Solved ? Examine the system, search for shutdown reason

Maturin

Maturin

Hi guys!

Yesterday evening I was playing HOMM3 on WinXP within VirtualBox (ose-6.1.50_4), using 14.1-RELEASE-p5.
Suddenly my monitors went black, then the machine shutdown, while the shutdown-process also seemed somehow strangely shorten.
I know playing on WinXP was anything but rock solid, but as far as I understand VM a crash within should not effect the machine outside the box, right? So the reason for the sudden shutdown does not necessarily has to come from VB.
I checked all updated entries in /var/log/,

/var/log/messages shows
when I started VB, and then the reboot after the crash again, but nothing about shutdown/reboot:

Code: 
[...]
Jan 12 15:39:31 THEIA kernel: vboxdrv: XXXXXXXXXXXXXXXX VMMR0.r0
Jan 12 15:39:31 THEIA kernel: vboxdrv: XXXXXXXXXXXXXXXX VBoxDDR0.r0
Jan 12 15:39:35 THEIA kernel: VMMR0InitVM: eflags=40246 fKernelFeatures=0x2 (SUPKERNELFEATURES_SMAP=1)
Jan 12 19:08:23 THEIA kernel: vboxdrv: XXXXXXXXXXXXXXXX VMMR0.r0
Jan 12 19:08:23 THEIA kernel: vboxdrv: XXXXXXXXXXXXXXXX VBoxDDR0.r0
Jan 12 19:08:27 THEIA kernel: VMMR0InitVM: eflags=40246 fKernelFeatures=0x2 (SUPKERNELFEATURES_SMAP=1)
Jan 12 19:47:29 THEIA syslogd: kernel boot file is /boot/kernel/kernel
Jan 12 19:47:29 THEIA kernel: ---<<BOOT>>---
Jan 12 19:47:29 THEIA kernel: Copyright (c) 1992-2023 The FreeBSD Project.
[...]

also looked a bit with last(1), which's one of the tools you might get, when you search the net, and recieve global answers, [also I'm always a bit paranoid being hacked], which showed me some entries 'crash' - 'oh-ha, that's new to me', set rc_debug="YES" in /etc/rc.conf, which of course produces an elaborated boot-log, but wasn't a real help to figure out what caused the shutdown before.
I like to have your advices.
Where else to look for, are there other logs then in /var/log/, particular shutdown logs?
Also the 'crashes' last(1) reported have no time-stamps fitting the shutdown, but I like to examine them anyway.
How to read/examine core dumps?
Any links to documentation on how I can learn to examine my system?

I like to find out, what happened.
Under Windows it was quite normal if the computer suddenly crashes.
But running an else rock solid FreeBSD makes me feel uncomfortable if it crashes, and I don't know what's going on.

Thanks in advance.
 
I also did a rkhunt over my system.
rkhunter(8) did not found any RK, but produced a lot of file property warnings within my
bin/ directories.

Didn't see anything conspicious within my Xorg logs.
 
Maturin said:
but as far as I understand VM a crash within should not effect the machine outside the box, right?
Click to expand...
That depends. If you're using vboxtools in theory there is a way to crash host (bug in vbox driver initiated by a guest).

VM should have its own flight log too, usually under <vmdir>/]Logs/VBox.log. Note they do get rotated fast. Have a look around the crash timestamp for possible clues.
As suggested above - make sure you are set for host crash (check the handbook/forums how to do it), test it too.

Try to reproduce the issue.
Make sure there's no HW issue (possible PSU issue).
 
cracauer@ said:
Likely hardware reset from overheating over power overload.
Click to expand...
_martin said:
(possible PSU issue)
Click to expand...
My system was hardly at its limits at this time; motherboard showed ~40°C CPU temp. Nothing else was running at that moment (no browser, no shell, nothing) except this only instance of VB with an almost vanilla XP, and an ancient 2D game - not a real load for my AMD Ryzen 9 5900X.

A reaction of my PSU to some powerloss in the net was my first idea, too, 'cause it looked that way.
But most powerlosses I experienced are complete outages: clack - all out, and there were no other recognizable powerlosses else; e.g. my desktop lamp didn't fluctuate.
But okay, after I gave it a thought:
Today's LED lamps may not flicker by a small power blip. Here in the rural mountains powerline fluctuations are not rare, especially when there's snow. And if my PSU is kind of sensitive, it may caused a shutdown because of a light fluctuation I did not recognized else.

If the system is going down by pressing e.g. the power button, and this causes no log entry, to me the case was settled.
I will test it.

Thanks.

Edit:
Of course there is no 'poweroff' entry in /var/log/message when you kill the system by pressing the powerbutton.
On the other hand I have no idea how to simulate not a complete outage but a powernet's voltagedrop (with my equipment) to examine if my PSU is shutting down in that case (one'd need some electronical load of about ~4..6kW to be switched within ~1/10th of a second to the circuit... - no toy at all.)
If there's someone expert on PSUs could give an answer I'd apreciate it.
Those are things you cannot learn from any documentation.

Insofar I set this thread to 'solved' and thank everybody who read it, especially to all answered:
Thank you, guys!
 
You must log in or register to reply here.
Back
Top