MichaelBernasconi

Hey ELP1, I'm currently running a Ryzen 5950X and a RTX 3070Ti on my Unraid server. I have 8 full cores (16 threads) and the 3070Ti passed through to my gaming VM (Windows 11). FS2020 runs fine for me. I have not compared against bare metal performance but I'm not noticing any lag that is bothering me. A few things to check in case you have performance issues. 1. Make sure the vdisk that you install FS2020 on is located on an SSD. This made a big difference for me. 2. Make sure you allocate enough RAM for the VM. I'm giving it 16GB. 3. In case your CPU has hyper threading (which it probably does) make sure that you are passing both threads that belong to the same core to the VM. I hope these tips help. Best, Michael

I'm hosting a wireguard server hub & spoke access vpn on a server I have. I was able to connect several machines to this vpn already and everything seems to work fine. Now I'm trying to connect my Unraid server to this vpn. I've tried using the "import tunnel" button in the GUI but it does not seem to work. Does anyone know if this is even possible with the GUI? Best, Michael

Great to hear that things are starting to work for you! I'd suggest moving the steam library to a separate vdisk. I've also encountered some problems in the past where games did not like being installed on a network share. Since you won't be able to use the same steam library from 2 VMs at the same time I don't see any reason to use a share instead of a vdisk. I've also had the best IO performance with vdisks.

How do these games not work? I've never encountered a game that does not work on my vm except for the anti cheat reasons I mentioned. I did encounter some sound problems a while back. If I remember correctly I was able to fix the problem after taking a look at this thread.

I've been running unraid for almost 2 years now. I have a windows 10 gaming vm and a linux vm for work. There is a bit of a learning curve in the beginning but I'm very happy with my setup now. I was even able to swap you my CPU and mainboard without any issues. So from my personal experience I can recommend using unraid with a windows 10 vm for gaming. Just be aware that you will not be able to play certain games because their anti cheat does not allow you to play in a vm! Valorant and Faceit (CSGO) don't work because of that. If you have any questions about my setup let me know. Best, Michael

Im running a Linux (PopOS) VM on my Unraid server and have noticed a large performance difference between shares and vdisks. The vdisk is passed to the VM as follows: <disk type='file' device='disk'> <driver name='qemu' type='raw' cache='writeback'/> <source file='<path_to_the_vdisk>' index='2'/> <backingStore/> <target dev='hdc' bus='virtio'/> <boot order='1'/> <alias name='virtio-disk2'/> <address type='pci' domain='0x0000' bus='0x03' slot='0x00' function='0x0'/> </disk> Note: The vdisk is stored on the array (not the cache drive) Shares are passed to the VM as follows: <filesystem type='mount' accessmode='passthrough'> <source dir='<path_to_the_share>'/> <target dir='<share_name>'/> <alias name='<share_alias>'/> <address type='pci' domain='0x0000' bus='0x05' slot='0x00' function='0x0'/> </filesystem> And then mounted with fstab like: <share_name> <path_inside_vm> 9p trans=virtio,version=9p2000.L,_netdev,rw 0 0 Note: The share has caching set to "Yes". Drive info: Array: 2x Samsung 870 EVO 4TB (one for parity) Array file system: xfs Cache: Samsung 970 EvO Plus 1TB Cache file system: btrfs I tested the read and write speeds by copying a large file. Here are my results. share -> share: 27 MB/s vdisk -> vdisk: 600 MB/s share -> vdisk: 56 MB/s vdisk -> share: 48 MB/s Why is the share performance so much worse than the vdisk performance? Do you have any suggestions on how to get better share performance?

Hi, I'm running an Unraid server which I use for deep learning. I use the Nvidia-Driver plugin to pass my GPUs to a docker container which then trains my models. Since my GPUs sometimes run hotter than I would like I'm looking for a way to control the GPU fan speeds. So far I have not been able to get nvidia-settings working in the docker container or on Unraid itself. Is it possible to manually control the GPU fan speeds on Unraid? In case it is possible, would it be better to control it from the docker container or from Unraid directly? Thanks in advance, Michael

First of all a big thanks to @dgaschk for posting this solution! After upgrading my CPU and mainboard I ran into this issue. Your solution fixed the problem without any further issues. Thanks a lot! I am, however, curious about the origin of this problem. Could you explain to me what these commands actually do? Which "console" us unbound? Why does this problem exist in the first place? Best, Michael

I know that these questions have been asked before but I found conflicting answers. So to be sure I decided to ask again. I've been using unraid for almost 2 years now. I use it to run VMs for work and gaming as well as some docker containers. Currently I have an array with HDDs and a single SSD cache drive. This works great but I could use some better read performance. Since I don't really need any HDDs I wanted to simply move to an all SSD array. I found conflicting information about whether or not this is save or not. So my questions are. 1. Is an all (SATA) SSD array possible / save to use? 2. Do I need to do something special to get it working? 3. Do I need some special SSDs? In case I do, do you have recommendations (~4TB per drive)? I am aware that I could simply keep the HDD array and add an SSD cache pool. In case the all SSD array is not an option I'll do that. However, an all SSD array would allow me to then run an NVME cache for even higher performance where it is needed. Thanks in advance, Michael

That is interesting. I followed the instructions in this video and got the image you can see at the bottom. Since there is only a single package I assumed that there is only a single die. Looking at the topology again I assume there are 3 cores per CCX sharing a single L3 cache. Is that correct? Just out of curiosity. Could you explain the difference between a CCX and a package?

