Hello,
i installed the watchdog service and tested it with a Shell Bomb, it restarts the Pi.
Service:
Dez 17 00:28:50 rpi4 watchdog[859]: interface: no interface to check
Dez 17 00:28:50 rpi4 watchdog[859]: temperature: no sensors to check
Dez 17 00:28:50 rpi4 watchdog[859]: no test binary files
Dez 17 00:28:50 rpi4 watchdog[859]: no repair binary files
Dez 17 00:28:50 rpi4 watchdog[859]: error retry time-out = 60 seconds
Dez 17 00:28:50 rpi4 watchdog[859]: repair attempts = 1
Dez 17 00:28:50 rpi4 watchdog[859]: alive=/dev/watchdog heartbeat=[none] to=root no_act=no force=no
Dez 17 00:28:50 rpi4 watchdog[859]: watchdog now set to 14 seconds
Dez 17 00:28:50 rpi4 watchdog[859]: hardware watchdog identity: Broadcom BCM2835 Watchdog timer
/etc/watchdog.conf
max-load-1= 180
max-load-5= 130
max-load-15= 80
min-memory= 1
allocatable-memory= 1
watchdog-device= /dev/watchdog
watchdog-timeout = 14
realtime= yes
priority= 1
But in a real case use the watchdog dont work.
Thats of course not user Software problems that makes just the RPI run out of memory thats
Kernel/OS Bugs. Like:
Nov 9 02:33:03 rpi4 kernel: [31409226.683098] rcu: INFO: rcu_sched self-detected stall on CPU
Today i have no log, the "/var/log/lastlog" file contains binary most 0x00 the only that is
readable is "gtty1". The Network LED of RPI4 and on the Network switch was blinking all the time,
but it transfers to nowhere(no Target, no Broadcast) cause no other Switch LED blinks.
Red&Green LED on RPI was active. No access to VNC like always. Have to unplug power.
Usual the Watchdog service should send signals to the CPU an if this signal stops the CPU
does a Reboot. Means the Watchdog Service should either running(but other Software dont like
vnc or my own Software that too trys to reboot on a error ret=system ("reboot");) or is the CPU is crashed inself.
Can i do anything against this or do i need to build a external Hardware Watchdog?
i installed the watchdog service and tested it with a Shell Bomb, it restarts the Pi.
Service:
Dez 17 00:28:50 rpi4 watchdog[859]: interface: no interface to check
Dez 17 00:28:50 rpi4 watchdog[859]: temperature: no sensors to check
Dez 17 00:28:50 rpi4 watchdog[859]: no test binary files
Dez 17 00:28:50 rpi4 watchdog[859]: no repair binary files
Dez 17 00:28:50 rpi4 watchdog[859]: error retry time-out = 60 seconds
Dez 17 00:28:50 rpi4 watchdog[859]: repair attempts = 1
Dez 17 00:28:50 rpi4 watchdog[859]: alive=/dev/watchdog heartbeat=[none] to=root no_act=no force=no
Dez 17 00:28:50 rpi4 watchdog[859]: watchdog now set to 14 seconds
Dez 17 00:28:50 rpi4 watchdog[859]: hardware watchdog identity: Broadcom BCM2835 Watchdog timer
/etc/watchdog.conf
max-load-1= 180
max-load-5= 130
max-load-15= 80
min-memory= 1
allocatable-memory= 1
watchdog-device= /dev/watchdog
watchdog-timeout = 14
realtime= yes
priority= 1
But in a real case use the watchdog dont work.
Thats of course not user Software problems that makes just the RPI run out of memory thats
Kernel/OS Bugs. Like:
Nov 9 02:33:03 rpi4 kernel: [31409226.683098] rcu: INFO: rcu_sched self-detected stall on CPU
Today i have no log, the "/var/log/lastlog" file contains binary most 0x00 the only that is
readable is "gtty1". The Network LED of RPI4 and on the Network switch was blinking all the time,
but it transfers to nowhere(no Target, no Broadcast) cause no other Switch LED blinks.
Red&Green LED on RPI was active. No access to VNC like always. Have to unplug power.
Usual the Watchdog service should send signals to the CPU an if this signal stops the CPU
does a Reboot. Means the Watchdog Service should either running(but other Software dont like
vnc or my own Software that too trys to reboot on a error ret=system ("reboot");) or is the CPU is crashed inself.
Can i do anything against this or do i need to build a external Hardware Watchdog?
Statistics: Posted by user12345 — Tue Dec 17, 2024 10:15 am — Replies 0 — Views 24