dresden-elektronik / deconz-rest-plugin

deCONZ REST-API plugin to control ZigBee devices
BSD 3-Clause "New" or "Revised" License
1.89k stars 498 forks source link

Lost of all network after Halt, Unplug/Plug and restart #1996

Closed gagadget closed 4 years ago

gagadget commented 4 years ago

Hello,

My Conbee GW is attached to my Rasberry Pi running HASSio using a USB extension cord.

Each time I reboot my PI, the Conbee GW loose the connexion to the Zigbee network : every devices appear in the VNC app but, they are not linked together and all the lights appears in grey in the conbee web app.

I tried everything I could, I am stuck.

Thx

Smanar commented 4 years ago

Hi, You have only xiaomi devices ? Or you have powered devices too ?

gagadget commented 4 years ago

Only TRÅDFRI devices : 21 lights, 2 power sockets, 3 motion sensors, 3 on/off switches and 3 remotes.

Everything is working nicely until I reboot.

bdickhaus commented 4 years ago

Does it eventually all start working? It takes a while for deCONZ to collect network information after a restart. For my network it usually takes about 5 minutes, but it could take longer. So I do not reboot my RaspBee gateways unless I really have to.

gagadget commented 4 years ago

Nothing, just one detector and one remote got back again after a short period of time.

gagadget commented 4 years ago

Deleted and Readded one device. It appears on the VNC app WITHOUT any link and does not appear in the web app ... rebooted and the new device was completely gone !

I suspect a corrupted DB, any advise ?

gagadget commented 4 years ago

Done a GW reset, reboot, add one device. Was working. Did a reboot. Not working. Giving up for now, I am going back to the Tradfri GW, can't afford to have no lights at home.

Here are the facts, if someone can help, I am ok to do more tests.

See screenshot of VNC app :

VNCapp

Here is the log in HA :

