Upgrade to v1.10.14: Can not send signals from coordinator #2351
Comments
|
Please do the following:
Note: do not run the adapter in debug, info is fine As additional information, provide output from Once this is done, update the firmware on the cc2538 to the newest version - yours is over 3 yrs old, and repeat the exact same steps. post the results here. As text, in code tags. A. |
|
Hey asgothian, the firmware seems to be the latest for this coordinator combo. I did not found a newer firmware for the cc2538 device. Output log with debugmessages for a osram plug: Output console: mod edit - proper code tags, so stuff gets readable |
|
Please run another test:
A. |
|
with the 2nd instance I can switch the Osram plug without any problems. Pairing new devices also works again. Do you also need log outputs here? |
|
No. But another test:
A. |
|
I got this error: |
|
I started again with the nvbackup.json and then the error message no longer appeared. Then I stopped it again and renamed the nvbackup.json. After that it started normally again and also found all 57 devices at startup and all of them were also available from the status. But control was still not possible :-( |
|
The network map is still not set after 30min. |
|
That's to be expected. Can you please do another test:
|
|
Starts as before: |
|
With or without the ability to control the devices ? And before you said that if you delete the nvbackup.json the adapter doesn't start immediately ? A. |
|
the adapter did not start without the nvbackup.json when switching from the 2nd instance to the first instance. Once the stick had the correct settings, it always started without any problems as shown in the last log. However, all devices are also directly available again. |
|
but control devices from coordinator did not work.... |
|
also pairing or deleting devices did not work |
|
I am not sure I follow you. Did you make that additional test I asked for ? Disconnecting the stick from the system to reset it, and then try to start it without any nvbackup.json available ? If that does not work, and you get the issue of being unable to do the network provisioning, please repeat the test, but disconnect the antenna of the stick (or wrap it in tin foil) to ensure that it does not see the current network. In this case, the adapter should start up without requiring any parameter changes. A |
|
"Disconnecting the stick from the system to reset it, and then try to start it without any nvbackup.json available ?" I disconnected the stick from the system, but i guess that the device list ist stored on the stick? |
|
It should not be, if the stick looses power. That's what the shepherd.db and nvbackup.json files are for. Removing the shepherd.db effectively removes the devices, so that is not a good idea, but the nvbackup.json file used to not be present, so the question is if we can get your adapter running without having one which does not seem to match in some specific information. Also - can you provide me with the 2 nvbackup.json files - the one from the test-instance where you can control the device and the one from your main instance where you cannot ? Maybe I can find something and patch it A. |
|
I have now done it again without the antenna and the log looks the same. |
|
if you are up for it, I am on the iobroker discord - channel 2useronly. It may be easier if we speak and/or use screen sharing in discord to run the tests and such. A bit of context:
What makes me consider that there is something off in this file is the fact that the identical hard/software combination will work when set up as new instance, but not operate your old instance. So there has to be something in the nvbackup.json file which causes this. A |
|
i printed the diff between both files: |
|
what are those numbers 7c7, 21c21 and 23c23. I assume the x's are edited in to hide some data from your side ? if so, can you please verify that the data matches the entries in your configuration. Also - do you have access to an nvbackup.json from before you updated from 1.10.13 ? A. |
|
yes i hidded the pan_id, extended_pan_id and tclk_seed. 7c7 is the line 7 of both files. ( it matched with the line :-) ) |
|
OK.. please stop the adapter, patch the nvbackup.json to say "is_child": false and then restart the adapter A. |
|
unfortunately no improvement :-( |
|
i compare it also with a old backup file. There are some changes on the devices, but the main settings are the same. |
|
did you soft reset the coordinator before the whole thing ? (unplug it, wait 20 seconds, plug it back in) ? |
|
no, in this case I let the coordinator sticked ..... Sorry |
|
what are the changes on the devices ? |
|
I removed some devices and a nwk_adress changed: |
|
also with soft reset the same behaviour |
|
is there any way to reset the zigee.0 instance and pair all devices without losing the objects settings in iobroker? |
|
or I can also buy a sonoff3 zigbee stick with TI Chipset and replace my cc2538. |
|
Yes, just delete the nvbackup.json and shepherd.db and then pair the devices again. It will seem as if all devices are still there, but they are not on the zigbee network |
|
but will this also reset the adapter behaviour, that i can´t pair actually? |
|
in my secound instance the pairing works as expected, but in the first instance the counter (60sec) did not appear. |
Yes, it should. You may need to modify the pan-ID / extPanID in order to get it to start without trying to reconnect to the existing network, as you cannot reset the devices up front A. |
|
Hey asgothian, |
describe the problem in Detail. What do you expect to happen ? What does happen ?
After upgrading to the latest version, my coordinator (cc2538) can not control any zigbee device. Receiving zigbee signals like temperature, current, etc works, but i can not switch any light or power socket. Testet with Osram, Blitzwolf, Tuya, Philips hue etc. devices.
Adapter version
1.10.14
Coordinator firmware version
20211222
Coordinator
cc2538+cc2592
Setup
Raspi4 8G
Protocol of the error (as text, encapsulated in multiline code tags <>). Separate different occurrances into clearly split log sections
Mod Edit - code tags. I wish people would read the instructions and follow them
The text was updated successfully, but these errors were encountered: