-
-
Notifications
You must be signed in to change notification settings - Fork 1.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Honeywell smoke detector (Xiaomi JTYJ-GD-01LM/BW) selftest fails with "No network route" #4285
Comments
Have the exact same issue after upgrading to 1.14.4. |
@guardiande would you be able to sniff the traffic when you execute the command? https://www.zigbee2mqtt.io/how_tos/how_to_sniff_zigbee_traffic.html @Sandriekus are you able to post the herdsman debug logging of this error on 1.14.4? To enable herdsman debug logging, see https://www.zigbee2mqtt.io/information/debug.html#zigbee-herdsman-debug-logging |
@Koenkk can I help? I'm using the cc2538 stick |
@Enzokot yes please sniff the traffic like suggested in: #4285 (comment) |
@Koenkk herdsman debug log? |
@Enzokot are you also able to provide a sniff? (https://www.zigbee2mqtt.io/how_tos/how_to_sniff_zigbee_traffic.html) |
Yep, but later. Need to find one more stick |
@Koenkk sniff synchro with herdsman? at the same time? |
Preferably both. |
Done |
sorry for the late response. current setup: when performing a self-test to the HoneyWell Smoke detector:
i can also upgrade the adapter firmware to the latest dev for testing if you want. |
@Koenkk |
ok, so for testing i've updated the firmware on the slaesh's CC2652RB to the latest dev version: current setup: herdsman debug logging when performing a self-test to the HoneyWell Smoke detector: |
I've did a lot of investigation in the smoke sensor lately and found out that many routers cannot handle this device (including our zstack coordinator). This is because the poll rate is very slow, causing commands to fail very often. I expect pairing this device to a Xiaomi router (e.g. xiaomi plug) should solve this. The smoke sensor does not switch parents once it's paired so it should also keep working. You can check the parent via the networkmap: https://www.zigbee2mqtt.io/information/mqtt_topics_and_message_structure.html#zigbee2mqttbridgenetworkmap |
@Koenkk Was there a major change in versions between 1.14.3 and 1.14.4? |
0x00158d00039a8358 - works great |
@Sandriekus yes there was some change. Can you try this fw: |
can you also check if it works when connected directly to the coordinator? Because when connected through the router the fw update shoulnt be necessary. Try repairing it close to the coordinator. |
Sure! Going to try tomorrow morning and let you know |
@Koenkk herdsman debug log: So i saw in my zigbee map addon that there was one other smoke detector without routes (0x00158d00052b7f72) herdsman debug log: |
@Sandriekus have you tried to force remove it first before pairing? https://www.zigbee2mqtt.io/information/mqtt_topics_and_message_structure.html#zigbee2mqttbridgeconfigremove |
@Koenkk i have force removed the device, removed the battery for couple of minutes and then re-paired:
see full log here: also setting the sensitivity doesn't work:
|
Hmm, seems it still doesn't work indeed. I would recommend pairing the smoke sensor to a router instead. |
an IKEA LED1842G3 bulb is oke for router? |
@Koenkk that is correct.
it is running stable now for 2 days (also after restarts of HA or complete host) i can upgrade Zigbee2mqtt to 1.14.4.2 if you want me testing that version with the znp_CC2652RB_LAUNCHXL_tirtos_ccs.hex.zip firmware? |
@Sandriekus thanks, published the firmware: https://github.com/Koenkk/Z-Stack-firmware/tree/develop/coordinator/Z-Stack_3.x.0/bin |
Hi guys, I already created the duplicate of this issue which has been closed #4502
Note: I didn't update zigbee2mqtt, I'm still on the latest official release 1.14.4, should I switch on dev? But the issue is still present even after those steps. Firmware is new for sure and it is still not possible to send messages to smoke detector and it still appears detached in the network map. Although reports correctly. One thing I noticed after re-flashing of the firmware of ZZH. All devices seemed to be working but all were displayed as detached in the network map. I needed to press the join button on each of them to be displayed correctly again. Even though no routers were involved. After re-joining they are displayed correctly on the map except for Sonoff humidity temperature and Xiaomi Smoke.. UPDATE: Additionally to my findings above, since I have the new firmware I'm experiencing unavailability of random sensors. I'm not sure if it is related, but I'm switching back to 20200805 :( |
One additional remark from my side. Do I understand correctly that "ping" is implemented internally as "node descriptor request".. If so, then it is really strange how it should work for sleeping devices. If I understand the specification correctly it advises to cache node descriptors on coordinators and routers but it consumes memory.. so probably there is some limitation. It would mean, that may be in practice only a limited number of descriptors is cached and it may result in unsuccessful "pings" for sleeping devices and it cannot be solved only by changing the timeouts.. Does anybody know how many node descriptors are cached by particular firmwares? If we cannot change this, then the whole network map and pinging do not make much sense for larger networks. But the impossibility to send the message e.g. to smoke sensor is a different question and might be related to timeouts if I understand correctly. |
* Z-Stack 3.x.0 20200831 firmware * 20200925 firmware. Koenkk/zigbee2mqtt#4285
@pedroke to check whether the issue is present, use 20200925 in combination with Zigbee2MQTT 1.15.0. What do you exactly mean with ping? If you mean availability you can read here how it works for EndDevices: https://www.zigbee2mqtt.io/information/availability.html#non-pingable-devices |