starting version 232 [11:17:45] INFO: Starting VNC server... [11:17:50] INFO: Starting the deCONZ gateway... [11:17:50] INFO: Running the deCONZ OTA updater... [11:17:50] INFO: Running the IKEA OTA updater... [11:17:50] INFO: Running Hass.io discovery task... {"uuid":"33084ef8af0b4fb1939293838f71ca6e"} [11:17:50] INFO: Success send discovery information to Home Assistant [11:17:50] INFO: deCONZ is set up and running! libEGL warning: DRI2: failed to open swrast (search paths /usr/lib/arm-linux-gnueabihf/dri:${ORIGIN}/dri:/usr/lib/dri) libEGL warning: DRI2: failed to open swrast (search paths /usr/lib/arm-linux-gnueabihf/dri:${ORIGIN}/dri:/usr/lib/dri) libpng warning: iCCP: known incorrect sRGB profile 11:17:52:730 HTTP Server listen on address 0.0.0.0, port: 40850, root: /usr/share/deCONZ/webapp/ 11:17:52:836 CTRL. 3.16.211:17:53:199 dev /dev/ttyAMA0 11:17:53:199 COM: /dev/ttyACM0 / serialno: DE2126914 11:17:53:199 COM: --dev: /dev/ttyACM0 (ConBee II) 11:17:53:199 ZCLDB init file /data/.local/share/dresden-elektronik/deCONZ/zcldb.txt 11:17:53:807 parent process bash 11:17:53:807 gw run mode: docker/hassio 11:17:53:807 GW sd-card image version file does not exist: /data/.local/share/dresden-elektronik/deCONZ/gw-version 11:17:53:807 sd-card cid: 035344535033324780ffffffff01329d 11:17:53:849 DB sqlite version 3.16.2 11:17:53:853 DB PRAGMA page_count: 30 11:17:53:853 DB PRAGMA page_size: 4096 11:17:53:853 DB PRAGMA freelist_count: 0 11:17:53:853 DB file size 122880 bytes, free pages 0 11:17:53:853 DB PRAGMA user_version: 6 11:17:53:853 DB cleanup 11:17:53:856 DB create temporary views 11:17:53:973 don't close database yet, keep open for 900 seconds 11:17:53:974 started websocket server at port 40860 11:17:53:995 found node plugin: libde_rest_plugin.so - REST API Plugin 11:17:54:002 found node plugin: libde_signal_plugin.so - Signal Monitor Plugin 11:17:58:435 found node plugin: libstd_otau_plugin.so - STD OTAU Plugin 11:17:58:466 dev /dev/ttyAMA0 11:17:58:466 COM: /dev/ttyACM0 / serialno: DE2126914 11:17:58:466 COM: --dev: /dev/ttyACM0 (ConBee II) PROTO: CRC error PROTO: CRC error 11:17:58:548 DEV config changed event 11:17:58:931 Device firmware version 0x264A0700 11:17:58:943 unlocked max nodes: 200 11:17:59:024 Device protocol version: 0x010B 11:17:59:037 new node - ext: 0x00212effff04f365, nwk: 0x0000 11:17:59:098 don't close database yet, keep open for 900 seconds 11:17:59:099 LightNode 1: Escalier added 11:17:59:278 dev /dev/ttyAMA0 11:17:59:283 GW update firmware found: /usr/share/deCONZ/firmware/deCONZ_ConBeeII_0x26490700.bin.GCF 11:17:59:283 GW firmware version: 0x264a0700 11:17:59:283 GW firmware version is up to date: 0x264a0700 11:17:59:287 Current channel 11 11:17:59:305 CTRL ANT_CTRL 0x03 11:17:59:336 Device protocol version: 0x010B 11:17:59:395 Current channel 11 11:17:59:414 CTRL ANT_CTRL 0x03 11:17:59:453 Device protocol version: 0x010B 11:17:59:515 Current channel 11 11:17:59:536 CTRL ANT_CTRL 0x03 11:17:59:565 Device protocol version: 0x010B 11:17:59:627 Current channel 11 11:17:59:647 CTRL ANT_CTRL 0x03 11:17:59:966 Announced to internet 11:17:59:966 discovery server date: Tue, 22 Oct 2019 09:17:59 GMT 11:17:59:966 local time seems to be ok 11:17:59:966 discovery found version 2.04.35 for update channel stable 11:18:04:363 don't close database yet, keep open for 900 seconds 11:18:08:303 New websocket 192.168.1.98:63856 (state: 3) 11:18:09:375 saved node state in 0 ms 11:18:09:587 sync() in 211 ms 11:18:30:584 0x086BD7FFFE55CA3A error APSDE-DATA.confirm: 0xD0 on task 11:18:37:152 scan finished 11:18:41:489 0x086BD7FFFE55CA3A error APSDE-DATA.confirm: 0xD0 on task 11:18:52:372 Current channel 11 11:18:52:379 Device TTL 3495 s flags: 0x7 11:18:52:386 0x086BD7FFFE55CA3A error APSDE-DATA.confirm: 0xD0 on task 11:19:03:184 0x086BD7FFFE55CA3A error APSDE-DATA.confirm: 0xD0 on task 11:19:14:078 0x086BD7FFFE55CA3A error APSDE-DATA.confirm: 0xD0 on task 11:19:14:078 max transmit errors for node 0x086BD7FFFE55CA3A, last seen by neighbors 75 s 11:19:14:151 0x5495 seems to be a zombie recv errors 6 11:19:14:151 LightNode removed 0x086bd7fffe55ca3a 11:19:14:151 Node zombie state changed 0x086bd7fffe55ca3a 11:19:52:368 Current channel 11 11:19:52:377 Device TTL 3435 s flags: 0x7 11:20:52:375 Current channel 11 11:20:52:386 Device TTL 3375 s flags: 0x7 11:21:52:362 Current channel 11 11:21:52:371 Device TTL 3315 s flags: 0x7 11:21:59:373 dev /dev/ttyAMA0 11:21:59:373 GW firmware version: 0x264a0700 11:21:59:373 GW firmware version is up to date: 0x264a0700 11:22:25:855 Remove websocket 192.168.1.98:63856 after error Unknown error 11:22:26:992 New websocket 192.168.1.98:63914 (state: 3) 11:22:52:368 Current channel 11 11:22:52:377 Device TTL 3255 s flags: 0x7 11:23:52:368 Current channel 11

Here is the DB export :

raspbee_gateway_config_2019-10-22.dat.zip

Hammertime84 commented 4 years ago

I think I am experiencing the same problem.. I have only 3 Xiami Aqara/Mijia Sensors. After a reboot of Hass.io, all sensors are grayed out in the Phoscon Web app.

