Z-Stack_3.x.0 coordinator 20240710 feedback

[Koenkk]

Update 2024-08-24:
The 20240710 firmware has been officially released. Feel free to continue discussing the firmware in this thread. Note that I will not actively monitor this discussion anymore. In case of issues with the firmware, report a new issue here. Make sure to include:

• The debug logging from starting Z2M until the issue occurs.
• Wether or not 20221226 worked correctly

---

Welcome to another firmware testing thread!

After testing a countless amount of different firmwares (many thanks to @nick4275 for all the testing!), we (very likely 🤞) found the root cause of the issues causing the 20230507 firmware to crash and managed to fix these (together with help from Texas Instruments).

Even more excitingly, this firmware is built on the latest TI SDK (7.41).

This is a continuation of #474, #478, #483 and #496.

Download: 20240710.zip

When providing feedback copy and comment the template below:

- Amount of devices in network:
- Bridge software (Zigbee2MQTT/ZHA):
- 20221226 worked fine: yes/no

If possible, provide herdsman debug log or even sniffing logs from the moment everything works fine until it fails.

If no breaking issues are found, this firmware will be officially released on 24 August.

[sjorge]

I tried flashing SLZB-06P7 on CC1352P2_CC2652P_other_coordinator_20240710.hex but that didn't seem to work, granted It was my first time trying to flash via USB. It just wouldn't verify for me. And the adapter wouldn't start afterwords. I did managed to flash 20240316 over USB again and it works again.

I'm not sure I did something wrong or not, as using a different fw image seemed to work fine.

[agriadsi]

Most likely your SLZB-06P7 uses CC2652P7 chip, so you need to use CC1352P7_coordinator_20240710.hex firmware.

[sjorge]

That is probably it, I'll try again later. My bash history showed me using CC1352P7_coordinator when restoring the old version but not when trying the new version. 😭

[simeon-walker]

deleted since I don't think it's a firmware issue

[Flavien]

Thanks @Koenkk, what are the improvements for ZigBee on SDK 7.41?

[hesham1216]

I've built 20dBm enabled firmware for sticks/gateways with RF-Star RF-BM-2652P4I module (CC2652P7 chip): UZG-01, HMG-01 Plus, Stick v4 etc. You can try it at your own risk. For me it's been working for more than a week now.

MOD information:

1. Made from the 202407010 sources published by @Koenkk (many thanks to) with minor modifications.

2. Built on Windows 11 computer, all firmware files are slightly smaller than ones published by Koen.

3. CCS 12.8 was used.

4. Changed CC2652P7 transmitter power settings table to allow 20 dBm TX power, according to the recommended method on the E2E TI forum.

5. Changed CC2652P7 chip settings based on RF-BM-2652P4I specifications and CC13xx/CC26xx Hardware Configuration and PCB Design Considerations:

   • DCDC converter built into the chip is enabled (as it is used according to the module schematic),
   • 48Mhz crystal capacitors built into the chip are turned off (as the module schematic shows external capacitors are used, as they are necessary for 20dBm),
   • 'Force VDDR' is enabled (necessary for 20dBm).

6. Internal build number changed to 20240711.

The TX Power can be set through standard advanced/transmit_power z2m setting up to 20.

Firmware and affected source files can be found here.

If you're using XZG firmware on your gateway, you can flash this mod directly from Tools setting page (thanks to @xyzroe).

On my setup coordinator-router LQI rises from 72 at 10dBm TX power to 109 at 20dBm.

Hello

Do you please confirm both router firmware and coordinator firmware you have on github is working based on CC2652P7 (RF-BM-2652P4I), as I am designing a new PCB router and Coordinator and I was looking for just that (I am upgrading the PCB design from CC2652P2 - to CC2652P7)

Thank you

[agriadsi]

Hi! Shortly - yes. This router firmware is based on latest available router firmware by Koen - 20221102, SDK 6.30.00.84. It does not have NPI interface (no communication with host MCU). But for me it works. To join the network, you have to enable coordinator join and power cycle router device. All other changes are the same, as for coordinator FW (+1 build number, 20 dBm, DCDC and crystal caps settings for RF-BM-2652P4I). As for now I didn't have any CC2652P7 dev board or spare gateway to add NPI and test it.

