Desktop computer self-restarting - hardware or qubes?

Using a new installation of 4.0.4 for a week now, fully updated, after a crash and self-restart left the system unbootable. The same crash and self-restart has occured three times since, fortunately without damage to the OS and I have been able to recover the openoffice documents, which were open at the time of the crash. No idea if the cause is hardware or software related. There may be a log file somewhere to offer some clues, but that’s beyond my level of knowledge. The most recent restart occured yesterday. Fortunately the system restarted back to square one and I just needed to restore the open office documents, which were open at time of crash.

When the crashes/restarts have happened, I have had a number of VMs running - typically let’s say three open office documents, Thunderbird, Firefox (dedicated to ProtonMail), two Firefoxes (one in untrusted VM and one in personal VM), a Keepass VM, and a Mullvad VM. In one of the firefox instances I have sometimes had two or three youtube videos stacked, watching one and the other one or two tabbed and waiting to be watched. So I’m not sure if I’m placing too much demand on the Qubes installation memorywise, if there is an update-related problem or if this is hardware related.

I always update my VMs the same day, that I see available updates. All VMs are debian based except for the as-installed dom0 and the various Sys VMs. The desktop system has 16GB ram, Intel i5-9600k cpu, Asus Prime B365M-A mobo, a Samsung 970 EVOPlus ssd plus 2TB drive and the usb keyboard and mouse are connected through a Tripp-Lite PS/2 to usb connector to the mobo, so I’m using the stock disposable sys-usb VM. I think, that the hardware should be sufficient to handle the load I’m putting on it, but cannot say for sure. Any ideas on how I can proceed to sort this out? Many thanks

Indeed, such crashes are not expected on your hardware (or at all!). You should post relevant logs, which you get by typing journalctl in dom0 terminal. More info: https://www.howtogeek.com/499623/how-to-use-journalctl-to-read-linux-system-logs/.

You can copy those logs into inter-vm clipboard using Qubes Clipboard widget on the panel.

There may also be logs accessible in Qube Manager by right clicking on dom0 and choosing Logs.

Thanks fsflover. Copying out of the dom0 Terminal to another VM through Clipboard widget is giving me problems. I cannot seem to get more than what’s visible in the Terminal window to copy, even if I copy to a file within dom0 and then use the qvm-copy-to-vm vm filename method. I tried the sudo journalctl -S -2d variation from the howtogeek link and got that data into the dom0 Terminal, which I can scroll through, but again I don’t seem to be able to get all of it out of the Terminal in it’s entirety. I think I need some help with that.
Thanks

Almost… System unexpectedly rebooted again, while I was watching a youtube video. Seems that watching a youtube video is a common factor. I have now isolated the latest time of difficulty to within a two minute period. I can get those two minutes of journalctl logs isolated and showing up in the dom0 Terminal with the command, "sudo journalctl -S “2021-03-19 19:21:00” -U “2021-03-19 19:23:00” This lists only those two minutes and I can scroll through them with the cursor, but how do I specifically copy all of the data to a file in another vm? I don’t seem to be able to get anything more than one Terminal window full of data copied out of the Terminal (shift-ctl-c doesn’t do it.)

You can run sensors in dom0 periodically to see if this might be related to overheating.

Then: Copying from (and to) dom0 | Qubes OS (but if possible, avoid it, since it could compromise the security of Qubes)

427F11FD0FAA4B080123: Thanks. I will keep an eye on the temps. At present they are high 20’s to low 30’s with a few VMs up and running.

fsflover: Thanks. That suggestion helped. I used:

journalctl -S “2021-03-19 19:21:00” -U “2021-03-19 19:23:00” > jrnl-ctl.log 2>&1

and then:
qvm-copy-to-vm vm name jrnl-ctl.log

jrnl-ctl.log (138.2 KB)

You could try updating the BIOS. Version 1203 is a bit old and 1605 seems latest.
https://www.asus.com/Motherboards-Components/Motherboards/PRIME/PRIME-B365M-A/HelpDesk_BIOS/

If there is still no improvement, I would suggest to run a memory test

Thanks kommuni. It just happened again. Temps running in the 30’s C. Then I started up more VMs and started watching youtube vids to test the temps. After about 30 mins the desktop system rebooted itself. I had installed the Sensor Plugin widget and had been monitoring the temps. They were ever so slightly increasing, but they must have taken a moderate upswing near the end. As soon as the system was back up and running, I checked the widget and all 6 cores were reading in the neighbourhood of 70 deg C. I will see about updating the bios. The mobo and cpu have only been installed for about 3 months now and I think I checked for a bios update, after I installed it (but not sure.) If there’s a newer one available, I will update. cheers

