I searched the tickets, but could not find this. I recently upgraded from 4.1.2 or 4.2. Since this update I noticed that if a “XY.items” file is updated in the configuration and reloaded into the system, new links are added, but no links are removed. E.g., if I have in the previous “XY.items” file:
Number Endpoint_Power "example [%.2f W]" {channel="zwave:device:xxxxxx:node162:meter_watts1" }
and put a new “XY.items” file in the config with
Number Endpoint_Power "example [%.2f W]" {channel="zwave:device:xxxxxx:node162:meter_watts2" }
then after openhab has loaded it, then in the channel list the item is linked to both channels, the previous and the new one. I think this is a bug, is it ?
Hi, there is something else which I noticed with 4.2, but this is harder to quantify. I am running OH on a small Raspberry Pi 3B. With OH 4.2 I noticed that a few times the system became unresponsive running a rule triggered by an event, only doing it with up to 30 sec delays. I don’t know how to debug or trace this to make a meaningful ticket, I do lack the knowledge of the internals of OH. It is also not very frequent, in recent days I restarted the system once to fix the responsiveness. But maybe this was noticed by others who know how to trace and make this reproducible ?
I actually noticed this awhile ago (prior to 4.2), but didn’t think to report it as a bug.
For anyone else who runs into this problem, the quick-fix solution is to comment out the item, save the file, then uncomment+save to recreate it without the old links.
Hi @hmerk, this is a normal RP3, the machine always max-out the 1GB, the importance is the swap space and how badly the machine is actually swapping. You see that the swap is used 4% and the machine is perfectly fine as a web-server and logging into it, or working with the web-interfaces of OH. This machine runs fine since OH 1.7 or so, OH 4.1.2 included. Only in 4.2 the responsiveness is sometimes bad. I am trying to monitor this. I have attached some cockpit screen shots on the amount of I/O, the machine is basically idle w.r.t. I/O as you can see. But this is a normal RP3, and the system in general is fine. This is why I said, I have a hard time quantifying what the issue is, if it would be a system issue it would be simpler, but it seems to be some sort of issue in OH or ZWave or something. I fully understand that a bug that is not reproducible is hard to fix. I was just wondering if my system is the only one reporting this.
Hi @hmerk, I agree, the note is outdated and should be replaced eventually. But as I said, there is no sign that it struggles with the workload from the system monitoring (4 CPU cores, all mostly idle). I’d say a large fraction of OH installations is still on RP3s, don’t you think ? (I am planning to do the shift to a more modern machine, together with other changes to my setup I am preparing to replace outdated zwave notes).
If there are no other reports on this, then it is probably my setup. Unless I can debug this further and give more information, it is hard to speculate.