dmk
-
Posts
24 -
Joined
-
Last visited
Content Type
Profiles
Forums
Downloads
Store
Gallery
Bug Reports
Documentation
Landing
Posts posted by dmk
-
-
FWIW: I'm using a $30 NVIDIA GT 640 1GB (the one with a Display Port on it) passed through in Mojave with native Metal support.
Not the fastest GPU out there but it works until the webdrivers are published.
-
@gridrunner
FYI: Clover 3974 (regular and patched version) fails to install on 10.12.3.
Anyone else having this issue?
What error are you getting installing it. As an alternative to installing with installer,you can just open the efi partition with EFI mounter and manually put the clover files into it.
My orginal video on installing sierra has those files in description (but not 3974) i cant link you 3974 to manually paste in as not at home
Sorry about that, user error. All fine.
Thanks for the video. Clover Configurator had an update so I went that route and configured it that way, mainly to ensure Nvidia drivers would stick. I also used Apple Remote Desktop instead of your solution, which works perfectly fine, no installs on guest required (just a share setting) and doesn't require me to register with a 3rd party. Note that I already had an ARD license otherwise it would have been a $80 expense.
-
@gridrunner
FYI: Clover 3974 (regular and patched version) fails to install on 10.12.3.
Anyone else having this issue?
-
Running UnRaid 6.2.0-beta20 with Mac OS X 10.11.3 (and Win 8.1).
It was kinda tricky to get it up and running. For anyone interested here is my XML:
<domain type='kvm' id='1' xmlns:qemu='http://libvirt.org/schemas/domain/qemu/1.0'> <name>OSX 10.11</name> <uuid>df84cac7-e3ff-4e0a-ed55-6cf3f79be06f</uuid> <metadata> <vmtemplate xmlns="unraid" name="osx" icon="osx.png" os="osx"/> </metadata> <memory unit='KiB'>12582912</memory> <currentMemory unit='KiB'>12582912</currentMemory> <vcpu placement='static'>4</vcpu> <cputune> <vcpupin vcpu='0' cpuset='0'/> <vcpupin vcpu='1' cpuset='1'/> <vcpupin vcpu='2' cpuset='4'/> <vcpupin vcpu='3' cpuset='5'/> </cputune> <resource> <partition>/machine</partition> </resource> <os> <type arch='x86_64' machine='pc-q35-2.5'>hvm</type> <kernel>/mnt/cache/OSX/enoch_rev2795_boot</kernel> <boot dev='hd'/> <bootmenu enable='no'/> </os> <features> <acpi/> </features> <cpu mode='custom' match='exact'> <model fallback='allow'>core2duo</model> </cpu> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>restart</on_crash> <devices> <emulator>/usr/bin/qemu-system-x86_64</emulator> <disk type='file' device='disk'> <driver name='qemu' type='raw'/> <source file='/mnt/cache/OSX/osx.img'/> <backingStore/> <target dev='hda' bus='sata'/> <alias name='sata0-0-0'/> <address type='drive' controller='0' bus='0' target='0' unit='0'/> </disk> <controller type='usb' index='0'> <alias name='usb'/> <address type='pci' domain='0x0000' bus='0x02' slot='0x01' function='0x0'/> </controller> <controller type='sata' index='0'> <alias name='ide'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x1f' function='0x2'/> </controller> <controller type='pci' index='0' model='pcie-root'> <alias name='pcie.0'/> </controller> <controller type='pci' index='1' model='dmi-to-pci-bridge'> <model name='i82801b11-bridge'/> <alias name='pci.1'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x1e' function='0x0'/> </controller> <controller type='pci' index='2' model='pci-bridge'> <model name='pci-bridge'/> <target chassisNr='2'/> <alias name='pci.2'/> <address type='pci' domain='0x0000' bus='0x01' slot='0x01' function='0x0'/> </controller> <interface type='bridge'> <mac address='52:54:00:00:20:30'/> <source bridge='br0'/> <target dev='vnet0'/> <model type='e1000-82545em'/> <alias name='net0'/> <address type='pci' domain='0x0000' bus='0x02' slot='0x03' function='0x0'/> </interface> <hostdev mode='subsystem' type='pci' managed='yes'> <driver name='vfio'/> <source> <address domain='0x0000' bus='0x0c' slot='0x00' function='0x0'/> </source> <alias name='hostdev0'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x05' function='0x0'/> </hostdev> <hostdev mode='subsystem' type='pci' managed='yes' xvga='yes'> <driver name='vfio'/> <source> <address domain='0x0000' bus='0x02' slot='0x00' function='0x0'/> </source> <alias name='hostdev1'/> <address type='pci' domain='0x0000' bus='0x02' slot='0x04' function='0x0'/> </hostdev> <hostdev