#Software Beta for Nanoleaf Essentials Matter
1 messages · Page 5 of 1
you use home assistant?
Yeah, I'm currently building this project for my NGO facility, the goal is to fully automated everything, lights, speakers, curtains. The facility is on a 10 acres land, and new construction is happening soon to increase the size of the building as well
In terms of technical stuff, infrastructure is up and running, I just need to watch for nanoleaf updates now and wait for things to improve before I start scaling Iot devices
yeah, good lucky standing near every bulb to update them 🤣
you can only do 3 at a time
Haaa yeah, I wonder how it's going to work out
you'll beat your head against a wall when an update fails for no particular reason soon enough
yep
Atm I don't have many nanoleaf bulbs, total of 7, 4 of them are the downlights and the 3 of them are the e27 matter bulbs
Hahaha I realised that, that's why I'm holding off on the scaling part
I was looking at placing a bulk order of 200 before I started encountering issues
you are super trusting in them
200? phwoar, I think the most someone has in here is ~60?
I spent like a week research, not much is available on the internet, until I saw a video of the developers at home assistant mentioning that there is a discord group, and then I ended up here
full send or no send
Thank god
ahaha yep, we try these days to keep all the nl complants to a single thread within the matter channel, as its just being mentioned too often
Haha yeah, so no one has tested matter over thread at this scale before?
I think I found 1 person on the home assistant forum with 160 devices
maybe outside of a dev enviroment, yeah
but you would need a few good and stable border routers lmao
Do software updates get pushed over Bluetooth? If so do you think that will change to being pushed over Thread?
So currently I just run openthread on a raspberry pi 5, I've attached and nvme ssd with and put it in a pc case to keep it extra extra cool, along with some other machines. I'm trying my best with limited knowledge
i doubt any time soon, issue is they they run their own stack as well as the matter stack
so pushing over BLE is what they know works, especally when their thread/matter is barley working atm in the first place
But theoretically it's possible right?
yes, for example eve does it already
But they dont run their own stack besides matter.
apple are currently the only ones that support OTA updates via matter, home assistant very well could but it isnt the easiest to add, especally with only 1 vendor officially supporting it
But that limits them to only being fully functional on apples ecosystem.
i think google can do it too
i havent seen any gui that would indicate it can? maybe thats changed recently tho
But also googles border routers disconnect from your wifi when you loose internet connectivity
not good for matter 🙃
nope, they will get kicked out of my home soon 😅
i mean, google seems to be killing fuchsia, so thats maybe not a bad idea 🤣
Any experience with Aqara matter products?
hmm yeah. what ones
Just the door sensors, I think those are the only matter over thread products they have
they are a bit rocky
Pretty reliable I hope? I need some sensors that's why I ask lol
look it up in the HA discord, its okish
just has a shitty soc from some random chinese vendor
forgot the name
eve ones honestly
Idk where you get this from but aqara has very solid devices. I've never seen anyone complain about them. While EVE is good I think they are overpriced
early fw versions of them suffered from the same issues as the eve ones did, could of changes since i left them at the old house i was at. but could very well have changed
I've got several P2s and they've been reliable. Their ZigBee stuff also works great
Will have to test the response time for these, I was originally leaning towards zigbee ones cuz they are a lot zappy, do we know what soc the philips hue bulbs/downlights use? I know philips have made there hub matter over thread compatible, but I guess the end devices itself won't be because of hardware limitations, at least that's what I've heard from someone
No they're matter over zigbee compatible
Ohhh got that totally wrong then 😆
Nanoleaf is the only matter over thread bulb, with the exception of one Chinese mfg who hasn't sold their product in months
I see
The thing is - I understand that Marcel from HA has found a possible issue, but what is Google doing differently to Home Assistant? Because I almost never lose my nanoleaf bulbs in Google Home, it's only ever Home Assistant that they go unavailable in. If this was a general bug in matter or even in Nanoleaf firmware, it should be affecting all ecosystems right?
I mean, would we be able to see anything in the logs? I haven't seen the loggs yet, I'll spend a good³ few hours tomorrow connecting and testing them out
Do you think marcel has tried chat gpt for a solution 😆
I did some of my home assistant automations on chat gpt
that wont really work lmao
Hahaha, for a personal who doesn't know yaml, it helped me a lot lol
It is affecting all ecosystems, just scroll up and you can see plenty of people saying they have issues in other ecosystems (perhaps not this particular FW that was just released but generally)
Yeah just updated to .165 on one bulb and that bulb has now gone from Google Home and Home Assistant.
Actually the Thread Border Router in the Nest Hub remains on even if internet is down.
I think further up someone linked to the actual CHIP issue Marcel opened that describes the issue
Yup I have verified that the bulb still has entries in both ltpdu and in matter. It is still publishing an IPv6 address and the IPv6 address it publishes is pingable. So it's defintiely not a connectivity issue, it is definitely firmware
Yeah the short is that when HA tries to subscribe to the bulb it doesn't return the data and just quits, from my understanding
But the nest hub is no longer on your local network
I think what you're saying is the hubs default to broadcasting a SSID essentially in a pairing mode when the Internet connection goes out?
yeah but the thread border router stays up - thats a different chip. But yeah the connectivity is basically gone at that point - because nothing can contact it
Exactly, so it’s completely useless once the internet goes
See this is what is so confusing about the whole thing. I updated my second bulb to .165 not only is it still available in Google and Home Assistant it never went unavailable at all in either of them. It is still controllable from both too. It's so bloody inconsistent
I'll actually have to test that in a bit - I am not convinced it disconnects from the main WiFi, because if it did it would never know when the internet comes back. I suspect it just broadcasts a WiFi network to connect to. Because if you use any decent WiFi scanner you will see that it is in fact always broadcasting that network for you to connect to, but when it has an internet connection, the network it is broadcasting is "hidden"
yeah, but without any connection i doubt it even "reports" to the lan, rather just spreads its own ssid so you can connect
it's a pity the GHub Gen 2s can't leverage LAN like the Gen 1s could. Would much prefer not to have to deal with WiFi when/if it packs up
yeah, it sucks
Like I say I will test it - bear in mind how many YouTube reviewers testing matter stuff pull their internet connection to demonstrate it working without the internet - and that matter is supposed to work completely offline - then Google would need to have addressed that
last time when i lost internet, my whole thread network started to act up and the google nest hubs where advertising their own wifi network.
So i might be wrong, but to me it seemed like they just disconnected from my wifi.
bear in mind one of the issues could well be that the IPv6 address that the hubs get are provided by the router, so if you have IPv6 on your network properly working, then the Hubs (like mine) will have a publicly routable IPv6 address. If the internet goes down then the IPv6 address will be dropped. Leaving only the locally routable IPv6 addresses - which might take time for other devices to start using - they are probably just retrying the primary IPv6 address and getting timeouts
Nope, im not advertising any IPv6 address on my router.
My ISP is too lazy to setup ipv6
Right OK so here's the final results of my update attempt. 2 bulbs updated and are now unavailable. 1 Bulb updated and still working in both Google and Home Assistant. Essentials strip played the game again of getting right to 100% on the update but not actually updating. @boreal surge
Has something significant changed about how the mesh builds? Since the 152, initial impressions are that it has become very slow and unstable again
I’m aiming to get all of my matter devices to 165 so that it’s a fair test (mix of 152 and 159 too, currently)… hopefully it’s the mixed estate which is causing issues
update all to the new version if you can
If only it were that easy 🤣
only thing I'm noticing consistently is HA shooting itself in the foot when the updates complete from the NL app. There is ALWAYS one light offline for an extended period, never the same light..
2024-03-08 06:01:35 (MainThread) ERROR [matter_server.server.device_controller.node_67]
ATTENTION: Node 67 (Essentials A19/A60) did not complete setup in 570 minutes.
This is an indication of a (connectivity) issue with this device.
IP-address in use for this device: unknown
Try powercycling this device and/or relocate it closer to a Border Router or
WiFi Access Point. If this issue persists, please create an issue report on
the Matter channel of the Home Assistant Discord server or on Github:
https://github.com/home-assistant/core/issues/new?assignees=&labels=integration%3A%20matter&projects=&template=bug_report.yml
62 of 63 updated to .165
Let’s see what happens now
One bulb refuses to do anything so I suspect it’s facing a hard reset later today
interesting... didn't see this kind of logging in recent Matter server updates in HA. Must have been added new? as I'm seeing the same in my logs
curiously the log can't ID A19 IP addresses, but can ID GU10s :\
Yeah, I'm not sure. I've posted in HA matter, as the instructions say to do. Hopefully Marcel sees it.
yep just noticed as well 😁
Something broke either on 5.4.1 or the last 2 firmwares here because they were working decently well before that now they're completely unavailable in HA.
fwiw, the availability problems i've had in HA recently were associated with multi-admin. unfortunately, if ha's not talking to the bulb, other ways of removing fabrics from the bulb are a fair bit more complex :(
it is kind of annoying that the subscriptions failing in HA causes the bulb to be completely unavailable, even tho HA can talk to the bulb and potentially run other commands on it or administer it.
i'll DM. there are 2 possible reasons
Great thank you. Marcel and balloob helping me troubleshoot this HA matter
#1049765219565576234 message
So if you find anything that would be great too
@pulsar rock said they made a discovery about the nanoleaf product and said he will contact you about this
that was actually @opal sparrow that did this discovery, and looks like he's active here as well.
fwiw, i've already provided more information (full chip-tool logs of the subscription failing) to nanoleaf in a jira.
Thanks all there's so many helpful people here it's hard to keep track and that's a good thing
Well, I have stared at this issue long enough to have a stomach feeling that you have found the cause.
@boreal surge the scalability of 165 is poor, sadly. Under 152 I had all 63 Matter bulbs available alongside the rest of my thread devices. With 165 it’s back to around 23 unavailable, which was the case with the pre 152 builds
152 also healed quickly when it hiccuped
All Apple ecosystem
*make that 24 unavailable 🤦♂️
For what it's worth - without having to power cycle the bulbs, the google hubs or restart the matter server - the 2 unavailable bulbs reappeared in Home Assistant (and Google Home) after 1hr and 28 mins
Wish I could say the same. 17 out of 25 have been unavailable for me for the last 6 hours. Reboots of all the things have occurred and nothing has improved. 😑
are any of the "offline" ones pingable?
Nope, as they have fallen off the matter fabric so HA doesn't have an IP for them in the integration.
right but can you get their IP address from am mDNS browser (I use Service Browser on Android) and then ping them with a normal command line terminal
I could, it's just a matter of whether I can be bothered as I've seen this all before. Power of at the mains for an hour usually brings them back.
Having said that
step 1 get last part of default device name
All the offline ones have lost their records
Yeah have done it before, dw
I'll look into more when I'm a little less dejected.
My device unavailable count seems to keep ticking up…
We've observed cases where matter mDNS services disappear unexpectedly To Reproduce Information to reproduce the behavior, including: Git commit id e6df00d (https://github.com/SiliconLabs/matte...
I feel your pain. It’s made even worse by the sheer amount of effort and time wasted updating bulbs all over the house.
