(Zooz 700 series) controller shows no neighbors #3810
Comments
|
Here is the controller's JSON. As mentioned above, the controller shows no neighbors. 😢 {
"id": 1,
"name": "",
"loc": "",
"values": [],
"groups": [],
"neighbors": [],
"ready": true,
"available": true,
"hassDevices": {},
"failed": false,
"inited": true,
"hexId": "0x0000-0x0004-0x0004",
"dbLink": "https://devices.zwave-js.io/?jumpTo=0x0000:0x0004:0x0004:7.15",
"manufacturerId": 0,
"productId": 4,
"productType": 4,
"deviceConfig": {
"filename": "/data/db/devices/0x0000/700_series_controller.json",
"isEmbedded": true,
"manufacturer": "Silicon Labs",
"manufacturerId": 0,
"label": "ZST10-700",
"description": "700 Series-based Controller",
"devices": [
{
"productType": 4,
"productId": 4
}
],
"firmwareVersion": {
"min": "0.0",
"max": "255.255"
}
},
"productLabel": "ZST10-700",
"productDescription": "700 Series-based Controller",
"manufacturer": "Silicon Labs",
"firmwareVersion": "7.15",
"protocolVersion": 3,
"nodeType": 0,
"endpointsCount": 0,
"endpointIndizes": [],
"isSecure": false,
"security": "None",
"supportsSecurity": false,
"supportsBeaming": true,
"isControllerNode": true,
"isListening": true,
"isFrequentListening": false,
"isRouting": true,
"keepAwake": false,
"maxDataRate": 100000,
"deviceClass": {
"basic": 2,
"generic": 2,
"specific": 1
},
"deviceId": "0-4-4",
"status": "Alive",
"interviewStage": "Complete",
"lastActive": 1638267648970,
"isBatteryPowered": false,
"statistics": {
"messagesTX": 10,
"messagesRX": 130,
"messagesDroppedRX": 0,
"NAK": 0,
"CAN": 2,
"timeoutACK": 0,
"timeoutResponse": 0,
"timeoutCallback": 0,
"messagesDroppedTX": 0
},
"_name": "NodeID_1"
} |
|
The devices that can see the controller at the end of the node list here are all battery powered (and thus won't forward messages): |
|
If the controller doesn't "see" any neighbors (or rather won't tell us), I'm afraid there's not much we can do. Other than that, you'll have to talk to the Zooz and/or Silicon Labs (who make the firmware). |
|
I have the same controller @Eriner - mine seems to be working fine. If you need me to be a control for any testing, let me know.
|
|
That is unfortunate. And I presume it's not helpful, but for the sake of completeness here is the log of the controller failing to heal: @AlCalzone I'm going to futz with this a bit more. Will send the nvm backup file along this evening, thanks. |
hints at connectivity issues. Z-Wave sticks in particular are prone to interference by USB ports, especially by USB3 ports. We recommend putting the stick in a suitable location:
|
|
As I mentioned in my original issue, i've tried a 6ft extension cord, ultimately moving the device into the middle of my basement. Just for kicks, I just pulled out a powered hub on an extension, now have it half-way up the staircase giving it verticality. Right above the staircase is an invoelli switch (just behind the thin ply door). However my observation is that the device seems to adequately perform. For example, I was able to upload firmware to one of my close-proximity nodes without issue, whereas the firmware updates fail from further-in-physical-proximity nodes, as expected. And in all fairness, my previous Aeotec 500 series stick performed great in these specific conditions 😈 :
I understand that the new 700 series devices don't have the same filters as the 500 series controllers. |
|
One other difference between this configuration and the previous configuration (with the Aeotec stick) is the whole reason I shelled out for the zooz stick: S2. All of my powered nodes are connected via Authenticated S2 except for one, node 4, which is not in close physical proximity to the controller (it's a zen15 - I moved it outside after setting up the inovelli switches). |
I hate to tell you, but you didn't have to switch for that. S2 is handled by the host software (Z-Wave JS). The only thing that was missing until recently is SmartStart support, but Aeotec included that with the 1.02 firmware update.
It is certainly not an ideal scenario, but firmware updates are not expected to fail on further away nodes. |
Then this may indeed indicate a degraded network. My understanding was that firmware upgrades couldn't be relayed, and that the device had to be in close proximity to the controller for (direct) upgrades. Regardless, my controller can perform high-throughput firmware upgrades to some nodes that are also not listed as neighbors.
