Correct me if I’m wrong here…
Watchdog watching an IP address will reboot if the IP address is not pingable after 6 consecutive pings.
The question is, what does the pinging? is it the active routeros kernel, the watchdog daemon or is it the hardware watchdog?
If I’m reading this right, then there is a hardware component to watchdog when running routeros on a routerboard. This hardware component is able to ping an IP address so theoretically even when a box locks up, the hardware watchdog will be able to restart the board… This includes kernel panics and
To have a component INDEPENDENT of the running kernel which is able to actually ping and take action is a very cool feature.
Kudos to routerboard. This is great news, I was always under the impression that it did not or could not run independent of the OS/kernel and still intelligently control the hardware.
it doesnt seem to work though, we have a random problem of locking up RB333s right now, i have ping watchdog enabled, but the box hangs up, needs power cycle to restore it. if ping watchdog was in hardware, it would always be up or rebooting, these RB333s are totally hung, dead OS, no console response.
Hmm what version are you running. we are running ROS 3.4 on a MT 333. IT has a similar problem I think it is related to the watchdog as well. Now in 3.5 they mention a pinging bug with an address of 123.123.12.12 Does that mean any where in the address if a 12 exists than this could be a problem? Well we changed our watchdog address and cleared up some other problems on a remote BH.. So tonight will be the night to test the 123.123.12.12 address theory. We were pinging an address of 192.168.121.253 (to me that looks like two consecutive 12’s) Any ways I changed my watchdog address that it was pinging and so far all is well. How this bug has occurred I may never know, but it was really frustrating because it was carrying a lot of bandwidth on our DS3 to our clients. Hopefully this is resolved if so I am not changed ROS for this RB333.