我正在Apache / Passenger下的服务器上运行一组rails应用程序。 其中一个应用程序不断进入状态，机器的负载变得非常高，机器变得没有响应，顶部显示大部分CPU时间花在IO等待上，但没有一个进程花费太多时间。 iowait和ifwait没有显示任何exception。 乘客状态是最有趣的，它显示了我怀疑是应用程序（下面的输出中的应用程序x）的每个应用程序进程被占用1个会话。 我可以杀死所有这些进程和乘客replace它们对服务器负载没有影响，只有当我重新启动Apache或服务器加载恢复正常。 这是输出：
----------- General information ----------- max = 20 count = 20 active = 16 inactive = 4 Waiting on global queue: 22 ----------- Application groups ----------- /home/A: App root: /home/A * PID: 5379 Sessions: 0 Processed: 6 Uptime: 3m 42s /home/X: App root: /home/X * PID: 5604 Sessions: 1 Processed: 2 Uptime: 1m 8s * PID: 5560 Sessions: 1 Processed: 3 Uptime: 1m 10s * PID: 5416 Sessions: 1 Processed: 5 Uptime: 3m 30s * PID: 5387 Sessions: 1 Processed: 19 Uptime: 3m 30s * PID: 5571 Sessions: 1 Processed: 2 Uptime: 1m 9s * PID: 5408 Sessions: 1 Processed: 6 Uptime: 3m 30s * PID: 5424 Sessions: 1 Processed: 3 Uptime: 3m 30s * PID: 5306 Sessions: 1 Processed: 18 Uptime: 3m 30s * PID: 5553 Sessions: 1 Processed: 1 Uptime: 1m 10s * PID: 5612 Sessions: 1 Processed: 1 Uptime: 1m 8s * PID: 5394 Sessions: 1 Processed: 4 Uptime: 3m 30s * PID: 5401 Sessions: 1 Processed: 14 Uptime: 3m 30s * PID: 5620 Sessions: 1 Processed: 1 Uptime: 1m 8s * PID: 5298 Sessions: 1 Processed: 7 Uptime: 4m 33s * PID: 5580 Sessions: 1 Processed: 2 Uptime: 1m 9s * PID: 5594 Sessions: 1 Processed: 1 Uptime: 1m 8s /home/B: App root: /home/B * PID: 5362 Sessions: 0 Processed: 3 Uptime: 4m 2s /home/C: App root: /home/C * PID: 5332 Sessions: 0 Processed: 5 Uptime: 4m 18s * PID: 5369 Sessions: 0 Processed: 10 Uptime: 3m 30s