I think that NL needs to add some more bulbs to its test rig. This is an obvious regression which really should have been caught in Alpha
Thanks
there is an update I sent to nathan. 2nd attempt updated the light strip and it came back online immediately in Google and HA
@boreal surge what’s the chance of another build this week which reinstates the previously fixed scalability issue, please?
My one matter A19 bulb started randomly turning on after updating to 3.6.161. In other news, after updating, I could finally get it installed to HomeKit, but once connected, it always shows as offline in Home.
note that 165 is effectively 152 with only some smaller changes that should not generally affect matter connectivity. i've reached out directly to understand if your scenario otherwise matches anything we're otherwise tracking
Ok thanks. That’s very strange then as 152 was immediately very scalable and stable, unlike all previous builds. I put it down to Matter 1.2
The builds since then have immediately felt bad again
Credit to @boreal surge and his colleague for persuading me as I thought it was hopeless. Restarting all of my border routers seems to have caused the mesh to fully rebuild. Wow 🥳
thanks @wild oracle
just an update; in the last 5 hours, 9 have recovered on their own but 8 are still off the Matter fabric entirely. I have no idea why it seems to always be 8 that decided to just stop playing ball, but that is the case.
2024-03-08 20:17:53 (Dummy-2) CHIP_ERROR [chip.native.CTL] Failed to perform commissioning step 18
2024-03-08 20:17:53 (Dummy-2) CHIP_ERROR [chip.native.EM] Failed to send Solitary ack for MessageCounter:83150102 on exchange 11465i:src/messaging/ExchangeContext.cpp:103: CHIP Error 0x00000002: Connection aborted
2024-03-08 20:17:58 (MainThread) INFO [matter_server.server.device_controller] Starting Matter commissioning with code using Node ID 111 (attempt 3/3).
2024-03-08 20:17:59 (MainThread) ERROR [matter_server.server.client_handler] [140737221705936] Error while handling: commission_with_code: Commission with code failed for node 111
When trying to share from Apple Home to HA I get this message
Also this as well:
2024-03-08 22:17:07 (MainThread) INFO [matter_server.server.device_controller] Interviewing node: 115
2024-03-08 22:17:09 (MainThread) WARNING [matter_server.server.device_controller] Unable to interview Node 115: Failed to interview node 115
2024-03-08 22:17:14 (MainThread) ERROR [matter_server.server.client_handler] [140737221705936] Error while handling: commission_with_code: Node 115 does not exist.
Some of my bulbs will not reset no matter how many times I do the 3 sec on off 5x
Do they never emit any light? I have some like that which I must exchange. They work fine on the network. Weird
Are u using home assistant as the matter controller?
I have the HA Matter integration, yes. I use 4x GHub Gen 2s as the Matter/Thread Controllers. No OTBR or Skyconnect etc.
I see
I'm now down to 2 lights being stubborn and not rejoining the fabric, with a third rejoining for all of 4 minutes and now showing 'unavailable'.
problem is if I restart HA for any reason, ALL lights will drop off the fabric and I'll be back to square one. Seems to be a fairly common trend for me, not sure why.
That's interesting 🤔
One way to put it 🙃
for context, by 'dropping off the fabric', the HA Matter integration shows this:
I thought that the lights would be connected to your Google hubs at all times as it's a border router,
if it's on the fabric but become unavailable* again, it'll show this:
Ohh it only drops form your home assistant, but still connected to your border routers right?
based on an mDNS browser, the records are all there for the light (have had issues where the records do disappear, but not in this specific example).
The NL app registers the light is on thread, and can be communicated with, but that's using NLs own Matter stack with the bulbs is my understanding
so for all intents and purposes, it should "just work", but not that simple it seems 🤷♂️
Hmm yeah
Aaaa that's no good 😦
I'm hopeful that nanoleaf will fix these issues, I guess because I haven't been here as long as u guys have 😆
Do the bulbs automatically send data to nanoleaf when problems occure?
I'm hopeful too, but doing the same thing over and over again expecting a different result is the definition of insanity
they're smart, but not that smart
Hahaha, so it's up to us I guess to provide logs and stuff?
welcome to beta testing
Hahahaha
So if we look at strictly in the last 6 months, have nanoleaf fixed a lot of issues? I know there are tons but have they made progress?
progress has been made, for sure, and all in the right direction
but when some of the reported issues we report they can't replicate, it makes it more difficult to resolve
I see, but what if we give them remote access? I would be happy to do that
wouldn't help I wouldn't think
Right
it's not like a software application where remote access could be provided
best we can do is provide logs of when things go kaput in a manner that Nanoleaf can interpret
Like if we SSH them in or something
Is the logs from home assistant enough? I want to contribute but not sure how to
The app doesn't have any logs I think
Referring to the nanoleaf app
#1184371187464290344 message would probably be a good starting point
So I can just provide logs here with as much details as possible?
The bulbs in production don't have the capability to provide debug logging on the device itself so there's a few things that are done to track issues
-
you report things you see qualitatively (my lights are unavailable)
-
report HA logs
-
if you have programming experience, you can do more in depth testing.
That said, NL is tracking multiple issues that are probably causing your problems so there isn't much more you can provide at the moment. They have a bug tracker here, PW CHIP00001
https://research.nanoleaf.me/upcoming-quests/chip0001-3-xx-thread-beta-software/
ie. the same thing I posted above.
Oh sweet, thanks for sharing that. I'll surely submit things as I see them
🤣 so the front door light of mine finally recovered... after 37 hours:
no other changes to network, mains power, pressing buttons and hoping for best. Just decided to appear now. Think that's a new record 😕
Is anyone able to successfully share from apple home to HA?
I always get errors
And it never successfully happens
I was in that situation with firmware 3.6.159, too. It was not possible to pair to HA directly and also not possible to share from Apple Home to HA. Did you try to pair the bulbs to HA directly?
At the moment I am on a holiday trip. So, I can’t do any further tests for more than a week. I am enjoying the time without resetting/readding bulbs. 😃👍
Well I cannot reset the bulbs lol
I do the 5 off on for 3 seconds and nothing
They won’t flash red on the 5th time
Ok I’m getting a couple to connect this morning
Man this is such a crap shoot. Like last night nothing no matter how long and this morning some of them share fine
Makes no sense
Did you try it very fast? Onoffonoffonoffonoffonoff…
what about offonoffonoffonoffon? (sorry)
That always worked for me. I also recognized that it doesn’t work, if I wait 2 or 3 seconds between the on/off switches.
I've had the 'something went wrong' error from GH to HA or to HA directly in the current firmware, but I don't use iOS (short of Flame for mDNS) when trying to share*
I had some instabilities in my Thread mesh. There was one EVE device that was defective. The new HA logs with the node id included, helped Marcel and me to figure it out. Now where this device and all Nanoleafs are not powered on, I do not have any timeouts in my logs. At least they are very seldom now.
But my holiday trip started, before I was able to complete my tests.
Would the new downlights get the same beta updates as the bulbs?
I think so, I have received 3.6.165 few days ago, it finally made my downlights connect to HA... Well, except one who worked once and then it stopped working, I can't even factory reset it by disconnecting the power 5 times. Fortunately, I can control the light through the NL app, just can't get that light back on HA, nor I can factory reset it. Big improvement though from the stock firmware.
Right, I hope that resetting issue gets fixed quickly lol
Hope so too, that's the light over my desk, sucks that I have to control it manually when the other three works with my HA automation 😅
Is it just that 1 light or you can't reset any of them?
I have only tried to reset that one as the other three are working. But, I could reset that light previously with other firmware versions, seems to be since .165 that either it's more finicky or it just doesn't work.
I see, I'm keen on updating my ones later this morning, I was going to sign up for beta and do it today but couldn't find time
It's actually 35 mins away from my house lol, so it's a hassel
It's night and day, on stock firmware it just doesn't want to work with Matter (at least for me), but with .165 it unlocked it. Encourage you, it's so much less of a pain to use your HA automation than to get your phone out, wait for the NL app to sync with BT and then open the lights.
lol, yeah, not something I would want to do on a saturday if it was away from home 😄
Just to confirm, how far is the official release away from the .165?
Like for beta it's 3.165, is the latest version availble to everyone other then beta 3. Something too?
I don't remember what version it was, it's pretty far 3.4 or 3.5 or something like this. Right now, I have 3.6.165 in the beta.
If I look in the app does it tell you the current version it's on? I'ma check later
Yes, in More -> My devices
latest version on the production firmware is 3.5.41 - https://helpdesk.nanoleaf.me/en-US/nanoleaf-essentials-matter-release-notes-255125
September 18, 2023. What's Changed:. August 17, 2023. What's Changed:. May 26, 2023. What's Changed:. Apr 12, 2023. What's New:. What's Changed:.
so the beta is a fair way ahead
Wow, let's see what to next 6 months hold, nice to connect with you all!
Matter evolves very quickly, which is good, I hope to move most of my smart stuff to it. I was on the brink of just removing the downlight because of all the frustration it was creating when I found the beta program. Good that NL listen to feedback and fix their products.
Yeah I found the home assistant discord by chance, watching one of the update videos, where they mentioned to join them and ask any question u may have, after that I ended up here, im happy because I would have probably returned the products also
Guys just to confirm I just gotta fill details here to join? https://nanoleaf.me/en-AU/beta-program/
No that's for the hardware beta
Anyone with a Nanoleaf account can opt-in to our public beta firmware programs through our forum. This allows your Nanoleaf lights to get access to new features and fixes available in the beta firmware updates, and helps us to gain valuable insights
Sweet thank you, do I need to do this everytime I reset the devices and add it to the app/account?
No once they are enrolled you don't have to visit the website again unless you want to unenroll them, which is kinda pointless ATM since you can't roll back from beta to general release until a new general release
@boreal surge fyi, flickering with gu10 is much much worse with the be 165 firmware compared to 159.
Sometimes get like 6 or so flickers on power on now before it settles
Ok so I updated all lights today, seems to be a lot more stable on my openthread network, however I still can't connect any of them to my home assistant
Did that just now. 👍 (BTW Discord is terrible. A forum would be much better. Discord keeps losing track of where you were and has terrible search. Plus it even scrolls forward when a new message arrives. This makes participating in this beta much more painful than it should be.)
Nothing wrong with discord, just don’t use a thread, make a specific channel for this
A thread like this is a terrible idea for over 3000 messages
I'm not complaining about threads. I'm saying it's impossible to just catch up on what's been said in the past few days if Discord keeps auto-scrolling to the last message … like your message just did for example! 😄
What a user-hostile UX 🤪
Aha, so this is similar to what I reported on March 4 and which Paul asked me to file a bug report for, which I just did. Does it happen for all bulbs for you? For me it's 3 out of my 9 EU GU10 bulbs. They're still on 3.5.41 (i.e. non-beta).
Ah … lucky you. I tried All The Things, no luck. (Except for factory reset. Don't want to do that, so that Nanoleaf engineers can ask me to test things to narrow down the root cause.)
It will need a factory reset. Your reasoning of not doing that doesnt make any sense to me. All the things i am reporting are for one bulb that has the latest beta on it. This is the beta thread after all …
I know a factory reset MIGHT fix it. That's not the point. The point is that 3 bulbs are still on production firmware. There's no reason that a firmware update would repeatedly fail to be applied. If that happens, that's a critical bug in non-beta firmware/non-beta Nanoleaf app. Nanoleaf engineers need to know about that and be able to fix it.
