Node is not rediscovered once it was offline

[rocktale]

I'm running version 2.0.0 of the plugin with Home Assistant 2023.6.0 (but the issue was already present with some of the previous versions of it) and it seems like the fix introduced with #42 does no longer work (at least for me). Maybe something was changed with the refactoring that went into the multi nodes setup in 2.0.0?

I have only one node configured that is not online all the time. When I reconfigure the integration it works like a charm. But once the node was offline and it comes back, all sensors show "unavailable" as if it was not there. Basically the same as when it was offline. I waited for at least half an hour but without any luck.

Would be nice if the integration could automatically detect that the node is back. Optimally, with a configurable retry interval.

[VeeeneX]

It's a bug but it has a workaround not a nice one but it works :D

I have a script called Reload proxmoxve:

alias: Reload Proxmox
description: ""
trigger:
  - platform: time_pattern
    minutes: /1
condition:
  - condition: state
    entity_id: binary_sensor.void
    state: "on"
action:
  - service: rest_command.reload_proxmoxve
    data: {}
mode: single

It calls rest_command that reload integration 😅 only when server is reachable over network and it does so every minute, you could preserve state somewhere to crate a simple State Machine and avoid repeated reloads but that's up to you. :)

[kumarunster]

same issue here

@VeeeneX could you share full example? where is the service 'rest_command.reload_proxmoxve' defined?

UPD:

found a good way to reload integration:

action:
  - service: homeassistant.reload_config_entry
    data:
      entry_id: <config_id>

where the config_id can be found via url (click on ProxmoxVE integration in integration/devices: config/integrations/integration/proxmoxve#config_entry=<config_id>

[VeeeneX]

Sure, this is my setup. But I like yours more 😄

rest_command:
  reload_proxmoxve:
    url: http://127.0.0.1:8123/api/config/config_entries/entry/89d77de9b91cb775b4bc24ceb646f208/reload
    method: POST
    headers:
      authorization: "Bearer <TOKEN>"
      accept: "application/json"
    content_type:  'application/json; charset=utf-8'
    verify_ssl: false

[porklander]

same issue here

@VeeeneX could you share full example? where is the service 'rest_command.reload_proxmoxve' defined?

UPD:

found a good way to reload integration:

action:
  - service: homeassistant.reload_config_entry
    data:
      entry_id: <config_id>

where the config_id can be found via url (click on ProxmoxVE integration in integration/devices: config/integrations/integration/proxmoxve#config_entry=<config_id>

sorry for that question but I can't find the <config_id>. Can you explain me where I can find it?

[kumarunster]

@porklander already explained at the end of my comment, where are you stuck?

[porklander]

Ich hoffe es ist im wenn ich auf deutsch weiterschreibe. Wenn ich die du schreibst auf ProxmoxVE integration in integration/devices gehe habe finde ich die config_id in der URL nicht. Irgendwas übersehe ich vermutlich. Sorry für die blöde Nachfrage. Danke Am 18. Juli 2023 10:57:40 MESZ schrieb Nikolai Raitsev ***@***.***>:

…

@porklander already explained at the end of my comment, where are you stuck? -- Reply to this email directly or view it on GitHub: #93 (comment) You are receiving this because you were mentioned. Message ID: ***@***.***>

[porklander]

[kumarunster]

hi, click auf Dienste unter "Intergrationseinträge" um auf der Hauptseite der Integration zu landen

[porklander]

Danke, mich hat das dashboard?historyBack=1 davor irritiert. Danke für deine Hilfe. Am 18. Juli 2023 11:30:22 MESZ schrieb Nikolai Raitsev ***@***.***>:

…

hi, click auf Dienste unter "Intergrationseinträge" um auf der Hauptseite der Integration zu landen <img width="720" alt="Bildschirmfoto 2023-07-18 um 11 26 50" src="https://github.com/dougiteixeira/proxmoxve/assets/2104820/9fa234ff-4bfd-45ac-8dbc-3cc97e4efd41"> <img width="798" alt="Bildschirmfoto 2023-07-18 um 11 27 26" src="https://github.com/dougiteixeira/proxmoxve/assets/2104820/2ed2c8de-4a33-4104-a9b9-a46603ef34ef"> -- Reply to this email directly or view it on GitHub: #93 (comment) You are receiving this because you were mentioned. Message ID: ***@***.***>

[kumarunster]

Kein Problem, viel Erfolg damit!

[VeeeneX]

As much as I like German 😅, for English speaking friends: Each integration has it's own id expressed as config_id. Thanks for pictures though!

[dougiteixeira]

I created a blueprint to work around this situation...

But it involves creating a sensor with Ping integration.

Since I only have one Proxmox host, I can't fully test it...

Follow the link, please, who can test it and give some feedback on its operation or bugs I appreciate it.

After getting some feedback here I include it in the integration documentation.

[AdmiralRaccoon]

Hi,
I´m using the blueprint in a 2 bode environment, and since the 2nd node is for testing only and not 24/7 online, I´ve tested your blueprint and can confirm it´s working!

Thanks for that
Philipp

[Munsio]

I do have the same problem since a few days that some nodes are still not there after I turned them on. What I found is the following error in the log maybe it has something to do with it:

Logger: urllib3.connectionpool
Source: custom_components/proxmoxve/coordinator.py:64
Integration: Proxmox VE (documentation, issues)
First occurred: August 27, 2023 at 14:33:23 (3 occurrences)
Last logged: August 27, 2023 at 14:55:23
Connection pool is full, discarding connection: pve-asus1.XXXXXXX.com. Connection pool size: 10

[dougiteixeira]

Can anyone confirm if this behavior still occurs in the latest version of the integration?

[Munsio]

Unfortunately for me it does not work 100%

HA: 2024.1.2
Addon: 3.3.0

• Disabled the reload Automation
• Updated Extension
• Rebooted HA
• Started the Node
• Entities are unavailable
• Run Automation without enabling it - which reloaded the extension
• Entities are shown as wanted
• Shutdown Node
• Entities are unavailable
• Started Node again
• Entities are shown as wanted
• Did the whole turn off/on of the node again
• Entities are shown as wanted

So it seems that if the node was not running before the HA reboot, then starting the node afterwards does not work. Everything else did work as expected.

Will try the HA reboot node shutdown part again.

Edit: Can confirm, if a HA reboot is in between turning on a node the data is not refreshed for it.

[e87131]

This still appears to be the case as of today, v3.4.4, using HomeAssistant 2024.10.4
Is the workaround the go-to solution or is there anything planned on integrating the reauth/reload into the integration itself?