Linux Watchdog?

Developer Discussion
1

I would like to have Linux watchdog in my Clear Linux. Is it possible?

1 Like
2

That would be nice, good idea.

3

I have been looking for some information and in bundle openstack-common | Clear Linux* Project watchdogd is added. But trying to deploy it I can’t. It is still unavailable?

4

systemd is the main system and process supervisor in Clear Linux. systemd has support for hardware and software watchdogs. Can that fit your needs?

5

I am able to launch watchdogd with systemctl, but it is necessary to use a “/dev/watchdog” to make it active/running. The issue is due to find this watchdog.

6

What I meant was that systemd has its own hardware watchdog functionality, separate from watchdogd.

You’ll need to create a /etc/systemd/system.conf.d/*.conf
See systemd-system.conf

To be honest I have never used either personally though. It looks like # CONFIG_WATCHDOG is not set in our kernel is not set. I’m guessing that is required here. @miguelinux is this something we can enable without impacting other things?

7

To be honest I have never used either personally though. It looks like # CONFIG_WATCHDOG is not set in our kernel is not set. I’m guessing that is required here. @miguelinux is this something we can enable without impacting other things?

Yes, CL disabled kernel watchdog, because the hardware watchdog will reset the system (causing a reboot) after the timeout occurs,
and sometimes is better to get the platform “alive” after an incident.

8

So It is not possible to enable kernel watchdog? Is there any way to obtain watchdog functionality?

9

In particular, I would like to implement iTCO_wdt module for Intel chipsets.

10

Well, I am trying to modify property

#CONFIG_WATCHDOG is not set to CONFIG_WATCHDOG=y

in .config file. After that I

make build

and the following error appears.

Finish: dnf install
ERROR: Exception(linux.spec) Config(clear) 0 minutes 23 seconds
INFO: Results and/or logs in: results/
ERROR: Command failed:
/usr/bin/systemd-nspawn -q -M d71b1149a2424c24bf085e25a3911213 -D /var/lib/mock/clear-linux/root -a --setenv=TERM=vt100 --setenv=SHELL=/bin/bash --setenv=HOME=/builddir --setenv=HOSTNAME=mock --setenv=PATH=/usr/bin:/bin:/usr/sbin:/sbin --setenv=PROMPT_COMMAND=printf “\033]0;\007” --setenv=PS1= \s-\v$ --setenv=LANG=en_US.UTF-8 /usr/sbin/useradd -m -u 0 -g 302 -d /builddir mockbuild
make: *** […/common/Makefile.common:117: results/linux-5.4.6-883.src.rpm] Error 4

What can I do to solve this error?

11

I created a request on GitHub to evaluate enabling CONFIG_WATCHDOG here so it doesn’t get forgotten about: enable CONFIG_WATCHDOG · Issue #1618 · clearlinux/distribution · GitHub

As far as compiling your own kernel with this change, are you following the steps here? https://docs.01.org/clearlinux/latest/guides/kernel/kernel-development.html I just tried with 5.4.6 and it worked.

12

I solved the problem related including watchdog module, using kernel modules. So now, /dev/watchdog exists.
Having it, I installed the bundle openstack-common which includes watchdogd daemon via systemctl
Before that, I included the file watchdogd.conf with the following content (I want to monitorize pinging to an IP address)

admin@clr-aa34762852fd4e658857ba0b2c6e5``5cf/etc $ cat watchdogd.conf
ping = 192.168.3.197
watchdog-device = /dev/watchdog
watchdog-timeout = 14
interval = 1

As a result, when I starts watchdogd the following error appears:

admin@clr-aa34762852fd4e658857ba0b2c6e55cf/etc $ sudo systemctl start watchdogd.service 
admin@clr-aa34762852fd4e658857ba0b2c6e55cf/etc $ sudo systemctl status watchdogd.service 
â—Ź watchdogd.service - Advanced watchdog daemon
     Loaded: loaded (/usr/lib/systemd/system/watchdogd.service; enabled; vendor preset: disabled)
    Drop-In: /etc/systemd/system/watchdogd.service.d
             └─override.conf
     Active: active (running) since Fri 2020-01-17 12:33:08 CET; 2s ago
       Docs: man:watchdogd(8)
             man:watchdogctl(1)
             man:watchdogd.conf(5)
             file:/usr/share/doc/watchdogd/README.md
             https://github.com/troglobit/watchdogd
             http://troglobit.com/project/watchdogd
   Main PID: 2714 (watchdogd)
      Tasks: 1 (limit: 4548)
     Memory: 308.0K
     CGroup: /system.slice/watchdogd.service
             └─2714 @usr/bin/watchdogd -nsx

ene 17 12:33:08 clr-aa34762852fd4e658857ba0b2c6e55cf systemd[1]: Started Advanced watchdog daemon.
ene 17 12:33:08 clr-aa34762852fd4e658857ba0b2c6e55cf watchdogd[2714]: /etc/watchdogd.conf:1: no such option 'ping'
ene 17 12:33:08 clr-aa34762852fd4e658857ba0b2c6e55cf watchdogd[2714]: Parse error in /etc/watchdogd.conf
ene 17 12:33:08 clr-aa34762852fd4e658857ba0b2c6e55cf watchdogd[2714]: watchdogd v3.3 starting ...
ene 17 12:33:08 clr-aa34762852fd4e658857ba0b2c6e55cf watchdogd[2714]: WDT does not support PWR fail condition, treating as card reset.

As a consequence, Watchdog behaviour is not expected, basically It does not work, but when I stop It, the system crashes and reboots after a few seconds.