. JVM1 not started/bound when other JVMs tries to connect to the test conductor. The reconnect logic was in place but it was bound to 10 client-reconnects (10 seconds). It's better to let it by default reconnect until the connect-timeout (20 seconds) is elapsed. |
||
|---|---|---|
| .. | ||
| java/akka/remote/testconductor | ||
| protobuf | ||
| resources | ||
| scala/akka/remote | ||