Yes nearly 80% of my bulbs are flickering now. I have 38
Regarding questions about how far behind production firmware is. Production firmware is running matter v1.0 in the beta we have gone through v1.1 (which was frankly terrible) and are now on v1.2 (matter v1.3 - the spec, not just NL stuff - arrives in June the last I heard).
are you sure this is matter 1.3? i swore it was on 1.2?
yup this is matter v1.2 - I said matter v1.3 - the actual spec - arrives in June
oh right, my bad
It's my fault really, I probably shouldn't have mentioned matter v1.3 - I'm just excited because - hopefully energy monitoring lol
hahaha yeah, will be nice to see vendors finally have a universial way of doing it
and HA wont have to use custom clusters (looking at you eve 👀 )
...if EVE decide to support it of course.....
eh, i expect them to. atleast a few months after the spec releases
So I've gone through some trouble shooting stuff, mdns definitely works within my network, ipv 6 is enabled. When I looked at logs in the home assistant matter server here is what I got:
-
Discovery time out
-
Secure Pairing Failed
3.mDNS Resolution Failures
can you provide an example of #3 ?
So in the logs where it says 2024-03-10 19:23:53 (Dummy-2) CHIP_ERROR [chip.native.DIS] Timeout waiting for mDNS resolution.
I interpretate that as the server couldn't resolve device address using mDNS
[matter_server.server.device_controller] Starting Matter commissioning with code using Node ID 2 (attempt 3/3).
2024-03-10 19:23:52 (Dummy-2) CHIP_ERROR [chip.native.CTL] Discovery timed out
2024-03-10 19:23:52 (Dummy-2) CHIP_ERROR [chip.native.ZCL] Secure Pairing Failed
2024-03-10 19:23:52 (MainThread) ERROR [matter_server.server.client_handler] [549277322448] Error while handling: commission_with_code: Commission with code failed for node 2.
2024-03-10 19:23:53 (Dummy-2) CHIP_ERROR [chip.native.DIS] Timeout waiting for mDNS resolution.
OK do you have any lines like -
2024-03-10 06:15:35 (MainThread) INFO [matter_server.server.device_controller.mdns] Node 2 re-discovered on MDNS
No just the above in sequence
do you have an mDNS browser - Flame for iOS, Service Browser for Android
You mean the avahi-browse? I used to make sure that my otbr is using mDNS
whatever you are comfortable using - you need to see if the Nanoleaf bulb is publishing an mDNS record - both on the ltpdu and the matter records
Ok let me see, what is ltpdu again?
Sorry im still very new to this stuff, learning as I go
its NLs propietary protocol
but it should be visible in an mDNS discovery
something like this:
OK are you trying to pair directly to Home Assistant or do you have it paired to another ecosystem first which you are then sharing to Home Assistant?
No just paired it to nanoleaf first then tried pairing it to home assistant directly, I used the connect matter through the app as well, let me see if that has any different logs one sec
ok Nanoleaf isn't a matter ecosystem - that's their proprietary thing. When you connect from Nanoleaf to somewhere else you are connecting your FIRST matter ecosystem
I would skip pairing to Nanoleaf app and go direct to HA after a reset. Can be added to NL app later
If you have a Google or Apple system - pair to that first.
or that
While some people have luck pairing directly to Home Assistant - the HA devs still advise against that.
true, but there are more people than not that have issues with the multi-admin functionality
to each their own, though
yes - but the HA devs say that OTBR is fully working for Thread devices, but they don't recommend OTBR with matter - until it exits beta
Ok, so you recommend, resetting it, adding to Google home first, then to home assistant, then to nl?
Yes
yeah that
once you add it to GH, you won't need the matter code again - you can just share the existing pairing to HA and the HA Matter integration will do the rest
they advise pairing it to Google / Apple - so it's on their thread router - and if you want after sharing to Home Assistant from Google or Apple, you can then remove Google / Apple from the matter fabrics if you want, which will not stop the device using Apple / Google for the actual Thread connection
How do I remove it from Google, just delete device?
yep
no you go in to the device - linked apps and services and remove it from in there
eh, not how it's worked for me 🤷♂️
Okie dokie
But the bulbs wills till be managed by the otbr and the matter sever?
Nothing got to do with google after that point right?
you don't remove it from in here? @main matrix
Yeah, I just like to delete it from everything possible, then factory reset it
Or unlink it from other vendors through Apple home
I haven't, no. I delete the bulb from GH entirely (when I had the bulbs connected through GH)
I didn't know if that would remove the thread connection too? So I just remove from inside the fabrics
removing just the 6006 vendor caused GH to complain that it wasn't paired to a Google ecosystem and refused to remove
which is why I deleted the bulb
YMMV
in which case just delete the bulb from Google after you have shared it to Home Assistant @wraith bear
idk, it's also quite possible my experiences have been outliers and not the general consensus
Got it!
Do you guys do this every time there is a new beta update?
😅
no just the once
Oh good lol
unless you have to factory reset the bulb
only when a bulb decides to go AWOL after an update 🙃
though I have never once had to factory reset anything other than the light strip up to now
must be nice
So do we know what the issue is at this point? Is the matter server on home assistant or NL? Like the commissioning part
Or both?
both
I think both right?
that's a loaded question
Haahhaha
and possibly also OTBR too right? Going on the opened issue that was posted the other day?
Ohh, forgot to ask, is there a otbr discord group as well, and beta program?
if there is, I'm not in it
Lol
I checked the last few updates on otbr, no bug fixes or anything, they just add stuff to it which isn't promising
I could be wrong
there's many moving parts, and I liken it to links in a chain. When something gets updated (NL FW, HA Matter integration, GH Fuchsia FW and whatever the Apple equivalent is), one of those links breaks and can throw the whole chain out
Jeez, are all of these developers in contact with each other to actively re establish these broken chains?
Stupid question, different companies probably not
Matter is a minor miracle. It's already come so far in such a short time - especially considering how many developers from different organisations are involved. It's pretty much unheard of for competing companies to get together and push out an interoperability spec in just 2 years.
The matter github has developers from all those different companies communicating with each other
Ooooo! https://github.com/project-chip/connectedhomeip/commit/c554f44a681e09e45724b604e27b6ecb69859590 that should help a lot.
These are all the matter server errors?
that specific error is one I got after the last beta update - some lights Home Assistant just could not speak to because there were numerous errors that the device was busy
yeah I also get them as well, usually after a HA restart
it took almost 2 hours to recover
🤣 mine are taking over 36 hours
So it basically keeps re retrying?
Well it does anyway - it's just that there is a backoff delay that constantly increases each time there is an error - so you very quickly get to the won't retry for like 6 hours territory. This change would give an indication of when to retry - instead of relying on the backoff timer
2024-03-10 10:07:13 (MainThread) INFO [matter_server.server.device_controller.node_4] Node could not be discovered on the network, returning cached IP's
2024-03-10 10:07:16 (MainThread) INFO [matter_server.server.device_controller.node_8] Node could not be discovered on the network, returning cached IP's
2024-03-10 10:37:19 (MainThread) INFO [matter_server.server.device_controller.node_4] Node could not be discovered on the network, returning cached IP's
2024-03-10 10:37:22 (MainThread) INFO [matter_server.server.device_controller.node_8] Node could not be discovered on the network, returning cached IP's
2024-03-10 11:07:25 (MainThread) INFO [matter_server.server.device_controller.node_4] Node could not be discovered on the network, returning cached IP's
2024-03-10 11:07:28 (MainThread) INFO [matter_server.server.device_controller.node_8] Node could not be discovered on the network, returning cached IP's
2024-03-10 11:37:31 (MainThread) INFO [matter_server.server.device_controller.node_4] Node could not be discovered on the network, returning cached IP's
2024-03-10 11:37:34 (MainThread) INFO [matter_server.server.device_controller.node_8] Node could not be discovered on the network, returning cached IP's
2024-03-10 12:07:37 (MainThread) INFO [matter_server.server.device_controller.node_4] Node could not be discovered on the network, returning cached IP's
2024-03-10 12:07:40 (MainThread) INFO [matter_server.server.device_controller.node_8] Node could not be discovered on the network, returning cached IP's
Now HA need to fix this
4 and 8 are nodes that were factory reset over Christmas when that very very bad firmware dropped
The current issue is understood thanks to someone in the HA community.
There's two principle issues
-
sometimes the bulbs straight up stop broadcasting that they're available
-
sometimes the bulbs straight up don't respond to the subscription request from HA but don't error out and it gets hung.
I'd say most current stability issues are understood but the solution is still TBD. Issue is mostly on NL not HA
Oh damn, I hope nanoleaf knows them if that's the case
They're aware
Does home assistant advice against using an otbr?, If there are issues with otbr, I feel like there are a lack of updates for it, I don't see much happening in there, I could be wrong
Well the absolute most stable is using only commercial BRs, but a lot of people, myself included, use one or more OTBRs. It's just generally a troubleshooting step to disable OTBR just in case. The current problems aren't related to OTBR.
Matter server sees a lot more updates (recently) because there's been a lot of work to triage and/or work around limitations/issues with NL bulbs.
I see
Has anyone tried to use andriod emulator to update nanoleaf firmwares, im thinking of trying that, only because I live away from the facility
Can't see I've seen anyone try that before 😅
FWIW very few (if any) matter devices support updates via matter, most require update via Bluetooth and proprietary app
So I planned on just running the nanoleaf app on the emulator and using the onboard Bluetooth, im going to try that later 🙃
Itl just be convenient, until I have to hard reset the devices to resolve issues lol, then I'd have to go in
More times than I want to be... Its always a case of doing the update and hoping everything comes back online. I typically need to reconnect/reset about 4 out of 24
You'll definitely have to hard reset sometimes until the FW is a bit more stable
Hahah yeah
You could always put a relay in front. One person here does that
Less then ideal lol
Wonder if I can add a remote switch to the circuit bord to hard reset lol
sometimes its been enough to remove from GH and HA and connect it from nanoleaf app again, but some times i have had to reset them
Since 161 and 165 all my devices unavailable in home assistant. However, interestingly, I can ping the unavailable devices from home assistant successfully
Does that make any sense?
The log said that there could be a connectivity issue with the device but I don't know if that's true
165 has been a step back from 159. Delays with turn on, more unavailable than there was (also on newer matter server so could be part to blame)
fwiw, the nanoleaf bulbs do have the matter ota firmware update clusters present, but no idea if they work - NL isn't providing firmware to use with them :)
that's probably a "down the road once other issues are solved" sort of thing.
Sorry what's ota?
"over the air" - just refers do doing firmware update over the same network interface you're communicating to the device with.
So without needing the nanoleaf software and Bluetooth?
Ohh that would be amazing!!!
I cant imagine trying to update over 200 devices....
Since you can only do 3 at a time with Bluetooth, that would take 66 attempts...
That's optimistic assuming the first attempt works. For me, Nanoleaf devices always take multiple attempts to update (but it eventually works).
