Ticket #11481 (closed defect: duplicate)
|Reported by:||Grimeton||Owned by:|
I know the topic isn't very clear, but that's what this whole thing is:
Sometimes processes in the guest are producing segfaults or exit without any reasons and fail. I can see this on Linux and on OS X hosts. Windows XP (guest) is sometimes affected, Windows 7 (guest) doesn't make it past the Windows Update installs without having the update service killed multiple times. It sometimes even produces a blue screen when you activate the guest (Windows license activation).
The main reason I didn't report the problem earlier was that I am unable to reproduce it. A few days ago someone else reported similar behaviour of his VMs on a Windows 7 Host, which could be solved by the same procedure I'm going to write about later, so I decided to file a bug report.
What happens is that the processes inside the guests fail. If you power the guest down and start it again (no reboot, new instance of Virtualbox), you still get the same error inside the guest. The only way around it, ist to power down the "broken" guest, start another VM, so that it allocates memory on the host, and then start the guest again (while the other VM is still running). After doing this, the guest starts to work properly again, as it now uses a different part of the host's memory (afaik).
In Windows 7 the update service fails, or .NET assemblies stop working from one second to another. On Linux guest you get all kinds of segfaults in all processes. Interesting is that you don't get a kernel core dump, all processes reside in user land. It's also independent of the memory the guest can use. I see it from machines with less than 256mb of Memory and up to 8GB and more of Memory.
The memory on my machines has been checked and everything is fine.
I'm sorry that I don't have any further information. If you have questions feel free to contact me.
Attachements: IRC log of #vbox