[hesham1216]

Hi! Shortly - yes. This router firmware is based on latest available router firmware by Koen - 20221102, SDK 6.30.00.84. It does not have NPI interface (no communication with host MCU). But for me it works. To join the network, you have to enable coordinator join and power cycle router device. All other changes are the same, as for coordinator FW (+1 build number, 20 dBm, DCDC and crystal caps settings for RF-BM-2652P4I). As for now I didn't have any CC2652P7 dev board or spare gateway to add NPI and test it.

Excuse me but I am not very much into Zigbee firmware so can you please tell me more about the NPI? What difference it will make when you add to the router firmware?

Also about the DCDC and capacitor settings these are already written in the firmware or does it need to be done manually somehow? Because I want to burn the firmware and make it work out of the box to save some time

I am buying couple of RF Star CC2652P7 boards soon
I can try if you want me to, just let me know what I should do and I will share the results

Would be nice if you have telegram or something so we can talk about it

Thank you for sharing your work and support 🙏

[agriadsi]

about the NPI

NPI is the communication interface between CC2652Px chip and host MCU, eg. on UZG-01/HMG-01 Plus gateway host MCU (ESP32) queries FW build number from CC2652 chip. This router FW has no ability to answer. So It works silently (like 'black box', plug and forget). The only communication can be done through Z2M to set TX Power ('Exposes' device menu).

I can try if you want me to, just let me know what I should do and I will share the results

Thanks, I'm waiting for RF-BM-2652P4I Dev Board to come from Ali any time soon.

In telegram my nick is the same, but you may be disappointed, as I'm not experienced developer (just building my smart home) and my English is not perfect).

[hesham1216]

about the NPI

NPI is the communication interface between CC2652Px chip and host MCU, eg. on UZG-01/HMG-01 Plus gateway host MCU (ESP32) queries FW build number from CC2652 chip. This router FW has no ability to answer. So It works silently (like 'black box', plug and forget). The only communication can be done through Z2M to set TX Power ('Exposes' device menu).

I can try if you want me to, just let me know what I should do and I will share the results

Thanks, I'm waiting for RF-BM-2652P4I Dev Board to come from Ali any time soon.

In telegram my nick is the same, but you may be disappointed, as I'm not experienced developer (just building my smart home) and my English is not perfect).

Thank you for your support much appreciated :)

When you receive dev board and test please share the results 🙏

[KoKolaj]

Coordinator Device: SONOFF Zigbee 3.0 USB Dongle Plus (CC2652P)

• Amount of devices in network: 97
• Bridge software: Zigbee2MQTT
• 20221226 worked fine: yes
• 20230507 worked fine: no
• 20230922 worked fine: ---
• 20230923 worked fine: yes
• 20240315 worked fine: yes
• 20240710 worked fine: yes

After upgrading to version 20240710, two devices showed an error and had to be again paired.
Around two hours some devices showed errors with time.
One example:
"error 2024-03-30 15:14:48Publish 'set' 'state' to 'Kuchyně zásuvka digestoř1' failed: 'Error: Command 0xa4c138764a954afd/1 genOnOff.off({}, {"timeout":10000,"disableResponse":false,"disableRecovery":false,"disableDefaultResponse":false,"direction":0,"srcEndpoint":null,"reservedBits":0,"manufacturerCode":null,"transactionSequenceNumber":null,"writeUndiv":false}) failed (Timeout - 63889 - 1 - 217 - 6 - 11 after 10000ms)'

Now the device seems to be stabilized.

[KoKolaj]

There was no problem after 4 days of operation.
Everything is still very stable, without failures.
LQI also strengthened slightly.
Thank you for your great job.

[LNKRCF]

Hi KoKolaj. Which firmware did you install? I found these two in the zip file
CC1352P2_CC2652P_launchpad_coordinator_20240710.hex
CC1352P2_CC2652P_other_coordinator_20240710.hex
Thank you.

[lucasteligioridis]

• Amount of devices in network: 115 [End devices: 56 - Router: 59]
• Bridge software (Zigbee2MQTT/ZHA): 1.39.0
• 20230507 worked fine: no
• 20240316 worked fine
• 20240710 worked fine
• Coordinator: TubesZB CC2652 P7 PoE

Flashed it successfully, will post a reply to my own thread for any issues, but didn't see any devices (that I've noticed) not pair correctly after a flash. So will wait and see, so far all looking good.

[lucasteligioridis]

Had the firmware for 2 weeks, has been rock solid and my network definitely feels overall more responsive!
I had 1 device I needed to pair again, but I reckon that was just from coming out of a new flash.

[RusselIce]

Amount of devices in network: 141
Bridge software (Zigbee2MQTT/ZHA): 1.39.0
20240409 worked fine
20240710 worked fine
Coordinator: UZG-01 (CC2652P7)

all devices work stably

[popy2k14]

Amount of devices in network: 132 [End devices: 90 - Router: 42]
Bridge software (Zigbee2MQTT/ZHA): Z2M 1.39.0
Coordinator: SONOFF Zigbee 3.0 USB Dongle Plus,Zigbee Gateway TI CC2652P + CP2102N
20221226 worked fine: yes
20230507 worked fine: no
20230922 worked fine: yes
20240316 worked fine: no
20240710 worked fine: ... flashed an we will see. At first, all devices online an reactive except 6x aqara door/window & water sensors. Repairing fixed them...

[popy2k14]

Over a month and the coordinator + mesh is rock solid stable with this version (as it was for 20230922).
Also no devices dropped or something like that.

[nbris95]

Would love to know more about your setup. Having a lot of issues w/ the Aqara sensors.

[popy2k14]

Most of my routers (if not all) are hue, they are working the best. I just have ~15 aqara door/window sensors and 2x P1 motion sensors (modded for faster response times).
I have tested a lot of firmware sind 20221226 and also on some of them aqara sensors got dropped. But not on 1226, 20230922 and 20240710.

I am using an: SONOFF Zigbee 3.0 USB Dongle Plus,Zigbee Gateway TI CC2652P + CP2102N Zigbee USB Stick

[edmonddc]

Can I use the baud rate 460800 on this firmware?

[mrmaximas]

Can I use the baud rate 460800 on this firmware?

what effect do you expect from such a port speed?

[fhloston]

To my understanding the 460800 version of 20240316 was intended to test if high device count or sensor rich deployments profit due to the then higher bandwidth between z2m and the transceiver.

I am running this x4 20240316 firmware currently.

Looking at the bindiff, it is a simple 1 byte edit.

However I fail to determine the proper place in the 20240710 firmware for that edit.

Can you help us out with a 460800 version?

Or should all on the x4 version revert back to 115200?

[edmonddc]

I changed my baud setting back to 115200 to test this. I never attempted to test it with 460800.

I was on x4_20240316 and recently most of my devices was offline. I was forced to "test" the 20240710.

[asychev]

Amount of devices in network: 23
Bridge software: Z2M 1.39.0
Coordinator: SLZB-06
20240315 worked fine: yes
20240710 worked fine: NO (unstable, does not work at all after ~ 1-2h)

[tonicb78]

Yes, no problems here with my SLZB-06. Just updated and then rebooted it.

[asychev]

My also works initially, but then suddenly stop after an hour or so. Requesting mesh map speed up the issue.

[tonicb78]

Tried requesting a mesh map without any issues. After 8h everything seems to work properly, even aqara devices are reporting properly without touching them. Quite strange. My SLZB-06 is conneted via POE, and no BT proxy anabled.

[Anto79-ops]

just updated. Time will tell if there is an issue. updated, rebooted and then waited a few minutes and asked for a map. So far, so good.
SLZB-06 (non-poe). No matter or BT or thread enabled.

[Anto79-ops]

I posted a separate response but so far 2 to 3 hours and everything was good the only thing that happened was all 4 of my Ikea tradfri Outlets were offline/not responding. all I did was unplug them and plugged them back in and they came back.

[tonicb78]

Amount of devices in network: 79 (End devices: 51 Router: 28)
Bridge software: Z2M 1.39.0
Coordinator: SLZB-06
20221226 worked fine: yes
20240315 worked fine: yes
20240710 worked fine: yes

[svmaris]

Amount of devices in network: 85 (End devices: 45, Router: 40)
Bridge software: Z2M 1.39.0-dev commit: [2dbbed1]
Coordinator: SLZB-06P7
20240315 worked fine: yes
20240710 worked fine: yes

[jchurchward]

Amount of devices in network: 45 (End devices: 26, Router: 19)
Bridge software: Z2M 1.39.0
Coordinator: SLZB-06
20240315 worked fine: yes
20240710 worked no: It became unstable shortly after and began losing some repeaters. Then, it started losing connection with devices altogether. So, I rolled it back to version 20240315, and it has been functioning flawlessly ever since.

[SmolikOne]

Amount of devices in network: 89 (End devices: 61 Router: 28)
Bridge software: Z2M 1.39.0
Coordinator: Sonoff Dongle-P
20221226 worked fine: not fully, timeouts after few days
20240315 worked fine: no
20240710 worked fine: no, it dies after ca. 2 hours with error ZCL command... BUFFER_FULL, so once again I'm returning to 20221226 firmware

Update: I've tried 20240710 one more time, it worked ok for two days, but on third Buffer full error, restarting RPi4 host repairs the issue, but it works for few minutes and again Buffer full error and nothing is working anymore.

Additionally I'm having random network slowdowns with error Read response to (address of the device) failed.

[Pferdebockwurst]

Via SLZB it's a 20240710 coordinator firmware with higher transmit power, that's why it's 20240716. There is also a router firmware with this date, but I've flashed coordinator.

Where is the router firmware from? Do you have a link?

[SmolikOne]

It's available directly via SLZB-06 coordinator management page, but I've tried also 20240710 and the results were the same.

[Pferdebockwurst]

It's available directly via SLZB-06 coordinator management page, but I've tried also 20240710 and the results were the same.

That's interesting. As far as I can see, @Koenkk didn't publish a router firmware based on the newest stack. So I really don't know if the firmware you are using is the same as the one we are discussing about, if you flash a firmware which originates from somewhere else.

[alienatedsec]

@SmolikOne I dont see anything newer than below - even the router version disappeared for me. I mentioned that I had to power cycle other routers here

[SmolikOne]

Now I see and you are right, 20240716 I'ts SMLIGT modification of Koenkk 20240710 firmware (apparently). But as said previoisly, I've tried also Koenkk 20240710 and errors were exactly the same.

[CultusMechanicus]

Coordinator Device: SONOFF Zigbee 3.0 USB Dongle Plus (CC2652P)

Amount of devices in network: 61 (Routers: 31, End devices: 30)
Bridge software: Zigbee2MQTT

• 20221226 worked fine: yes
• 20230507 worked fine: no
• 20240315 worked fine: no, devices dropped off in batches at random, pairing often required a restart of Z2M
• 20240710 worked fine: yes, impressively every device that was offline rejoined overnight, apart from one elderly Aqara motion sensor.

[Anto79-ops]

Amount of devices in network: 43 (End devices: 36 Routers: 7)
Bridge software: Z2M 1.39.0
Coordinator: SLZB-06
20221226 worked fine: n/a
20240315 worked fine: yes
20240710 worked fine: yes

Updated and rebooted the device, and that went well.

2 hours in, and I noticed all 4 of my Ikea Tradfri outlets devices were not responding, so I just unplugged them and plugged them back in, and they came back. Perhaps if I left them, they would have come back by themselves, but considering they are critical to my network, I wanted to make sure they were working first

[Anto79-ops]

its been 2 weeks now on 20240710 and overall, this has been really good. There is less lag and more preformance increases.

thank you devs and team!

[Triarii1]

Z-Stack 20221226 has been working flawlessly. After seeing the mixed comments here, not keen to update.

As always with updates, if the older version is working and the new release does not bring anything new, no point in updating.

[kevincaradant]

Yes. I tried to update to fix an issue with a switch roller device and mostly to increase the range signal (LQI) of the network.
But the New firmware version didn't work as expected so stay on a tested and approved previously version which is in my case 20221226 is still a good idea 😁

[rwjack]

After seeing the mixed comments here, not keen to update.

Feel the same after reading through the comments on the past few releases.

I'm still on 20210708 - Sonoff ZB Plus thingamajig, the same firmware that was loaded when I bought it, works flawlessly.

Are there any actual new features in the firmware since the 20210708 release? Or is it all 'improved connectivity', etc.

[almirus]

Zigbee2MQTT version
1.40.0 commit: unknown
Coordinator type
zStack3x0
Coordinator revision
20240710
Coordinator IEEE Address
0x00124b0022810ef5
Frontend version
0.7.4
zigbee-herdsman-converters version
20.8.4
zigbee-herdsman version
0.57.1
Amount of devices in network: 49
Coordinator: egony stick v4 (CC2652P)
Getting error after goto to pairing mode

error 2024-09-01 22:13:43z2m: Failed to configure 'Датчик температуры на кухне', attempt 1 (Error: Bind 0xa4c1380b27b90459/1 msTemperatureMeasurement from '0x00124b0022810ef5/1' failed (ZDO error: bind failed with status 'TABLE_FULL' (140)) at /app/node_modules/zigbee-herdsman/src/adapter/z-stack/adapter/zStackAdapter.ts:1126:36)
error 2024-09-01 22:13:53z2m: Failed to configure 'Датчик температуры в спальне', attempt 1 (Error: Bind 0xa4c1385678f731eb/1 msTemperatureMeasurement from '0x00124b0022810ef5/1' failed (ZDO error: bind failed with status 'TABLE_FULL' (140)) at /app/node_modules/zigbee-herdsman/src/adapter/z-stack/adapter/zStackAdapter.ts:1126:36)
error 2024-09-01 22:15:23z2m: Failed to configure 'Датчик температуры на кухне', attempt 2 (Error: Bind 0xa4c1380b27b90459/1 msTemperatureMeasurement from '0x00124b0022810ef5/1' failed (ZDO error: bind failed with status 'TABLE_FULL' (140)) at /app/node_modules/zigbee-herdsman/src/adapter/z-stack/adapter/zStackAdapter.ts:1126:36)

[agriadsi]

20240710 firmware has lower limit for directly connected devices (50). May be this is the problem?

[habitats-tech]

Flashed SLZB-06P7 with 20240710 FW a couple of weeks ago. Since the new FW was applied, the device randomly stops receiving updates from devices, while Zigbee2MQTT reports no issues except communication errors.

Restarting ESP32 or Zigbee (the last time I restarted Zigbee, was not enough, I had to restart ESP32 for the device to start responding; it could have been I did not give enough time for the devices to start reporting - gave it a couple of minutes which was my prior experience with the failure) the SLZB-06P7 starts processing packets again.

Time between failures, one took 2 days, the other 6 days. The last 2 weeks the device failed with the same issue 3 times.

Initially I thought it was an isolated incident, but now I am more confident is a FW issue.

Are these type of issues related to TI chipsets only?

[santanar00]

I have exactly the same scenario, using the same hardware, however, I don't know if it was a coincidence or not, I noticed that it started happening when I increased the number of devices on my network. Currently I have 80 devices connected, 42 of which are routers.

I took care to avoid directly connecting the devices to the coordinator, connecting most of the devices between the routers and only a few routers to the coordinator (about 15).

[rodrigogbs]

Flashed SLZB-06P7 with 20240710 FW a couple of weeks ago. Since the new FW was applied, the device randomly stops receiving updates from devices, while Zigbee2MQTT reports no issues except communication errors.

Restarting ESP32 or Zigbee (the last time I restarted Zigbee, was not enough, I had to restart ESP32 for the device to start responding; it could have been I did not give enough time for the devices to start reporting - gave it a couple of minutes which was my prior experience with the failure) the SLZB-06P7 starts processing packets again.

Time between failures, one took 2 days, the other 6 days. The last 2 weeks the device failed with the same issue 3 times.

Initially I thought it was an isolated incident, but now I am more confident is a FW issue.

Are these type of issues related to TI chipsets only?

Same here, rolling back to 20240315 everything starts working again.

[habitats-tech]

I have exactly the same scenario, using the same hardware, however, I don't know if it was a coincidence or not, I noticed that it started happening when I increased the number of devices on my network. Currently I have 80 devices connected, 42 of which are routers.

This might well be the issue. I have close to 150 devices, one third is routers, plus one SLZB router.

[Mrsash2020]

I also agree with this 20240315 for me works. The moment I update the SLZB-06P7 to 20240710. Once I update it the devices go offline and dont connect for a long time. This I have tried twice previously and the only way to get everything working is re flash 20240315 and re -pair every device. Am I meant to be doing anything else?

I thought the firmware update is meant to be seamless.
I am on Z2M.
Also am I meant to update the config etc?

[l-mb]

Just upgraded (53 devices, SonOff ‎ZBdongle-P), and it does appear more stable (especially during the OTA updates).

Nice, thank you so much!

[habitats-tech]

You can interchange SLZB-06P7 with UZG-01 without changing anything. Just shut down Z2M replace the coordinator and fire up Z2M. The same applies to any FW upgrade/downgrade. There is no need to re-pair anything. You need to be patient as it takes a couple of hours for the network to stabilise and at least a day for everything to be fully connected.

If using 20240710 with 1.40.0 I have noticed when something goes offline, just try to re-interview. It has worked for me in one instance, but took almost a day. I have tried on a second device and waiting to see if it has worked.

[chris-1243]

You can interchange SLZB-06P7 with UZG-01 without changing anything. Just shut down Z2M replace the coordinator and fire up Z2M. The same applies to any FW upgrade/downgrade. There is no need to re-pair anything. You need to be patient as it takes a couple of hours for the network to stabilise and at least a day for everything to be fully connected.

You must adapt the ieee-address from the old coordinator to the new one

https://www.zigbee2mqtt.io/guide/adapters/flashing/copy_ieeaddr.html

[habitats-tech]

You must adapt the ieee-address from the old coordinator to the new one; this is correct sorry I missed this step.

Re-interviewing, even if successful, does not resolve offline issues.

[alp-by]

zigbee2mqtt addod at Home assistance doesn't work.
Zigbee2mqtt has the latest update: 1.40.1-1
Coordinator: UZG-01 20240707

[mhoogenbosch]

zigbee2mqtt addod at Home assistance doesn't work.

Zigbee2mqtt has the latest update: 1.40.1-1

Coordinator: UZG-01 20240707

Personally I don't think it's related to the firmware. If this would be the case I'd think there would be a lot more issues reported. Maybe investigate a bit more and open a new thread.

[cloudbr34k84]

i get alot of these now since upadting to this firmware. I just resett the zigbee module through the UI

[cloudbr34k84]

Another complete crash for me. here. Worst version to date unfortunately

[dankocrnkovic]

Zigbee2mqtt has the latest update: 1.40.1-1
Coordinator: UZG-01 20240707

Again last night total adapter crash... (Web gui still works, and zigbee reset on UZG does not help). After PoE reset, and I let it for 1 hour to stabilize, network was a mess.. Ikea bulbs were reporting no network route. had to reset all router devices with appartment elec. braker. Yesterday befor the crash I noticed that everything is very laggy.

And it is like this every 1-2 weeks from the day this firmvare version is released. Already rejoined all devices.
As I can see SONOF dongles are the winners here, but this newer devices are simpy not working.

We can not say it works but sometimes devices disconect. That is not Producton ready then. My wife will kill me soon :)
And yes, I rolled back to last stable for me, but its like this firmware messed up something on devices, as older firmware behaves the same now.

[cloudbr34k84]

Zigbee2mqtt has the latest update: 1.40.1-1 Coordinator: UZG-01 20240707

Again last night total adapter crash... (Web gui still works, and zigbee reset on UZG does not help). After PoE reset, and I let it for 1 hour to stabilize, network was a mess.. Ikea bulbs were reporting no network route. had to reset all router devices with appartment elec. braker. Yesterday befor the crash I noticed that everything is very laggy.

And it is like this every 1-2 weeks from the day this firmvare version is released. Already rejoined all devices. As I can see SONOF dongles are the winners here, but this newer devices are simpy not working.

We can not say it works but sometimes devices disconect. That is not Producton ready then. My wife will kill me soon :) And yes, I rolled back to last stable for me, but its like this firmware messed up something on devices, as older firmware behaves the same now.

im on HG POE Plus which is Hamgeeks version of the UZG-01 20240707 and im now just reflashing the firmware when this happens..

[habitats-tech]

We all need to assist Koenkk get to the bottom of the issues. Each with their positive contribution. The process is frustrating and long but we need to get things resolved one way or another.

Let's all embrace a positive attitude and assist in getting a working solution acceptable to everyone/most.

Please assist any way you can. Reverting to older FW does not speed up the process of getting a handle of what the issues are.

IMHO 20240710 is reliable enough for production use, although not the most desirable solution.

I believe at this point nobody has a clear understanding of the full extent of the issues and therefore cannot attempt to provide a final solution. To me it seems a trial and error process until we fully understand the underlying issues.

[chris-1243]

What kind of brands do you have in your network ?

I would avoid Awox and some Sonoff router (Zbmini...). They are unfortunately known to be bad routers and may mess up your network completely.

I would check for interference (wifi) as well.

20240710 is based on SDK 7.40 and 20240315 on an older SDK (6.10). 20230507 is known to have some bugs (SDK 7.10).

CC2652P7 are not supported on SDK below 7.x. If I am not wrong, 20240315 should work on P7 chip even if based on SDK 6.xx. @Koenkk did a hard job to support P7 chip on older SDK as it seems more stable for some users.

[dankocrnkovic]

Network was rock solid stable for 2 years with same devices. Mostly Ikea an aqara.. No changes in network devices.
then I changed coordinator to UZG-01 on 2040315 and was stable about 2 months until i updated to 20240710. Then all went down.

