Windows 10 guest on Ubuntu 14.04 host freezing to the point that Workstation will not close

[danslay]

I've been having problems over the past few weeks with my VMware Workstation 12.5 Pro guest session freezing up completely. I believe this started happening after upgrading to a  Windows 10 guest, although that was several months ago, and I've only really noticed the freezing more recently.

• Host machine is Ubuntu 14.04, 64 GB RAM, single 12-core Xeon CPU.
• Guest: Windows 10, and now 10 anniversary edition. 12 GB RAM, 4 cores. 20 GB allocated in Workstation to all guests.
• This has occurred with both Workstation 12 Pro, and with the 12.5 recent 12.5 update - no change.
• Upgrade to  Windows 10 Anniversary Edition (after upgrading to 12.5) did not affect this.
• It never occurs when I'm using the VM actively. It occurs when I'm working on some other task / not interacting with the VM, perhaps for a few minutes, or maybe only after idling for several hours.
• It does not always occur - I can leave the VM up for days with no problem. Sometimes.
• When it occurs, you cannot power off the VM (let alone a graceful shutdown) - Workstation responds saying you cant close it when the guest is active (something to that effect). I've let it run for hours, but it never recovers. You must kill the vmware process to escape.
• CPU utilization is often (but I think, not always) high on the guest when this occurs. (as monitored in the host). Host still has plenty of CPU resources and functions normally.

Thanks for any suggestions on either how to diagnose, or what to try to avoid this. 

[chinaren0001]

Please upload more detail info, like collect support data(Help-> Support-> Collect Support Data...).

And I've read some posts, some solutions were used to avoid this problem, you can have a try:

1. Close the browser on host.

2. Downgrade to HW 10

3. Launch vmware using "vblank_mode=0 vmware"

[danslay]

Thanks for your suggestions. After downgrading to HW10, it no longer freezes as it had been. Instead, the screen fills with weird noise (vmware_freeze1.jpg attachment). When you then click on the screen, it mostly returns to normal. However the start menu becomes variously unreadable (one time, text was in red, but was functional. Another time text was black on black = very difficult to find shut down option...) and some window graphics (Close box in upper-right for example) disappear (but are there). In the attached vmware_freeze2.jpg, its not too bad, but has some garbage. There's prob

So, this is an improvement as I no longer need to be root to kill all vmware processes and restart!

In the meantime since my first post, I upgraded the Ubuntu host system to 16.06 from 14.04, but this had no impact on this issue.

I can take and send to vmware the support data file, if you are with vmware support.. but let me know what configuration you'd like the support data collected from (HW10, as now, with the graphical problems, or HW12, when I get the full freeze. The file I collected from a full freeze was 75MB.

[Ippy2]

I have the same problem (I think) on Windows hosts, running both Windows and Linux guest OS. I have tried with Windows 8 & Windows 10 guests, as well as Ubuntu 16.04, Debian and Zorin Linux guest installs.

In trying to isolate the problem, it seems to only occur when the VMware Workstation window is running 'Maximized'. Not 'Full Screen', but just maximized to the host monitor (happens on different hosts). If I run VMware Workstation in Windowed mode with guests running at a lower screen resolution than host-screen resolution, everything seems to be stable. Guests all have latest version of VMware Tools installed; though in the case of the Ubuntu guest, it's running OpenVM Tools as recommended by VMware...

I haven't done extensive testing, but I left my Ubuntu and Debian guests running overnight in Windowed mode, and both were still responsive in the morning. I maximized the Workstation window on one host, and the application (and guest OS) become unresponsive less than 30 minutes later. I had to use a remote taskkill command from another system to terminate the running process to restore functionality to the host console.

Sooo... any chance you're running your VMs in a 'Maximized' Workstation window?

[danslay]

No, I actually never run in maximized (or full-screen) mode, but thanks for the suggestion. My host Ubuntu system has a workspace switcher setup, so I run vmware in one of those separate workspaces, but still not maximized. It only seems to have the issue when I'm on another workspace...but I assumed that was because it only happens when I'm not actively using the VM. But maybe its related....who knows. Yesterday, the graphical corruption happened after maybe just 10 minutes on another workspace. But then overnight, it remained uncorrupted.

Running with HW version 10 does prevent the complete lockups of the vmware workstation application that otherwise require root to kill (or reboot), so at least that is a step forward, as I can then easily reboot the VM (if the reboot/shutdown icon is marginally visible anyway).

I also did the minor upgrade to 12.5.1, but issue persists.

I've been meaning to try running workstation in the background and then using rdesktop to connect to it, to see if that avoids the issue, but I havent been able to figure out how to get rdesktop connections to it working (despite following online postings regarding setting up port forwarding in the vmnet editor, and adjusting windows firewall (or turning off) and allowing incoming connections...still no dice...).