Ubuntu 22.04 freezes randomly on VMWare Professional 17

[toukay]

Update 2023/03/20

The latest  revision of VMWare has not solve the issue: 17.0.1 build-21139696

Environment:

OS: Windows 11 with Virtual Machine Platform installed. All patches installed.

Software: VMWare Professional 17

VM: Ubuntu 22.04 with all patches.

Windows Host Power savings config: Never put computer to sleep.

Symptoms:

No TTY response. No response to CTRL-ALT-DELETE

Have to hard reboot the VM to get things working again.

Window does not resize and no interaction keyboard or otherwise works.

I have to hard boot the VM to restore functionality.

Frequency of problem:

When I have logged out for a long period of time, when I return to the VM again it is dead.

Hardware:

Intel i9 10900K

Graphics: Nvidia 3060

64 Gig of RAM

[toukay]

Thanks for the reply. Unfortunately when I tried removing WSL, the problem was not resolved.

[Mejato]

Similar thing happened here, got WSL recently deployed in March 2023 and will remove and see if it works.

[MagicDragon]

I removed WSL from my host windows 10 PC. But it did not resolve the issue.

[joshmcorreia]

Not that it's a great solution or anything but I did a fresh install of Windows 11 and VMWare 17 and now it works perfectly. I have WSL installed as well so we can eliminate that from being the cause, at least on Windows 11... Hopefully a solution is found that doesn't require a complete OS reinstall though 😞

[aussieklutz]

We had some success fully disabling hyperv virtualisation through BCDEDIT, but think the actual cause in the Virtualisation Based Security feature which relies on HyperV. This worked until recently when policies started enforcing VBS.

[vm71jpuser]

Here is an interim update. I was reviewing the error logs after the recent latest up date to 17.02.

My last frozen session did not have the processor freeze messages. Instead, I noticed that as I forced a reboot there was an SMBUS error stuck on the screen.  I further recalled in the 17.02 patch notes something about legacy Linux support fixes.

After verifying that that the system boot logs did indeed contain an SMBus initialization error. I looked for a way to avoid trying to stand up that device.

On the hypothesis that there is a regression caused by recent changes to Windows I have applied an old fix from 2013 for this from: https://whitehatty.com/2012/09/30/vmware-fusion-5-fix-ubuntu-linux-12-04-host-smbus-controller-not-e...

So far so good. I will update after a few more days of not having a system freeze.

How to fix it:

• sudo nano /etc/modprobe.d/blacklist.conf
• add the line:
  blacklist i2c-piix4

[jeromeoneil]

For anyone digging into this, I've noticed that it doesn't seem to impact kernel v2.  I experience this with RHEL-8 (v3) and RHEL-9 (v4) but do not experience it with RHEL-6 (v2).   

[MagicDragon]

I tried using a previous version of Ubuntu kernel as suggested by   jeromeoneil

The kernel version that I was using that caused system freezing was 5.19.0-43-generic.

So I installed a previous version 5.19.0-41-generic.
I used grub to boot using the older kernel version.

The result.
Well not only did it not solve the freezing problem but now there is no audio.

For Ubuntu, using an older version of Ubuntu does not work!

[toukay]

Previous attempts to fix this issue have all failed for me. Including the driver fix previously discussed.

Debian continues to go to **bleep** very fast on VM Ware 17... reproducible in a few minutes.

Ubuntu will stay up for days at times, other times freeze much more quickly.

Having just paid for 17, I'm starting to feel rather salty about the lack of a statement from VMWare or a fix.

[louyo]

Well, on another symptom: I had not seen any of this on either 16 or 17 until yesterday. I was doing some browser investigation using Firefox running on a throw-away Debian 11.5 VM. I was running on Debian 11.5 host with 16. As I was looking at results with Brave versus Firefox, the system would slow to a freeze. I put the system monitor and the Mate system monitor on the desktop and watched. As I connected FF to streaming services, like youtube, it consumed more and more memory until it froze. In several cases, I had to reset the VM to recover. It did not affect the host or other VM's. Duckduckgo'ing the problem found others with FF hogging memory. Probably not the same as you folks but the fix is simple. I guess another fix would be throwing a bunch of RAM at it, I only gave it 4GB

Have you tried watching system monitors, including good old simple top?

Lou

[toukay]

I tried: boosting the ram to 20 gig. Adding more CPU cores on top of the other solutions.

I went nuclear. My previous removals of WSL and associated components may have been insufficient.

So, I went nuclear.

Remove all things Windows Hypervisor, every single Windows feature including sandbox, and every component related to it. What a pity since this problem only ever affected Linux and not my Windows VMs.

Completely reinstall VMware Workstation 17 with the repair function.

Select the keyboard driver, because for God's sake I can't handle how that previous keyboard setup was so bad.

