Repeated UI configuration mismatch with multiple z-wave devices

The moment he gets involved it’ll become clear I’ve ignored some basic instruction written in 18 point bold text somewhere… :confused:

I bet he’ll ask for some debug logs

1 Like

Already got them saved off, I’ll post one of them, the second one is for the device I just updated (the one that the UI lost the association group for), if the second device doesn’t clear itself up then I’ll post that one.

1 Like

Log for Node 17 in here, I did spot an “invalid node ID” message in there but don’t know the significance of these particularly as the config went through. I’ve yet to try updating it again as I didn’t want to disturb the crime scene… gzipped renamed to XML to allow upload.

Node17log.txt.xml (936.0 KB)

Looking at node 17, the SET to set the wakeup interval, and the GET, to read it back, are getting reversed when they are sent, and I think this is causing the issue (at least this one).

Here we see that we read the value back (it’s initially 660) then the binding sets it to 600. This all works fine, but the UI and reality are now not aligned and I think this is the issue you are reporting (correct?).

The binding is creating them in the correct order, but I guess they are being reversed somehow in the queue. This shouldn’t happen - both have the same priority, and in this case it should use the transaction ID to ensure they maintain order.

I will take a look at this.

1 Like

Yep that’s what’s been happening, I’ve been a bit busy so haven’t played with it any more yet, and also I really need to get that transaction visualiser, I’ve seen it before so I’ll track it down.

1 Like

I can’t find any problem here :confounded:. I’ve added some more tests to replicate the issue, and again it works fine -:

I’m not sure what’s happening. This code (to sort on insertion order if priority is the same) was only added recently, but it seems you’re on the latest snapshot, so it should be fine.

What system are your running on, and what JRE are you using?

It’s a virtual Debian box running on KVM, the JRE is Oracle’s rather than the default one, I’ll see if I can remember what to type to get the version numbers out:

ianr@openhab:~$ java -version
java version “1.8.0_181”
Java™ SE Runtime Environment (build 1.8.0_181-b13)
Java HotSpot™ 64-Bit Server VM (build 25.181-b13, mixed mode)
ianr@openhab:~$ javac -version
javac 1.8.0_181

I chose the Oracle one on the grounds that I hadn’t tried it before and it was recommended on the openhab install page whereas the openjdk version I used in the past seemed to be a bit problematic at times.

It’s using the Aeon Labs z-wave USB stick appearing as a serial device on ttyACM0

I’ve got another log with this kind of thing in it, reads followed by writes, is that what you’re looking for? This is for a completely different device.

Does this only happen on battery devices, or all devices?

All of them, the above screenshot was from a mains-operated node.

Thanks. I’ve found the reason for this - now I just need to work out how to eliminate it!

1 Like

Great, and sorry for the extra workload :wink:

That’s ok - it’s a significant issue that needs to be resolved so thanks for your help with useful feedback.

1 Like

Could this be the cause of the setpoint problem I am having with my CT30 thermostat posted here

At first glance, I don’t think so - with this bug, the commands would still be sent correctly and the thermostat would be updated, but the UI may be out of sync. Please provide the log (in the other thread) and I’ll have a look.

Thans @chris I did put the log files (and the log viewer rendering) in that post. If you could take a look at it I would be much obliged.

After the hair-tearing session I had today getting some motion sensors to work properly, here’s what I did to get everything into a known state.

  • close down paperui, do not use it! Even changing one configuration parameter makes it re-write all parameters and it makes a bit of a mess of it all.
  • with habmin, set it up the way you want it and “save”
  • use “advanced settings” and click on “reinitialise device”

This got my motion sensors going again and they’re now quite happy, as am I.

I’d read somewhere that a new build is incoming soon-ish, is this issue resolved in that build? I’ve been having quite a few problems with association groups vanishing during nightly heals or system restarts, and also configuration parameters on some devices being reset back to old values.