SharePoint 应用程序池重新启动 - 效果如何？
我们有 SharePoint 2007，运行 II6...
目前我们遇到了一个麻烦的 Web 部件问题，开发人员正在调查该问题。
基本上它经常崩溃，并且主 SharePoint 站点的应用程序池重新启动可以修复此问题（直到它再次崩溃）。
当我们重新启动应用程序池时，SharePoint 将关闭 30 秒左右，但我不知道的是......
- 回收应用程序池对使用 SharePoint 的用户有何影响？任何（超过 30 秒不可用的情况
- 回收应用程序池经常会导致任何问题吗？
- 我猜不推荐这样做？
我知道我们应该完全删除 Web 部件，直到其修复，但不幸的是，该业务表示它不可行 。
任何信息都会很棒，
谢谢
We have SharePoint 2007, running II6...
Currently we have a problem with a troublesome webpart which is being investigated by the developers.
Basically it crashes frequently and an App Pool restart of the main SharePoint site fixes this (until it breaks again).
When we restart the app pool it takes SharePoint down for 30 secs or so, but what I dont know is...
- What effects does recycling the app pool have on users using SharePoint? Any (above its un-availability for 30 ish seconds
- Could recycling the app pool often cause any problems?
- Im guessing its not recommended?
I know we should take the webpart out altogether until its fixed, but its not feasible unfortunately, the business says its stays.
Any info would be great,
Thanks
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（2）
正如您所猜测的，这并不理想，崩溃的应用程序池肯定会导致其他用户与 SharePoint 进行中的交互中止。
As you surmise, it is less than ideal and the crashing application pool could certainly cause other users with inflight interactions with SharePoint to be aborted.
重新启动应用程序池会强制重新启动 w3wp 进程。正如您在任务管理器中看到的那样，每个应用程序池有一个 w3wp 实例。
应用程序池重新启动后的第一次命中总是需要很长时间，因为必须加载每个程序集并且将发生每个 aspx 的第一次命中编译。
缺点是，在终止 AppPool 时，可能会丢失当前正在传输到服务器的数据。
因此，重启突击队有一个 noforce 标志，该标志会阻止进程终止，直到处理完所有请求为止。
A restart of an Application Pool enforces the w3wp process to be restarted. As you can see in your taskmanager you've one w3wp instance per Application Pool.
The first hit after the app pool restart always took long because each assembly have to be loaded and the first hit compilation for each aspx will occur.
The drawnback is, that while terminating the AppPool it's possible to lose data which are currently being transferred to the server.
Therefor the restart commando has a noforce flag which blocks the process from being terminated until all requests are processed.