[raveit65]

We all need to assist Koenkk get to the bottom of the issues. Each with their positive contribution. The process is frustrating and long but we need to get things resolved one way or another.

Let's all embrace a positive attitude and assist in getting a working solution acceptable to everyone/most.

Please assist any way you can. Reverting to older FW does not speed up the process of getting a handle of what the issues are.

IMHO 20240710 is reliable enough for production use, although not the most desirable solution.

I believe at this point nobody has a clear understanding of the full extent of the issues and therefore cannot attempt to provide a final solution. To me it seems a trial and error process until we fully understand the underlying issues.

Normally i would do a git bisect action to find the culprit commit for weird and unknown issues with a software package.
But than you need to know how to build the firmware from git to test bisect results.

For the record i got mentioned issues with zigstar UZG-01 bridge. I tried to update 2 times to 20240710 firmware but everytime i run in weird connection issues. I will roll back to stable 20240316 for the moment

[LeoeLeoeL]

We all need to assist Koenkk get to the bottom of the issues. Each with their positive contribution. The process is frustrating and long but we need to get things resolved one way or another.
Let's all embrace a positive attitude and assist in getting a working solution acceptable to everyone/most.
Please assist any way you can. Reverting to older FW does not speed up the process of getting a handle of what the issues are.
I wrote this FW wasn't OK for my network but Koenk ignored my issue so that the only thing I could do has been to fall back to 20240922.
I tested everything between 20230922 and 20240710 and nothing works..
It's useless to test again 20240710 for me.

[LeoeLeoeL]

We all need to assist Koenkk get to the bottom of the issues. Each with their positive contribution. The process is frustrating and long but we need to get things resolved one way or another.
Let's all embrace a positive attitude and assist in getting a working solution acceptable to everyone/most.
Please assist any way you can. Reverting to older FW does not speed up the process of getting a handle of what the issues are.

I wrote this FW wasn't OK for my network but Koenk ignored my issue so that the only thing I could do has been to fall back to 20240922.
I tested everything between 20230922 and 20240710 and nothing works..
It's useless to test again 20240710 for me.

[habitats-tech]

If my observations are correct the fact that a device goes offline does not mean it is not operational or not communicating. Devices can show offline, yet their last seen can always show a value of a few minutes (not hours of days). So when you see a device showing offline check the last seen value as well.

Until someone can articulate the availability algorithm we will not be able to contribute in a meaningful way. For my understanding, the issue is what availability is checking and how (someone posting the code snippet is fine as well). Why using 20240315 router FW shows the router always online, while using 20240716 router FW alternates between online/offline. What has changed between the two FW releases which makes availability behave in a different way.

[dankocrnkovic]

This is new....

Permit join just disables after fiew seconds....

[2024-09-09 12:04:42] error: zh:controller: Failed to keep permit join alive: Error: SRSP - ZDO - mgmtPermitJoinReq after 6000ms
Error: SRSP - AF - dataRequestExt after 6000ms
at Object.start (/app/node_modules/zigbee-herdsman/src/utils/waitress.ts:59:23)
at /app/node_modules/zigbee-herdsman/src/adapter/z-stack/znp/znp.ts:300:45
at Queue.execute (/app/node_modules/zigbee-herdsman/src/utils/queue.ts:36:26)
at ZStackAdapter.dataRequestExtended (/app/node_modules/zigbee-herdsman/src/adapter/z-stack/adapter/zStackAdapter.ts:1245:9)
at /app/node_modules/zigbee-herdsman/src/adapter/z-stack/adapter/zStackAdapter.ts:681:13
at Queue.execute (/app/node_modules/zigbee-herdsman/src/utils/queue.ts:36:20)
at GreenPower.permitJoin (/app/node_modules/zigbee-herdsman/src/controller/greenPower.ts:261:13)
at Controller.permitJoinInternal (/app/node_modules/zigbee-herdsman/src/controller/controller.ts:311:13)
at Timeout._onTimeout (/app/node_modules/zigbee-herdsman/src/controller/controller.ts:301:25)

[chris-1243]

error: zh:controller: Failed to keep permit join alive: Error: SRSP - ZDO - mgmtPermitJoinReq after 6000ms
Error: SRSP - AF - dataRequestExt after 6000ms

Your adapter has crashed unfortunatelly.
As suggested by @Koenkk, I would open a new issue. Follow the link in first post.