IIRC, the stick I had was the older gen5 aeotec stick, not the gen5+. IDK if that would have mattered, but something I read must have made me think I needed something newer. Maybe a firmware update would have sufficed... regardless here I am. |
They are using the same communication method as all the other commands. The only downside is that you're sending roughly 2-3000 commands in quick succession which can get problematic for the rest of the communication in the network while it is going on. Being closer means this is over quicker. |
|
Ah, good to know, thanks! |
|
@Eriner I was reading your last comment in the other issue... How many nodes are in your network? Do you think you have a busy network? I came across a Reddit post from the HomeSeer developers, and they claim that the 700-series are broken for "networks that have a lot of traffic". They are currently not recommending 700-series yet, even their own G3 controller. That latter statement is kind of hard to quantify (what is "a lot" of traffic?), but since you've said you didn't have any problems with the 500-series controller, I'm wondering if you're hitting that problem. Not sure how to even confirm if that's the case. I personally have no issues with my 700 controller, but my network is pretty small (less than 20 nodes) and mostly idle. |
That's consistent with some of the things I've seen. Earlier firmwares failed to discover neighbors after a certain size (part of healing), maybe this is a different manifestation of the same problem. @jmgiaever your network is pretty large too, right? |
|
I would think the amount of traffic would vary with connectivity - more devices that require relaying means more traffic. To emphasize my plight, when using my aeotec stick in the back of my server rack, in my basement, connected without an extension cable straight into a USB3 port.... all but a few devices were neighbors of the controller and connected directly, no hops. @AlCalzone After using multiple extensions, powered hubs, etc, I can get my zooz 700 controller to see the invovelli switches that are now only ~15 ft away. But it can't see/neighbor with the switches that are 20ft away. Nothing about this seems right - as I've stated before my aeotec stick worked great a worse environment (no extensions, server rack, basement) and this zooz stick is just eating dirt. Again, a difference here being that the inovelli switches (my powered relays) are now all using authenticated S2. My only other powered relay (the zen15) is connected insecurely. To directly answer your "large or not" question, I have: 5x BE469ZP Schlage locks 53 total devices. |
That doesn't make a difference. Encryption happens one level higher than we're looking at here.
That's on the larger side already (definitely larger than the majority of Z-Wave JS users). |
Interesting. I'm a pretty early adopter, have grown the network over two moves 😅 Does the number of nodes really matter that much if the majority are powered, zwave plus nodes? With my aeotec stick everything was working great, zero problems ever. |
It shouldn't, but apparently it does. |
|
~90 second video showing my layout that should explain why things are surely amiss somewhere... https://noagendatube.com/videos/watch/f5a38945-8b33-4e6d-8a80-6123a78fee02 |
|
If it were me, I would consider migrating back to the Aeotec, and see if the problem is solved. If you upgrade the firmware to 1.02 first, you'll get SmartStart support. At that point, there's not much difference in functionality between it and a 700-series controller. If that works, then you'll have a better user experience and it will probably prove that there is an issue with the 700-series. If it doesn't work, it was a big waste of time. It's no picnic to migrate a network of your size, so I feel a little bad even suggesting it. 😢 At least SmartStart makes it easier. Otherwise, if it really is a problem with the controller, you'll have to wait for some firmware upgrade with no ETA. If you can tolerate the occasional reset of the controller, then maybe it's worth it to avoid the migration and tough it out for now. Unless there are any new revelations, those seem like the options for now. You could also try to remove some nodes and make the network smaller, and see if the problem goes away at some certain node count. Of course, you can't use those nodes unless you include them on the Aeotec in a separate network. I've watched the product page for the HomeSeer stick bump release dates from October, to November, and now December 14th. Maybe wait until then to make a decision. If they go on sale, then perhaps the problem is fixed. If it's delayed again, I would assume they are waiting for a fix. You can signup for in-stock notifications. 😄 |
Yeah, I have 37 smartstart entries. I assume there's a JSON file/section somewhere I can pull to back that up locally just in case... THAT part (getting those entries) isn't something I'd like to do again, as it involves removing switchplates.
I vaguely remember reading somewhere that zooz isn't/won't offer firmware upgrades for controllers, but that may or may not be accurate. Or it could have been for the 500 series controller. C'mmon zooz, giving the same model number to your 500 and 700 series controller? Seriously?
I was going to wait for the Aeotec stick, after I had such great success with their Gen5 stick. But a friend said "just buy the zooz one on amazon, it'll be fine..." and I didn't listen to my gut, caved, and here I am. I'm going to stick around to see if @AlCalzone's revert of some things suddenly fix it. If not, assuming I can back-up the smartstart entries, I'll try the following (in order):
|
that's in the
All 700-series controllers use the same firmware, which is available from Silabs.
You can try building your own container (or just run zwavejs2mqtt natively and update zwave-js there): |
|
Great on all counts, thanks @AlCalzone. I'll get that backup handled and see if that branch makes a difference. |
Yeah, 56 nodes and I have still maybe 20 to add, but I've waited to do that. Seems to me that the battery devices is causing most problems.... |
I meant use the "in stock notification" as a signal that maybe the issue has been fixed. I'm not suggesting you buy a new stick. If HomeSeer starts selling their 700-series stick, then that would indicate they think the issue is fixed. Apparently it was for sale earlier, but not now. I found the post in the HomeSeer forums, here is the start of the conversation from the developer. There is a lot of information in that and following posts, it's nice they are providing updates. Here is the latest status where the issue has been confirmed by SL.