@Koenkk Thanks for the hint, I will try with the newer version. By "ping" I mean the mechanism of how a network map is built. What I can't understand is that when I trigger the network map build, all my battery-powered devices, which are sleeping for sure are displayed as connected. So it is probably only decided based on the received messages in the last 24 hours - as you mentioned. But e.g. smoke detector which is also SED is displayed as detached, although messages come regularly.. so I suppose that for that specific device different approach is taken. And what is even more interesting, that even Sonoff humidity and temperature which is almost the same as Aqara temperature sensor is also sometimes displayed as detached, even though it reports periodically... |
@pedroke the way this works is that not the end device itself, but the parent of the end device (a router) report what childs it has, this is what you see on the network map. |
Let's take an example. I reflashed firmware on ZZH..All devices were reporting correctly and had strong signal. But on the map all devices which are reporting to coordinator were detached. So coordinator was accepting messages from them, it was aware of them and still was not able to name them for networm map... strange. As soon as any of the device forcibly re-joined, it was again showing attachment in the map. Even stranger is that devices behind router were displayed correctly. |
Ok..now I see.I checked the code of network map so the connections are based on routing table of the device. So in case there are detached devices there is no route defined. That also explains why smoke detector returns 'no route' error while it is detached in the map. And it also explains that flashing clears the map from RAM. But why are some routes cleared during runtime remains mistery for me. ED like xiaomi battery sensors keep the connection on map for ages even though they sleep and not listen at all..but e.g. smoke detector keep detaching and also e.g Sonoff temperature which is basically very similar to aqara...I need to find out what are the rules for routing table. What kicks some records out of it and why some of the sleepy devices remain untouched without any problem. |
Hi @Koenkk, You mentioned also a new firmware along with the Z2M update https://electrolama.com/projects/zig-a-zig-ah/#flash-firmware
Greetzzz, Gerben |
Just released version Zigbee2MQTT 1.16.0 so make sure to use that. |
Hey @Koenkk , don't want to steal the conversation, but I noticed I also have a "No network route" error with a SilverCrest Motion sensor when pairing to zzh with the latest firmware. I created a forum post here: And here's the relevant log file when pairing that device fails: Search for "No network route". Not sure if related, just though worth pointing it out. |
@andrej-peterka For me it seems more like a failed interview - many 240 errors.. and subsequent removal from network and thus no route is found anymore. Have you tried to pair closer to the coordinator? Btw didn't you press pairing button while the interview was still in progress aprox after 17 seconds? Maybe Koen have better understanding, from the logs it seems to me like unfinished interview and device leaving the network. |
@pedroke I paired it right next to the coordinator and did not touch anything. Maybe some motion triggered something, not sure. I'm a newbie when it comes to the zigbee network, so my knowledge is lacking in this regard, sorry. Seems I jumped to conclusion too fast. |
@andrej-peterka As it turned out with Honeywell smoke detector, the problem was with enrolling it to IAS zone. It is a special feature of zigbee specification for some security sensors.. in the end we found that coordinator was not triggering the enroll procedure in a way that sensor expected.. and it was fixed. But there were logs about enrolling.. In your case I see some data confirmation errors code 240 which is "transaction expired" that's why I thought that interview failed because of weak signal..but now..I don't know.. Btw..good to know that Lidl is also in Zigbee business ;) |
I had the same difficulty with pairing by just holding it close to my ZZH adapter. |
@Koenkk Result was that 2 of the 3 JTYJ-GD-01LM/BW smoke alarm devices suddenly had no network route. This the log after restart Z2M...totally different values then the 3rd “Rookmelder_zolder”.
Update: the Smoke alarm which I have paired now with the 2530 router now has an LQI |
@Anjerlaan I had also problems after update. On Koen's suggestion I deleted coordinator backup BEFORE I plugged freshly flashed ZZH and that helped. Obviously the ZZH NVRAM included zigbee related variables such as indirect message timeout and thus even with the new firmware old values were restored from backup. |
@pedroke So let met get it straight:
|
This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 7 days |
Bug Report
What happened
Redoing my whole setup I just recently paired the Honeywell smoke detector. I was paired directly to the CC2531 coordinator. To make sure the device works I pressed the button on the detector and Z2M received a normal payload message.
Then I decided it would be good to do a self test. That failed with the following error:
I read the other issues on the similar problems but they all refer to the device being connected to a router.
I once again had problems with pairing the device (see #4271). I had to resort to the secondary device and resetting the adapter configuration. This time I was not able to finish the interview sucessfully. I fixed the database to say interview successful true and the device seems to work with the exception of the selftest.
What did you expect to happen
To hear a beep of the device after performing the selftest.
How to reproduce it (minimal and precise)
See above.
Debug Info
Zigbee2MQTT version: 1.14.4
Adapter hardware: CC2531
Adapter firmware version: 20190619
The text was updated successfully, but these errors were encountered: