474 字
2 分钟
ProxmoxVE ( smartd ) 单独忽略某块硬盘的 SMART 报警

背景#

PVE 的系统盘是三块 NVMe 组的 mirror,其中两块寿命已经到期,但还能正常使用,暂时不打算更换。smartd 每天会发邮件提醒:

Device: /dev/nvme3, Critical Warning (0x04): Reliability
Device info: SAMSUNG MZ9LQ128HBHQ-000H1, ...

需求是单独屏蔽这块盘的报警,其他盘继续监控。

配置方法#

默认在/etc/smartd.conf中,会检测所有设备

DEVICESCAN -d removable -n standby -m root -M exec /usr/share/smartmontools/smartd-runner

感谢Arch手册smartd.conf 要排除其中一块,有两个关键点:

  1. -d ignore 指令可以让 smartd 跳过指定设备。
  2. 写在 DEVICESCAN 前面的配置会优先生效。

所以在 DEVICESCAN 上面加一行即可:

/dev/nvme3 -d ignore
DEVICESCAN -d removable -n standby -m root -M exec /usr/share/smartmontools/smartd-runner

使用smartd -q onecheck 2>&1 | grep -i monitor检查看到/dev/nvme3被排除了就没问题。

root@serverhub:~# smartd -q onecheck 2>&1 | grep -i monitor
Device: /dev/nvme0, is SMART capable. Adding to "monitor" list.
Device: /dev/nvme1, is SMART capable. Adding to "monitor" list.

补充#

大家都知道/dev/nvme3这样的写法是阴间的,因为设备变动导致在开机时重算会生成不同的id,应该使用/dev/disk/by-xxx/xxx的方式。

但是,/dev/disk/by-id/下的设备没有/dev/nvme3然而用/dev/nvme3n1这样带namespace的就不行。

root@serverhub:~# ls -l /dev/disk/by-id/
total 0
lrwxrwxrwx 1 root root 13 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448741 -> ../../nvme3n1
lrwxrwxrwx 1 root root 13 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448741_1 -> ../../nvme3n1
lrwxrwxrwx 1 root root 15 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448741_1-part1 -> ../../nvme3n1p1
lrwxrwxrwx 1 root root 15 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448741_1-part2 -> ../../nvme3n1p2
lrwxrwxrwx 1 root root 15 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448741_1-part3 -> ../../nvme3n1p3
lrwxrwxrwx 1 root root 15 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448741-part1 -> ../../nvme3n1p1
lrwxrwxrwx 1 root root 15 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448741-part2 -> ../../nvme3n1p2
lrwxrwxrwx 1 root root 15 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448741-part3 -> ../../nvme3n1p3
lrwxrwxrwx 1 root root 13 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448748 -> ../../nvme0n1
lrwxrwxrwx 1 root root 13 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448748_1 -> ../../nvme0n1
lrwxrwxrwx 1 root root 15 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448748_1-part1 -> ../../nvme0n1p1
lrwxrwxrwx 1 root root 15 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448748_1-part2 -> ../../nvme0n1p2
lrwxrwxrwx 1 root root 15 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448748_1-part3 -> ../../nvme0n1p3
lrwxrwxrwx 1 root root 15 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448748-part1 -> ../../nvme0n1p1
lrwxrwxrwx 1 root root 15 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448748-part2 -> ../../nvme0n1p2
lrwxrwxrwx 1 root root 15 Jan 8 18:26 nvme-SAMSUNG_MZ9LQ128HBHQ-000H1_S5MJNX0R448748-part3 -> ../../nvme0n1p3
ProxmoxVE ( smartd ) 单独忽略某块硬盘的 SMART 报警
https://www.homelabproject.cc/posts/proxmox/proxmoxve--smartd--单独忽略某块硬盘的-smart-报警/
作者
Channing He
发布于
2026-05-07
许可协议
CC BY-NC-SA 4.0