April 4, 201412 yr Hi, Just reading a lot in the last weeks about "unRAID as Host - Xen". Also managed to set up running Arch and win 8.1 VMs. I have read some forum members do have "passthrough" enabled and that they play .... as seen in Ironics video ... with the same performance as in a desktop. I am going to try "passthrough" too, but since I do have my win 8.1 VM installed, I was wondering if I should setup/fine tune windows in a specific manner to obtain best performance for gaming? Rgds.
April 4, 201412 yr Setup of the os should be no different than if it was a physical machine, I believe Sent from my SCH-I535 using Tapatalk
April 5, 201412 yr Nothing special required, just passthrough the card with pci line in your cfg file. Then on reboot Windows will hopefully install the drivers automatically or you can grab them yourself from ATI's website (assuming you have an ATI card). If you have a Nvidia card you will likely need to flash it is a quaddro card. With ATI you want to avoid catalyst control centre as it causes issues in the VM. Just use windows update to install the drivers is the easiest way I found. Then, you may need to adjust for overscan issues. If you do, see my blog for how to do so without CCC installed. Enjoy!
April 6, 201412 yr You can check here which cards can be passed through without any hassle: http://wiki.xen.org/wiki/Xen_VGA_Passthrough_Tested_Adapters I finally was lucky with a NVIDIA QUADRO 2000 with Windows (7 or 8 isn't making any difference). However I didn't succeed in passthrough this card with similar cfg settings in UBUNTU 12.04. Might be a Kernel issue, something for a separate thread. Sent from my iPad using Tapatalk
June 20, 201412 yr Author I am going to try "passthrough" too, but since I do have my win 8.1 VM installed, I was wondering if I should setup/fine tune windows in a specific manner to obtain best performance for gaming? Hi again, it took some time to buy my new hardware, but finally I got it. My specs: UNRAID 6.0-beta3 > Motherboard: ASUSTeK COMPUTER INC. - M5A99FX PRO R2.0 CPU: AMD FX-8350 Eight-Core, Speed: 4 GHz, Cache: 384 kB, 8192 kB, 8192 kB Memory: 16384 MB (max. 32 GB), Network: br0: - Duplex, eth0: 1000Mb/s - Full Duplex I am trying to finish my 8.1 VM for gaming and fine tune it. Still some issues to solve: 1) Passthrough my (old) nvidia 9800GT card: I wil try to explain what I did. I created my win8 VM and I installed the GPLPV drivers. After that I installed the latest nvidia drivers and fixed rebooting the VM to work following Ironic´s post. Now If I check in "Windows Device Mgr > display adapters" this is what I get: http://i236.photobucket.com/albums/ff75/razzta01/win%2081%20VM/20140620_devicemgr_zpsf45beb4e.jpg Is it normal that I get this yellow icon warning? Maybe there is a conflict between both gpu drivers and I need to unistall the video gplpv driver? I tried to uninstall the gplpv drivers but the effect was not to be able to boot back into my VM. Hopefully I have a backup of both my.cfg and .img files. Also, I cannot select more than 1280X1024 resolution although my screen handles 1680X1050. Why is this happening? 2) HDD transfer speeds/network I think my transfer speeds (a and d) are too poor. Any idea how to solve this? a)I can transfer a file from my other NAS (unraid1) to a share in my new unraid2 (where I installed 6b3) at 17MB/s. b) I can transfer a file from a share in my new unraid2 to the VM folder in unraid2 too (VM folder is in my SSD; my SSD is not a cache drive within the array) at 52MB/s. c) I can transfer a file from my main-PC to the VM in my new unraid2 at 112MB/s.? d) I can transfer a file from my other NAS (unraid1) to my main-PC at 25MB/s. e) I