Yeah that's an issue for sure...I've been lucky, I can update in the first try, but honestly I haven't had the chance to test the system as long as you guys have
I still can't figure out why nanoleaf wants to join a pre determined thread network instead of a custom one you would make...
Nanoleaf doesn't have anything to do with which thread network is joined, that's part of the commission flow for matter
I had to change the thread network details to exactly what's on the nanoleaf app for it join
That's before I joined beta though
Change the details where?
If you have a look at the screenshot, the 259 tower hill drive was the thread network I made on my otbr, and the bulbs wanted to join NEST-PAN-E88F
So I changed my networks name, channel, pan ids and other things to exactly match NEST-PAN-E88F, then it suddenly joined
This was the network it was actively trying to join
I see you're on a Samsung. Android always tries to commission to it's own network if one exists (in your case nest pan). I actually saw someone else have an issue with this recently but it's not like NL or HA can do anything (until HA does it's own commissioning but that's down the line)
Both the NL app and HA use built in services to commission / add to network
Ohh so this is an issue only on Samsung's?
No you'll see that behavior for all android phones with thread devices.
Your use case is pretty atypical from what was imagined I think hence you have some headaches. Most people are using deviceswhere they live
I started the build at my home to begin with, I almost gave up on fixing it, until I saw a reddit post, the guy just joined this network instead of using his own, which was pretty confusing because nest-pan-e88f didn't exist, so I figured just to copy the details from the above screenshot, and it suddenly joined
I think for people who have tbrs, it will just let you make whatever network like nest pan, but for people who use otbrs it's different
If this issue is an andriod one, I suppose the same behaviour will apply to other products like aqara, evee?
Yeah this particular issue isn't device specific
I see 👀
You can make a new network using OTBR but getting android to use it requires some workarounds (I'm not familiar with these first hand)
The easiest way would to be have the sky connect just join that network
I mean to say that itl make it automatically with whatever details the andriod/nanoleaf likes, but with otbr that's not the case
The issue here isn't the settings or something with the thread network, it's that, if Google has created a thread network, then android will try to always use those regardless of it not having any border routers
I was initially thinking to go to the skyconnect route, though a couple reasons why I didn't
- It sounds like skynet 😅
- I was worried about the range
- Also wasn't sure how much processing power it would use on my home assistant, I tend to use it for long term and for a big facility, many things I want to do with it
Ah I think I see and you just extracted the thread credentials using the NL app and set your dongle to use those
Sounds like it worked out okay
Yes! Sorry im very bad at explaining I see
The short is then that fix should work for all your thread devices and that problem wasn't NL specific
It's just how android works under the hood for thread devices for now
Yeah that's good to know, I thought it was nanoleaf before speaking to you on here. Though if I use an evee door sensor and andriod decides it wants it to join a different network then I'm screwed 😆😅
I only have nanoleaf products as of now
I think it should work the same
for what it's worth - in future -
go in to the bulb settings page when connected to it over bluetooth
scroll down to "Thread Network" and tap it - it will list thread networks available and you can join the bulb to a different thread network
no need to faff about changing all your thread details @wraith bear
I never knew about that
Huh right, neither
Oh it doesn't let you join the different network
Trust me I tried like a thousand times
Before crying, then I changed the credentials, and cried again 😂
huh maybe the app has changed - you could still do it last week. @boreal surge - comments?
If you could do it, that means NL probably fixed it and it wasn't an andriod issue as me and @brazen canyon thought
I'll test it today hopefully and share with you all 🙂
No that's not true android will always prefer/try to commission to the Google thread network, if present. I think what @bitter compass mentioned was that you could change it after it joined a thread network. In your case, the Google network was empty so this feature wouldn't help as the bulb would just fail to join the empty network
See here
#1065584078906216500 message
Sorry im confused, is ajm saying to change the thread details after nanoleaf joins?
I assume it is a feature to cope with situations where for example you might have factory reset the Google Hub which would obviously change the name of the Thread network. This feature lets you change the network a bulb is connected to, without having to factory reset them all
He couldn't add it to a network at all because Android was trying to commission with thread credentials to a network with no BRs
Oh damn, @brazen canyon I just read that link you sent, so apparently I have 11 google audio speakers, so as longnas I have them on my network, andriod will always prefer to join Nest pan networks, even if it doesn't exist
The new "update all" button worked really well for me, went through 16 without an issue
That's shady as hell 🤣
theres an update all button?!
Update all button? I didn't see that yesterday when I was updating for the first time in beta
on the my devices page I have an update all button for available firmware updates
it just queued them all up and processed through
Yea on the play store you can opt into beta
it will usually take a little while till the play store offers the update to switch it to the beta. You can try killing play store and app, and opening and checking for updates
I don't think Google speakers work as a BR
I usually give it 15 minutes and then go to the app info page on Android and then tap the app version at the bottom of the screen to take you to play store and the open button will have changed to an update button
they don't Only Nest WiFi Pro and Hub Gen 2
@brazen canyon they don't, but I guess andriod will still try to join one unless delete other google stuff? Like Google wallet, the app itself that's what I understood from the link you sent me
I'm not really sure when Google generated a network -- maybe you're right and it generates one if you have ANY Google nest device if it's a BR or not... Sounds like something Google would do lol
Yeah, let's see how things go
On positive news, i reset a bulb and could add it direct to HA without doing NL or google home first. Strangely it completed in HA, but android kept saying connecting to home assistant even though logs showed completed and it was showing up. Then it went to something went wrong. but it did complete
yeah I've been getting something similar (something went wrong in android, but HA logs think its fine). Classic task failed successfully 🤷♂️
I've seen the same thing on non-NL devices lately as well
So 2 in a row, ends with the something went wrong, 3rd one completed successfully. All show the same logs in matter server and all added.
Im just happy they are adding first go now 😄
Well after giving it a couple days there was 4 out of 24 that were annoying with their connection, so reset them and added again and currently everything is up. so see how it goes for the next few days
will be interested to see if you start to get what I am experiencing over time; lights in HA "look" like they're available, but manually toggling doesn't do anything as it times out. Yet the same lights on automations (motion sensors in my example) will trigger without issue.
Will let you know, seem to get a mix of lights nicely turning on and other just doing an instant change :S, nice if they all faded on haha (im asking too much, ill start with everything just being stable and consistent connections)
Jinxd it, now 1 of my hallway bulbs just dropped off for no reason
yeeeeeep. I also have one that has been offline close to a day now. Just tested the other lights on my wallpanel, and 8 out of 24 will turn on and off, the others just go "nope" but don't error out.
problem is if I restart the HA integration or HA itself, they'll all fall off and take upwards of 12 hours to rejoin
but the same "non-working" lights work when you move past the motion sensors that are tied to them
yea im scared to restart HA at all 😛
When?!
I think they got bigger issues to fix at the moment rather than new features 😅
I wouldn't be surprised if that syncing support will be coming as part of the work that's being done on the "Nala" hub. (Which is supposed to be enabled on NL devices with thread border router capabilities in a future firmware update).
I agree. I think Discord (or similar) has a place as a part of running an overall program, but not as the only tool. This beta's been running longer than most that are currently active, we'll be evolving the format and tools used over time to improve.
No changes to the Nanoleaf app; you can specify which available Thread network the device should communicate on in the settings of the device (tap the network name, if more than one network is visible there will be a small arrow beside the network name).
There already is a forum: https://forum.nanoleaf.me/forum/ but it doesn't have a section that seems appropriate for public beta stuff. It also seems to have some moderation issues - there's quite a lot of spam present (and no way i can find to report it)
whoops, thought i'd run into a new and exciting matter failure, but turns out that I just accidentally turned off the switch that controls an outlet one of my lamps was plugged in to ;)
had me confused since the MDNS entries were still there - but i guess the thing about those is that they're cached by the thread border router for a while.
#1184371187464290344 quick update for you, especially those that are facing home assistant unreachable issues. we have at least 2 known causes:
- Matter services disappearing
- Subscription timeouts
For 1, we tested a fix over the weekend with select beta testers who had high degrees of reproducibility. We'll look to release that to the rest of beta ASAP.
For 2, we dont have definitive reproducibility with Home Assistant where we can routinely develop a fix for, but we've identified what we expect is equivalent steps with chip-tool (thanks @opal sparrow ). Hoping to zero in on a resolution on that ASAP and subsequently release.
Thanks to all for your patience and input.
yeah, they time out after 30 minutes (which is the default TTL)
@pulsar rock @opal sparrow this is a feature of SRP, which allows a (thread) device to register an entry with a server with a timed lease. The server then is responsible for multicast portions of mDNS. The default lease time in matter SDK is 2 hours. For HomeKit over Thread, we previously used 60 minutes. there are IETF efforts to take SRP beyond thread. there are camps that wish they had never build the "m" part of DNS - i'm not an expert on all the places it really kills networks (perhaps our head office dev network has excessive amounts of devices between thread and wifi), but I can definitely appreciate there's a ton of overhead when every wifi device is multicasting its identity willy nilly (ttl for wifi is order of 120s typically)
Indeed. getting mdns working well in larger office networks usually requires separating things into vlans and putting mdns repeaters between vlans to reduce the amount of multicast/broadcast traffic. took a bit for me to figure that out when setting up my office lan so people on wifi could use mdns to find the wired printers :)
getting thread border router stuff working with that sort of network would be a pain tho, need to convince your inter-vlan router to accept RAs sent by the TBR and add routes for those. (or just manually configure the routes).
Anyways, good to hear you seem to be able to reproduce the subscription issue with chip-tool. I just provided some logs of when I noticed the issue, and helped a dev out with a hint on how to do multi-admin matter with chip-tool :)
Today - I am reconsidering my decision to have smart lights.......
Underground power cable failure overnight - multiple attempts by the electric company to reconnect power. Nanoleaf bulbs very unhappy.... Was going to factory reset them, but Nanoleaf app claims they are still connected to Thread - and after a few hours of everything being offline, everywhere - a few popped up online in Google.
I have had ok results after something being out, turning everything off (inc border routers) start border router, then turn on groups of the lights and wait for them to all connect, then turn the next group on.
Yeah, I think it's the rapid on and off cycle when they were trying to reconnect the supply that's caused the issue. I'm guessing not factory reset because they still know how to connect to thread - but it's definitely upset them
I can still do this, but only for my HomeKit-only Essentials E27 bulbs, not the Matter-only GU10 bulbs 🤔
I don’t think the bulbs can just change networks freely within the matter spec, there would be a reason behind it not working
What's odd is that my E27 HomeKit-only bulbs get that disclosure triangle, and upon tapping it I see that only one Thread network is available, and they're already on it. But for all the GU10 bulbs, It doesn't show the disclosure triangle and just shows the actual Thread network name right away. Seems like a small inconsistency? 🤔
Would I have to do that if I scaled upto 200 nano devices?
A ton of border routers, for stability
More routes that go through a bulb, more chance of it crashing and nuking most of your bulbs
So far I have 4 rcp with 4 raspberry pi 3's I initially planned on running open thread on all and make it join the existing network
Is that a good number?
Hard to tell, only way is really to just go for it and find out, but at 200 mesh routers within the network, placement of the border routers matters
Yeah, I'd distribute it as best as possible
I think i will wait a couple more months before I scale this up 😅
🙈
…just thinking about the updating process for so many devices makes me dizzy.
