Ruby 进程吃掉 100% cpu [关闭]答案

【问题标题】：Ruby process eat 100% cpu [closed]Ruby 进程吃掉 100% cpu [关闭]
【发布时间】：2011-12-25 02:40:48
【问题描述】：

当我试图从我的应用程序中注销（使用设计）时，rails 应用程序挂起并且 ruby 吃掉 100% 的 CPU，在我重新启动服务器后它仍然冻结（无法加载页面）。只有当我删除并重新创建我的数据库时它才会消失。我正在使用 PostgreSQL，试图切换到 MySQL，但没有帮助。使用 mongrel（也发生在 webrick 上）。 Ruby 1.9.2，rails 3.0.9

跟踪

futex(0xe1be68, FUTEX_WAKE_PRIVATE, 1)  = 1
futex(0xe1c3fc, FUTEX_WAIT_PRIVATE, 1, NULL) = ? ERESTARTSYS (To be restarted)
--- SIGINT (Interrupt) @ 0 (0) ---
rt_sigreturn(0x2)                       = -1 EINTR (Interrupted system call)
futex(0xe1c3fc, FUTEX_WAIT_PRIVATE, 1, NULL) = 0
futex(0xe1c458, FUTEX_WAKE_PRIVATE, 1)  = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = -1 EAGAIN (Resource temporarily unavailable)
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0
futex(0xe1be68, FUTEX_WAIT_PRIVATE, 2, NULL) = 0

控制台输出：

Started GET "/signout" for 127.0.0.1 at 2011-11-09 10:11:13 +0200
  SQL (1.0ms)   SELECT a.attname, format_type(a.atttypid, a.atttypmod), d.adsrc, a.attnotnull
 FROM pg_attribute a LEFT JOIN pg_attrdef d
 ON a.attrelid = d.adrelid AND a.attnum = d.adnum
 WHERE a.attrelid = '"jokes"'::regclass
 AND a.attnum > 0 AND NOT a.attisdropped
 ORDER BY a.attnum
  Processing by SessionsController#destroy as HTML
Redirected to http://localhost:3000/
Completed 302 Found in 1ms

我不确定那里会发生什么？

UPD* 这冻结不仅发生在注销时，它是随机发生的 :(mac 上的其他开发人员也没有这个问题，我正在运行 Ubuntu 11.10，在 10.04 上尝试过，仍然存在。

【问题讨论】：

100% cpu 听起来像是某种无限循环，并且有无数可能的罪魁祸首可能导致这种情况。将模型保存在 after_save 钩子中是一个，无限重定向是另一个。这可能是其中之一，但也可能完全是其他东西（就此而言，甚至不一定是无限循环）。
我检查了两次，钩子上没有无限循环
你能告诉我们用户模型的模型代码吗？您给我们的信息还不足以让我们开始查明错误的来源。
请看更新，我认为这与 linux 相关
Ruby 吃 100% 可能只意味着一件事：您的服务器是单核 :-)

标签： ruby-on-rails ruby cpu mongrel

【解决方案1】：

尝试在 development.rb 中开启 cache_classes

【讨论】：