can transfer a file from my other NAS (unraid1) to the VM in my new unraid2 at 23MB/s. f) b) I can transfer a file from my VM folder in unraid2 to a share in my new unraid2 at 23MB/s. I attach some info of unraid2: root@Tower2:/mnt/vm/windows# ifconfig br0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500 inet 192.168.0.110 netmask 255.255.255.0 broadcast 192.168.0.255 ether 40:16:7e:20:90:4a txqueuelen 0 (Ethernet) RX packets 2080592 bytes 29144367961 (27.1 GiB) RX errors 0 [b]dropped 106[/b]6 overruns 0 frame 0 TX packets 2315993 bytes 406822415 (387.9 MiB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 eth0: flags=4419<UP,BROADCAST,RUNNING,PROMISC,MULTICAST> mtu 1500 ether 40:16:7e:20:90:4a txqueuelen 1000 (Ethernet) RX packets 20583258 bytes 30208514802 (28.1 GiB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 2402629 bytes 411557706 (392.4 MiB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 lo: flags=73<UP,LOOPBACK,RUNNING> mtu 65536 inet 127.0.0.1 netmask 255.0.0.0 loop txqueuelen 0 (Local Loopback) RX packets 81 bytes 7060 (6.8 KiB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 81 bytes 7060 (6.8 KiB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 root@Tower2:/mnt/vm/windows# ethtool eth0 Settings for eth0: Supported ports: [ TP MII ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Half 1000baseT/Full Supported pause frame use: No Supports auto-negotiation: Yes Advertised link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Half 1000baseT/Full [color=red] Advertised pause frame use: Symmetric Receive-only[/color] Advertised auto-negotiation: Yes Link partner advertised link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full [color=red] Link partner advertised pause frame use: Symmetric[/color] Link partner advertised auto-negotiation: Yes Speed: 1000Mb/s Duplex: Full Port: MII PHYAD: 0 Transceiver: internal Auto-negotiation: on Supports Wake-on: pumbg Wake-on: g Current message level: 0x00000033 (51) drv probe ifdown ifup Link detected: yes root@Tower2:/mnt/vm/windows# And for my old unraid1: root@Tower:~# ifconfig eth0 Link encap:Ethernet HWaddr 00:22:15:a1:f6:30 inet addr:192.168.0.102 Bcast:192.168.0.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:9876573 errors:0 [b]dropped:25[/b] overruns:25 frame:25 TX packets:22190077 errors:0 dropped:0 overruns:0 carrier:4 collisions:0 txqueuelen:1000 RX bytes:3362198320 (3.1 GiB) TX bytes:2520104365 (2.3 GiB) Interrupt:42 lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:67 errors:0 dropped:0 overruns:0 frame:0 TX packets:67 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:5848 (5.7 KiB) TX bytes:5848 (5.7 KiB) Settings for eth0: Supported ports: [ TP ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full Supports auto-negotiation: Yes Advertised link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full Advertised pause frame use: No Advertised auto-negotiation: Yes Speed: 1000Mb/s Duplex: Full Port: Twisted Pair PHYAD: 0 Transceiver: internal Auto-negotiation: on MDI-X: Unknown Supports Wake-on: pg Wake-on: d Current message level: 0x00000000 (0) Link detected: yes As you can see I have dropped packets. Can anyone explain why what I marked in red is not showing up in my unraid1? 2) Audio As seen in previous image of device mgr, audio seems to be ok. I installed the realtek drivers an no error was shown. I still have to check it anyway.
June 22, 201412 yr Author Nothing special required, just passthrough the card with pci line in your cfg file. If you have a Nvidia card you will likely need to flash it is a quaddro card. Hi, can anyone explain how to do this? Or point me out to a guide describing the steps needed? Rgds.