You can only do max 3 at a time, it’s kinda annoying
I’m in the TestFlight beta for the app, and the last beta I couldn’t
Max 3 at once
And the rest would be unable to update/queue
Yeah probably that
I don't know if this is a HA bug or a nanoleaf issue but I can't control my A19 bulbs via HA anymore. The lights show as connected in HA and the nanoleaf app but anytime I try to change the state in HA nothing happens.
I see this in my HA logs.
2024-03-12 06:58:54 (MainThread) ERROR [matter_server.server.device_controller.node_2]
ATTENTION: Node 2 (Essentials A19/A60) did not complete setup in 450 minutes.
This is an indication of a (connectivity) issue with this device.
IP-address in use for this device: unknown
Try powercycling this device and/or relocate it closer to a Border Router or
WiFi Access Point. If this issue persists, please create an issue report on
the Matter channel of the Home Assistant Discord server or on Github:
https://github.com/home-assistant/core/issues/new?assignees=&labels=integration%3A%20matter&projects=&template=bug_report.yml
This bulb is the closest to my border router so I'm not sure what else to do. I have tried power cycling a few times with no luck.
You are doing something wrong
Is the bulb already in another ecosystem?
Or you pairing it fresh to HA?
It's in both Google home and home assistant. I paired to Google first. This was working fine before
Yes, so share the pairing code from google home to HA, don’t rescan the code
To add it to HA
Yeah I did all that a long time ago. I was able to control the lights in home assistant without problem. This issue is new
Hmm, you on the beta firmware?
I wouldn’t remove it from HA, just update the matter server add-on and enable the beta toggle (then restart the add-on) and see if it fixes it
They all on the same thread network? The skyconnect and google nest displays?
Yes. I only have the one thread network
Oh I also have the nanoleaf lines that act as a border router as well. Those are on the same thread network. I did just update those last night
I haven't restarted anything besides the light yet in case there are logs I should be getting
I can control some of my other lights. It seems the issue is only with specific bulbs
Yeah I'm doing that now. I've been doing that the past couple days and the issue keeps coming back.
I'm guessing the only logs I can get for the light is from the home assistant matter logs?
Yes
After restarting the matter server I can control that problematic light now
Right EVERYTHING was factory reset and everything is talking now.
does that persist after a restart?
I reinstalled the app it works
But then I turned off my lights at night and it happened to all disconnected bulbs
Hey guys, does aqara p2 door sensor support over the thread software updates like evee?
Oh shoot, your right wrong forum
3.6.170 Release Notes
Thanks again for your patience #1184371187464290344
- Expected solution for disappearing Matter services, which should enhance reliability when using multiple fabrics
- Removed custom cluster temporarily, which should resolve subscription timeouts and errors. We'll look to add this back in before going to prod, just need a bit more time to find a true solution. This won't affect any features you currently use.
We're hoping this build lifts any major reliability with Matter connectivity that many of you have reported. But if you continue to see things, please reach out to Paul
GU10 flickering update: we've sent affected bulbs to our China team for in depth hardware investigation. If anyone knows for sure whether they've seen this before coming to 3.6 beta, that would be much appreciated (DM or mention Paul)
oooh, a wild Update All button appears. 😁
any chance of a hint about what that custom cluster was for anyways? whenever i looked, it always had one empty attribute, and one attribute containing 256 bytes of 0s, which doesn't explain a whole lot :)
it provides us with the ability to perform magic that the spec or slow moving giants otherwise get in the way of 😄
Thank you this firmware update immediately brought back all my devices to home assistant.
better setup UX, triggering certain modes that don't fit in a (boring) box, etc.
🎉
yep, did to mine as well. Was at 16 of 25 unavailable, now just 2, but those 2 are still processing the beta updates.
Have to say, best firmware update yet
Agreed. All lights are reachable and communicable in HA, and for funsies I spammed my wallpanel with a bunch of on/off commands, and they all responded!
Recovery in HA after unavailability is markedly improved too (spamming on/off for 25 lights ~20 times triggered 3 timeouts on 3 lights, but they recovered within 60 seconds).
opposed to 37 hours 🙃
That is exciting!! Do we still have to add to Google before home asisstant?
Nope
Hasn't skipped a beat for me so far. Half tempted to swap from the ghubs to full otbr and retest. Only have the hubs because they were the "supported" ones from a HA dev perspective.
i would slowly do it, but honestly seems like it has a strong chance of working
Agreed.
Though I'd try and avoid mixing the two environments, so not sure slowly is an option.
Tomorrow job though.
this new firmware update allowed me to commission all my devices into HASS without any issue - thank you so much!!
Don’t forget to pin message
Just updated and everything instantly came back to HA, that part usually takes a bit so that's awesome 😎👍🏻
well, no problems with the update for me, guess I'll try the multi admin stuff again just to confirm :)
Did the update last night and my 6 bulbs seem solid still with Apple Home - no regressions so far.
Multi admin stuff?
I wont get to try it out till tomorrow
multi admin refers to when a matter device is simultaneously joined to and can be controlled by multiple separate matter fabrics, e.g. both google and apple - or in my case, home assistant and chip-tool (a debugging tool).
Ohh, Theoretically would multi admin put load on the light bulbs hardware itself?
yes, the light bulb has to manage separate secure connections for each fabric, and some of the attributes need to be stored separately for each fabric too. As well, the bulb needs to separately send updates for subscribed state changes or events to the coordinator for each fabric.
Ps: lately almost none of my bulbs appear connected in the app, even though i can control them through GH and ST. The 3 that are showing as connecting wont update either
I have a lot of rooms. Does the ’update all‘ work over Bluetooth or over Thread?
In my understanding Nanoleaf Essentials get their updates through Bluetooth. So, I think that some bulbs aren’t reachable for Bluetooth and I have to go into the room with the respective bulbs. Is this correct?
It is a nice addition, if you have a lot of bulbs in one room and you do not have the 3 bulbs barrier.
I've managed to be a good 20-30 feet away from a bulb and have it update no problem even through walls It's just slower for further away ones and sometimes they have to try a few times if they time out or what not. I say go for it and see what happens.
Thanks, I only wanted to understand how this works. So it’s still Bluetooth.
I am still on a holiday trip. At the moment I do not have any Nanoleafs installed, because I had a lot of issues lately and not a lot of time available. Next step when I am back at home is to install some bulbs and look how they behave. 😉
@boreal surge Im not sure why but since the last update I have not been able to reset my bulbs. I turn them off/on 3 sec 5x and nothing. I have 1 bulb that sais it needs to be reset to add to apple home and I just cannot get that to happen, hence im stuck.
Did you try it on/off 5 times without 3 seconds pause? Try it without pause, if not already done? (Maybe we already discussed that some days ago… 😃)
Why is it so red?
Idk I just made it red
It's really hard to tell what's happening. But I think the second video shows really fast on/off switches, which always worked for me.
if your light is set to red, it might be hard to notice it blinking red after the reset sequence tho :)
I think it was stuck in the power state mode for some reason? I had to turn that on and off.
What is the power state mode and what did you turn on and off?
Sorry "Power Loss Recovery"
You can see my bulb remained red every time I turned it back on. I had the option turned off in the settings of the nanolead app. However, the bulb was not abiding to that setting. Turning the setting on and off again I saw the bulb had switched its natural warm white when turned on again
Are you removing power from the bulb and re-applying it? Ie, turning it off and on at the wall?
Ah ok. Yes, that could be the case. I haven't activated this because I want to see when the lights switch on by themselves.
Yeah
But why…
So, you found a new bug. 🍻
Possibly lol
Is there any other way to factory reset a bulb?
Oh, is that what you’re trying to do 🫠
What would be a good spot to ask about two homes showing as Currently on Cloud? I only have one showing in Apple Home and the second one in the NL shows no devices
Contact NL support to delete your cloud and build it again, that was the only way I synced my current home
Yeah sounds full on, thanks for explaining 🙂
ok, so looks like OTBR is still a cause of upset for Matter devices. All lights were solid for the last ~24 hours after the FW update, but added a single OTBR device to the mix and timeouts started within 2 minutes.
I do want to do away with the GHubs at some point, but guess it won't be anytime soon.
i'm honestly kinda surprised at that, openthread was released by google so I assumed they were using at least some of the same code in their own devices.
(fwiw, seems fine in the single border router and it's running otbr case)
Yeah, I am also, but that said the OTBR in my case is a GL-iNet GL-S200 (https://www.gl-inet.com/products/gl-s200/). I do have a Skyconnect with OTBR flashed somewhere, but have misplaced the USB. The second Skyconnect I have has Zigbee only on it, and that's working fine so not touching it.
I'd have thought the device shouldn't have mattered given OTBR should be an open standard, but I could be wrong 🤷♂️
hmm, you've been making sure to keep the firmware on that updated, right? if it's running an outdated otbr that could be a source of issues :/
also, they're running a "proprietary firmware based on OpenWrt" which sounds like a GPL violation to me :/
seems like it's a point of contention, but to me they are running a fork of OpenWRT with their own closed source spec on top: https://forum.gl-inet.com/t/gl-inet-false-advertising-not-really-using-openwrt/39143/28
yeah, google's allergic to gpl, so openthread (and anything based on it) doesn't have to be open sourced when used in a product like this. but at a minimum they need to be providing the kernel they're using and a few other parts of openwrt. It's a really old openwrt too :(
it actually would be neat to see a thread border router package available for running on stock openwrt tho, i wonder if anyone's been working on that. A lot of wifi routers have usb ports which could be used to connect thread radios, and we'll probably start seeing them builtin soon.
running the thread border router on your main router makes dealing with setting up the ipv6 routing quite a bit easier :)
yeah, for sure it would.
in any case, removing the OTBR from the mix and leaving it GHub only has brought the lights back into the fold without the need for reboots, so I'm happy the recovery side of things for lights that fall away is much better.
anyways, since you can't see anything about the OTBR version which is on that device, you might as well call it a "closed" thread border router ;)
need to complain to them about any network instability, not to the nanoleaf or openthread folks.
wasn't a complaint, just a general observation
and there was an OTBR version on the S200, but it's unplugged now
It's gpl2 so they can get away with a lot of stuff like saying it's open source but only if you register for our platform or something
A lot of that stuff was remedied in gpl3
Where is the link to add devices to the beta?
I got a new lightbulb, but I want to add it under the beta
I got the same one! I started if with that, then I discovered the andriod nest pan issue, thought it was the gl net box, so I shifted to otbr on raspberry pi
nest-pan issue? do you mean the GL-iNet not joining the existing HA Thread network, and making its own?
No, so initially I was trying to get the Nl to join Gl-iNets thread network, but I kept encountering the Nest-Pan issue. I thought it was the device itself so I decided to do it on a raspberry pi with open thread
Encountered the same issue, long story short its an andriod problem as mentioned previously 😅
I see. The issue I had initially was getting the S200 to join the existing Thread network that HA generated, as there was no indication of what the PSK was. Turns out from the OTBR doco, the HA Thread network uses the generic PSK as noted in the documentation...