Bios successfully flashed. Took a while to find the virtualization settings again in bios. Asus does a pretty good job of hiding the VT-d setting in this board’s bios. Up and running and temps at 30 degs C for all cores. If there are no other suggestions in the meantime, I will monitor the temps with the Sensors Temperatures widget in the Panel and load up the system again with VM’s to see if there is a reoccurance of problems. Thanks to all, cheers

Still no joy I’m afraid. Watching the second of two three-minute videos and the video “sampled” three or four times, repeating about one or two seconds of video, then screen went blank, then power down and self-restart all without any input from me. This exact behaviour has occured at least three times now. I was watching a saved video this time and not a streaming video over firefox. This behaviour has occured with multiple VMs, so it’s definitely not limited to one VM. I’m wondering if possibly adding a video card to replace the Asus mobo’s on-board video might solve the problem. Any views? I’m at a loss. Thanks

Try a newer kernel. You can install it by entering below command in dom0 terminal, and then reboot.

sudo qubes-dom0-update kernel-latest

I’m wondering if possibly adding a video card to replace the Asus mobo’s on-board video might solve the problem.

Indeed, it makes sense. Or you could try 4.1.

1 Like

Thanks kommuni. I have just updated my 5.4.98-1 kernel to the latest 5.11.4-1 and will try viewing some videos again today, to see if problems recur as a first step. I looked up forum posts on 4.1 as well, as I had been waiting for the official release, before deciding when to install it. Does it seem stable enough yet to use as my day to day OS and is it more stable video-wise than 4.0.4? My 4.0.4 is always fully up to date with qubes-suggested updates. My mobo is using the onboard Intel graphics, but I do see some reference to getting nvidia to work in the 4.1 posts. cheers

Just my suggestion, I try to always download the YouTube videos instead of viewing online.
Other folks are curious as to why you are running Qubes.
Are you trying to view using TOR?
Also, I use Brave browser to cut the communications clutter, works great.
Just a message from a paranoid person!
Cheers

I have tried downloading videos from the net and playing them. I’ve had the problems with those as well unfortunately, but it’s mostly been with the streaming ones. As to why I run Qubes, I don’t have any work or political reason, etc for needing qubes - just my main home computer. I refuse to do any social media. But I seriously value any privacy I can protect from those, who want to compile my life into databases to monetize me. I know I can be monitored by governments or others with the reources for snooping, etc but as my Dad used to say, they had a saying in the RAF during the war, “nolite te bastardes carborundorum”, which I understand loosely means, don’t let the **stards grind you down. I virtually never run Tor, because although I like the idea of Tor, I find it slow. Just the vpn, password manager and try to segregate my VM useages according to general “throw-away” browsing, more stringent browsing, financial uses, personal emailing, etc. I do sometimes use Brave as an alternate browser. Firefox has always been my daily driver though. Yours truly, paranoid :slight_smile:

1 Like

Especially with newer CPUs, it might be more stable.

If you are buying a new one, I would recommend sticking with AMD.

Thank you kommuni. There have been a couple of developments here. Temps-wise I wanted to make sure the cpu paste was adequate, so I pulled the fan to check. The store from which I buy my parts, insists on installing the cpu and paste to the mobo, before the parts leave the store, in order to stand behind the return policy. I assumed all was ok. But the other day, when removing the fan, I found only two adjacent hold-down locks to be locked, so one side of the cpu may have been somewhat lacking in contact. Replaced the paste and relocked the fan to the cpu. Also with your kernel replacement suggestion done, so far I am not seeing rising temps past low 40’s (and once to 50 deg C.) Too early to say for sure, but the problem may have been alleviated.
Also as it happens I did VM backups on the weekend with the new 4.0.4 installation. So if I do see the temps climbing again, I will indeed try a new 4.1 installation.
I will keep the nvidia option in mind as well. I had thought Intel was the preferred video option. I had always preferred to use nvidia prior to switching to qubes for my linux mint installations.
Thanks again. I will leave the thread as unsolved for a bit longer to make sure there are no recurring problems.

1 Like

Just about time to try 4.1 as my daily driver I guess. The self-restart occured again yesterday. You tube again (as always) at the time. If that doesn’t work, maybe a video card installation, rather than the onboard graphics.

i had a similar problem with random restarts (qubes 4.0.4, kernel 5.4.107).
they were caused by my graphics card. it’s a AMD Radeon RX 580.

the problem was solved by creating (as root on dom0) the file ‘/etc/environment’ containing just ‘AMD_DEBUG=nodma’ (without the ‘’). now reboot and enjoy :slight_smile:

since then - no more unnecessary reboots.

Thanks duke. I would like to try your suggestion. However I’m presently using the onboard video on my Asus Prime B365M-A board with the integrated graphics support. Would your command work with, “Intel_DEBUG=nodma” in a new “/etc/environment” folder? And if so, would I use Intel, intel or INTEL? I’m unfamiliar with the effects of this level of tinkering and a bit leery of restarting the system to find no video at all. Cheers