June 22, 201412 yr Nothing special required, just passthrough the card with pci line in your cfg file. If you have a Nvidia card you will likely need to flash it is a quaddro card. Hi, can anyone explain how to do this? Or point me out to a guide describing the steps needed? Rgds. OK here is what I did: [*]Install unRAID beta (still on 6b5a myself so if below has changed for 6b6 - sorry). [*]Look in "utilities" tab for "PCI devices" - open it. [*]From the list presented make note of the number of the device you want to pass through (something like 05:00.0). [*]If you haven't already registered it with unRAID then open a command line to unRAID and "xenman register [configfile]" so that your vm is registered with unRAID - makes editing the config file easier. [*]Now go to "Extensions" tab and look for your VM in the list and click on the name which is an html link. [*]Now find the box for the config file labeled "Xen config file:" [*]Now add a line like this "pci=['05:00.0']" where 05:00.0 is the number you noted in #3 to your config file in that box. If you have multiple devices it would look like this "pci=['05:00.0', '04:0.0']". [*]Now edit the "syslinux.cfg" file in the syslinux directory on your flash drive. If doing so remotely use an editor that understands linux line endings like notepad2 or notepad++ in Windows. [*]Find the boot line for the Xen/unRAID boot option - it will look like this: "append /xen --- /bzimage --- /bzroot" [*]Add your PCI devices to it so that it looks like this: "append /xen --- /bzimage xen-pciback.hide=(04:00.0)(05:00.0) --- /bzroot" if you only have one device then it might look like this: "append /xen --- /bzimage xen-pciback.hide=(05:00.0) --- /bzroot". [*]Reboot your unRAID server and take the Xen/unRAID boot option. [*]Start up your VM with the devices you attached to the VM and if necessary install drivers in the OS.
June 22, 201412 yr Author Thx for your help. I was running 6 beta3 which didn´t have a "PCI devices" option within the "utilities" tab. Also, there was no "extensions" tab. I will follow the steps you provided, but I will start from scratch (doing a new windows VM install (i guess the GPL PV drivers messed it up). I will post the results. A "noob" question: When I run 1 VM (windows vm) I access it using VNC with no issues. But how should I do it if I run 2 VM´s simultaneously? I discovered VNC handles only 1 IP (the unraid host IP). But typing the actual Ip of each VM does not work. How can I change between both VM´s (actually how can I access both in parallel)? Rgds.
June 22, 201412 yr Thx for your help. I was running 6 beta3 which didn´t have a "PCI devices" option within the "utilities" tab. Also, there was no "extensions" tab. I will follow the steps you provided, but I will start from scratch (doing a new windows VM install (i guess the GPL PV drivers messed it up). I will post the results. A "noob" question: When I run 1 VM (windows vm) I access it using VNC with no issues. But how should I do it if I run 2 VM´s simultaneously? I discovered VNC handles only 1 IP (the unraid host IP). But typing the actual Ip of each VM does not work. How can I change between both VM´s (actually how can I access both in parallel)? Rgds. I only have one VM myself and haven't needed to figure that out yet. There is a setting in the config file that sets (a port value?) that you change. It is likely this one "vnclisten = '0.0.0.0'" and you would change it from the zeros. But I easily could be wrong too. I would figure it out today but I'm thinking about upgrading and want to use KVM but I want to let the dust settle on 6b6 first since I may only be able to activate Windows once more before I have to buy another license.
June 22, 201412 yr Author I only have one VM myself and haven't need to figure that out yet. There is a setting in the config file that sets (a port value?) that you change. It is likely this one "vnclisten = '0.0.0.0'" and you would change it from the zeros. But I easily could be wrong too. I look forward to hear how to solve this since I want to create more than 2 VM´s running at the same time.
June 23, 201412 yr A "noob" question: When I run 1 VM (windows vm) I access it using VNC with no issues. But how should I do it if I run 2 VM´s simultaneously? I discovered VNC handles only 1 IP (the unraid host IP). But typing the actual Ip of each VM does not work. How can I change between both VM´s (actually how can I access both in parallel)? Rgds. all you need to do is specify the port number for the vm in your vnc connection by default port 5900 is used for the first vm booted, this will increase by 1 each time you boot a vm for example, if your unraid tower was on 192.168.2.20 and you have 2 vms booted, the connection strings would be; 1st vm - 192.168.2.20 or 192.168.2.20:5900 2nd vm - 192.168.2.20:5901 using the defaults can mean the vms grab a port number based on what order they were booted you can specify a specific port for a vm in its cfg file, but off the top of my head I cannot remember the entry needed if you have a search through the forum im sure you'd find it though hope that's of use
June 23, 201412 yr Author Escellent mossiemosforth! I just tried it and it worked. I don´t have the experience to figure out how to do it in the .cfg file. I tried changing 0.0.0.0 to for example 192.168.2.20:5900 and it does not work. Maybe someone can post how this should be done. Rgds.