I've got 2 of the GL-iNets...not sure what to do with it now
Ohhh
Oh dang, I was thinking of getting the (new) S20
Hahahaha
I've got 2 S20s and an S200. I'll keep experimenting with them, just not in an prod environment per se
had to email their Sales team to allow me to buy the new ones though :\ as at the time they weren't publicly available (at least not through their store or Aliexpress etc)
that was ~2 months ago though. things could have changed since
I see they still only ship from China
Will wait until they ship from the EU, who knows what surprises customs will add on to the price 🫠
Latest beta looking good FYI, devices in HA have been online 100% of the time since installing, more than a day already.
💪🏼💪🏼💪🏼
FWIW, it was ~$75AU (~€45EU) including shipping, which wasn't too bad considering
Im in AU 😄
dude same
i doubt we will ever see it, but it was said to be a 4-6 weeks wait time
- 2 or more weeks coz we all know how fast aus post are 🤣
also same 🙃
Do we pay for the products? I have no idea how it works lol
Yeah, it's not perfect—OTA overThread is on our roadmap. Right now, OTA is over Bluetooth for these devices. The Update All UI just creates a queue managed by the app that updates 3 at a time so you folks don't have to tap-tap for what feels like an hour 🙂
Afraid that the registration is closed for this one, sorry! But there are quite a few more for the same and other products happening over the next few months. If your device and home profile is a match, you'll see an invitation by email 🙂
hmm, all my bulbs appear to have gone unresponsive over thread about an hour ago, that's weird. no time to debug it right now, i'll have to take a look later.
might be my border router rather than the bulbs
Will start in 3-4 weeks
Cheers mate 👍
I haven't been able to do much/any digging into it but this .170 firmware started strong (was able to upgrade most of my lights without needing to retry) and the bulbs were seeming solid in HA. Then they started to drop off. I went through and repaired a lot of them (by removing bulbs from HA and NL, and resetting through powering on and off repeatedly). That seemed to restore connectivity for a while but now I'm at a point where zero of 15 bulbs are available in HA.
170 so far is doing well since it was installerd here. Have you tried rebooting the HA matter server and your TBRs? I did this just after updating to 170 to get a fresh start. I keep telling myself this is the last time to do this...so now, I hope this is the last time 🙂
I just updated from the production version to .170 yesterday. All lights were available the whole day. This morning a lot dropped of. Had to restart the tbr and ha matter server, since then all came back again
Mine have been running for 2 days straight now with zero issues in HA and Apple Home, 100% availability.
alright, pretty sure this was my border router. the ethernet cable i was using had a broken clip and appeared to have come loose. I've replaced the cable :)
I upgraded 2 bulbs and 3 light strips to 170. All upgraded successfully but Thread network is not working.
Currently in the NL app 3 bulbs are only visible over Bluetooth.Only 2 are showing on Thread. I can control all 5 via the NL app but not via Apple Home; all show as "no response". The upgrade seems to have affected the broader Thread network as my Onvis switch is not available in HomeKit either (shows as "updating" then "no response". Also, Eve Home app's Thread Network view normally shows lots of end-points which I think represent the NL devices; they appeared briefly then disappeared again.
I've tried power-cycling the bulbs and both border routers (HomePod mini, Apple TV); one or two NL devices appeared briefly in Apple Home then disappeared. I won't try any more troubleshooting tonight but I guess next step would be resetting the bulbs and re-adding.
Hat-tip to @remote goblet above - restarting my HA matter server did the trick, all bulbs now available over Thread in Apple Home. And the Onvis switch started working again.
On 3.6.170, commissioning fully reset devices to home assistant has never been better! It works all the time every time.
Hey everyone, did anybody lost all the devices from Apple home ? Now they are only available in NL app only :/ I am not talking about 1 or 2 devices … my 35 devices are out of Apple home
Do you think resetting NL devices and pairing them again would help ?
sure it's not just Apple Home having a moment? seems odd to have them all go walkabout
I don't know enough about the Apple ecosystem (only have an iPhone specifically for Flame), but if you try and add a new accessory and go to more options, do any lights appear there? Seems Homekit devices apparently disappeared in a similar way (from the quick Google search on it) and the devices were just "hidden", but bringing one back brought them all back.
failing that, not really sure. sorry.
sounds like a @boreal surge / @wild oracle query then.
I have mentioned to Nathan previously.. just checking if someone else had it
do you see any of the lights in Flame?
It started with couple@of devices
makes me think the services have disappeared
They went to default room
And now everything
Now it’s too late
Already removed them all and going to reset before the night … to ensure I’ll sleep .. I have the switches everywhere
Mess
How I can see that
Oh yeah , already having the fusion.
Flame is the mDNS browser to check if lights have lost their .matter or .ltpdu services
for iOS*
install the app, tie your phone to the same WiFi network as your Apple TBRs and see if anything Nanoleaf appears
example:
if you drill down into one of the lights, do you see the same number of records as above?
NL app uses the ltpdu protocol, Apple Home uses the .matter
seems normal, though that IPv6 address looks a bit odd
I don't know enough about your network topology though so that could be fine, but from my testing most of the IPv6 addresses used for thread communication usually start with f xxx, as it's a local address. 2a02 suggests a public IPv6 address, but some routers do assign public IPv6 as private, so again could be normal.
Before yes
your call to go nuclear, but I think other things could be tried first 🤷♂️
personally, I'd wait to see whether it's something others have seen/Paul or Nathan's input, but that's me
dang it; after 3 days of solid connectivity, one light has fallen off my HA Matter integration 😦 curiously though, that light is reachable by BLE only and there are zero entries for it in Flame...
Is anyone using the ESP Thread BR-Zigbee GW as their border router? I was thinking of picking one up to extend my thread network to my garage
The s3-h2? I've got one and works fine.
Yeah. I saw they were only $10 and thought it might be too good to be true
If I run into issues setting it up later do you mind if I bother you?
Just get a project box to stick it in, after you config it runs fine.
Only issue I had was I couldn't load the thread creds during setup, had to do it after using the web-ui
Yeah sure it's been a few weeks since I set mine up so might be a bit rusty on details though
Sounds good! It should get here Monday so we'll see when I have time. I might have to 3d print a box of for it too. I've been trying to learn 3d modeling
The rest didn’t help
I was still loosing the bulbs from Apple home@
I deleted the Nanoleaf app
So far so good
🙈
... deleting the Nanoleaf app added the bulbs back into Apple Home?
You can buy one off Amazon or similar
No I reset the bulbs to factory settings and continued adding them only in Apple home@
Hopefully will not loose them anymore
Certainly seems like we are getting close to a stable world!
Going to be away for the next 4 days, so nothing should be happening in the house, hopefully all 24 bulbs will keep connected 🙂
@boreal surge Still getting some flickering (seems to be the same bulbs). 1 flickered then the light went off, the other flickered and ended up on the correct colour and brightness still, so this is an improvement
It works
Sometimes they disconnect if you connected it by the NL app, just reconnect it
Ah you fixed it, nice!
After the update to 3.6.159 it was not possible for me to pair any Nanoleaf Essentials bulb to HA and I also had problems with my Thread network one week ago. I removed all currently installed Nanoleaf bulbs from power, but the Thread network with 32 EVE devices was still unstable. With the new HA logs, where the node id is also logged, we found out that one of my EVE devices had a defect. After removing it from current, my network was stable since one week. Today I updated HA, Matter server (5.9.0 BETA 0), tvOS, homepodOS, iOS, ipadOS and macOS to the latest and greatest. I rebooted my 7 Apple Thread Border Routers and the HA Yellow. After that I waited until everything was connected again.
Then I did the firmware update to 3.6.170, I paired one of my (unpaired) bulbs to HA directly and shared it with Apple Home Home. It worked at the first attempt. Great!
Now I updated one of my bulbs, that was already paired to HA and Apple Home, to 3.6.170. Update procedure was smooth and without issues. But when it was ready, the bulb was removed from Apple Home. It was still controllable from HA. I opened the Nanoleaf app, went to the bulb settings, disconnected it from Apple Home and connected it again. Now I see and can control the bulb in HA, Apple Home and in the Nanoleaf app. Maybe it’s the same issue that @shell vigil had.
If it stays stable I am going to add more bulbs.
In a previous firmware, that happened to 3x of my devices. The _matter service for AppleHome was no longer on the device. Still in HomeAssistant though. I fixed it via the NL app, and just selecting the “Connect to Matter” button for those 3 lights.
As of this morning though, all my lights disappeared out of the NL app on my iPhone, in that they are still seen but they need to be “setup” again, and can therefore no longer see the thread details etc. But I think that’s more of an issue with the app that lost all the information, rather than an issue with the lights or this firmware. Also, I checked the NL app on my iPad, and everything is shown correctly there still. So definitely just an issue with the app.
They are still connected via AppleHome and HomeAssistant without any problems.
I have deleted the app, the night passed and didn’t loose any of the bulbs in Apple home
And works great
It's funny how 1 bad device can nuke the whole topology..
An update on 5.6.170:
I updated 51 lights yesterday. Many were updated with the use of the new "update all" functionality, although I had to use it 4 times in various places of the house (Bluetooth range) and, then I was left with 12 lights which I had to update manually (restart, turn off the circuit breakers of other lights, ...). It's definitely an improvement, but I wonder - with understanding of Bluetooth discovery limits - there should be a possibility to connect to a specific device if Bluetooth address is known. Can the app save it and reuse, so that discovery is not needed?
Regarding Thread connectivity, it's better than the previous beta, but after 10 hours since the update I have only 31/51 lights connected. Unfortunately, they seem unresponsive - I can't turn them on or off even with multiple retries. I was able to control them right after the update though.
How many BR's do you have, seems like you have way too much traffic going through a few and its constantly crashing those
There's a pattern to this though - it's mostly bulbs distant to the border router which are offline. Seems like meshing might not work too well. I'm also not able to display the Thread network in the Nanoleaf app, to check which devices were elected as leaders/routers.
yeah you cant, only can if you have the OTBR web server (nanoleaf doesnt expose it, HA does if you have a thread dongle on the network).
its hard to debug without much info honestly
1, Nest Hub gen 2. There was a previous beta which has worked well for me, I had all of them connected and functional.
I was able to see this information on all previous versions. I thought it might be related to the custom cluster which was removed?
huh? i mean this
yeah, i didnt mean that
I mean this page used to display the entire network with these details.
yeah, that doesnt show the mesh and and all the details around it tho
which is why i suggested adding that ability to their border routers (shapes/BR) for beta users but hasnt seem to gone anywhere
It does display device role though
most of the bulbs would just show mesh extender? no?
Maybe 🙃 I used to have some as "child" which is basically end device
huh, yeah right
Presumably if you had enough BRs strategically placed you'd minimize mesh extender roles
I have 7 bulbs connected. 6 of them are a child.
Do I understand it right, that you have only one TBR for 51 bulbs? Theoretically it shouldn’t be a problem. But in the past I had 35 EVE devices connected to one hardwired AppleTV 4K 3rd Gen. Everything worked, but the most distant devices were slow. So I bought a second AppleTV for the other floor of my house. Immediately everything got faster. After that I decided to exchange some of my Sonos speakers (without TBR) to HomePod Minis (with TBR). That was a great decision.
