Stop the HashedWheelTimer before shutting down EventLoopGroups #123

Motivation: Active timer tasks can request interaction with a EventLoopGroup that is signalled for shutdown. Tasks in that state are likely to fail and cause exceptions which are logged.

Result: The timer is stopped before the EventLoopGroups are shut down so there are no more active timer tasks when the EventLoops are stopped

RELEASE-NOTES.txt

@@ -199,6 +199,7 @@ Fixes
 -----
 * Synchronization/cross thread visibility of variables #94
 * Check channel state before calling Channel.close.syncUninterruptibly #113
+* Stop the HashedWheelTimer before shutting down EventLoopGroups #123
 
 Other
 ------

src/main/java/com/lambdaworks/redis/AbstractRedisClient.java

@@ -254,6 +254,8 @@ public void shutdown() {
      */
     public void shutdown(long quietPeriod, long timeout, TimeUnit timeUnit) {
 
+        timer.stop();
+
         while (!closeableResources.isEmpty()) {
             Closeable closeableResource = closeableResources.iterator().next();
             try {
@@ -296,8 +298,6 @@ public void shutdown(long quietPeriod, long timeout, TimeUnit timeUnit) {
                 throw new RedisException(e);
             }
         }
-
-        timer.stop();
     }
 
     protected int getResourceCount() {