Via the VNC viewer all sensors are gone, except for the temperature/humidity sensor. But this sensor is not connected (green line) to the coordinator (Conbee)

Is this problem introduced recently for you? Because I just bought the Conbee II last week..

Smanar commented 4 years ago

Have you try to take a look in the hidden configuration > https://github.com/dresden-elektronik/deconz-rest-plugin/issues/1849#issuecomment-530287817 to check if something happen to your network.

But it like you have connection problem, have you a USB DDR, a metal case, ... ?

gagadget commented 4 years ago

Yes I did check the hidden conf and it was showing only one line.

Hammertime84 commented 4 years ago

Yes I did check the hidden conf and it was showing only one line.

Same here, just one line. After updating HA to 100.3 and HassOS to 2.12 still the same problem. I did find another problem what is possibly related. In the Hassio logging I've found the line:

ERROR (MainThread) [hassio.supervisor] Supervisor does not run in Privileged mode. Hassio runs with limited functionality!

Maybe this causes problems.. I will start to reinstall HASSIO on another RPI. I will let you know the results..

Hammertime84 commented 4 years ago

Ok, reinstalled Hassio from scratch, integrated 3 sensors, rebooted HASSIO from VNC I only see the temp/humidity sensor. Motion and door/window sensor are gone.

Smanar commented 4 years ago

Deconz is one thing HASS is another, it's not the same thing deconz can work even HASS is broken. How much time are you waiting after pairing and before reboot ? I noticied that xiaomi sensor have curious reactions during the firsts hours after pairing. And xiaomi devices leave the network if the gateway or their parent are off, it's normal but it takes hours ...

But you have too connection problem, when you have the green line in deconz, try enabling LQI to have some values (on top, under the menu). It's not reliable but can indicate problems.

schwatter commented 4 years ago

The same for me. I have only 4 lights (tint). They work rocksolid till i reboot my pi3 with conbee stick. Solution is always, light switch off, wait some seconds and light switch on. After that, all lights up and working immediately. All updates done before.

Smanar commented 4 years ago

BTW when you have made your last tries, have you used too a direct connexion with a remote or a switch ? cf https://github.com/dresden-elektronik/deconz-rest-plugin/issues/1261#issuecomment-545543147

And power cycle do nothing for you ?

Hammertime84 commented 4 years ago

And xiaomi devices leave the network if the gateway or their parent are off, it's normal but it takes hours ...

That was it. I was too impatient. The motion and door/window sensor took their time to be visible. Thank you!

gagadget commented 4 years ago

Back to the original issue with the POWERED tradfri light, any advise ?

Should I try again with a "debug mode on" ( which I don't know if it exists ) ?

Smanar commented 4 years ago

@gagadget Yes there is log, but I m not able to understand them. If you don't have support here, try direclty the support. But it seem you are not alone with this kind of bug.

BTW can you make a try just with bulbs, without remote and switch pls ?

dandrzejewski commented 4 years ago

If you "reset" the devices by selecting them in deCONZ and pressing "F5" - does that restore your connectivity? That seems to help for me. Does not fix the actual problem, but at least it's a quick way to get things back. It doesn't look like there's an API for that function.

TheWizz commented 4 years ago

We've had the same problem for several months. After restarting the Deconz server, all lights fail to control. A power cycle of the lights fixes it. Note that even though Deconz can't control the lights, a Philips remote can, so there seems to be nowthing wrong with the light or the zigbee network.

Mimiix commented 4 years ago

It seems this issue is resolved or otherwise inactive. If it is not, please re-open!