Yes, I have only one at the moment. In a few months I'm gonna have more, just waiting for a release of border routers I'm planning to use long term.
Ok, maybe you want to buy a used one as a workaround. But yeah, you said, that everything was fine with an older firmware, if I remember right. Too many posts… 🤔😃
Yes, there was a build where I had 100% connected at all times, no disconnects. They were sometimes slow to respond, and sometimes I had to retry commands 2-3 times. Not ideal, but not too bad either 🙂
Do you remember which build it was? Maybe this helps the devs to find the issue.
I think it was 3.6.94. I posted about it Jan 19th, from what I can see this was the latest version back then.
Ok, that’s long ago. Matter 1.2 was implemented with 3.6.152…
I read somewhere that the Theoretically limit on thread devices with one RCP is 32, and I'm assuming that will be lower on the Br's. I think for a large network like that, your probably better off useing otbr. It's really inexpensive compared to a Br device. I can send you some Linux commands to install and get it up and running if you need
Just need a good Nordic based dongle, and you’ll be set
while that may be the case, OTBR is as unstable as all heck at the moment, so it's currently a trade off until either Matter or OTBR matures further.
@main matrix your probably right, I haven't had the chance to test mine, I had to move my system to the other side of the building, and I did something stupid, upgraded some kernals on the Linux machine that was running my otbr, had to reset everything, hopefully I'll get it up and running today though
Typically Linux kernels are not removed, when you install a new kernel. Did you check your boot manager, e.g. grub?
I lost 0 bulbs over the weekend and everything worked smoothly 🙏 💪
This beta release is just fantastic. All of my 32 nanoleaf bulbs stayed connected 99% of the time. It's huge improvement comparing to the previous releases. Well done team!
Agreed that this is the best firmware yet. ALL 9 of my lights are controllable in home assistant and Google home. That has never happened before. And they've been connected all weekend long.
I only have my Google nest hub Gen 2 and 2x nanoleaf lines as border routers
@wraith bear Sending you a PM about installing otbr
You got a thread dongle at all?
I picked up an Espressif Thread Border Router. Should be here today
Ah dope, should be good!
They're finally in stock on Amazon so I figured why not
Yeah I did, I couldn't revert back to the old one which was annoying
Hmm. so there is definitely some sort of issue with service registration on OTBR. I've seen a few times that the bulbs are correctly sending srp requests for all of the services, but that OTBR is reporting an error and fails to register a service. Seems to happen most often when quickly pairing several devices over matter (pairing a single NL device via android to HA requires creating as many as 4 registered services, then deleting all but 2 of them). I'm thinking it might be a resource exhaustion issue in OTBR especially on small or memory limited devices.
I (76620409) OPENTHREAD: update host 1A35D748C486FA6D
I (76620409) OPENTHREAD: Set addr of host to FDA9:D862:9078:1:AF32:18BA:BA46:26AD
I (76620409) OPENTHREAD: Service 23BB6362BF69F1B5-0000000000000012._matter._tcp
I (76620409) OPENTHREAD: Add new service
E (76620409) OPENTHREAD: Failed to add service
W(76620409) OPENTHREAD:[W] SrpServer-----: Send fail response: 5
tho i haven't ruled out a malformed request being a possible cause either, i should see if i can use one of my spare esp32-h2 boards as a thread traffic sniffer :)
I’ve upgraded my two A19 bulbs I have online and the light strip. Since the upgrade the flood of CHIP ERRORS in my Home Assistant Matter logs have disappeared. This looks promising.
I’m relatively low key right now, waiting on some Inovelli White switches to ship. Hardware is two AppleTvs and one HomePod mini as boarder routers, and two Echo Dots waiting for the next release that allows easier thread network merging. Right now they are just dots. 🙂
Resource issue on the r p itself or the machine that's running openthread?
3.6.173 RC Release Notes
A few minor fixes you likely haven't run into. Almost otherwise the same as 3.6.170.
#1184371187464290344 This is our release candidate, so if you haven't already upgraded to 3.6.170, we'd love to get your feedback here. And for those that have, we'd love to get any verification you can spare.
Thanks for all your input so far!
Has anyone shared a device yet from AppleHome to HomeAssistant on 3.6.173?
I’m setting up a new device, but the light doesn’t show up under “Nearby Devices” after copying the code and trying to add it in HA.
Also, the power loss recovery doesn’t seem to work for the same light.
you can manually add the code
Yes, but it needs to show up as a nearby device, then you select it, and you paste the code.
No.
Click on that, then click on enter code
Nowadays I usually do it the other way around. First pair with HA and then share with Apple Home. Worked great with 3.6.170. Always right on the first try. I didn’t try it with 3.6.173.
After a good experience with 3.6.170, I updated all my 16 bulbs from 3.6.170 to 3.6.173 this morning. The update process was smooth. All bulbs disconnected briefly after the update, but became available again after a few seconds to a minute. They were then connected in HA, Apple Home and in the Nanoleaf app and can be controlled. It's too early to say anything about stability. But so far everything looks good.
I will check my HA logs later today.
This has not worked for me whenever I tried it. Granted, it was on the older firmwares. Anyway, it's not working now either, just spins round and round looking for the device
Ah, the other way around worked without issues: HomeAssistant --> Apple Home
Yeah, I also had this issue with 3.6.165. but it works great since 3.6.170.
#1184371187464290344 message
Under known issues the following is documented:
3.6.165 Release Notes
Known Issues:
3. HA subscription error can block pairing or connectivity: https://github.com/project-chip/connectedhomeip/issues/32493
This GitHub issue is still open.
The only issue I had updating was that some of the devices dropped off of my Apple home matter connection, but were fine in HASS
Three of my devices completely disappeared from HomeKit somehow after I updated last night.
Reconnect them from the Nanoleaf app
Delete the NL app
I lost all my devices once
Or eventually you can install an older build of NL app from test flight
I can’t reconnect them in the Nanoleaf app, it says they are already in another Home.
Had to reset the bulbs to get them to reconnect
same
Will DM you to collect more details, thanks for noting it. Same @eager frigate and @brazen heart.
That is from a Espressif OTBR or Home Assistant add-on?
that's the expressif OTBR. I should try the openthread posix border router at some point to see if it has the same behaviour
Yeah would be interesting. Also there have been some mDNS/SRP related fixes/improvements recent weeks in ot-br-posix at least.
From the logs it seems that the platform mDNS API refused to register the service.
Not sure what Espressif is using there. The HA add-on is using the Apple mDNS implementation, aka mDNSResponder
@boreal surge almost everything went smoothly in the upgrade, the only hiccup was that the nanoleaf app claimed that the essentials light strip was only on bluetooth despite all the A19 bulbs immediately reconnecting to thread. I thought I was going to have problems, but checked Home Assistant which was still able to speak matter to the light strip despite the app claiming it was bluetooth only. So functionality everything was immediately back on matter and controllable - the app was just slow to recognise that - but only for the light strip. Which incidentally is a bug I have seen before. All lights were still available in Google after the update too.
the RC is stable enough that I feel comfortable spending £36 in the Amazon spring deal to get another nanoleaf essentials light strip - put it that way 😉
Thanks! I might ping you a bit later on DM for some details about the lighstrip / BLE issue. This is something I've noticed occasionally on devices in my own test environment, but nailing down a pattern with reproducible steps has been elusive.
@boreal surge Yeah I am also happy with that firmware. My 32 EVE Matter over Thread and 15 of my 16 paired Nanoleaf bulbs are rock stable. No unavailable messages for the 15 bulbs in HA for round about 30 hours. But there is one bulb that became unavailable 4 times in HA. Don’t know what’s wrong. The bulb always gets available by itself within seconds.
I can have a look at my HA logs, if that helps.
The device is surrounded by 4 EVE Energies (1-5m away), a hardwired Apple TV 4K 3rd Gen (4m away) and 7 Nanoleaf bulbs (2-4m away). All these devices are in the same room, my living room. WiFi Channel 11 is not used in my setup anymore to avoid interference to Thread channel 25, used by Apple Thread networks. My ZigBee network is on channel 20.
Before this latest firmware update all bulbs were disconnected from power. I wanted to sort out issues in my Thread network. I already wrote about that above. However, I reconnected and updated one bulb after the other, always with the HA log in the view.
In the first round I updated 7 of my bulbs and in the second round 9 bulbs. In the first round I also had some devices that were disconnected from Apple Home after the update, some showed that they were connected via Bluetooth in the Nanoleaf app, while they worked fine in HA and AH. The Bluetooth issue sorted itself. For some bulbs it was enough to connect them to Apple Home via the Nanoleaf app. Some bulbs had to be resetted and readded to Home Assistant > Apple Home > Nanoleaf app.
The second round was absolutely smooth. No need to reset or reconnect to Apple Home. All bulbs came back to HA immediately.
So Multi-admin also seems to work now. At least it got better.
Mine also disappeared from Apple home (12 out of 14 bulbs)
@boreal surge Here is the HA Matter server log for the Nanoleaf Essentials A19-E27 bulb (Node 213) that has 4 unavailable logs in the HA logbook. Its not my complete log file, only the parts, where HA logged that it became unavailable:
As already said it always became available in HA 'immediately'.
An update from my end - the situation is relatively stable with 34/51 lights connected.
Thread network menu in the Nanoleaf app works again for me. With that, it's pretty clear why some lights/rooms remain disconnected - lights which should be mesh extenders are children. At the same time, I have a room with 12 lights where 8 of them are extenders, one leader, three children.
Is there an option (planned?) to force a bulb to be an extender?
A map would be useful. The Eve app shows you which devices are connected to each other (if you have some Eve plugs in the mix), but of course the names are all weird so it doesn’t say much in the end.
#1184371187464290344 we've released 3.6.173 to production. Thanks to you all for your patient testing, as this has definitely been rockier than we would have liked.
We'll keep you posted via email about upcoming tests we've got bubbling. And of course, we love your feedback and help to always keep improving.
Well done guys !!
Yeah!!! Even though the testing phase was too short in my opinion, I understand the decision to launch it now. There are a lot of people out there with unreliable Nanoleaf bulbs. This is definitely the best firmware for the Matter Essentials we ever had. But what do we do with our free time now? 😂 Well done, thanks
Congrats for shipping the new firmware. I think a lot of folks will be quite happy with the results.
How can I get rid of these ghosts in the Nanoleaf app? This is my problematic light bulb. Maybe it's a coincidence. I only see this bulb once in Apple Home and HA.
Probably cache
What do you mean by cache? Where can I delete this cache?
When I disable the cloud sync, the 2 devices can be deleted. But as soon as I enable the cloud sync again, the devices come back.
I used the beta Nanoleaf app and went to the stable app. It’s the same situation. The devices come back after deleting them. So, it’s definitely a cloud sync issue.
@boreal surge What can I do? Thanks
I meant your apps cache, there's an option to delete cache.
I am an iOS user. At least I don’t see any possibility to delete a cache.
I think such things are not possible with Apple. 😂
Correct; in Apple land, everything just works and nothing ever breaks. /s
However… I think, it has something to do with the cloud sync.
