Is there any method to debug the watch dog restart problem?

[saccleo]

I can not connect to the router by ssh and webgui. There are a lot of messages like this"check_watchdog: [check_watchdog] restart watchdog for no heartbeat"
This problem occured accidentally, and can not get direct reason from the system log.
Any tools for tracking this?

 

[L&LD]

What router? What firmware? What recent/current changes have you made?

What device(s) are you using to access the router? How are you seeing the messages?

What happened, accidentally?

Can you attach or link to the system log for others to have a look?

 

[dave14305]

The immediate cause is that the router does not detect a recent update to the file /tmp/watchdog_heartbeat so it assumes watchdog is dead. The issue is to understand why watchdog isn't updating that file regularly. Maybe it died and therefore cannot update it, which is the most likely scenario. If you don't have any further syslogs, and cannot access the Admin interface via web or SSH, you're rather limited in troubleshooting besides a factory default reset (who do I sound like?).

But if you cannot connect to the admin interface, how do you know those messages are occurring? A separate syslog server?

 

[saccleo]

My model is rt-ac66u_b1, and the firmware is 386.1-2. I upgrade to this version with a factory reset(formate jffs and reset to defautl). I did nothing recently.
when this happened, what i can do is restart the router, so i can see the log.
i will upload syslog later.