virtualbox.org

Running Version 5.1.16 r113841 (Qt5.6.2)
OSX 10.12.3

Until now:
- various machines that have been suspended using the previous version of VirtualBox crashed after resume - but not directly. It seems that they crash on CPU load / activity on the mounted virtual file systems.
- Guest machines running: Windows 7 64bit, Debian Stretch 64bit

What I did:
- Started the machines from the "aborted" state and installed the newest Virtual Box Guest Additions

I'll update this post, if further crashes happen.

=== Update

Crashes also happen after the restarts and VBoxAdditions update.

Mapache23 wrote:various machines that have been suspended using the previous version of VirtualBox crashed after resume

Of course they will!!!

Do you realize what "suspend" and "resume" do? It's like putting your computer in hibernation. And while your computer is hibernating, you change the hardware. That's what in a sense changing VirtualBox versions does. That's not even a bug, it's expected behavior. Unwanted perhaps, but expected none the least.

There are two ways you can avoid crashing a suspended VM after you've upgraded VirtualBox:

1. Discard the saved state, potentially losing some work you've done since you last started the VM. It's up to you what can get lost.
2. Downgrade your VirtualBox to the last version that you were working with. This will allow you to resume the VM properly. You shutdown the VM completely, upgrade VirtualBox and reboot your VM.

1. They didn't do that in previous versions and I wanted to point that out if it is somehow important
2. They also crash after a restart - not directly, but with some activity on the virtual file systems

I downgraded to 5.1.14 and everything works fine again.
I am waiting to see others report similar crashes before I will dig into this any deeper.

And BTW ... This was just a revision number switch from 5.1.14 to 5.1.16.
I do not expect that the "hardware changes" in a bugfix revision change.

1. The crashes after a suspend/upgrade/resume are not mandatory. It depends on the parts of the code that might have changed and the areas that have been affected, from RAM, to HD, to network, to audio, to USB, to ... The certain thing is that they don't come as a surprise if they happen.
2. Crash after a restart is a completely different issue. If so, please provide some more details, like logs, error messages, etc.

There's no "similar reports" that you have to wait for. Just search for "suspend, upgrade VirtualBox, crash". Plenty of reports, tickets and stories. The answer is the one that I gave you. You may not like it, but if you think 2 minutes about it, it's the only logical one...

You're too fast for me

Mapache23 wrote:This was just a revision number switch from 5.1.14 to 5.1.16. I do not expect that the "hardware changes" in a bugfix revision change.

Why? Who told you that? You want to know a little secret that you wouldn't know unless you searched for and compared the source code between 5.1.14 and 5.1.16? They changed completely the audio system, and it is 99.9999% certain that in your case you'll have a crash.

Bottom line is one: do NOT upgrade VirtualBox while your VM is suspended. You're not going to get anyone's sympathy if you do not follow best practices. It doesn't even qualify as a bug, do you realize that?

Ah - someone already opened up a ticket. Thanks a lot for your help.

viewtopic.php?f=8&t=82141

I am fine with my 5.1.14 until the bux is fixed

5.1.14 and 5.1.16? They changed completely the audio system

Professional software development for the win

Mapache23 wrote:I am fine with my 5.1.14 until the bux is fixed and you can just gfy

And you can enjoy the 1-week cool down period/ban for swearing. Directly...

EDIT: I see that you edited out the offending language while you were getting the warnings...

Mapache23 wrote:Ah - someone already opened up a ticket. Thanks a lot for your help. viewtopic.php?f=8&t=82141

Just to clear any potential misunderstanding for other users, the linked ticket does NOT involve restoring from a saved state after an upgrade. Yes, it contains the keywords "OSX", "crash", "5.1.16". That's where the similarities end. Not the same issue I'm afraid...

We found a regression in the 5.1.16 Shared Folders code (host side) when running Windows guests. This regression was fixed in the latest 5.1.x test build which can be found here. Could you install this test build and verify if the crash is still reproducible?

Hi,

I'm encountering exactly the same problem with the 5.1.16 release of VirtualBox and a Win7 VM : an access to a hosted mounted drive increase considerably the load of the host and finally crashes the VM !

I confirm that after replaced my installed 5.1.16 release of VirtualBox by the latest 5.1.x test build (5.1.17 r113974), I've no more problems !

Thanks for the fix