AppFabric缓存重启超过5分钟 [英] AppFabric Caching restart more then 5 minutes

问题描述

环境：Windows 2008 R2（MultiCore NUMA，64GB RAM）上的AppFabric v1.1，SQL Server 2008 R2商店提供商。单个群集（小）和群集中只有一个主机。

AppFabric重启非常慢，重启需要5分钟以上。我注意到它在日志中几乎没有警告（见下面的错误）。在配置中设置它不是leadhost（默认情况下），错误消失，但
启动时间是相同的。

当AppFabric缓存大小减少时（默认值为32GB）到4GB或1GB - 重启减少到30秒。这是为什么？在4GB的情况下，分配的内存为350MB。当处于默认设置（32GB）时，Cache会很快跳到4GB分配的内存 - 在
a几秒后，但开始很慢（5分钟）。

日志中的错误：

{3e43dc3c000000000000000000000000}无法刷新查找表，异常：{Microsoft.Fabric.Common.OperationCompletedException：操作已完成，异常--- > Microsoft.Fabric.Federation.RoutingException：目标节点明确中止
操作

   ---内部异常堆栈跟踪结束---

  在Microsoft.Fabric.Common.OperationContext.End（）

  在Microsoft.Fabric.Federation.FederationSite.EndRoutedSendReceive（IAsyncResult ar）

  在Microsoft.Fabric.Data.ReliableServiceManager.EndRefreshLookupTable（IAsyncResult ar）}

Thanx

Srecko

解决方案

嗨Srecko，

启动时收到的警告群集是良性的，不应导致任何问题，包括启动时间。 

 "在配置中设置它不是leadhost（默认情况下是，）错误消失，
但启动时间相同。"

我有点意外。如果没有主机主机，那么群集不应该启动。可以
你能验证一下吗？我假设你正在使用管理命令重启。这是对吗？

现在进入启动持续时间，内存量不应对启动时间产生重大影响。你是如何测量启动时间的？是时候服务处于运行状态还是所有节点都显示为健康状态？

谢谢

阿伦

Hi,

Environment: AppFabric v1.1 on Windows 2008 R2 (MultiCore NUMA, 64GB RAM), SQL Server 2008 R2 store provider. Single cluster (small) and only one host in cluster.

AppFabric is very slow to restart, it takes more then 5 minutes to restart. What I have noticed is that it has few warnings in log (see below the error). After setting in configuration that it is not leadhost (by default it is), the errors gone, but the startup time is the same.

When AppFabric cache size is reduced (default is 32GB) to 4GB or 1GB - the restart is reduced to 30seconds. Why is that? In case of 4GB, allocated memory is 350MB. When was in default settings (32GB), Cache jumps to 4GB allocated memory very quickly - after a few seconds, but start is slow (5 minutes).

Errors in log:

{3e43dc3c000000000000000000000000} failed to refresh lookup table, exception: {Microsoft.Fabric.Common.OperationCompletedException: Operation completed with an exception ---> Microsoft.Fabric.Federation.RoutingException: The target node explicitly aborted the operation
   --- End of inner exception stack trace ---
   at Microsoft.Fabric.Common.OperationContext.End()
   at Microsoft.Fabric.Federation.FederationSite.EndRoutedSendReceive(IAsyncResult ar)
   at Microsoft.Fabric.Data.ReliableServiceManager.EndRefreshLookupTable(IAsyncResult ar)}

Thanx

Srecko

解决方案

Hi Srecko,

The warnings you are getting while starting the cluster is benign , and should not lead to any issues including the startup time. 

 "After setting in configuration that it is not leadhost (by default it is), the errors gone, but the startup time is the same."

I am bit surprised by this . If no lead hosts are there cluster should not be starting . Can you please verify this ? I am assuming you are doing restart using the administrative command-lets. Is this right ?

Now coming to the startup duration , the amount of memory should not have major impact on the startup time. How are you measuring the startup time ? Is it the time for service to be in running state or the all nodes are shown as healthy ?

Thanks

Arun

这篇关于AppFabric缓存重启超过5分钟的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！