主页 / server / 问题 / 1064123
Accepted
Energya
Asked: 2021-05-20 07:51:49 +0800 CST

munin-node 工作正常，现在“启动操作超时”

  • 772

我已经让 munin-node 在我的机器上成功运行了一段时间，但最近它不再启动了。没有 munin-node 日志可供我检查，systemctl status munin-node也没有提供很多有用的信息：

[root@host /]# systemctl status munin-node
● munin-node.service - Munin Node
   Loaded: loaded (/usr/lib/systemd/system/munin-node.service; enabled; vendor preset: disabled)
   Active: failed (Result: timeout) since Tue 2021-05-18 23:35:16 CEST; 1h 8min ago
     Docs: man:munin-node(1)
           http://guide.munin-monitoring.org/en/latest/node/index.html
  Process: 7710 ExecStart=/usr/sbin/munin-node --foreground (code=exited, status=0/SUCCESS)
 Main PID: 7710 (code=exited, status=0/SUCCESS)

May 18 23:33:44 host systemd[1]: Starting Munin Node...
May 18 23:35:14 host systemd[1]: munin-node.service start operation timed out. Terminating.
May 18 23:35:16 host systemd[1]: Failed to start Munin Node.
May 18 23:35:16 host systemd[1]: Unit munin-node.service entered failed state.
May 18 23:35:16 host systemd[1]: munin-node.service failed.
service munin centos7

1 个回答

  1. Best Answer

    问题原来是插件花费的时间太长，特别是nvidia_gpu_*插件，因为它是一台多 GPU 机器。没有明确的指标表明插件导致了超时。

    为了加快nvidia_gpu_*插件速度，我使用了以下命令，基于https://forums.developer.nvidia.com/t/nvidia-smi-is-slow-on-ubuntu-16-04/50416：

    nvidia_smi --persistence-mode 1

    只需运行nvidia_smi命令即可测试其效果，因为不必先唤醒 GPU，因此加载速度会更快。

    • 0

相关问题