RetryPolicy: After failing the last connection attempt, the retry time is still applied

[jwaeab]

When ecChronos is trying to connect through CQL and those attempts fail, ecChronos will retry a set number of times (according to retry policy settings, which is 5 by default). After each attempt it will wait a bit before the next attempt. However, after the last attempt, it still waits before moving on. This should not be the case. It should fail and move on immediately.

Log example:

07:19:59.039 [main] WARN  c.e.b.c.e.a.DefaultNativeConnectionProvider - Unable to connect through CQL using localhost:9042, retrying.
07:19:59.040 [main] WARN  c.e.b.c.e.a.DefaultNativeConnectionProvider - Connection failed in attempt 5 of 5. Retrying in 30 seconds.
07:20:29.045 [main] WARN  o.s.b.w.s.c.AnnotationConfigServletWebServerApplicationContext - Exception encountered during context initialization - cancelling refresh attempt: 
...

After the "attempt", notice the delay is still in effect. It should move on to cancelling refresh attempt immediately.

The code:

        LOG.warn("Connection failed in attempt {} of {}. Retrying in {} seconds.",
        attempt, retryPolicy.getMaxAttempts(), TimeUnit.MILLISECONDS.toSeconds(delay));
        try
        {
            Thread.sleep(delay);
        }

This is recreated by just trying to start ecChronos, without any reachable Cassandra node being present.

Perhaps splitting the log output, so the failed/retrying parts are output as applicable.