mode='subsystem' type='pci' managed='yes'> <driver name='vfio'/> <source> <address domain='0x0000' bus='0x02' slot='0x00' function='0x1'/> </source> <alias name='hostdev2'/> <address type='pci' domain='0x0000' bus='0x02' slot='0x05' function='0x0'/> </hostdev> <hostdev mode='subsystem' type='usb' managed='no'> <source> <vendor id='0x05ac'/> <product id='0x020c'/> <address bus='3' device='5'/> </source> <alias name='hostdev3'/> </hostdev> <hostdev mode='subsystem' type='usb' managed='no'> <source> <vendor id='0x0a5c'/> <product id='0x200a'/> <address bus='3' device='6'/> </source> <alias name='hostdev4'/> </hostdev> <hostdev mode='subsystem' type='usb' managed='no'> <source> <vendor id='0x0556'/> <product id='0x0001'/> <address bus='3' device='4'/> </source> <alias name='hostdev5'/> </hostdev> <memballoon model='none'> <alias name='balloon0'/> </memballoon> </devices> <seclabel type='none' model='none'/> <qemu:commandline> <qemu:arg value='-device'/> <qemu:arg value='usb-kbd'/> <qemu:arg value='-device'/> <qemu:arg value='usb-mouse'/> <qemu:arg value='-device'/> <qemu:arg value='isa-applesmc,osk=youcanfindthisanywherebutpeoplegetangrywhenpostingithereforsomereason'/> <qemu:arg value='-smbios'/> <qemu:arg value='type=2'/> </qemu:commandline> </domain>
It is now possible to pass USB devices per devices instead of a full USB controller. This is handy for me as I wanted to pass the full Renesas USB controller to Windows as Oculus doesn't like per device passthrough for some reason. Before I had another USB controller passed to OSX but that excluded the Renesas controller as there was the UnRaid USB thumb connected. Switched it all around and it's a go
Few issues:
- high cpu on a kernel_task, maxing out 1 core
- About this Mac mentions 4x Core 2 Solo, it mentioned my Xeon before and I would like it back
Maybe some of you know the answers to these issues?
-
So i have resolved the usb issue i believe by just repeating the process. I now have an entirely different issue lol. OSX 10.11.2 installs beautifully and works correctly when using VNC, now when i attempt to pass through the GPU, the system starts to boot however the loading bar stops at 25 percent on the apple logo. Has anyone seen this before?
In reply to your keyboard issues, I use <bootmenu enable='no'/> to skip the first boot menu. This focuses the keyboard on the Enoch boot loader.
For passing the 25% mark on boot you need to add some boot flags at the Enoch bootloader. I used nv_disable=1 to fully boot. Then install the Nvidia webdriver. The webdriver won't work for 10.11.2 so you need to install and modify the driver with an application called NVIDIA® WebDriver Updater.app. On the NVDAStartupWeb.kext patch tab enter 15D9c as fake OS build. After installing the driver the driver thinks you're still on 10.11.1. I didn't touch anything else in the app.
After installing the drivers, reboot with nvda_drv=1 at the Enoch boot loader. I need to enter this at every boot which is annoying. I hope you don't have to.
Check out this post for the correct files in the /Extra folder: https://lime-technology.com/forum/index.php?topic=44908.msg429831#msg429831
ps: I also use a GTX 960 for the OSX VM. It drives 4 monitors (1*4K, 1*HD, 2*1600x1200) perfectly fine.
-
So as a recap...
I've got the Nvidia webdriver working by adding org.chameleon.boot.plist and smbios.plist to the /Extra folder on the Mac HD.
This is the org.chameleon.boot.plist I used:
<?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd"> <plist version="1.0"> <dict> <key>Timeout</key> <string>5</string> <key>EthernetBuiltIn</key> <string>Yes</string> <key>PCIRootUID</key> <string>1</string> <key>KernelBooter_kexts</key> <string>Yes</string> <key>nvda_drv</key> <string>1</string> </dict>
And this is the smbios.plist:
<?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd"> <plist version="1.0"> <dict> <key>SMbiosdate</key> <string>10/07/2013</string> <key>SMbiosvendor</key> <string>Apple Inc.</string> <key>SMbiosversion</key> <string>IM141.88Z.0118.B00.1309031249</string> <key>SMboardproduct</key> <string>Mac-27ADBB7B4CEE8E61</string> <key>SMfamily</key> <string>iMac</string> <key>SMmanufacturer</key> <string>Apple Inc.</string> <key>SMproductname</key> <string>iMac14,2</string> <key>SMserial</key> <string>C02LH068FLHH</string> <key>SMsystemversion</key> <string>1.0</string> </dict> </plist>?
