运行StartChaosAsync方法似乎对服务结构集群没有明显影响

Question

我期待利用Service Fabric附带的混乱测试功能。我已经按照文档中的描述设置我的代码：https://docs.microsoft.com/en-us/azure/service-fabric/service-fabric-controlled-chaos

我遇到的问题是我几乎总是看不到我的集群显示的故障事件。以下是我运行混沌示例代码的控制台输出示例;

[StartedEvent], Timestamp=03/02/2017 16:32:21 
 
ChaosParameters: maxClusterStabilizationTimeout=00:00:30, waitTimeBetweenFaults=00:00:20, waitTimeBetweenIterations=00:00:30, maxConcurrentFaults=3, timeToRun=01:00:00, enableMoveReplicas=True, Context: 
 
ClusterHealthPolicy=ClusterHealthPolicy: MaxPercentUnhealthyNodes=0, MaxPercentUnhealthyApplications=0, ConsiderWarningAsError=True 
 

 

 
[ExecutingFaultsEvent], Timestamp=03/02/2017 16:32:26 
 
0 Faults: 
 

 
[ExecutingFaultsEvent], Timestamp=03/02/2017 16:33:00 
 
0 Faults: 
 

 
[ExecutingFaultsEvent], Timestamp=03/02/2017 16:33:33 
 
0 Faults: 
 

 
[ExecutingFaultsEvent], Timestamp=03/02/2017 16:34:06 
 
0 Faults: 
 

 
[ExecutingFaultsEvent], Timestamp=03/02/2017 16:34:40 
 
0 Faults: 
 

 
[ExecutingFaultsEvent], Timestamp=03/02/2017 16:35:13

我失去了任何配置？

我得到这些结果针对我的本地和Azure群集。我也尝试了C＃和PowerShell的例子，它们都有相同的结果。

我只看到过这个工作一次（本地），每个[ExecutingFaultsEvent]都重新启动一个节点。我是否也应该在这里看到多种类型的错误？

预先感谢

Answer 1

问题是与所提供的样本代码（和缺乏有用样本，通常的，在这个领域...）。另外，为了立即满足（看到混乱发生而没有等待太久......），你需要比文档样本更具攻击性（这又是一种真正的工作，因为你已经发现了......） 。

你是更好的服务使用不同过载为ChaosParameters构造...

试试这个（更换样品代码与此IMPL）：

  var startTimeUtc = DateTime.UtcNow; 
      var stabilizationTimeout = TimeSpan.FromSeconds(30.0); 
      var timeToRun = TimeSpan.FromMinutes(60.0); 
      var maxConcurrentFaults = 7; 
      var timeBetweenFaults = new TimeSpan(0, 0, 10); 
      var timeBetweenIterations = new TimeSpan(0, 0, 10); 
      Dictionary<string, string> _context = new Dictionary<string, string>(); 
      //Aggressive chaos... 
      var clusterHealthPolicy = new System.Fabric.Health.ClusterHealthPolicy() 
      { 
       MaxPercentUnhealthyApplications = 90, 
       MaxPercentUnhealthyNodes = 100 
      }; 

      var parameters = new ChaosParameters(
       stabilizationTimeout, 
       maxConcurrentFaults, 
       true, /* EnableMoveReplicaFault */ 
       timeToRun, 
       _context, 
       timeBetweenIterations, 
       timeBetweenFaults, 
       clusterHealthPolicy); 

注意：我建议你在新的静态异步任务返回功能中执行此操作...

Full（working）示例：

public static async Task RunChaos() 
    { 
     var clusterConnectionString = "localhost:19000"; 
     using (var client = new FabricClient(clusterConnectionString)) 
     { 
      var startTimeUtc = DateTime.UtcNow; 
      var stabilizationTimeout = TimeSpan.FromSeconds(30.0); 
      var timeToRun = TimeSpan.FromMinutes(60.0); 
      var maxConcurrentFaults = 7; 
      var timeBetweenFaults = new TimeSpan(0, 0, 10); 
      var timeBetweenIterations = new TimeSpan(0, 0, 10); 
      Dictionary<string, string> _context = new Dictionary<string, string>(); 
      //Aggressive chaos... 
      var clusterHealthPolicy = new System.Fabric.Health.ClusterHealthPolicy() 
      { 
       MaxPercentUnhealthyApplications = 90, 
       MaxPercentUnhealthyNodes = 100 
      }; 

      var parameters = new ChaosParameters(
       stabilizationTimeout, 
       maxConcurrentFaults, 
       true, /* EnableMoveReplicaFault */ 
       timeToRun, 
       _context, 
       timeBetweenIterations, 
       timeBetweenFaults, 
       clusterHealthPolicy); 

      var token = new System.Threading.CancellationToken(); 

      try 
      { 
       await client.TestManager.StartChaosAsync(parameters, new TimeSpan(0, 30, 0), token); 
      } 
      catch (FabricChaosAlreadyRunningException) 
      { 
       Console.WriteLine("An instance of Chaos is already running in the cluster."); 
      } 

      var filter = new ChaosReportFilter(startTimeUtc, DateTime.MaxValue); 

      var eventSet = new HashSet<ChaosEvent>(new ChaosEventComparer()); 

      while (true) 
      { 
       var report = await client.TestManager.GetChaosReportAsync(filter); 

       foreach (var chaosEvent in report.History) 
       { 
        if (eventSet.Add(chaosEvent)) 
        { 
         Console.WriteLine(chaosEvent); 
        } 
       } 

       // When Chaos stops, a StoppedEvent is created. 
       // If a StoppedEvent is found, exit the loop. 
       var lastEvent = report.History.LastOrDefault(); 

       if (lastEvent is StoppedEvent) 
       { 
        break; 
       } 

       Task.Delay(TimeSpan.FromSeconds(1.0)).GetAwaiter().GetResult(); 
      } 
     } 
    }