Final 1.0.0 errors and crash report....
-
I use mycontroller with HTTP and not HTTP and with esp8266 mqtt-gateway .. no problem so far with Final 1.0.0
-
Using the OrangePI Zero, nrf24l01 is directly attached to GPIO for a couple of months now. I have not seen any of these issues.
@skywatch, are you using the latest version of mysgw?
if so, what is the output ofmysgw -d
if the above command does not provide any output that shows it connected the radio successfully, check the wiring, another thing i can imagine is that your radio is dead...
-
I realized that i was still using the previous release.... just upgraded to Final-1.0.0.. will keep you posted
-
pi@HAMASS:~ $ mysgw -d mysgw: Starting gateway... mysgw: Protocol version - 2.2.0-beta mysgw: Couldn't create a symlink '/dev/ttyMySensorsGateway' to PTY! (17) File exists mysgw: Failed to open serial port.
Yet it worked for more than a week until the crash....Replaced radio today, still the same....
-
Strange can you post the output of
ls -la /dev/ttyMySensorsGateway
It should show something like this:
root@orangepizero:~# ls -la /dev/ttyMySensorsGateway lrwxrwxrwx 1 root root 10 Oct 7 13:11 /dev/ttyMySensorsGateway -> /dev/pts/0
if not just stop mysgw, and delete the file. mysgw will create a new once mysgw starts.
-
here it is....
ls -la /dev/ttyMySensorsGateway lrwxrwxrwx 1 root root 10 Oct 9 21:19 /dev/ttyMySensorsGateway -> /dev/pts/1
I changed the radio module this evening and rebooted, no change at all.....Puzzled am I.....
-
@skywatch I believe you might have done these steps already. If not, kindly check it.
Have you attached capacitor for your NRF24L01+ radio? It works on 3.3 Volts. Kindly check all the details once again from https://www.mysensors.org/build/raspberry Seems the original repo deprecated https://github.com/mysensors/Raspberry
-
Weird, if you kill mysgw, is the device file removed?
(will dig into this a little later today) -
So yes capacitor on nrf and 3.28v seen across it, so no problem there. The radio has a single 2x4 connector so all leads stay in position (much quicker)
Anyway, today brings a new mystery for me as I did the following.....
sudo systemctl daemon-reload
sudo systemctl restart mysgw.service
sudo systemctl stop mysgw.service
sudo systemctl start mysgw.serviceThen to see what was going on....
tail -f /var/log/syslogThis showed data, so I connected to the web server on the pi and graphs were showing of data received since last night.... It looks like it's all back to normal, except it isn't as I still get exactly the same error message from mysgw -d.
Weirder still!
-
I have tried your commands:
sudo systemctl restart mysgw.service sudo systemctl stop mysgw.service sudo systemctl start mysgw.service
This is my output:
root@orangepizero:/var/log# tail -f syslog Oct 10 22:25:09 localhost systemd[1]: Started MySensors Gateway daemon. Oct 10 22:25:09 localhost mysgw: Starting gateway... Oct 10 22:25:09 localhost mysgw: Protocol version - 2.2.0-beta Oct 10 22:25:39 localhost systemd[1]: Stopping MySensors Gateway daemon... Oct 10 22:25:39 localhost mysgw: Received SIGTERM Oct 10 22:25:39 localhost systemd[1]: Stopped MySensors Gateway daemon. Oct 10 22:25:56 localhost systemd[1]: Starting MySensors Gateway daemon... Oct 10 22:25:56 localhost systemd[1]: Started MySensors Gateway daemon. Oct 10 22:25:56 localhost mysgw: Starting gateway... Oct 10 22:25:56 localhost mysgw: Protocol version - 2.2.0-beta
Does this match your output?
looking at the /dev/ device directory, the device file /dev/ttyMySensorsGateway is gone once stopped. and recreated when the mysgw daemon started again...
(i also had to re-enable the GW in mycontroller...since it got disabled...)
-
the link seems okay. now i am wondering what is behind the link...
root@orangepizero:~/MySensors/bin# ls -l /dev/ttyMySensorsGateway lrwxrwxrwx 1 root root 10 Oct 12 18:45 /dev/ttyMySensorsGateway -> /dev/pts/2 root@orangepizero:~/MySensors/bin# ll /dev/pts/2 crw--w---- 1 root tty 136, 2 Oct 12 18:46 /dev/pts/2
if you do an
ls -l /dev/pts/1
what does that look like?
did you use the correct flags to build mysgw? (as described here)
-
@Tag
thanks for the support but the mystery is solved (at least this bit) - I rebooted the pi before giving the mysgw -d and of course mysgw was already started at boot and so was trying to start a second session and that is why the strange messages happened.I have the system working again now and hope that it will stay that way for many weeks! ......
-
@skywatch said in Final 1.0.0 errors and crash report....:
but the mystery is solved (at least this
ah okay!! good!,
It will stay runninggood luck!
-
It didn't stay running - It crashed again.
Nothing in the myc log.
The MYSGW log shows data still being received from the sensor nodes.Now I am stumped.....really stumped, as to how to get this sorted out......
-
Sad to hear.....
Have you tried setting up an arduino with an NRF21L01 attached to the arduino?,
Arduino can be connected to USB and the gateway can be pointed to i.e. /dev/ttyUSB0 -
Sorry for the delay,
I am trying some other solution now and will report back if successful.
-
I tried powering the rf module via the 5V pi rail via a 3.3V regulator just incase the 3.3v from the pi wasn't very clean. Still the same.......
So...... Today I have made a promini (3V, 8MHz) gateway adn connected to the pi via serial connecting.
I got Final-1.1.0-SNAPSHOT and installed that fresh and reconfigured it all from scratch. So far all sensors bar one have come up and behaving as they should. So now it is a waiting game (with a little investigation as to why one node is not showing - it seems to be registering as node0 when it should be node1 - I may reflash it and force the node id to see if that has an effect).
It will be annoying if it is a broken pi gw from mysensors, but not everything works as you expect it to I have discovered.
Cheers!
-
Okay great!,
About the node, it might be that it has a value in the eeprom where the node id is stored..
A sketch exists wich "clears" the eeprom (it is in the examples) , maybe upload that sketch, reset the promini and re-upload your original sketch (without the forced node id...)see if that will fix the node id issue
-
Thanks!
Somehow it sorted itself out while I was tweaking my dashboards and now seems to working as expected (still some 'extra' sensors on that node showing in the topology page - but I can live with that).
So now onto some bootloader testing and low power stuff for battery sensors. Then onto signing and automation - you can bet I'll be back with more requests for help soon!
-
keep building!!
(maybe nice to explain a little on your setup in this topic, it may spark inspiration for the mycontroller/mysensors community ;))