使用高可用性功能时遇到延迟 [英] Delay encountered when using high availability feature

问题描述

我有三个AppFabric缓存服务器（服务器A，B和C）。我创建了一个小型Web应用程序，它有两个按钮，一个用于将系统日期存储在会话变量上，另一个用于检索相同的。

I have three AppFabric cache servers (servers A, B, and C). I have created a small web application that has two buttons, one to store the system date on a session variable and the other one to retrieve the same.

这是我在页面上执行的操作。

Here is what I do on the page.

1。我点击"保存"按钮按钮将当前日期保存到会话变量"日期"。

2.我点击"检索"按钮按钮以检索会话变量"日期"上的数据。

3.数据当前显示在页面上。

4.我关闭服务器A上的缓存服务。

5.我点击"检索"再次按钮。

6.我收到错误ERRCA0017，SubStatus ES0001。

1. I click "Save" button to save current date to session variable "Date".
2. I click "Retrieve" button to retrieve the data on session variable "Date".
3. Data is displayed currently on the page.
4. I turn off the caching service on server A.
5. I click "Retrieve" button again.
6. I get error ERRCA0017, SubStatus ES0001.

如果我在步骤4和5之间引入几秒延迟，则不会发生这种情况。

This doesn't happen if I introduce a few seconds of delay between steps 4 and 5.

问题是，一旦该数据的主要（或辅助）存储器出现故障，数据在另一台服务器上复制之前的延迟有多长？根据我的经验，一旦我遇到上面的错误，即使我再等几秒钟并刷新
页面，问题似乎也没有得到解决。但话说回来，这个问题似乎是间歇性的。

Question is, how long is the lag before the data is replicated on another server once the primary (or secondary) storage of that data goes down? Based on my experience, once I encounter the error above, even if I wait a couple of seconds more and refresh the page, the problem doesn't seem to get solved. But then again, this problem seems to be intermittent.

谢谢。

推荐答案

谢谢这个问题。由于必须将对象复制到新的缓存主机以创建新的"备份"，因此可能存在延迟。但是这种延迟通常与不同的NoWriteQuorum异常相关联（参见
http://msdn.microsoft .COM / EN-US /库/ ff921025.aspx ）。使用AppFabric缓存，有时您会获得"RetryLater"。由于
缺乏数据，因此异常并不容易确定原因。我建议用较短的等待时间为这些错误实现重试逻辑。服务器最终应该恢复，并且希望延迟对于影响用户并不重要。这是关于异常处理的主题，但它没有提供
a建议的重试逻辑代码（抱歉）：
http://msdn.microsoft.com/en-us/library/ff637704.aspx 。

Thanks for the question. There can be a delay due to the fact that objects must be copied to a new cache host to create a new "backup". But this delay typically is associated with a different NoWriteQuorum exception (see http://msdn.microsoft.com/en-us/library/ff921025.aspx). With AppFabric caching, there are times that you get the "RetryLater" exception and the cause is not always easy to determine due to lack of data. I would suggest implementing retry logic for these errors with short wait times. The server should eventually recover, and hopefully the delay is not significant to affect users. Here is the topic on exception handling, but it does not provide a suggested retry logic code (sorry): http://msdn.microsoft.com/en-us/library/ff637704.aspx.

Jason Roth

Jason Roth

这篇关于使用高可用性功能时遇到延迟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！