Aeotec devices with wrong names and other issues

[lockytaylor]

Hi all,

I have a few issues with my zwave network that I was hoping someone smarter than me might be able to offer some advice.

I recently migrated my HA from a raspberry pi to docker in a container on my QNAP NAS, mostly for stability reasons. After creating a new instance of HA on my NAS, I copied my HA config from the rPi to the new HA instance and then moved my whole network over to a new Aeotec Z-stick plus (to work with the NAS) by excluding on the rPi and including every device on the newly created HA instance/z-stick. After finally adding all the devices, I then went through my config and updated all references to the new device names. PITA but had to do it.

Since doing that I am finding that I am having all kinds of issues with devices becoming unresponsive after about 3 days of use and the only way I can get them to work again is to power cycle all the aeotec devices via my meter board breakers. Heals, refreshes, reboots etc don’t make any difference, only a power cycle or exclude/re-include fixes them. They don’t go dead or anything, they just get very high RTTs and stop working. Directly after a reboot their RTTs are quite normal. I’m not positive but I suspect my network becomes slightly slower each day and is its fastest after a power cycle. Could just be my imagination tho.

Although many of my devices have energy metering, all of them have it turned off and is using basic report type to minimise chatter. I have no devices that require polling and the network is about 30-40 devices, nearly all of them powered, all in close proximity. The Zwave ‘graph’ panel shows nearly all of them as only 1 hop and most have RTT between 40 and 120ish. When newly rebooted and not having issues, the devices respond nearly instantly all over the house.

If I try trigger a switch that has stopped responding in HA, it locks up the rest of the zwave network for about 10 seconds. Once it has timed out, all the other devices work fine.

In trying to work out what is happening I have noticed that a bunch of my single nano switches are named/referenced as Dual nano switches which is very weird. For example, one nano single switch has an entity name of zwave.aeotec_zw132_dual_nano_switch_5 in the states tab of developer tools. This device is definitely not a dual unit. All I can think of is that when I copied my config over before adding all the new devices, HA has gotten confused data after adding the new devices and is confusing its name with whatever device had that nodeID in the old network. This might be a separate issue and may not be having and impact on my zwave network but some of these wrongly named devices are the ones which are becoming unresponsive on the network.

Does anyone know whether there is a chance that HA or OZW has a file I can delete that rebuilds or something that might clean up the entity data? I tried deleting zwcfg_XXXX.xml but all that did is make half of my devices ‘unknown’ and I had to exclude and re-include again.

Hope this all makes sense and I really appreciate any help people might be able to offer.

Thanks all!

[lockytaylor]

gentle bump

[lockytaylor]

Anyone with any ideas?

As an update: I went and excluded and re-included all the devices which had false names and all the names are now correct. So that part seems to have been fixed by these devices having new node ids. ie, the newly included devices are using a node id number range not used before. The old network topped out at about node id 70, I am now into the 80s on the new aeotec network. This is presuming the old network data was the issue.

However I am still having big issues with random devices just stopping working after a few days. This morning my Kitchen Left light suddenly was stuck on (noticed last night the lights started misbehaving in the kitchen) and was unresponsive. After a HA reboot, the mesh graph shows 0ms response time from it. As long as I don’t try use it, all the other lights seem ok. One thing to note, the kitchen left device is literally 10 feet from the zstick so distance or signal strength is not related to this issue

All my devices are aeotec switches and I am using an aeotec zstick gen 5 plus.

Here is my network graph:

The device 79+ that I have hovered over is the one that is suddenly dead.

The only thing I can think of that might be an issue (not convinced it is the issue) is that when I recreated my network on the aoetec from the razberry, for some reason my network key was missing from my configuration.yaml. It used to be there but somehow I must have deleted it. When I added all my aeotec devices I used the secure add node button in HA (NOT via the button on the stick) and all the devices responded with blue flashing light as though they were successfully added as a secure device. So it would seem they were added as a secure device but without the network key. I only realised this the other day so I have since added my network key and the exclude/reincludes I did above to fix the name issues was using the new network key. I don’t think this is related (they are essentially non-secure) but if anyone has any ideas on why this might be an issue please let me know. Not keen on excluding/reincluding over 40+ devices but if need be I will - was my own mistake. One important thing to note, the kitchen left that has gone unresponsive this morning was one which had the wrong name and that was reincluded 2 days ago with the network key included.

Thanks all. Really hope someone can give me a hand here - it’s causing real headaches. Losing the battle with the wife on the value of home automation!

PS: Aeotec suggested I try another system like HomeSeer - not because they have any problems with HA, just to test whether it is related to the software of the network. Not keen to have to do that yet.