Page 2 of 2

Re: Same image different host OS problem?

Posted: 30. Dec 2016, 16:35
by scottgus1
OK the .MAN files are .ova's renamed (why?) So I renamed them to OVA and imported them. Both the Windows and the Mac versions have the same settings, though slightly different byte counts in the files within the MAN-OVAs, and both run on my Windows 10 host, as in I can log into the host-only-networked guest web server on each. After that I powered them down.

I believe I have solved the case! A search on the error message "counter constraints enabled...etc." in the guest screenshot was the clue.

I tried a RedHat-oriented search on this error, since the guest type came in as RedHat-64 and got this: https://bugzilla.redhat.com/show_bug.cgi?id=921724. I then saw in the guest boot windows that this was really a CentOS 6.6 guest and reoriented my search, and got this: http://www.gossamer-threads.com/lists/l ... el/1716237.

Both indicate the error appearing due to either an error in AMD family 15h "Bulldozer" processors, or Linux's possibly-too-firm method of interpreting the AMD processor info. (Or possibly running in a virtual machine - see note below.)

Your Mac runs an Intel I7 Sandybridge, and your monster erroring server runs AMD FX-8150 Bulldozer family 15h. So the Mac runs fine and the AMD server doesn't. QED. :)

Solution? Unknown. Hunt down that error more. See if Linux has made an update to the issue, upgrade the Mac-Intel-based guest, then copy the guest to the AMD-based server (see Moving a VM for a super easy transfer method).

(Note: the above links, and another that might have to do with the same error https://communities.vmware.com/thread/5 ... 0&tstart=0 do all seem to implicate running in a VM: Virtualbox, kvm, and VMware, as an error vector. So maybe the developers might want to hear about this on the Bugtracker even if the actual issue isn't specifically Virtualbox-caused.)

Re: Same image different host OS problem?

Posted: 30. Dec 2016, 22:03
by socratis
Great detective work Scott! I was about to say that it's not the CentOS, it's the AMD, and I suspected it, because besides the forum I've been keeping a track on the bugtracker as well. I believe that there is one or more AMD related bugs already available, so please search before creating another one. I'll also check myself.