Websocket reconnecting Zigbee2mqtt not starting after migration to new Rpi4

[hocomadvies]

I started my problem in an other topic #27391 but did not get any feedback so start a new topic here.

Main problem is I moved to a new rpi4 and installed zigbee2mqtt. taking the usb-controler from the previous (same version) rpi3 zigbee2mqtt and added it to the rpi4.
Adapted and used some of the config files from the rpi3 version. And gave it a go. The good news is I get the dashboard and see all the old devices listed. Only thing is that they do not work and the message "Websocket Reconnecting" is shown. Also I see that ping the devices is not succeeding.

Some info about my system (more at the other topic):

So after few days of searching and trying how to get this going I would be happy to get some advise how to solve this. Thanks.

zigbee2mqtt_configuration_29mei2026.yaml

reading the log I do not see a reason why it would not work? Or is the Coordinator firmware outdated?
`[2026-05-29 16:52:22] info: z2m: Starting Zigbee2MQTT version 2.10.1 (commit #16fd422c)
[2026-05-29 16:52:22] info: z2m: Starting zigbee-herdsman (10.0.7)
[2026-05-29 16:52:22] info: zh:zstack:znp: Opening SerialPort with {"path":"/dev/serial/by-id/usb-Silicon_Labs_slae.sh_cc2652rb_stick_-_slaesh_s_iot_stuff_00_12_4B_00_23>
[2026-05-29 16:52:22] info: zh:zstack:znp: Serialport opened
[2026-05-29 16:52:22] info: zh:zstack:znp: Writing CC2530/CC2531 skip bootloader payload
[2026-05-29 16:52:23] info: zh:zstack:znp: Skip bootloader for CC2652/CC1352
[2026-05-29 16:52:33] info: z2m: zigbee-herdsman started (resumed)
[2026-05-29 16:52:33] info: z2m: Coordinator firmware version: '{"meta":{"maintrel":1,"majorrel":2,"minorrel":7,"product":1,"revision":20211217,"transportrev":2},"type":>

error: z2m: Could not init sd_notify: Could not locate the bindings file. Tried:
→ /opt/zigbee2mqtt/node_modules/.pnpm/unix-dgram@2.0.7/node_modules/unix-dgram/build/unix_dgram.node
→ /opt/zigbee2mqtt/node_modules/.pnpm/unix-dgram@2.0.7/node_modules/unix-dgram/build/Debug/unix_dgram.node
→ /opt/zigbee2mqtt/node_modules/.pnpm/unix-dgram@2.0.7/node_modules/unix-dgram/build/Release/unix_dgram.node
→ /opt/zigbee2mqtt/node_modules/.pnpm/unix-dgram@2.0.7/node_modules/unix-dgram/out/Debug/unix_dgram.node
→ /opt/zigbee2mqtt/node_modules/.pnpm/unix-dgram@2.0.7/node_modules/unix-dgram/Debug/unix_dgram.node
→ /opt/zigbee2mqtt/node_modules/.pnpm/unix-dgram@2.0.7/node_modules/unix-dgram/out/Release/unix_dgram.node
→ /opt/zigbee2mqtt/node_modules/.pnpm/unix-dgram@2.0.7/node_modules/unix-dgram/Release/unix_dgram.node
→ /opt/zigbee2mqtt/node_modules/.pnpm/unix-dgram@2.0.7/node_modules/unix-dgram/build/default/unix_dgram.node
→ /opt/zigbee2mqtt/node_modules/.pnpm/unix-dgram@2.0.7/node_modules/unix-dgram/compiled/24.15.0/linux/arm64/unix_dgram.node
→ /opt/zigbee2mqtt/node_modules/.pnpm/unix-dgram@2.0.7/node_modules/unix-dgram/addon-build/release/install-root/unix_dgram.node
→ /opt/zigbee2mqtt/node_modules/.pnpm/unix-dgram@2.0.7/node_modules/unix-dgram/addon-build/debug/install-root/unix_dgram.node
→ /opt/zigbee2mqtt/node_modules/.pnpm/unix-dgram@2.0.7/node_modules/unix-dgram/addon-build/default/install-root/unix_dgram.node
→ /opt/zigbee2mqtt/node_modules/.pnpm/unix-dgram@2.0.7/node_modules/unix-dgram/lib/binding/node-v137-linux-arm64/unix_dgram.node
[2026-05-29 16:52:37] warning: z2m: Failed to ping '0x00158d00035659ed' (attempt 1/1, ZCL command 0x00158d00035659ed/1 genBasic.read(["zclVersion"], {"timeout":10000,"disab>

`

[hocomadvies]

Bit disappointing no one is responding on my help request.

I saw some older posts where the problem with "Could not init sd_notify: Could not locate the bindings file" was discussed. So this might be the same problem here as well? Has to do with pnpm somehow I think? Not sure how to fix hat.
My system is down for days now and and really want to have up and running again asp.

What are the options to get it working again?

• revert to previous version?
• run the updater ?
• flash the usb stick with new controller software (not my favourite thing). Or even get a new stick.
• other things to try or make it work?

node -v : v24.15.0 which node : /usr/bin/node pnp -v : 11.5.0 (when run this command, it says that it wanted to download it. so I agreed. Not sure if this has any impact need to test).

