Change select timeout from 100ms to 1. This mitigates a bug on some systems where the select doesn't process new data until the next call. (Spinning faster doesn't increase load on an idle system.) patch by jbellis; tested by goffinet

git-svn-id: https://svn.apache.org/repos/asf/incubator/cassandra/branches/cassandra-0.3@782095 13f79535-47bb-0310-9956-ffa450edef68

src/java/org/apache/cassandra/net/SelectorManager.java

@@ -94,7 +94,7 @@ public class SelectorManager extends Thread
         {
             try
             {
-                selector.select(100);
+                selector.select(1);
                 doProcess();
                 synchronized(gate) {}
             }