Hi,

I'm running a NEO2 with Ubuntu at a remote location, so watchdog would be very important to have. Didn't use it before, but afaik H5 SoC supports hardware watchdog. 

How should I turn it on and do I need to install and configure anything?

 

I read about H3, but not sure about H5.

 

Thanks!

So not exactly an answer as you are running ubuntu but the armbian image for the neo2 does have /dev/watchdog and it can trigger a reboot. 

I found a page describing basic watchdog stuff and cat >> /dev/watchdog followed by a couple of carriage return and waiting = reboot which indicates its reachable and usable.

Thanks. I'm running Armbian Ubuntu image. 

But how do I activete and configure watchdog? 

I too am running H5 based board and want to get hw WD functions as well as others working. I have Nano PI NEO 2 ,  NEO plus2, and NEO Core2. I need to be able to use the WD features on all of these boards.

 

I have read all in the fora here as well as what was linked with particular attention to this

and have tried to implement the example provided by gnasch but when I do, I get this:

#  ps aux |grep watch
root        44  0.0  0.0      0     0 ?        S    22:39   0:00 [watchdogd]
root      1773  0.0  0.1   2824   648 ttyS0    S+   22:41   0:00 grep watch

Honestly, I haven't figured out what it means but it differs from what gnasch posted in particular, nothing for each core.  I also tried:

# systemctl status watchdog
* watchdog.service - LSB: Start software watchdog daemon
   Loaded: loaded (/etc/init.d/watchdog; generated)
   Active: active (exited) since Sat 2019-12-28 22:39:33 EST; 1h 10min ago
     Docs: man:systemd-sysv-generator(8)
  Process: 659 ExecStart=/etc/init.d/watchdog start (code=exited, status=0/SUCCE

Dec 28 22:39:33 Millheim systemd[1]: Starting LSB: Start software watchdog daemo
Dec 28 22:39:33 Millheim systemd[1]: Started LSB: Start software watchdog daemon
lines 1-8/8 (END)

 

Also, I can do:

#echo 1 > /dev/watchdog

and the system does reboot. This was suggested earlier in the thread but did not indicate that the reaction was supposed to be a reboot 

 

However, I don't understand what is being watched! In particular, I don't know how to verify the HW WD is working. There is a watchdogd process running which is no surprise since it is possible to cause a reboot as above.   I do see /dev/watchdog and /dev/watchdog0 but do not know what exactly is going on here. I think I understand that this is associated in some way with the HW WD in some sense but why two?  I see the "watchdog-device = /dev/watchdog" in watchdog.cfg but do I need an additional line that says "watchdog-device = /dev/watchdog0" ?

 

I think I understand how to test cpu loading but the question arises, how are multiple cores handled? Is it possible for one of 4 cores to get hung up with a runaway process or some such? This would seem (in part) to be the reason for gnasch's comment:

Quote

i have 4 watchdog processes, one per processor.

As well as the reason I think I want to see something for each core. 

 

Any additional insights would be appreciated.

 

Regards to all,

 

Q