Integration broken 2025.10.3

[StrandmonYellow]

Hi,

Already a while ago I found this integration. At the time, I found the installation a bit too much work, and everytime i rebooted HA it stopped working. Then recently I revisited this project, and to my surprise it was supported in HACS and the installations was very easy! I followed the instructions and got it all working. Then the next day, I updated to version 2025.10.1 and I could remove the proxy container. This also worked flawlessly for me, and it ran completely fine for a couple of days.

Then I updated to version 2025.10.3 and it completely stopped working. First, it did not initialise properly. I reconfigured the integration using my created login credentials. This fixed the initialisation error, but all the entities where unavailable and marked as "no longer supported by the integration". After trying a bunch of things, I removed the complete integration, rebooted HA and installed it again. Now I only have 1 instance left with only two entities:

The only error I could find in the logs:

Bron: helpers/entity_platform.py:451
integratie: Sensor (documentatie, problemen)
Eerst voorgekomen: 20:43:19 (1 gebeurtenis)
Laatst gelogd: 20:43:19

Error while setting up ugreen platform for sensor: invalid literal for int() with base 10: ''
Traceback (most recent call last):
  File "/usr/src/homeassistant/homeassistant/helpers/entity_platform.py", line 451, in _async_setup_platform
    await asyncio.shield(awaitable)
  File "/config/custom_components/ugreen/sensor.py", line 34, in async_setup_entry
    UgreenNasSensor(hass, entry.entry_id, config_coordinator, entity, nas_model)
    ~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/config/custom_components/ugreen/sensor.py", line 65, in __init__
    base_device_info = build_device_info(hass, self._entry_id, self._key, nas_model)
  File "/config/custom_components/ugreen/device_info.py", line 56, in build_device_info
    p = int(part0[4:])
ValueError: invalid literal for int() with base 10: ''

Thankyou in advance!

[Tom-Bom-badil]

Thanks for the log, appreciated! I received 2 more error reports of the same kind, yet not sure what is causing this (currently investigating). Could you please roll back to v2025.10.2 - it seems before I 'improved' the displaying of devices, everything was working fine. Thanks!

/tom

[buchnoun]

got exactly the same error here :
`Enregistreur: homeassistant.components.sensor
Source: helpers/entity_platform.py:451
intégration: Capteur (documentation, problèmes)
S'est produit pour la première fois: 5 octobre 2025 à 23:40:48 (1 occurrence)
Dernier enregistrement: 5 octobre 2025 à 23:40:48

Error while setting up ugreen platform for sensor: invalid literal for int() with base 10: ''
Traceback (most recent call last):
File "/usr/src/homeassistant/homeassistant/helpers/entity_platform.py", line 451, in _async_setup_platform
await asyncio.shield(awaitable)
File "/config/custom_components/ugreen/sensor.py", line 34, in async_setup_entry
UgreenNasSensor(hass, entry.entry_id, config_coordinator, entity, nas_model)
~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/config/custom_components/ugreen/sensor.py", line 65, in init
base_device_info = build_device_info(hass, self._entry_id, self._key, nas_model)
File "/config/custom_components/ugreen/device_info.py", line 56, in build_device_info
p = int(part0[4:])
ValueError: invalid literal for int() with base 10: ''`
hope you'll find a solution soon as I don't know how to roll back to previous version :)
thanks by advance

[Tom-Bom-badil]

Rolling back: Go to HACS center, uninstall, re-install, choose v2025.10.2

On the problem itself: Commenting out lines #53ff (the entire #pools block) of custom_components/ugreen/device_info.py will solve this. Note: The root device at botton needs to remain uncommented, as it' s needed: return DeviceInfo(identifiers={(DOMAIN, root_id)}).

This fix affects only the 'Device' builder' of HA - entities will work as supposed. There seem to be some NAS configurations which I do not understand right now that return invalid response that does not match the pattern 'pool1', 'pool2' etc - so commenting out the whole pool building fixes this.

/tom

[StrandmonYellow]

Thankyou for the reply! I rolled back to 2025.10.2 and for now it is working again. Lets wait until the problem is found! I was just about to write some instructions on how yo roll back. @Tom-Bom-badil you were too fast :p

[buchnoun]

I dunno if I understood well what you're saying but I've noticed a lot of entities which are not working since I've update... Would make sense to get only one apparel as @StrandmonYellow

[StrandmonYellow]

I had the same, then I reinstalled the integration and got as i posted, so there is no difference there.

[Tom-Bom-badil]

All your entites should still be there. Just the displaying on 'Appareils' screen will be broken till next version.

/tom

[buchnoun]

Ok I did reinstal 10.0.2 and it's working again, except I got a lot of indisponible entities

those appears when I've updated from v9, and all of the sudden in my dashboard all the card are red in the lower right part...

[Tom-Bom-badil]

All entities are missing it's 'Type', which is used for the color coding:

That's why they are all displayed in red.

Regarding the top-right red circle - is it pointing to the right entity sensor.ugreen_nas_ram_total_size? YAML code of the entire section should be:

type: heading
heading: Details
heading_style: title
grid_options:
  columns: 24
  rows: 1
badges:
  - type: entity
    show_state: true
    show_icon: true
    entity: sensor.ugreen_nas_cpu_model
  - type: entity
    entity: sensor.ugreen_nas_cpu_cores
  - type: entity
    entity: sensor.ugreen_nas_cpu_threads
  - type: entity
    entity: sensor.ugreen_nas_cpu_speed
  - type: entity
    entity: sensor.ugreen_nas_ram_total_size
card_mod:
  style: |
    ha-card {
      background: transparent;
    }
    ha-card .content.title p,
    ha-card .content.subtitle p {
      margin: 0;
      font-size: 4vh;
      color: white;
    }

At first, I would try another restart plus clearing the browser cache (Ctrl-F5) to check if that solves the problems.

/tom

[buchnoun]

could you tell me how i got the type back in my entities, thanks for the name of the sensor it's working now but I have no clue as why it changed ?? I'm not very good at it though (and french on top of that ;-))

[Tom-Bom-badil]

To be honest: Not really sure what is causing this. Maybe you are running an older version of the example page? That would also explain why the top-right was pointing to the wrong entity.

But let's check first if the attribute is there. Click the 'NAS Model' card again - does it show the 'NAS Part Category' in the attributes section?

If it is there, I would suggest you run through the installation instructions for the page again and replace everything that you have right now. The page code changed a couple of times meanwhile (while further developing), so maybe your code for the dashboard is just outdated.

/tom

[buchnoun]

thank you tom, it is working after I updated the code from the installation instruction. I thought everything will be updated through the integration, good to know I may have to check from time to time the code for the interface...
Thanks again great work from you !

[Tom-Bom-badil]

No - unfortunately not. Editing the Lovelace dashboard is still a bit ... well ... let's say: 'Special'.
Glad everything works again for you. :)

/tom

[Tom-Bom-badil]

Good news - I was able to track down the root cause of this error. The UGOS API is full of surprises:

• The error is affecting all users who have exactly A SINGLE disk or pool or volume.
• Normally, the API reports disk1, disk2 ... - volume1, volume2 ... - pool1, pool2 ...
• If only one device is available, UGOS strips the number at the end: Then it's only 'disk', 'pool' or 'volume'.
• As the device builder is expecting a number at the end (e.g. for building Pool 1 | Volume 1, which would now be Pool | Volume), it crashes at this point - simply because the expected number is not found.

A fix for this behaviour is available here. I would appreciate if some of you who reported this error could help with testing:

• install latest v2025.10.3,
• replace the custom_components/ugreen/device_info.py with the above linked code (you can use the 'Copy raw file' button top-right on Github, then replace the entire content of device_info.py in your HA editor by pasting),
• restart HA.

After that, everything should be working again as expected.

/tom

[StrandmonYellow]

I can only turn them off, do you know how I can completely remove them?

[Tom-Bom-badil]

I can only turn them off, do you know how I can completely remove them?

These are so-called 'ghost devices' - a common problem in HA. Quite often they are refrerenced a second time by another Integration (e.g. a 'device tracker' of your router - due to same MAC address, HA autoconnected it). Thats likely also the reason why they didn't disappear automatically.

There a different ways how to solve this, depending on the cause. You should Google 'remove home assistant ghost devices' to find out which one is applicable here.

/tom

[StrandmonYellow]

Okay, thanks for the explanation

[ctgiant]

Tested here and it is now working correctly. The one odd thing that happened though is that it added the NAS twice. Once deleting the second instance all is working as it should.

@StrandmonYellow Easiest way to get rid of the ghost devices is to delete the entire NAS and add it back in.

[gyuris]

Tested and works here also.

[Tom-Bom-badil]

v2025.10.4 is out - thanks everybody for reporting and testing!
In case you find any further errors, please create a new discussion.

/tom