bastl
-
Posts
1,267 -
Joined
-
Last visited
-
Days Won
3
Content Type
Profiles
Forums
Downloads
Store
Gallery
Bug Reports
Documentation
Landing
Posts posted by bastl
-
-
14 hours ago, Whizadree said:
Having already created a script what increase and decreases the img file its easier and more productive not to turn a running vm , time lost booting . shutting down and reloading all the programs up from boot , could take 10 mins or so , a simple increase by decrease by would be so beneficial to say the least time taken 5 seconds , thats productivity ... and no down time if running servers on vms
Serious question, how often do you change the size of your VMs? 100 times a day? 😂
The majority of users of Unraid, let me guess 95% or even more using Unraid in their home environment or in small offices. That small portion of "office people" will never use that feature as a "productivity" feature as you called it. To have that feature available via the GUI for everyone would be great as long as it is save. Don't get me wrong, but I have read a lot of people here in the chat that already broke there VMs by reducing the size of a vdisk.
-
Go to Settings > VM Manager and enable PCIe ACS override, reboot your server and check if the devices/groupings are split up even further. Without having a device that you wanna pass through in it's own group, passthrough won't work.
-
@Pducharme You can resize a vdisk via the gui for a long time already if you not noticed it. But only if the VM is turned off. Leftclick the name of the VM in the VMs tab and you can see the actual size of the VM. Click the number and you can change the value. Last time I tried to reduce the size this way it worked. The issue is, unraid maybe knows how much is allocated, but it doesn't know where blocks are so to say. If your data is at the end of the file, unraid will cut it of. I know, you can reduce the partition size inside the VM and than reduce it, but still thats not a accurate way to do it.
-
In most cases you need a physical device pluged into the GPU. There are dummy HDMI sticks available to emulate a connected display. In some circumstances without a device connected the GPU wont initialize correctly and you can't use it in the VM. If the card works inside the VM you have different options to connect to the VM remotely such like RDP, VNC or Teamviewer. But for gaming for example, don't expect smooth frame rates like on a usual monitor if you connect remotely.
-
If you only use a dedicated GPU there is no option for a "console". You have to plug in a display to the GPU to see the output.
-
Does it work for reduce the size off the vdisk? If so, than thats the easiest way to break your VM. 🤨
-
2 hours ago, limetech said:
We can add this patch to upcoming Unraid 6.8, however note this caveat
As long as you don't break something else with this patch I think the users will benefit from it. Having a VM in a state where you always have to restart the server to use your GPU again vs. a VM where "maybe" a driver crash can lead to the same issue?! Even if it's not the final solution and only a small improvement, it's still an improvement. As an early adopter of the TR4 platform I know this feeling, every slight improvement in stability is better than none. I'am not actual sure how many people in the Unraid community are still using the vega cards but every now and then people reporting this issue in the forums. I guess there is a demand for this especially for people with MacOS VMS. Nvidia web drivers for Mojave for example are still not available and maybe never will be and AMD cards kinda the only up to date cards that work on newer MacOS VMs. I don't know if you guys have some insights on the numbers how many people using these cards on unraid but I guess there are a couple.
Btw. is there any timeline for the 6.7.3 release and more important for the 6.8? 😁
-
Hey guys,
everone who`s trying to passthrough a Vega 10 card to a VM will know this issue. Restarting a VM isn't possible without resetting the whole server. The card will stuck in a D3 power state until Unraid is rebootet. The early adopters of TR4 and AM4 might already know this guy from fixing some bugs which made it into the kernel and a lot of people using Unraid benefit from his work. He did it again and fixed the reset bug together with AMD for the mentioned cards.
This might be a useful information for a couple users and might also be something @limetech is interested in to implement into the next releases.
https://forum.level1techs.com/t/vega-10-and-12-reset-application/145666
Here is a short video from him showing that it is working right now with a MacOS Mojave VM.
-
1
-
-
Remove the following part in the textfile. I think you have to restart Unraid to take this changes in place.
