Improved two EC2 waiters.

The `:instance_running` and `:instance_state_ok` waiters were
raising an unexpeced error, i.e. `Aws::EC2::Errors::InvalidInstanceIDNotFound`
instead of retrying. This could happen only newly launched instances.

See #859.

CHANGELOG.md

@@ -1,6 +1,14 @@
 Unreleased Changes
 ------------------
 
+* Issue - Aws::EC2 - Updated two `Aws::EC2::Client#wait_until` waiters to
+  retry instance ID not found errors. The two waiters are:
+
+  * `:instance_running`
+  * `:instance_status_ok`
+
+  See [related GitHub issue #859](https://github.com/aws/aws-sdk-ruby/issues/859).
+
 2.1.11 (2015-07-30)
 ------------------
 

aws-sdk-core/apis/ec2/2015-04-15/waiters-2.json

@@ -186,6 +186,11 @@
           "matcher": "pathAny",
           "state": "failure",
           "argument": "Reservations[].Instances[].State.Name"
+        },
+        {
+          "matcher": "error",
+          "expected": "InvalidInstanceIDNotFound",
+          "state": "retry"
         }
       ]
     },
@@ -199,6 +204,11 @@
           "matcher": "pathAll",
           "argument": "InstanceStatuses[].InstanceStatus.Status",
           "expected": "ok"
+        },
+        {
+          "matcher": "error",
+          "expected": "InvalidInstanceIDNotFound",
+          "state": "retry"
         }
       ]
     },