June 23, 201412 yr Author 1) Passthrough my (old) nvidia 9800GT card: As I previously posted, and after following all the steps in Ironic´s passthrough guide this is what I get: Pls, can anyone confirm to me that passthrough is done? I am not sure because: -in "Windows Device Mgr > Display Adapters" geforce 9800gt shows up , but the generic microsoft adapter too; is it normal that I get this yellow icon warning in the Windows Device Mgr? -after running GPU-Z only geforce 9800gt shows up (the generic microsoft adpater does not show up) -in Control Panel\Appearance and Personalization\Display\Screen Resolution: I can select only max. resolution 1280X1024, but you can see the black bars on the left/right (which is the actual resolution capacity of my screen : 1650X1050). I actually have 2 screens. What you see at the right of the black bar in the middle of the screen is not the VM, it is my other pc (from which I access unraid). Any way to check it I am missing? Rgds.
June 24, 201412 yr 1) Passthrough my (old) nvidia 9800GT card: Pls, can anyone confirm to me that passthrough is done? I am not sure because: -in "Windows Device Mgr > Display Adapters" geforce 9800gt shows up , but the generic microsoft adapter too; is it normal that I get this yellow icon warning in the Windows Device Mgr? -after running GPU-Z only geforce 9800gt shows up (the generic microsoft adpater does not show up) -in Control Panel\Appearance and Personalization\Display\Screen Resolution: I can select only max. resolution 1280X1024, but you can see the black bars on the left/right (which is the actual resolution capacity of my screen : 1650X1050). I actually have 2 screens. What you see at the right of the black bar in the middle of the screen is not the VM, it is my other pc (from which I access unraid). Any way to check it I am missing? Rgds. I'm not an expert, but I had a similar triangle before installing drivers for the graphics card. Have you installed any drivers, either by using Windows Search or from the manufacturer? Do you have a display on your screen? I know there has been trouble with Nvidia and Xen, but if it's working you might just need updated drivers.
June 24, 201412 yr Author I'm not an expert, but I had a similar triangle before installing drivers for the graphics card. Have you installed any drivers, either by using Windows Search or from the manufacturer? Do you have a display on your screen? I know there has been trouble with Nvidia and Xen, but if it's working you might just need updated drivers. Hi, yes I installed the nvidia drivers after a clean install of the OS. I tried to run a GPU benckmark (Furmark) and only windows generic drivers are recognized. It´s a pity but I am stuck. If can´t get any help I will give up.
June 25, 201412 yr In your case I'd probably just move to trying a KVM build. It's a little newer for unRAID, but GPU passthrough appears to work more reliably. There have been mixed results with those using the Xen part of Beta 6 it seems. Some of the updates compared to the ver. 5 have caused issues.
June 25, 201412 yr Author Hi, I think I will try this KVM route since not much support available at the moment for Xen. it´s a pitty since I think I am close to it. Pls, can you pass me a link to the best post describing the KVM way in unraid. I will have a look at it, but if it demands a high level of knowledge I may try esxci and have unraid as guest. Thx anyway.
June 25, 201412 yr The issue with Xen and PCI Passthrough is unRAID's implementation of it. I have already spoke to Jonp and made him aware of several Xen and Linux Kernel Patches that are needed for both KVM and Xen. These were not included in the last beta. Once he applies them and rolls out the next beta, Xen PCI Video Passthrough will work better for everyone.
June 26, 201412 yr The issue with Xen and PCI Passthrough is unRAID's implementation of it. I have already spoke to Jonp and made him aware of several Xen and Linux Kernel Patches that are needed for both KVM and Xen. These were not included in the last beta. Once he applies them and rolls out the next beta, Xen PCI Video Passthrough will work better for everyone. I actually wanted to post and ask just that. Is there anyway for us to compile Xen on our own for unraid? The reason I ask is that some of us may need particular patches that not everyone needs. It makes me wonder if compiling our own Xen would help us there? -Marcus
Archived
This topic is now archived and is closed to further replies.