For the first time in longer than I can recall, the VM has survived the weekend without freezing.

Conclusion:

It is readily apparent that VMWare has turned into garbage for Linux. At least for my hardware. This is the only explanation for why my Windows VMs worked well but everything Linux kept dying.

VMWare: get your act together for Linux. It's an important OS and I am buying VM Workstation so that I have reliable VMs with it.

Right now, you just suck.

[Technogeezer]

This is a user-to-user forum. VMware technicians and developers sometimes visit here, but it's on their own time.

If you just paid for Workstation Pro 17 (within the last 30 days), open a support request with VMware through the Customer Connect portal. You have free complementary support  for 30 days after purchase. 

[toukay]

Sorry. I don't think this should require a support request. This Linux. The lastest release. It should work as well on VMWare as Windows.

End of story. No excuses.

I have a work around by not using VMWare's absolutely broken support for Windows Hypervisor.

I am not interested in 10 layers of support requests that will give me answers the community has already documented.

[Technogeezer]

---

@toukay wrote:

Sorry. I don't think this should require a support request. This Linux. The lastest release. It should work as well on VMWare as Windows.

End of story. No excuses.

I have a work around by not using VMWare's absolutely broken support for Windows Hypervisor.

I am not interested in 10 layers of support requests that will give me answers the community has already documented.

---

I understand, but if there aren't any good answers from other users in the forum on how to get around this, then having the ability to contact VMware directly might not be a bad idea.

And I would not place the blame on "broken support for Windows Hypervisor" solely on VMware. VMware's support is only as good as the API set that that Microsoft provides. Ask yourself: would Microsoft would rather that you use Hyper-V or VMware?. Do you think they have a vested interest in making it easy for third party hypervisors to run with Hyper-V, or to expose the full capabilities of their native hypervisor to third-party applications? 

[dicalleson]

I am experiencing the problem.  Can you provide more details on what you did to get rid of the issue.  What is "guest update"?

[toukay]

Status: after the purge of all things hyper-v and even related software my Ubuntu 22.04 VM was more stable.

It is, however still freezing, albeit it with less frequency. As stated earlier, my Windows VM on the same machine has no issues.

Also, by way of reminder, this device is always on. Therefore the issue is not a function of sleep or power savings.

Bottom line: there is no fix or workaround. The best you can do is to remove hyper-v completely and grin and bear your way through the occasional virtual processor lock ups.

--

FYI: guest update means updating the guest OS running on your VM.

[prplshroud]

This issue has been affecting me for far too long as well.  I have tried this as a latest attempt at a fix.  It worked once for a good while, but it has eventually reverted back to It's old ways of locking up after about 20-30 minutes. 

I have found this is not an exclusive Ubuntu issue.  I also have the same issues with Fedora (34 and up), but the time the VM runs is far shorter on Fedora.

Linux VMs get slower and slower then locks and video goes out.  No other option exists but to "reset" the VM.

I'm seriously considering dumping VMWW and moving to virtualbox instead to see if I can get better results.

The level of frustration with this is insane and I can't believe that VMW has not affected a fix for this yet.

[MagicDragon]

I used to use Virtual Box, but moved to VMWare 3 years ago.
VirtualBox was having the same issues as VMWare is currently having.
Plus VirtualBox requires much more manual setting up than VMWare as follows:
1. You have to manually apply the new version of addon's every time Virtual gets a version update.

2. If you want to share files with the host, you'll need to manually change user permissions to allow write privileges for your Linux user. (off the top of my head I can't remember which user account you'll need to change permissions)

I don't know if VirtualBox has improved since then, but be warned, you may be jumping out of the frying pan into the fire.

[wdliii]

I am hosting 17.0.2 build-21581411 on  Windows 10 enterprise Build 19044. It's not addressed as of that release.  I have three guest operating systems that all freeze, they are all Linux based: Centos 7, RHEL 8, and OpenSuse Tumbleweed. Plus  Windows server running: W2K12, W2K16, and W2K19  with no issues.  Only the Linux servers are freezing.  I read another post where someone mentioned an with USB and another with Suspend/Auto-protect. I have refrained from Suspending any of the servers.   I  seemed to be running fine once I removed all of the USB devices except for the Kingston xs2000 2tb drive.  I moved all of my windows servers to the Kingston and left the Linux servers on the internal HD. It works fine but every  now and then it comes back.  It's back now. This time I am going to make a copy of the servers, remove the snapshots, restart them and see what happens.  This is a royal pain to deal with. I hope it gets addressed soon. 

[prplshroud]

All,

Something else I'm discovering on my end in this.

I have run a Linux VM successfully for hours on end.  When I engage a VPN connection within the VM, things go off the cliff for me within 20-30 minutes.  Sometimes less than that.

Does anyone have any thoughts on the Linux VM locking up while using a VPN but running reasonably well when not?