If I delete them local, they get synced from the cloud to local immediately.
Hehe more flexibility more errors here in Android land. Lol
Oh I know. My comment was 100% pure sarcasm (hence the /s). I'm on Android too, but at least we have basic features like cache clearing 🙃
Ha! Went into settings of the 3 dot menu for these 2 bulbs and there I pressed the delete button 3 times. After that I went back to the device overview and now the devices are deleted. 🤔😂
i thought IAR studio was trash, then i started using xcode 😅
Will DM you
after 4 days, 1 out of my 24 GU10's has dropped off and gone unavailable from home assistant.
Best result so far
ah, figured out what was going wrong with service registration on my esp thread br. it turns out that when you build the firmware, it gets configured with a static maximum number of services that can be registered… which defaults to only 10. Which isn't very many :)
10 is lots for one device, but the thread border router needs to hold services for all of the thread devices; and with nanoleaf matter devices, you could hit as many as 5 services per device (which would occur during the process of commissioning a 3rd matter fabric)
So after 5 days every single one of my nanoleaf gu10's have dropped offline. Restarting ha to see if they come back or if i have up flick the power switch on then
some things I do (and do less offten now-awadys) is reboot not only the TBRs but also your router and/or access points. If that does not solve the issue, then roboot the devices, if that does no work then I report it.
Turns out real estate came in for an inspection and turned off the switches haha
I typically restarted the Matter server, when I had issues. If that doesn’t solve the issue, I rebooted all my TBRs. That typically solves my issues. In 1 or maybe 2 very crazy situations I rebooted my router. But typically I don’t do this. Nowadays where HA Matter server log file is very communicative and shows the node ids, it’s a must to have a look at the log file. My problem is that I do not always understand the messages or error codes in the logfile and some log lines do not include a node id. 😃
Has the reliability poll been shut down?
I have 5/24 GU10's on .173 that are unavailable for over 6 hours now 😦
Seems like they lost nanoleaf app and HA. adding htem again fixed the issue :S hopefuly something just weird and one off
My lights are all connected.
But tonight my wife noticed that one of my bulbs was glowing very dimly. She pressed the smart light switch (Hue Dimmer Switch), but it stayed on. This appears to be less than 1%. In the Nanoleaf app, in Apple Home and in Home Assistant, this bulb is in the off status.
Here is a photo:
And here is another photo with 3 seconds exposure time:
looks like I'm in a similar boat; 2 of my GU10s have also gone non-responsive for the last 14 hours 😦
I have 4 in total so I've only lost 2, FWIW
Though it must be hard replicating issues
the HA matter logs are... weird though..
the HA matter docker container has been up for the last week, but the logs only go back for the last 40 minutes?
Can you control them through nanoleaf app?
yeah they can be controlled through the NL app
in my case they are 4 lights fixed to the front of the house outside that turn on and off* with a HA automation at sunrise and sunset, so it's not a huge issue per se
but this is all I have in the logs, which is basically nothing: https://www.dpaste.org/OH4Hw
the services are still there (both ltpdu and matter), NL can see the lights in the* app, and it's only caused 2 of my 25 lights to go all weird. not really a lot to go on :\
well, the beta firmware here is now the production firmware, so unless they have had a bunch of reports through their normal CS channels, I wouldn't bet on them working on it.
But your not useing home asisstant os right?
that's all I'm using'
not tied to Google Home anymore
restarting HA didn't fix it, so I'll restart the addon and give it a few minutes
other than what is in the dpaste above, no
Just the ones you provided
none of the A19s were affected, so... idk
I wonder if nanolead has its own logs that we can access?
based on previous conversations in this thread, it's unlikely
ok, have gone from losing 2x GU10s to 3x GU10s
all are reporting node busy
sigh
As you might know, we think it’s so prevalent to HA users because we use google/apple home API to commission devices, which technically turns them into multi-admin and hence these issues
it only turns them into multi-admin if you commission them to one ecosystem then share the same device
So if we can commission these devices directly to home assistant a lot of the issues could go away?
Yeah I think you mentioned you have them directly in HA
No, because it uses Apple/google for the commissioning (phone’s BLE and their API)
Unless you do it VIA websockets, which isn’t recommended
Websockets?
So there is another way to commission without involving andriod or iphone?
I had a myriad of issues going GH to HA (even with this FW), as GH could see them but HA couldn't and vica versa.
technically yes, but undocumented and unsupported
yep. and it will be undocumented for a long while, thre is a reason its not meant to be used by anyone 😉
So currently if we only use HA directly to commision we are still using iPhone or andriod right?
How long is a while lol
how long is a piece of string
you cant use HA directly, you are doing it via the google or apple API
I'm interested to try this websocket thingo out
until all the issues with bluetooth can be resolved
Are we pretty close currently?
you running a bluetooth adapter on your HA instance?
Or another year or so
I've done the websocket method before, but it was hit and miss and quite a while ago
again, how long is a piece of string
I did buy a Bluetooth adapter for this
Well not for this, I wanted to run andriod emulator on the otbr so I can commission it remotly and manage the devices like updating and stuff through the NL app
Still haven't been able to run the emulator though which is annoying, but I have an adapter I can use
@plain kettle are you a developer?
as in what way? dev for HA?
Just any dev
just a uni student 😛
Cool
@wraith bear https://community.home-assistant.io/t/power-plugs-cant-connect-to-thread-network-skyconnect-thread-matter/638332 - if you're still keen on the websocket of things
Hello, I am struggling to get the SkyConnect stick running with Thread + Matter power plugs (both ONVIS and EVE) and I would be very thankful if someone could help me figuring out what the issue is. I tested this with two systems (an AMD mini pc and a Pi 3) and each time I setup a fresh install of HA and have the same result. The following in...
but don't expect any support from the HA matter devs 🙂
you need your BLE dongle to be connected, but not set up
even with the bluetooth intergration
Hah, wait skyconnect isn't supported 😮
correct
But they advertise it
so it's unsupported on unsupported on undocumented
not "officially"
they say thread in a future update*
people like to poke around though
Ohhhh
so you can "try" it, and if it fails, well 🤷♂️
So good thing I'm running an otbr then
@plain kettle so your saying running skyconnect will solve a lot of issues (in the future when its supported)
but every network is different
its upstream issues mostly
😄
FWIW, fixed the light issue by turning off the lights at the switch for ~5 minutes and back on. all 4 of them came back into HA as responsive.
That's weird, 5 mins. The normal cycle is like 10 seconds isn't it?
You can use the following command on the command line to get the last 10000 lines of your Matter server log:
ha host logs --identifier addon_core_matter_server -n 10000
You can modify how many lines you want to see. One of the HA devs (Stefan) posted this command some time ago.
fair point, should have remembered that. The restart (and subsequent flip on/off the lights at the switch) fixed it, but I'll remember that for future use.
normally yeah. I've only ever seen HA Matter integration fully lock up when it can't get its bearings (transport write close or similar), but I assume the running logs have a finite amount of time before you need to define an amount of logs to retrieve older ones
I don’t think the logs are wiped that often?
well there was no activity prior to what I pasted earlier, so... not sure.
and I've been running the integration in debug mode
Did you look in the CLI using the line above?
so would have expected a LOT more traffic
bit late now; the previous logs are gone
they're non-persistent on a restart of the integration
Hmm right
yeah, just run it - nothing prior to the restart (first entry is the client_handler connecting to my HA instance, which is consistent with a restart)
Huh right
That’s not right. I can go back for weeks in my log. Did you try the command on the command line?
only place I can try it
Are you able to ssh into your HAOS and type this command? I am unsure, sorry.
I tried it in both
it's all good; might be something to do with my log retention, I don't know
Ok, that’s strange.
Move the discussion to the HA Matter discord channel. Maybe someone else can help you.
🙈
Didn't see that in the 8b messages 😣
Your trying to access matter logs on ha right?
if that was directed towards me, yes I am.
lol all good.
seems the same GU10s that fell out of sync yesterday, continue to fall out of sync every 24 hours or so. workaround is a physical off/on at the switch (after 2 mins of being off). Annoying, but oh well.
What is the problem with the command line command above? What exactly does not work?
Install the community addon ‘Advanced SSH & Web Terminal‘.
https://community.home-assistant.io/t/home-assistant-community-add-on-ssh-web-terminal/33820
Read the warning in the info tab. 😉
- Configure your username/password and authorized_keys (it also works without the keys) and set the
compatibilty_mode: falsein the configuration tab. - Start the addon
- SSH into your HAOS
ssh username@password - type the following command
ha host logs --identifier addon_core_matter_server -n 10000
You can also modify the value 10000 to set the lines your Matter server log should show.
That’s it. 😃
This add-on is provided by the Home Assistant Community Add-ons project. About This add-on allows you to log in to your Home Assistant instance using SSH or a Web Terminal, giving you to access your folders and also includes a command-line tool to do things like restart, update, and check your instance. This is an enhanced versi...
I don't recall saying anything was broken with it? In any case, I rotate my logs after a day due to space constraints (found that in the config file). If I have a need to expand that, I will.
Yep I'm having issues on .173 too
they consistently keep dropping - a few for days and still not been able to reconntect them after restting
Whilst .173 is great improvement I don’t think it fixes all woes?
Is anyone else getting delays with HA and turning hte lights on? My automations are regularly taking about 5 seconds to turn them on with this firmware.
Motion detected > Turn light group on, 3-5 seconds, bulbs turning on.
Which sucks when you walk into a dark kitchen and then have to wait for them to turn on. (Wondering if its more a nanoleaf issue or HA issue, and given all the dramas we have had to get to this point, im figuring NL :P)
Same automation with zigbee bulbs and the automation is completing in <1 sec
I had that with older firmware on the NL lights, but not recently.
Ah that, we it hasnt been shipped out to anyone outside of the US yet
You know exactly what I mean 🤣
Yep, most of us here outside CA/US are still awaiting it unfortunately
Also have a delay issue with my HA setup only with the Nanoleaf bulbs. And still have some blinking with the GU10. So I hope there will still be some updates for that...
for those running a GL-S200 (from GL-iNet), looks like they just released an update to improve Thread performance and update their SDKs to more recent ones, as well as update the topology map for Thread devices. It actually looks accurate now for the Nanoleaf lights on my network (vs what the OTBR addon seems to think is "accurate" in HA):
not sure, but I'm all for it
in terms of their SDK updates:
fairly sure the previous version was using ~9 month old SDKs
I assume they are just using the API with a custom wrapper, might be worth buying one
How much AUD was it from them? They don’t seem to have it on Amazon yet (so will have to contact them)
~$100AU off the Aliexpress store last I knew
na, the contact one was for the GL-S20 (this one is the S200)
hmm, no longer there... I'll find my receipt, 2 secs
ah right, I bought it directly off the Gl-iNet store
Didn’t SL20 not get an update as well?
https://store.gl-inet.com/products/gl-s200-multi-protocol-iot-gateway-thread-border-router?_pos=1&_sid=0b305c36d&_ss=r - $69AU (nice - looks like they're on special at the moment) + postage via DHL.
it didn't as it's built on a different architecture
@plain kettle https://dl.gl-inet.com/release/iot/stable/s200/4.1.5-0400 - release notes