If 20240315 is working fine for you, I would remain on this firmware for the time. I have some Ikea devices (router and end devices) and they all work fine

[dankocrnkovic]

error: zh:controller: Failed to keep permit join alive: Error: SRSP - ZDO - mgmtPermitJoinReq after 6000ms
Error: SRSP - AF - dataRequestExt after 6000ms

Your adapter has crashed unfortunatelly. As suggested by @Koenkk, I would open a new issue. Follow the link in first post.

If 20240315 is working fine for you, I would remain on this firmware for the time. I have some Ikea devices (router and end devices) and they all work fine

Its not crashed... its working with devices, and afer 3-4 retrys it stayed in pairing mode.
I cant go to back to 20240315 (it ws working without problems) as the problems with disconecting (or rather frozen devices) is the same after I once upgraded to 20240710.

[chris-1243]

Its not crashed...

https://www.zigbee2mqtt.io/guide/faq/#common-error-codes

I let you read. Z2M was able to recover but it crashed. What procedure do you follow while flashing your adapter ? It is important to cleanly stop Z2M then you may flash your coordinator.

[cloudbr34k84]

This was common for me over that last week. Reflash the firmware to that coordinator

…

On Mon, 9 Sept 2024, 8:13 pm dankocrnkovic, ***@***.***> wrote: This is new.... Permit join just disables after fiew seconds.... [2024-09-09 12:04:42] error: zh:controller: Failed to keep permit join alive: Error: SRSP - ZDO - mgmtPermitJoinReq after 6000ms Error: SRSP - AF - dataRequestExt after 6000ms at Object.start (/app/node_modules/zigbee-herdsman/src/utils/waitress.ts:59:23) at /app/node_modules/zigbee-herdsman/src/adapter/z-stack/znp/znp.ts:300:45 at Queue.execute (/app/node_modules/zigbee-herdsman/src/utils/queue.ts:36:26) at ZStackAdapter.dataRequestExtended (/app/node_modules/zigbee-herdsman/src/adapter/z-stack/adapter/zStackAdapter.ts:1245:9) at /app/node_modules/zigbee-herdsman/src/adapter/z-stack/adapter/zStackAdapter.ts:681:13 at Queue.execute (/app/node_modules/zigbee-herdsman/src/utils/queue.ts:36:20) at GreenPower.permitJoin (/app/node_modules/zigbee-herdsman/src/controller/greenPower.ts:261:13) at Controller.permitJoinInternal (/app/node_modules/zigbee-herdsman/src/controller/controller.ts:311:13) at Timeout._onTimeout (/app/node_modules/zigbee-herdsman/src/controller/controller.ts:301:25) — Reply to this email directly, view it on GitHub <#505 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AOB2WBF5TMXW43EJ56KO2C3ZVVYGHAVCNFSM6AAAAABK2RNKGCVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTANJYHEZDMMQ> . You are receiving this because you were mentioned.Message ID: ***@***.*** com>

[raveit65]

Normally i would do a git bisect action to find the culprit commit for weird and unknown issues with a software package.
But than you need to know how to build the firmware from git to test bisect results.

For the record i got mentioned issues with zigstar UZG-01 bridge. I tried to update 2 times to 20240710 firmware but everytime i run in weird connection issues. I will roll back to stable 20240316 for the moment

[Joylancer]

Amount of devices in network: 26 (End devices: 22 Routers: 4)
Bridge software: Z2M 1.40.1 commit 403d3c0
Coordinator: Sonoff Zigbee Bridge Pro (TCP Bridge)
20220219 worked fine: yes
20230507 worked fine: no
20240710 worked fine: yes

Sonoff ZBMini worked fine: yes
Sonoff SNZB-02 worked fine: yes
Aqara RTCGQ14LM & RTCGQ11LM worked fine: yes
Tuya TS0043 & ZY-M100-S_2 worked fine: yes
Develco SMSZB-120 worked fine: yes
Xiaomi WXKG01LM worked fine: yes
Sonoff ZB Dongle-P as Router worked fine: yes

[habitats-tech]

@Koenkk

On 4 occasions I have attempted to access the webUI of UZG-01 the coordinator rebooted - lost and re-established connection within a few seconds. The last time it happened UZG-01 was on 23 hrs uptime.

I suggest you take a look at possible issues when coordinator busy a connection over the network for management could cause the connection to reset?

I have not tried to setup a continuous ping to test what happens when I try to access the webUI, but this is something I intent doing and will provide feedback.