So, I've spent HOURS getting programs reinstalled, and re-setting up my Windows 8.1 VM in XEN after it broke when switching to beta6.  Now, I just reinstalled the gplpv drivers because whenever it would go to sleep the VM would show as shutdown, so I had to restart it from unRAID.  I also just installed the ATI drivers for my HD5550 video card.

 

When I start the VM, all is fine, but after a few minutes, I lose my internet connection.  It may be related to me starting my SABnzbd docker, but I've not tested it fully yet.

 

Any thoughts on how to assure myself network access to my windows VM?

 

I'm pretty unhappy with the VM situation on beta6 right now, and have been hoping we'd see a beta7 that addresses some of these issues, but it doesn't sound like we'll be seeing that anytime soon, sadly.

It appears there is a problem with the xen nic driver in beta6, there has been a few of us with network issues on Windows vms. I and a few others have rolled back to beta5a as its perfectly stable. Sadly now I have lost any chance of playing with docker because of this though.

 

Your other option would be to go the kvm route as it sounds more stable for Windows vms and has docker. I'm looking at this for the future myself but I have only tried out a libvirt xml config once and it's a little more complex than xen and I haven't had the time recently to get back into it, hence the rollback.

 

Either way you go, good luck ;-)

 

Sent from my LT26i using Tapatalk

 

 

hmmm...

 

Too bad.  I've tried the KVM route also, but it's not working for me either, and there's almost no documentation of any use for that either.

 

I've had to take the memory out of my HTPC to put into the unRAID server when I upgraded the motherboard, so I can't just go back to using the HTPC either.  Since I've converted my cache to btrfs, I can't easily go back to v5 either, so I'm pretty much screwed.

 

I have been hoping that LT would release a new beta with patches for KVM and XEN, but since they haven't even released the promised 'roadmap' for version 5 that 3-4 people keep demanding, I don't expect a new beta anytime soon.

 

Sadly, it's hard to find DDR3 memory in small-town mexico, so I guess I just don't get to play for a while.  Oh well, such is life on the cutting edge.

I'm pretty sure it's the gvplv drivers causing the problem, but I can't uninstall them.  When I do a system restore, I can get back into windows, then I uninstall the gvplv drivers, and it says I need to restart.  When I restart, windows won't start again.  I have to restore to a point before I uninstalled them, and then windows starts again.

 

I tried booting into safe mode, but I'm not able to use the installed in safe mode, so I'm sort of stuck with either geting into windows, but not having network access (useless), not being able to start windows at all (more useless), or refreshing the install, which removes all my programs, including these drivers (works, but will cost me several more hours AGAIN).

 

or, tell the family there is no move or TV watching for the next who-knows-how-long until a new beta resolves this issue.

 

None of the choices appeals to me very much.  I've tried to just so a KVM install of windows, but I can't get GPU passthru working, which does me no good either.

 

I'm afraid I have to give beta6 a failing grade

I just uninstalled the drivers again, then told windows to shutdown.  the VNC connection stopped, but windows still shows as running in unRAID.  I checked the log and see this...

 

Jun 28 19:40:04 media kernel: xen_netback:set_backend_state: backend/vif/25/0: prepare for reconnect
Jun 28 19:48:35 media kernel: mdcmd (74): spindown 7
Jun 28 19:48:36 media kernel: mdcmd (75): spindown 8
Jun 28 19:48:36 media kernel: mdcmd (76): spindown 9
Jun 28 19:51:33 media kernel: qemu-system-x86[7937]: segfault at 0 ip 00002b429eeb2a56 sp 00007fff0197f588 error 4 in libc-2.17.so[2b429ee21000+1bf000]
Jun 28 19:51:33 media avahi-daemon[1703]: Withdrawing workstation service for vif25.0-emu.
Jun 28 19:51:33 media kernel: br0: port 3(vif25.0-emu) entered disabled state
Jun 28 19:51:33 media kernel: device vif25.0-emu left promiscuous mode
Jun 28 19:51:33 media kernel: br0: port 3(vif25.0-emu) entered disabled state
Jun 28 19:53:42 media php: /usr/local/sbin/xenman shutdown windows8

 

hopefully this helps pinpoint the problem, so it can be fixed.

If reverting back to 5a would still be an option for you there is a method, without formating your btrfs drive.

You could mount the drive outside of your array, editing the paths to you vms in your cfg files, and if you edit you go script then it will mount the drive on boot too.

 

This is what I have done, admittedly I have a separate ssd for my vms (outside the array) which I formatted btfrs for docker.

 

That would stop your drive being a cache though, so you would either have to start writing to your array directly, or fit another drive as cache (it doesn't have to be big, all your apps/vms would be on the btrfs drive).

 

I'm currently on my phone, but if this method interests you I could provide some links to threads that would help you achieve this when I go on my pc later if you like?

 

EDIT - here's the thread that helped me mount the drive outside the array (obviously skip the ext4 formatting section)

http://lime-technology.com/forum/index.php?topic=30496.0