openHAB 5.1 Milestone discussion

Ah OK, thanks. I was only looking in the Milestone discussion.

Please try this URL in your WEB browser.

http://192.168.1.119:8080/rest/items?recursive=false&fields=name,type,groupType,state,stateDescription

It works for me. What error shows the browser in your case ?

This request run by the binding is apparently not returning status code 200 for you, that is what I can determine with your logs.

The browser page returns:

{ā€œerrorā€:{ā€œmessageā€:ā€œAuthentication requiredā€,ā€œhttp-codeā€:401}}

Tried addon parameters no identification. Fails.

User/password. Fails.

Token. Fails.

Ok, all my fault.

The target remote server must have the following API Security parameters:

Implicit user Role ON.

Authorize basic authentication ON.

Thank you for your advices.

Don’t know how you setup your remote OH server but this request requires no authentication by default.

Check your thing (linked to your remote OH server) parameters and how you setup the security parameters for your remote OH server.

I remember someone complained about binding no more working with 5.0 and finally he discovered that one of his thing parameters was not set properly. I hope it was not you already.

And there was of course no recent changes in the binding.

Fixed with a recent snapshot.

Hi, Lolodomo.

I’m a bit lazzy and I activated both Basic Auth and Implicitly User Role in API Security parameters on the remote OH server GUI.

But you’re right, Basic Auth is not necessary.

For your information, I just discovered that the Sonos binding is fully broken in 5.1 milestone 2. This is due to a change very recently applied in the core framework related to UPnP stuff. I will not be surprised if other bindings are also impacted. I hope this change will be reverted to make bindings working again soon.

We have this thread which at least touches on the subject, although I’m not sure if the problems initially reported there properly matches the symptoms:

I currently have UPNP working, but I’m afraid to touch Openhab as it’s so flakey. I’m not on Milestone 2 yet though (daily snapshot from a few weeks ago).

I don’t know why it keeps stopping working though. Initially, it starts normally, then a short time later it just stops responding - but sometimes it keeps working. Hence ā€œflakeyā€.

If UPNP has become unstable/broken, all bindings that rely on it will not work anymore (like my SamsungTV binding).

I’m not sure if I’m doing it right.
I’m on M2 and auto-injection is on for all scripts.
I switched on debug for Javascript but no message with ā€œtransformationā€ or similar.
The manual way works, e.g.:

|(() => { const { items } = require('openhab'); return Math.round(0.1 * parseFloat(input) + 0.9 * items.GoodweInv_PV_pbattery1_smooth.numericState); })()

For information, the potentially (probable) other impacted bindings are : wemo, onkyo and samsungtv. But this was not yet confirmed they are really broken like the Sonos binding (depends on how is structured their UPnP devices). The upnpcontrol binding could also be impacted but as the change was done for this binding, I assume it is not.

I already proposed to revert the wrong change in core framework and we also discuss a solution that will satisfy everybody including the upnpcontrol binding.

1 Like

My Onkyo is not affected, and I doubt any of the Onkyos are. I’m also relatively sure that the Samsung TVs aren’t affected, even though I don’t have one, I’ve worked with them in the past, and don’t think I’ve ever seen more than one UPnP device in their implementation.

My Samsung TV has always been the least reliable device connected to my system (to the point that the primary rule that drives interation with it now simply includes disabling and re-enabling the thing every time which increases the odds the command gets through by 15-20% or so).

Subjectively, it seems like it has, in fact, been even worse than usual since i moved to a recent snapshot (#4859), but I cannot say for sure that is the case.

1 Like

Yes, we’re dealing with two issues it seems:

My comment was about the latter, the fact that UPnP devices that have ā€œembedded devicesā€ (very few) are now broken when it comes to receiving GENA events. I don’t think that applies to Samsung TVs, but the former issue probably applies to UPnP in general.

They are simply ignored with no errors. The value of the button is ok but metadata is not updated accordingly.

Probably this is the reason why blockly can’t change metadata in M1. Any idea on when it will be fixed ?

The PR was merged last week.

1 Like

Hi all Im on 5.1.0M2

Has anyone else noticed items/linked channels dissapearing one a week?

I will have errors saying the Trigger will not work due to missing Item.

When i go into the thing all the linked channels are gone. This happened on the Zwave binding, but now its also happening on the Zwave-Js binding.

I have this same issue (recurring and maybe related to an OH restart or system restart). The Health Check pops up with 100+ orphaned links. Looking deeper it was all Zwave-JS things. A quick disable/enable of the Zwave-JS bridge corrects this problem for me, or going into each thing and clicking Save brings back the links. I was thinking about opening an issue, but I am in the middle of a system rebuild and move to a new zwave stick, so wasn’t sure if it was something I was doing.