[SOLVED] Interesting observation Bullseye

[ticojohn]

I have upgraded to Bullseye on two systems, a NUC5i5ryh with M.2 NVME SSD and a Gigabyte MB with a HDD. I have seen a different behavior on the two systems when I reboot. On the NUC I see a watchdog message when the reboot occurs but do not see that message on the Gigabyte board with HDD. I have to assume the the reboot procedure is identical on the two systems so i am wondering, is this due to some hardware thing on the NUC or is it because the HDD is so much slower than the SSD on the NUC?

I have looked at journalctl but obviously see nothing as the message likely occurs AFTER the drive is unmounted, so no messages can be written to the logs. And I have seen some board messages, I think on Reddit, that seem to indicate that this message is normal. Weird, as I've never seen the message with Buster.

Just to be clear, I am NOT concerned since the message does not seem to cause any issues. I'm GUESSING that some change in systemd is the culprit and will likely be resolved if it is an issue. I'm just inquisitive.

Thanks.

[Bulkley]

Inquisitive is good.

My guess is it's an hardware issue but, as you say, it probably isn't important.

[Head_on_a_Stick]

http://0pointer.de/blog/projects/watchdog.html

[ticojohn]

http://0pointer.de/blog/projects/watchdog.html

Interesting read. Thanks.

[ticojohn]

Inquisitive is good.

My guess is it's an hardware issue but, as you say, it probably isn't important.

I would tend to agree with your assessment. Just thought is was interesting that I never noticed the message before upgrading to bullseye.
I won't worry about it unless the system starts doing something undesirable, like hanging during reboot. Thanks for your feedback.

I'll mark this post as [SOLVED].

[ticojohn]

Just wanted to add this one last observation. I found an additional bit of info. Following the link that was provided by HOAS, and if I understand what I read, it would appear that the NUC has a hardware timer and the Gigabyte board does not, or is not being used.
For the Gigabyte board:

Code: Select all

root@debian:~# wdctl
wdctl: cannot open /dev/watchdog: No such file or directory

And for the NUC:

Code: Select all

root@debian:~# wdctl
Device:        /dev/watchdog0
Identity:      iTCO_wdt [version 0]
Timeout:       30 seconds
Pre-timeout:    0 seconds
Timeleft:      30 seconds
FLAG           DESCRIPTION               STATUS BOOT-STATUS
KEEPALIVEPING  Keep alive ping reply          1           0
MAGICCLOSE     Supports magic close char      0           0
SETTIMEOUT     Set timeout (in seconds)       0           0

Interesting !