With <key>nvda_drv</key><string>1</string> in org.chameleon.boot.plist it still won't boot to the Nvidia webdriver so I need to add nvda_drv=1 as a boot option every time I boot the VM. This is annoying but at least it works.
Real world performance is awesome. Compared to my late 2014 Core i7 MacBook Pro Retina it feels much more responsive and snappier.
Being able to work on my "HackVM" while simultaneously run Windows to test games and VR applications on my Oculus is great. No more rebooting.
-
You should have created the /Extra folder if you followed my guide?
Yeah, sorry about that. I did follow the guide but I directly used passthrough and somehow I skipped that part as it was right below the absolute mouse positioning.
Anyway, thanks a bunch. Now I've got an updated 10.11.2 working with a GTX 960 passthrough and a 4 display setup (incl a 4K display). The 4K display doubles for gaming on the other VM with a GTX 980Ti.
All working fine. Great holiday project.
-
Succes, I've added org.chameleon.boot.plist and the iMac14,2 smbios.plist.
org.chameleon.boot.plist:
<?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd"> <plist version="1.0"> <dict> <key>Timeout</key> <string>5</string> <key>EthernetBuiltIn</key> <string>Yes</string> <key>PCIRootUID</key> <string>1</string> <key>KernelBooter_kexts</key> <string>Yes</string> <key>nvda_drv</key> <string>1</string> </dict>
However, the nvda_drv setting isn't working. I have to add it manually as a boot flag during boot.
-
I've ditched the Radeon cards and installed a GTX 960 for OSX.
Now I can't get the Nvidia webdriver to work, probably due to the model identifier of the VM is iMac (Summer 2001). I've asked a question about how to change the model identifier in the Mac OS X 10.11 El Capitan – VM on unRAID guide:
https://lime-technology.com/forum/index.php?topic=43816.msg429679#msg429679
-
Its been at least 5 years since first saying "winter is coming" and it hasnt arrived yet. Lets hope LT isnt on HBO GoT time table.
Change winter for zombies and it all makes sense
-
I've been trying to get the Nvidia webdriver to work but I won't work. The Nvidia driver page mentions the driver is for "Mac Pro 5,1 (2010), Mac Pro 4,1 (2009) and Mac Pro 3,1 (2008) users".
Following this guide I the resulting model identifier is iMac (Summer 2001). I suspect the Nvidia webdriver won't work because of the wrong model identifier.
Every guide on the net mentions smbios.plist in the /Extra folder, except I don't have that folder.
How do I change the model identifier?
-
Thanks for the reply rhinoman,
I have it set up this way now:
- BTRFS cache pool of 4 * 120GB SSDs (see performance stats a few posts back).
- Storage array of 1TB (parity), 750GB, 500GB.
The cache pool is in the default as UnRaid sets it up in BTRFS RAID1 (which is somewhat different than regular RAID1).
I run the VMs from the cache pool and while the 4K performance is less than perfect the real world performance is great.
- BTRFS cache pool of 4 * 120GB SSDs (see performance stats a few posts back).
-
The Radeon GPUs aren't working out. I'm selling them to get some cash for a GTX 960, that would also be powerful enough for OSX to drive a 4K monitor and two 1600x1200 side monitors which I currently drive from my rMBP with the help of a Kensington USB3 docking station via DisplayLink.
-
Upgraded 10.11.1 to 10.11.2, no issues but GPU still stuck at 1024x768 (somehow also thinks it's a projector).
Did a disk speed test from OSX, pretty good although the 4K blocks are relatively slow.
Results 800.30
System Info
Xbench Version 1.3
System Version 10.11.2 (15C50)
Physical RAM 4096 MB
Model Macmini2,1
Drive Type QEMU HARDDISK
Disk Test 800.30
Sequential 501.97
Uncached Write 2103.98 1291.81 MB/sec [4K blocks]
Uncached Write 1970.72 1115.03 MB/sec [256K blocks]
Uncached Read 151.69 44.39 MB/sec [4K blocks]
Uncached Read 2541.54 1277.36 MB/sec [256K blocks]
Random 1972.62
Uncached Write 688.20 72.85 MB/sec [4K blocks]
Uncached Write 3760.43 1203.85 MB/sec [256K blocks]
Uncached Read 5612.70 39.77 MB/sec [4K blocks]
Uncached Read 7657.28 1420.86 MB/sec [256K blocks]
-
I've got Win8.1 and OSX 10.11.1 running both.
