[BUG] Not seeing disk space usage with docker

[FaizVisram]

Go Hass Agent Version
v14.6.0

Describe the bug
There are sensors for the only disk drive in my machine but it's only reporting read/read rates/write/write rates but not disk space free/used/total.

To Reproduce
Steps to reproduce the behavior:

1. Do first time registration
2. Start container with this docker-compose.yml

services:
    go-hass-agent:
        container_name: go-hass-agent
        network_mode: host
        privileged: true
        volumes:
            - /services/go-hass-agent/config:/home/go-hass-agent
            - /proc:/host/proc:ro
            - /sys:/host/sys:ro
            - /run/dbus:/run/dbus:ro
            - /:/host:ro
        environment:
            - PROCFS_ROOT=/host/proc
            - SYSFS_ROOT=/host/sys
        cap_add:
            - SYS_ADMIN
            - SYS_RAWIO
            - MKNOD
        image: ghcr.io/joshuar/go-hass-agent:v14.6.0

4. Go to Home Assistant device page
5. See entities

Expected behavior
Sensors for disk space free/used/total.

Screenshots
If applicable, add screenshots to help explain your problem.

Logs

Container Logs:
2:11AM WARN Unable to add desktop portal to context. Some sensors requring it may not be available. error="no portal present"
2:11AM WARN Unable to determine user session path from D-Bus. Some sensors requring it may not be available. error="could not determine session path"
2:11AM WARN Unable to set up pipewire monitor. error="error starting pw-dump: exec: "pw-dump": executable file not found in $PATH"
2:11AM WARN Could not init worker. worker=smart_status error="check capabilities: capability check failed: missing cap_sys_rawio"
2:11AM WARN Could not init worker. worker=running_apps error="get desktop portal: no desktop portal in context"
2:11AM WARN Could not init worker. worker=desktop_settings error="get desktop portal: no desktop portal in context"
2:11AM WARN Could not init worker. worker=mic_in_use error="no pipewire monitor in context"
2:11AM WARN Could not init worker. worker=webcam_in_use error="no pipewire monitor in context"
2:11AM WARN Could not init worker. worker=screen_lock error="get session path: no session path in context"
2:11AM WARN Could not init worker. worker=chrony error="find chrony executable: exec: "chronyc": executable file not found in $PATH"
2:11AM WARN Could not initialize input device monitoring. error="no input devices found"
2:11AM WARN Could not init worker. worker=abrt error="get abrt problems: SystemBus: unable to get data org.freedesktop.problems.GetProblems from org.freedesktop.problems: The name org.freedesktop.problems was not provided by any .service files"

I see no such file ~/.config/go-hass-agent/go-hass-agent.log in the config volume

Desktop (please complete the following information):

• OS: Linux
• Distribution: Debian
• Version: 13

Additional context
Add any other context about the problem here.

[joshuar]

Hey there, thanks for trying out Go Hass Agent and reporting this issue. That is indeed a bug! I've made a fix with df2e8e8 which will be in the next version which I'll release in the next few days.

Thanks!

[joshuar]

Hey there, v14.7.0 should fix this issue. Please try it out and report back.

Cheers!

[FaizVisram]

Hi @joshuar, thanks for the fix! I definitely see disk usage now but I have a couple questions.

1. Should I be seeing all of the following and manually add them to the ignore list or should the default ignore list include some of these?

   Mountpoint /dev Usage
   Mountpoint /dev/shm Usage
   Mountpoint /etc/hostname Usage
   Mountpoint /etc/hosts Usage
   Mountpoint /etc/resolv.conf Usage
   Mountpoint /home/go-hass-agent Usage
   Mountpoint /host Usage
   Mountpoint /host/boot/efi Usage
   Mountpoint /host/dev/shm Usage
   Mountpoint /host/run Usage
   Mountpoint /host/run/credentials/systemd-journald.service Usage
   Mountpoint /host/run/lock Usage
   Mountpoint /host/run/user/112 Usage
   Mountpoint /host/run/user/1000 Usage
   Mountpoint /host/tmp Usage
   

2. Is there a way to see the actual free or used space and total space in (T|G|M|K)bytes? I only see percentages here.

Side note, thanks for your work! The agent is great 😄

[joshuar]

Hey there, thank you for the kind words, I'm glad it is useful!

I deliberately chose not to ignore /dev/ by default, some people may want to track /dev/shm usage. /etc is ignored by default, but because you have an existing configuration, it won't be added to it so you'll need to add it manually to the ignore list. Good point on ignoring /host, I'll add that to the defaults (but you'll need to add it manually).

The sensor attributes will have the block usage and the block size, which you can use to calculate the usage in bytes. But I'll add attributes that does it for you (likely in bytes which you can convert to whatever unit size you want).

Cheers!

[FaizVisram]

That sounds great, thank you! :)

And sounds good, I'll add these extra to my ignore list. Thanks for the support!

[joshuar]

Hey there, I'll close this issue off, as I believe the disk usage issue is fixed and the latest version also has the extra attributes mentioned in #747 (comment).

If there is still an issue, please re-open and we can continue to troubleshoot.

Cheers!