Statistic SampleCount works incorrect if heat engine restart

[JunJie Nan]

Steps to reproduce:
1. heat stack-create -f WordPress_Single_Instance_With_IHA.template iha -P "..."
    "HeartbeatFailureAlarm": {
     "Type": "AWS::CloudWatch::Alarm",
     "DependsOn" : "WaitCondition",
     "Properties": {
        "AlarmDescription": "Restart the WikiDatabase if we miss a heartbeat",
        "MetricName": "Heartbeat",
        "Namespace": "system/linux",
        "Statistic": "SampleCount",
        "Period": "60",
        "EvaluationPeriods": "1",
        "Threshold": "1",
        "AlarmActions": [ { "Ref": "WebServerRestartPolicy" } ],
        "ComparisonOperator": "LessThanThreshold"
      }
    }
2. heat resource-list iha
    | CfnUser | AWS::IAM::User
    | WebServerKeys | AWS::IAM::AccessKey
    | WikiDatabase | AWS::EC2::Instance
    | HeartbeatFailureAlarm | AWS::CloudWatch::Alarm
    | WaitCondition | AWS::CloudFormation::WaitCondition
    | WaitHandle | AWS::CloudFormation::WaitConditionHandle
    | WebServerRestartPolicy | OS::Heat::HARestarter
3. Stop heat engine, wait more than 60 seconds and start heat engine,
4. Check result
Expected result:
Resources WebServerRestartPolicy, WaitCondition, HeartbeatFailureAlarm and WikiDatabase should NOT be destroyed and recreated.
Actual result:
Resources WebServerRestartPolicy, WaitCondition, HeartbeatFailureAlarm and WikiDatabase were destroyed and recreated.

Get `WordPress_Single_Instance_With_IHA.template' at
https://github.com/openstack/heat-templates/blob/master/cfn/WordPress_Single_Instance_With_IHA.template

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/37837

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/37839

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/37839
Committed: http://github.com/openstack/heat/commit/07d1a9c90b2c96f231f9bd8fe6d59bfec8c1d22d
Submitter: Jenkins
Branch: master

commit 07d1a9c90b2c96f231f9bd8fe6d59bfec8c1d22d
Author: Angus Salkeld <email address hidden>
Date: Fri Jul 19 16:06:20 2013 +1000

    Reset the the watch_rule.last_evaluated on start up

    This is so we don't fire off alarms when the engine has not been running.

    bug 1202552

    Change-Id: Ief258fd601b48e2495ac69091a942f23a019e92c