beszel icon indicating copy to clipboard operation
beszel copied to clipboard

[Bug]: Disk I/O Monitoring Unreliable After Reboot Due to Non-Persistent Device Names

Open Korveck14 opened this issue 2 months ago • 3 comments

Component

Agent

Description

After system reboot, Disk I/O monitoring tracks the wrong physical disks. I have multiple SATA disks, and the Linux kernel assigns /dev/sda, /dev/sdb, etc., randomly based on detection order during boot. This causes I/O metrics to be assigned to incorrect physical disks. Could you look into it?

Expected Behavior

I/O metrics across all disk consistent after system reboot.

Steps to Reproduce

  1. Set up Disk I/O monitoring for a specific disk (e.g., /dev/sdb)
  2. Reboot the system
  3. Observe that metrics now track whatever disk happened to get assigned that device name

Category

Metrics

Affected Metrics

Storage

OS / Architecture

linux/amd64 (agent)

Beszel version

0.13.2

Installation method

Docker

Configuration

    volumes:
      - /mnt:/extra-filesystems/sda1__Media:ro
      - /backup:/extra-filesystems/sdb1__Backup:ro

Hub Logs


Agent Logs


Korveck14 avatar Oct 17 '25 07:10 Korveck14

Have you tried not specifying the device name in the mount? In most cases it should be able to pick up the correct device for I/O automatically:

    volumes:
      - /mnt:/extra-filesystems/media__Media:ro
      - /backup:/extra-filesystems/backup__Backup:ro

This doesn't always work (for example if the device is encrypted it probably will not) so let me know if you don't have luck with this approach.

henrygd avatar Oct 19 '25 15:10 henrygd

It is not working, for reference these are my disks and they are not encrypted

NAME     SIZE FSTYPE MOUNTPOINT LABEL
sda    476.9G
├─sda1   976M vfat   /boot/efi
├─sda2 460.2G ext4   /
└─sda3  15.8G swap   [SWAP]
sdb      1.9T
└─sdb1   1.9T ext4   /mnt       data

I also tried specifying the label with no luck:

    volumes:
      - /mnt:/extra-filesystems/data__Media:ro

Korveck14 avatar Oct 19 '25 19:10 Korveck14

Also having this issue, I have my disks defined as such:

    volumes:
      - /mnt/disk01/.beszel:/extra-filesystems/disk01:ro
      - /mnt/disk02/.beszel:/extra-filesystems/disk02:ro
      - /mnt/disk03/.beszel:/extra-filesystems/disk03:ro
      - /mnt/disk04/.beszel:/extra-filesystems/disk04:ro
      - /mnt/disk05/.beszel:/extra-filesystems/disk05:ro
      - /mnt/disk06/.beszel:/extra-filesystems/disk06:ro
      - /mnt/parity01/.beszel:/extra-filesystems/parity01:ro

And I get no disk I/O stats and this shows up in the agent logs:

2025/11/17 00:20:26 WARN Device not found in diskstats name=disk06
2025/11/17 00:20:26 WARN Device not found in diskstats name=parity01
2025/11/17 00:20:26 WARN Device not found in diskstats name=disk01
2025/11/17 00:20:26 WARN Device not found in diskstats name=disk02
2025/11/17 00:20:26 WARN Device not found in diskstats name=disk03
2025/11/17 00:20:26 WARN Device not found in diskstats name=disk04
2025/11/17 00:20:26 WARN Device not found in diskstats name=disk05

com6056 avatar Nov 18 '25 05:11 com6056