IFNAME[1]="br1" BRNAME[1]="br1" BRNICS[1]="eth1" BRSTP[1]="no" BRFD[1]="0" PROTOCOL[1]="ipv4" USE_DHCP[1]="no" IPADDR[1]="192.168.2.2" NETMASK[1]="255.255.255.0" GATEWAY[1]="192.168.2.1"-
1
-
-
Click the trashcan should do the trick. At least it did for me as I had that issue upgrading to 6.7.2. This is the corresponding line:
192.168.2.0/24 br1 1 -
It is used by the bridge br1. If you don't use it by any docker or VM remove it and you should be able to remove the 192.168.2.1 route or it is removed by removing the bridge. Can't test it right now. You can also find the network config in the config folder on your Unraid flash drive in one of the text files. Make sure you are removing the right bridge.
-
You have 2 default routes. There should be only one defined.
-
11 hours ago, nuhll said:
make it avaible to the internet
Please don't make RDP directly accessible from the Internet. Please don't!!! There were a couple flaws in the RDP protocol the past months and I have the feeling this won't be the last ones.
-
1
-
-
Tried with a 13GB file and it also works 🤨
-
21 minutes ago, CobaltAUS said:
Maybe throw it back in, restart the docker and see if it fixes it?
I did a quick test with a 1.4GB file. It uploads without issues without that line in the proxy.conf. Weird.
-
@CobaltAUS Thanks for the info. I only have the client_body_buffer_size in that config at the top. Not sure if I changed something in the past. I'am kinda sure I didn't.
-
@CobaltAUS Where exactly you changed it in the let's encrypt docker? I can't find any setting in /config/nginx/proxy.conf inside the docker releated to max_body_size
-
6.8? RELEASE DATE? 😁
Just kiddin. Nice to see that feature will be in soon 😍
-
1
-
-
In general it's a small hidden file. You shouldn't see it either on Windows or MacOS if you doesn't have "show hidden files" enabled. Maybe there is a option on MacOS to prevent it to create these files. What are the downsides of this? I don't know. I'am not using that OS 🤨
-
-
That file is a proprietary Mac/OSX system file that holds attributes / meta-data about the folder it resides in. I don't know if it's possible to stop apple devices to create these files.
-
One option would be to setup a scheduled user script to delete this files.
#!/bin/bash echo "Searching for (and deleting) .DS_Store Files" echo "This may take a awhile" find /mnt/user -maxdepth 9999 -noleaf -type f -name ".DS_Store" -exec rm "{}" \; -
17 hours ago, thatnovaguy said:
Should changing the drive type not help would I be able to pass through the nvme controller like you mentioned without reinstalling Windows?
First of all, have a backup of all your important data and settings, in case something breaks in the VM. As I switched to unraid, my main OS was already installed on that NVME I wanted to pass through. It was kinda easy to have that install working without reinstalling the OS. In your case this should be even easier, because I guess you did the install in an VM and the OS don't have extra chipset drivers or motherboard tools installed. All that I had to clean up. Start with a new VM template for the NVME controller passthrough to not messup your current settings. All your manual edits you can add to the new config like you had set it up before.
17 hours ago, thatnovaguy said:Windows eventually found my nvidia drivers but it took a few minutes.
Are you using the Nvidia driver windows installs itself? If so, switch to the one directly from Nvidia and you should see a performance gain in most games.
17 hours ago, thatnovaguy said:background processes eating up my resources.
As I said earlier, if you pin down dockers to the core 0 and they are under heavy load this can decrease the performance of the whole server. Unraid itself always uses core 0 for itself. If that core is packed with transcoding operations from PLEX for example, everything else will slow down.
Using cores from both dies for a single VM might not be the best option. Sure it will work, but if you have heavy latency dependend operations you might be better off by using only cores from 1 die. In my case the first 8 cores Unraid has access to and all the dockers without pinning them down to specific cores. The other half 8 cores are isolated from Unraid for specific use in one VM only. This way none of the background operations on the server touches the cores used by the VM.
-
I never attached a USB device this way. Only SSDs or harddrives I attached by-id. For USB devices like mouse and keyboard, the VM settings page option to add them never had any issues with that. The Libvirt-Hotplug Addon also does it's job if I needed it. Best option still is to passthrough a dedicated USB controller to a VM to have plug and play function like on a real system.
Low GPU (2080ti) Performance on Win10 VM (Unraid 6.2.7)
in VM Engine (KVM)
Posted
Check your windows power plan. Switch to performance instead of balanced(default). What are the CPU clocks Unraid reports running under load?
watch grep \"cpu MHz\" /proc/cpuinfo