Long running kvm suddenly thinks virtualization isn't supported in CPU

Hello,

I have had virt manager up and running for years. I shut down my VM’s before going to bed last night and attempted to start one in the morning and was met with the following error:

unsupported configuration: Domain requires KVM, but it is not available. Check that virtualization is enabled in the host BIOS, and host configuration is setup to load the kvm modules.

I seldom shut the system down and did not last night. In short it was working fine when I went to bed and wasn’t in the morning.

I thought I’d just ask if anyone has ever heard of anything like this before I expend a lot of energy loading config and error data…
That said, I did see “kvm_amd: SVM not supported by CPU 2” error in the logs with the core number varying between 2 and 8. I did a few diag steps like modprobing kvm_amd (it’s a 12 core Ryzen 9). And /dev/kvm doesn’t seem to exist.

In short it seems to match every detail of not having virtualization enabled in the BIOS…but it was running a few hours before, had been running for years… I’ve never heard of a CPU failure but unless someone here has knowledge of anything similar I’m thinking my best bet before chasing this down further from an OS perspective is to see what I can find from AMD 1st.

Thank you in advance.
Dee

First thing I’d have done would have been to check the BIOS settings - whether it is enabled there - since the error message seems to imply that it is not.
… kind of based upon the old saying:
“Have you tried turning it off and on again?” :slightly_smiling_face:

1 Like

Means: BIOS Update switches to default = no kvm…

Lol…

I just came back to this post to post an update and saw your response. My update was that I had checked if somehow the SVM setting had changed… is was disabled… I didn’t check before because it was on and working fine since the ast reboot about a week ago or so… but I thought that a mother board failure isn’t any less lkely as a CPU failure so… taking a look isn’t going to hurt…

So now I’m wondering how in the world that can change on it’s own… no reboot, no anything other laying down for a few hours shut eye and boom… doesn’t work…I’ve never heard of such a thing… I didn’t have anything running over night so the system was just idling… I don’t know if BIOS settings can even be changed and made effective from a running system… particularly without a reboot… but I’ve never research the possibility…
It all just seems so unlikely… Before bed I had just thrown up a fresh Arch install and installed/tested Enlightenment which went well… everything was working as it had for a long time…then in the morning is wasn’t…
A few weeks ago a upgraded from 32GBs of memory to 64GB…and did have to make a BIOS change to get the system to see the whole 64GB… required turning on XMP… which I think changes the frequency of the memory.
Sorry for the false alarm I suppose I should have looked 1st but I kept thinking as my searched the net for a report of a similar issue that it isn’t the BIOS that every item I found indicated it to be…
If anyone has every heard of a BIOS setting changing on it’s own… I’d like to know…guess I’ll see if I find something on the net now… again my apologies… very odd

Thanks

2 Likes

… a cosmic (or some other) ray hit the EPROM and flipped a bit :man_shrugging:
It’s a plausible explanation with only as much speculation involved as with any other …

3 Likes

Well, BIOS settings don’t change ‘on their own’ or for no reason. However, it is possible for a BIOS to lose settings – this can happen if the BIOS battery is failing, in which case some BIOS can revert to default settings. A tell-tale sign that this has happened is often that the time/date is suddenly wrong.

From the description you’ve given, this doesn’t seem likely.

Otherwise, I’d suggest a setting had been manually changed without realising it; this does seem more likely; especially if it’s only one setting we’re talking about.

Regards.


A bit of what? :crazy_face:

1 Like

I agree with the likelihood of fat fingering unintentionally the BIOS setting… but I hadn’t been into the BIOS since changing the XMP memory setting (The XMP setting was still set fyi). While it seem like it should be possible to change a setting while the system is booted and running I would think it unlikely it would take effect without some kind of retest or…something… which surely would have caused Linux to crash…
I’m mystified… I did dig a little and I do see quite a few claims out there about a “BIOS changing on it’s own”… but I didn’t see one in which the system was continuously running making, as you put forward, an unintended fat fingering while in the BIOS more likely.
It is a Gigabyte X870 motherboard so later today I’m going to see if there’s any info on their site. Perhaps they have a forum where I might make an inquiry… donno… but I kinda feel like I’m claiming to have seen Bigfoot…but at the same time if I didn’t… how does that happen on a running system…
Thanks
Dee

2 Likes