Network binding shows wrong online status since Build 1020

I am using the OH2 network binding to check if some devices in the network are online. This worked perfectly for months now, but since I switched from OH2.2 1003 to 1020 the items always show the status as online, even if the device (my smart phone) is not in the network.
Here is my thing definition:

network:device:192_168_178_66 "IP: Handy " [ hostname="", port="0", retry="1", timeout="5000", refresh_interval="60000", use_system_ping="false", dhcplisten="true" ]

and the related item:

Switch ON_MVHAND "Handy " { channel="network:device:192_168_178_66:online" }

When I ping the device in the networtk from shell, I get only timeouts and lost packages. But the binding for some reason states that the switch is on, which is wrong…

Any ideas?

As a workaround I installed meanwhile the old OH1.x NetworkHealth-Binding as legacy binding additionally in OH2. The old binding works correctly, but the OH2 Network binding is just telling nonesense about the online state of my devices.
Is anyone experiencing the same problem?

1008, settings like you, everything is fine.
Updated to 1023, everything works.

Which OS are you using? I am on Windows.


The network binding interface changed. Please re-read the readme. It is pingdevice now, port is ignored and the other parameters are binding config instead of thing config now. So it will default to use system pings while you want java pings instead.

Please be aware that snapshots may contain those breaking changes. It is not yet finally decided if pingdevice will be renamed again or the entire bindi g is renamed.

Cheers David

1 Like

So the conclusion is that the binding is not reliably usable at the moment…where are these breaking changes announced?
I noticed that you apparently changed the binding description accordingly in the docs today. BUT: Does this make sense? All the people using a stable build are lost now (as I was before with Snapshot 1020).
I would propose to include a description starting from build 1020 (or whenever things were changed) and a description for all earlier builds with a clear reference to which version needs what kind of thing/item definition.

Breaking changes in snapshots are not announced anywhere but tagged with a breaking change tag. The release notes of each released version contain detailed information of breaking changes.

The readme in the repository got updated with the change. The doc is automatically build every week I think so may be delayed.

And no it does not make sense to show this to stable build users. But what would be the alternative.

Cheers David

Breaking changes could perhaps be announced on the forum with buildnumber? I dont have any numbers but my impression is that a lot of people use the snapshot release and it could save them alot of time figuring out whats wrong.

Just a suggestion…:slight_smile:

Well, documentation should always be related to a relase/version/build especially if the described functionally changes.
What is so difficult in adding several lines in the docu file like

From OH2.1.0 Stable - OHSnapshot2.2.0 Build 1003
—Docu for these versions–

From OHSnapshot2.2.0 Build 1003 onwards
—Docu for these versions–

Not really rocket science towards my mind and helpful for everybody…