[MSL-DA]

Were data/database.db and data/coordinator_backup.json copied from the old installation?
Do pan_id, extended_pan_id, and network_key match the old installation?

[hocomadvies]

@MSL-DA yes in both the data/coordinator_backup.json those are the same. I just checked that.

[habitats-tech]

Use a USB 2.0 extension cable (2+ meters) to move your Zigbee dongle away from the Pi’s body. Helpful if you also share your config yaml.

[hocomadvies]

@habitats-tech config is the link above zigbee2mqtt_configuration_29mei2026.yaml. I can try with extension cable but have been using the stick for many years without any power issues. But sure will give it a try see if it matters.

[MSL-DA]

I can try with extension cable

A waste of time... it has nothing to do with a cable.

Do pan_id, extended_pan_id, and network_key match the old installation?

Check this.

[MSL-DA]

@hocomadvies

Bit disappointing no one is responding on my help request.

If you want help, you can start by doing what you're being asked to do, instead of wasting other people's time!

[hocomadvies]

@MSL-DA sorry there was a bit of delay in me seeing the different messages. So I only saw some later when already replied to them or others.

[habitats-tech]

missing from configuration.yaml
pan_id:
ext_pan_id:

Additionally I suggest for readability and organisational purposes you move all devices to devices.yaml file and replace devices in configuration.yaml as follows:

devices: devices.yaml

You can do the same in configuration.yaml for:

groups: groups.yaml

and in groups.yaml just have one line:
{}

I also suggest you use the USB cable to move your zigbee away from Pi.

One last suggestion you add the following in configuration.yaml:
report: false
cache_state: true

[hocomadvies]

Great :) Ok I have those id's generated. Still getting an error and not starting. Will see if a cable bring improvement.

[habitats-tech]

pan_id: and ext_pan_id: come from your old system. If you cannot find them then you need to create a new zigbee network and pair all devices again. To create a new zigbee network just replace the following:
network_key: GENERATE
pan_id: GENERATE
ext_pan_id: GENERATE

[hocomadvies]

I did not have a pan id in my old network the config is still around so I can check again.

[hocomadvies]

But the thing is that I don't want to pair all since there are several devices which are not very easy to reach (build in a case or so outside). The network keys are there. So why not the pan_id . Can I found those in the log files somewhere? I'm 100% certain I did not have them in my config before now.

[hocomadvies]

I have a backup of the coordinator. In which I found the pan_id etc. So i can use those to test.

[hocomadvies]

Tried that. Now it complains that the

part of the coordinator backup from the old rpi
{ "metadata": { "format": "zigpy/open-coordinator-backup", "version": 1, "source": "zigbee-herdsman@4.0.0", "internal": { "date": "2026-05-27T06:45:05.808Z", "znpVersion": 1 } }, "stack_specific": { "zstack": { "tclk_seed": "ca48ec393489d00a334b70995a306a71" } }, "coordinator_ieee": "00124b0021cc394a", "pan_id": "1a62", "extended_pan_id": "00124b0021cc394a", "nwk_update_id": 0, "security_level": 5, "channel": 11, "channel_mask": [ 11 ], "network_key": { "key": "daf1bddb096ec2fb6826a82bc06afe0d", "sequence_number": 0, "frame_counter": 64510397 },

this what the system give as error:
Refusing to start because configuration is not valid, found the following errors: May 30 16:29:48 zigbee2mqtt4 node[5502]: - advanced.pan_id: should be number or 'GENERATE' (is '1a62')

So why is it not using my pan_id "which is 1a62" according the coordinator_backup.json. and it is inserted without ' '

[hocomadvies]

Ok, made some progress. Thanks to Google AI I found out that, the configuration is(was) missing a valid pan_id. You can either set it to a specific 4-digit hexadecimal number (e.g., 0x1a62) and that is almost the same as my old 1a62 ... So using 0x1a62 made it working again. All my devices back and working.
I have even been able to switch some devices on off.

Only thing is that I still get this error and the system crashes after some minutes. So still something wrong here :)

/opt/zigbee2mqtt $ sudo systemctl start zigbee2mqtt Job for zigbee2mqtt.service failed because a timeout was exceeded. See "systemctl status zigbee2mqtt.service" and "journalctl -xeu zigbee2mqtt.service" for details.

For now zigbee2mqtt is in a loop (I think because of watchdog restarting it again). Not sure why.

[hocomadvies]

logfile attached.
log.txt

Any idea why zigbee2mqtt stops working?

[hocomadvies]

this is the part that pops up in the logs, over and over again.
[31-5-2026, 09:45:43] z2m: Stopping Zigbee2MQTT (restart=false, code=0, signal=SIGTERM)
than after short period it is starting again, and repeat. So it is partly working and I can use the devices if I am lucky it was not stopped.

[hocomadvies]

With a little help from google AI I have it working completely:
Had to remove the node_modules and install again the right ones.

/opt/zigbee2mqtt $ sudo rm -rf node_modules .pnpm-store
sudo pnpm install --frozen-lockfile --force
and
sudo nano /etc/systemd/system/zigbee2mqtt.service
Type=notify change into Type=simple

sudo systemctl daemon-reload