USB and GPU passthrough working on both but OSX has issues with the GPU.
HD7750 results in garbled display, can't find a solution anywhere that works with enoch. Tried to figure out how to change to clover but I'm lost here.
HD6570 does work but is stuck to 1024x768. Not sure how to fix it.
-
No, rebooted the VM.
-
Please let us know it you are successful to install OSX, we have found some issue on 10.11.2 iso where we don't get into the install GUI
I installed 10.11.1 as that was the one already in my Applications folder. It didn't get to the GUI as well but after a reboot it did. Installed fine after that. HD7750 passthrough also working, at least to the boot screen. I need to add keyboard/mouse to get past it. Maybe later tonight.
For a Christmas project in between family obligations everything is going very well so far.
-
Well, that was easy.
Had to use ACS override for the passthrough to work but after that, easy peasy, Windows is installed. Even the Nvidia driver installed without issues. Color me impressed! Thanks everyone, much appreciated!
Now to install OSX...
-
Well, I've accumulated about 150GB in the last 15 years. I've had a Freenas zfs system but I just wasn't using it. It was way more time efficient to just buy a Synology for what I needed it for in the first place: backups.
Right now my interest is running a passthrough OSX VM with (relatively safe) data protection, the Windows VM I care less about and I consider it as a throwaway VM just for games. Having the protected data array as a user share to save games and my 150GB of data to is fine for me. If I need more space I'll delete some games, no biggie, I can always download them again. The 150GB of important data is also synchronized to a rMBP, a Synology in RAID1 and an offsite backup.
-
Thanks @garycase
Final question: does the cache pool support btrfs raid-5 or only btrfs raid-1?
-
So @garycase, what you're saying is to basically use btrfs raid5 with 4 120GB SSDs to create a 360GB cache pool and run the VMs from the cache?
Would you consider btrfs raid5 as safe right now? Any considerations I need to make regarding the write hole issue?
Just remembered I have a few 2.5" HDDs (1TB + 750GB + 500GB). I might as well use those as an array to move games to when not playing them, with the 1TB drive as parity drive. Sounds about right.
I wrote "I'm not interested in storing large quantities of data." a few posts back. What I mean is I don't have a need for tens of TB of storage, the only reason I can think of is storing loads of video and music files but services like Netflix and Spotify do that for me. Stuff I would store are photo's, documents, git repos, etc. To be honest even that isn't more than 150GB. Games would add to it though.
-
Thanks for the reply.
I've done the research, this is not my first go around
The purpose for this build is having a workstation and a gaming machine all in one, with the added benefit of docker and securing the data I work with. Hence the choice for UnRaid. I'm not interested in storing large quantities of data.Besides that I don't have HDs, only SSDs. My HDs are in a Synology, which is mainly used for backups.
As I understand it, it is possible to run nightly TRIMs.
-
I'm rebuilding a machine I have because I want to run two VMs with passthrough.
Current hardware:
Motherboard: SuperMicro X10SAT
CPU: Xeon E3-1275 V3 - 3.5Ghz
RAM: 32GB ECC unbuffered
PSU: 1500W Silverstone Strider
GPU 1: GTX 980Ti
GPU 2: HD7750 or HD6570
SSDs: 4 * 120GB + 1 * 64GB
The passthrough VMs I want to run are OSX El Capitan as workstation and Win10 for gaming.
Questions:
1. Which ATI gpu should I use for OSX? The HD7750 or the HD6570?
2. What setup for the array/cache and/or dedicated VM drives should I use?
So far I've come up with these scenarios:
Scenario 1
Array: 3 * 120GB
VM OSX: 1 * 120GB
VM Win: 1 * 64GB
Scenario 2
Array: 2 * 120GB + 1 * 64GB
VM OSX: 1 * 120GB
VM Win: 1 * 120GB
Scenario 3
Array: 4 * 120GB
Cache: 1 * 64GB
Scenario 4
Array: 4 * 120GB + 1 * 64GB
I'm hoping running VMs from an array as in scenario 4 in combination with a virtual 10Gb network would perform reasonably. I can accept some performance loss of the drive speeds but not all the way down to hard disk speeds.
3. Would this work or do I need to invest in some more and/or higher capacity SSDs?
Thanks in progress and: Hi all, first post
**VIDEO GUIDE** How to Install MacOS Mojave or High Sierra as a VM
in VMs
Posted
Yeah, the GT 640 isn't fully supported (no boot screen) but it does boot to login and Metal is supported!