@Eriner Are you seeing command timeouts when your problem starts? Do one of the Z-Wave JS timeout driver settings correspond to the HomeSeer transmit timeout? I'm not sure what any of this has to do with the neighbor listings though... is that another symptom resulting from aborted commands? |
It could be related, not sure. The heal of the controller often simply times out. Sometimes it tries all 5 neighbor updates (actually tries), other times it tries once, then spams the log saying it tried 4 more times (but clearly didn't in ~20ms). Could be the stick is just locked.
@jmgiaever what makes you say that? I set up all my powered devices first, without issue. Everything worked. Then I stared adding battery devices and things became a problem. Have you had an identical experience? |
Same. I also had no big issues with healing until then, but now I have the same problem as you. It fails instantly until the z2m is restarted. Sometimes I have to unplug the stick too, but that might be just on a heavy load. I do have many devices, but now barely a drop (or duplicate command) on a «normal» network. It happens every now and then during heal or interview, but I guess that is normal. |
To issue the healing commands, the controller must be able to send. If it can't (or thinks it can't), it will call back with a failure. I think both issues are pretty related. |
|
@jmgiaever on my old network I had all of my devices connected as insecure. For the 700 stick rebuild, I tried adding all of my battery devices as S0. It "worked" but their inclusion state always returned "added node XX as undefined", followed by a message "added node XX with security None" (or something like that). I'm not sure if that ^ is related to the issue at hand at all or not, just trying to find more potential commonalities. Edit: Oops, I tagged jamarzka but meant to tag @jmgiaever (which GitHub doesn't tab complete, for some reason). Sorry jamarzka! |
|
@Eriner That seems to me like the communication issues were already present directly after inclusion. The exchange of security info is pretty sensitive to dropped messages. Side note: we don't recommend using S0 unless required by the device type. |
|
@AlCalzone interesting.
These devices are all zwave plus without a DSK. I'm adding all of the devices as S0 because I want to force encryption. Is this an incorrect approach? |
|
It's your choice. S0 adds Encryption but triples the messages. If a network isn't 100% reliable, you'll see a higher rate of message failures or longer delays. |
|
Understood. Even if it triples the messages, given my large number of powered plus nodes (which are well distributed) I wouldn't think it would be much of an issue for my use-case. I can try adding them as default/insecure when I rebuild to see if it matters. |
|
I've followed what's been posted in #3842 and #2545, cheers @kars85 and @kpine. Anyway, I updated to the latest release (with the revert) and no dice - nothing changed. Tried pulling the stick, healing, etc. Same symptoms of the stick becoming unresponsive. I upgraded to 7.16.3... firmware upgrade worked, but the network is remains borked. Tried manual healing of devices, no dice. :( |
|
Duplicate of #3906 |
Is your problem within Home Assistant (Core or Z-Wave JS Integration)?
NO, my problem is NOT within Home Assistant or the ZWave JS integration
Is your problem within ZWaveJS2MQTT?
NO, my problem is NOT within ZWaveJS2MQTT
Checklist
I have checked the troubleshooting section and my problem is not described there.
I have read the changelog and my problem was not mentioned there.
Describe the bug
Issue is similar to #2335: flat graph and controller has no neighbors.
Device information
How are you using
node-zwave-js?zwavejs2mqttDocker image (latest)zwavejs2mqttDocker image (dev)zwavejs2mqttDocker manually built (please specify branches)ioBroker.zwave2adapter (please specify version)HomeAssistant zwave_jsintegration (please specify version)pkgnode-red-contrib-zwave-js(please specify version, double click node to find out)Which branches or versions?
Homeassistant zwavejs2mqtt (addon docker wrapper thing).
zwavejs 8.8.2
zwavejs2mqtt 6.0.2
Did you change anything?
yes (please describe)
If yes, what did you change?
Previously was using an older aeotec 500 stick, no issues.
Excluded devices, added new Zooz 700 series stick, reincluded devices starting with the powered devices in closest physical proximity. I understand that RF doesn't behave in a straight line, but I did the best I could. This went fine.
Some point down the line, when adding the battery powered devices, the controller seems to have "lost" its neighbors. The network seems to (mostly?) function, but I've stopped adding devices pending investigation of this issue.
I've tried enabling and disabling soft reset. I've tried healing the controller. I've tried re-interviewing the controller. I've tried soft-resetting the controller. I've tried healing the nodes in proximity to the controller manually (rather than batched "heal all"). I've tried different USB ports. I've tried a 6ft extension cable (which my aoetec stick never used or needed).
Log attached includes one false-start (I fatfingered) and then a full start.
Did this work before?
Yes (please describe)
If yes, where did it work?
No problems when I was using the Aeotec 500 series stick.
Attach Driver Logfile
zwavejs_2021-11-30.log
The text was updated successfully, but these errors were encountered: