billington.mark

Members
  • Posts

    362
  • Joined

  • Last visited

Everything posted by billington.mark

  1. If this was me, id back up what you wanted to put back on cache post pool change into the array temporarily, reduce to one device, reformat the cache drive and copy data back. Obviously beforehand, disable VMs and/or docker so nothing tries to write to cache while you're reformatting and copying data back. The GUI has always seemed to be geared to adding more disks to the array, or more disks to cache, but not the other way around. EDIT: See linky below
  2. Thankyou for posting this! You've saved me a fortune! I was ready to pull the trigger on 3 new drives!
  3. see here: https://wiki.archlinux.org/index.php/KVM Seems its possible, but not sure if the way its implemented inside UNRAID will allow this to be done, proceed at your own risk
  4. Please post your syslinux, VM XML and IOMMU groups
  5. You could possibly try and add multiple Virtual NICs and then team them inside the OS?
  6. you'll be hitting the read speed on your VM Disk and\or write speed inside the array well before maxing out the virtual NIC. I dont think there's a way to change it to something "faster".
  7. Does it give a reason why it doesnt like those cards? Another option to test if thats the case is to not stub the controller remove the pass through, and try pass through each device (probably add the new IDs of the oculus to your stub list in syslinux too)? not sure it'll work, but if its something funky with the windows driver for that usb controller, that issue is removed if you pass through the device with this method instead. Its worth a shot. Also, worth giving Q35 a try over i440x with this method too, (you'll most likely need to reinstall windows if you try doing this) all things you can try for free before spending money on the new expansion card. Also, when you're playing VR pool.. remember not to lean on the virtual table! https://www.youtube.com/watch?v=gZCAi0MLxAE
  8. Sorry, was reading though your post on my phone, so missed the usb controller you passed through! Im not familiar with the Oculus (although i hope to be after xmas!), so ive done a quick google on the device detecting usb2 on a usb3 port and it seems to be a common problem when using a built in controller on the motherboard. Depending on how its deciding if its a USB2 or USB3 port (some sort of built in speed test?). Saying that, a lot of the posts are saying the Oculus functions fine, regardless of it being shown as USB2 in a USB3 port? Is that not the case? Either way, the add on card route seems to be solving this issue. My next step if I was you would be to get a USB3 add on card, pass that through and take it from there (assuming you have a spare PCIe slot?)
  9. Passing through individual USB devices tends to be quite troublesome and doesn't always work as expected... The quick fix is to install a USB 3 add on card, stub that and have all the devices plugged into that which you want passed though to the VM. I'm not familiar with how oculus plays when it's plugged into a hub off a single port, so have a read up before buying one on case you need more ports.. It does look like you've stubbed your usb3 controller and passed that through... Is that detected properly in the VM? If yes, I'd try a USB 3 hub with all your USB devices you need, plugged into one of the ports on that controller. You don't need to stub ever USB device in that case too, only the USB controller.
  10. It was that long ago it was flashed, i actually cant remember. (sorry!)
  11. You'll be able to control fans if you use a USB fan controller and have that plugged into a USB controller you've passed through, but to overclock on the fly you'd have to be able to pass through other components on the motherboard to the VM which isn't possible.
  12. Well that was quick! Drop me a pm with your PayPal address and I'll send through a PayPal invoice and get it shipped to you. Mark
  13. Dell Perc H310 no longer required due to me upgrading my motherboard. flashed to IT already. Necessary sata cables included. £42 with free postage in the UK. Would prefer not to post internationally.
  14. Thats interesting... did you modify the drivers in any way or just run through the normal setup?
  15. I had this battle this weekend. it seems the latest drivers have more checks in them which detect you're running a VM and refuse to start. If you're exclusively using the GUI to manage VM options, make sure you've disabled hyper-v in the advanced features. After that, add this to your XML (repace the <features> part with the below). you can set the vendor_id value to whatever you want. avoid using anything KVM,QEMU,UNRAID related as this value is checked by the nvidia driver once windows starts up.: <features> <acpi/> <apic/> <hyperv> <relaxed state='on'/> <vapic state='on'/> <spinlocks state='on' retries='8191'/> <vendor id='none'/> </hyperv> <kvm> <hidden state='on'/> </kvm> </features> After that's done, driver version 368.81 was the latest i could get to install and start. Later drivers didnt work as they must have additional checks to see if you're running a VM. Worth mentioning that i was experimenting with fresh installs of windows each time to work out exactly what i needed to do. I didnt have the battle of having to remove old or downgrade nvidia drivers each time I made a change, so just adding this to XML and installing the driver in a current "code 43" VM might not work. Good luck
  16. Easy peasy. on current VM, do a backup using the built in backup feature ('Backup and restore (windows 7)' in control panel) and back it up to a network share on unraid create new VM with passed through SSD and boot from a windows 10 ISO (you'll need the virtio drivers iso mounted too, to install drivers). On the first screen, instead of install, select "repair my computer" and select the option to restore from an image. cant remember the exact wording, but its pretty obvious. you'll get a pop up asking for the network location and if you want to install any drivers. Install the network and storage drivers, then give it the location you sent to backup to previously. Once its finished, it should boot from the SSD. done. worth mentioning you cant switch from i440x to Q35 using this method, that has to be the same.
  17. At the moment QEMU and libvirt are quite far behind in the latest rc (rc3 when this was posted). RC3 versions: Libvirt version: 1.3.1 QEMU version: 2.5.1 (cant find anything anywhere about this release?) Current: Libvirt version: 2.1.0 (7 releases behind) QEMU version: 2.6.1 \ 2.7.0 rc3 (2\3 releases behind) Are there plans to keep more up to date with these releases now unraid is trying to market itself as a hypervisor\NAS combo OS? or at least allow parts like this to be more modular and able to be upgraded outside of a major\minor OS update? sources: https://libvirt.org http://wiki.qemu.org/ Mark
  18. How have you passed through the drive? by disk name (/dev/sdx) or by ID? If its by name, it could have grabbed a different disk on startup if unraid decided to name the disk differently for whatever reason. Worth checking your xml against the disk name in the GUI to make sure.
  19. sounds like your VM is trying to boot form the new Vdisk. post your XML, or try adding this into vdisk code you want to boot from: <boot order='1'/> so it'll look like: <disk type='block' device='disk'> <driver name='qemu' type='raw' cache='writeback'/> <source dev='/dev/sdf'/> <backingStore/> <target dev='hdc' bus='virtio'/> <boot order='1'/> <alias name='virtio-disk2'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/> </disk>
  20. Not really looked into it, but would this not be more possible to use chromeOS as the base now it can run android apps? Is androidTV just a fancy launcher which could be set to launch on boot? Just throwing ideas out there, not sure if its any help!
  21. 1. Does "Info" in the GUI report that HVM and IOMMU are enabled? 2. are all Visualization settings enabled in the BIOS? (eg VT-d) 3. Post your syslinux.cfg file (click "Flash"on the main tab). Seems you may need to add a couple of options to get pass-through working