Thank you for the quick responses! I am aware that using cores from different dies is bad but this should be a problem for the 1920X, right? It has only a single die as far as I can see. Interesting. Is there any hope that this problem will be fixed in the future or is it an inherent problem with Zen/Zen+?

I did not. I just installed 'tip and tweaks' and set the CPU governor to performance. I now get ~25 fps. So thank you for the suggestion! Unfortunately the performance now is still below what I would expect it to be. Do you have any other suggestions?

Hey everyone, I recently installed Flight Simulator 2020 on a Windows 10 VM and the performance is quite bad. Naturally I would like to find (and hopefully fix) the problem. The VM has 2 vdisks attached. Vdisk1 is stored on the cache (fast SSD) and contains the OS + the game's cache. Vdisk2 is stored on the array (HDD) and contains the rest of the game. Here are my findings so far. - FlighSim2020 runs at 15-20 fps - Changing graphics settings does not influence fps at all - The game's dev tools report the main thread to be the bottleneck - GPU utilization is at ~25% - Task manager shows that 1 core is at ~75% utilization while all other cores are at ~40% - RAM does not seem to be a limiting factor. No performance difference could be observed between 16GB and 32GB - The number of cpu cores assigned to the VM does not influence performance (Tested with a minimum of 4 cores / 8 threads). My system: CPU: Threadripper 1920X Mainboard: ASRock X399 Taichi GPU: RTX 2080Ti RAM: 64GB (3000Mhz) From the above I can see two potential problems. First, single core performance could be bad on the 1920X which would explain the bottleneck. This, however, is unlikely as I have seen the game running much better on and i7 6700 HQ. Second, memory latency could explain everything. Is it possible that Unraid somehow introduces a lot of memory latency to the VM? In case it is, how do I fix it? Does anyone have an explanation or even better a fix for my problem? Best, Michael

Hey guys, I trying to run some GPU heavy tasks (AI training) inside a docker container. So far it works great except that the GPUs run very hot. I had a custom script that controlled fan speeds to keep the temps low on Ubuntu but I'm unable to run that script inside the docker or directly on Unraid. The problem is that the nvidia-settings tool does not seem to work. I get the following message "nvidia-settings: error while loading shared libraries: libXxf86vm.so.1: cannot open shared object file: No such file or directory" Does anyone know how to get around this problem? Best, Michael

Cool! Thanks! Intelligent caching would for sure be a very cool addition to Unraid.

Thanks! I will try to setup some kind of backup process to my cache drive. Just out of curiosity. Is there a reason why the kind of caching I described is not a standard feature of Unraid? To me it seems that having fast read access to data while at the same time having a secure copy of it on the array would be something many people are interested in. I understand that adding a second cache drive for redundancy kind of does the same thing but it seems very expensive and unnecessary given that the array already exists. Another advantage would be that in case of a cache drive failure Unraid could just continue using the (slow) data on the array until the cache drive is replaced. This way the data would not have to be restored manually from some kind of backup.

I took another look at the description in the Unraid gui and i also watched SpaceInvaderOne's video about cache. Either I am really missing something here or we are not talking about the same thing. What I would like to have is a share which is stored on the cache AND the array. I want the data to be on the cache drive at all times for the read speed and I want it to be on the array for the parity protection. The available cache settings are the following. This one is clearly not what I want. The cache drive will be ignored. This is also not what I want because when the mover is invoked, files and subdirectories are transferred off the Cache drive. This means they will no longer be on the cache drive, right? This would give me the read speeds I want as the data is always on the cache drive but it would never be written to the array. So it is also not quite what I want. This is basically the same as Yes but it does not complain when the cache is full. So the files are still transferred off the cache drive to the array when the mover is invoked. So my question remains. How do I make a share that is both on the cache AND the array? Essentially I would like something like the Yes setting but the mover does not remove the data from the cache drive. I realize that I could set my share to Only and have a script periodically backup the share to the array. However, this would not allow Unraid to automatically use the share on the array in case the cache drive fails. Best, Michael

Thank you for the quick reply. About the cache pool. As I understand it it's main objective is increasing the perceived write speed. This is done by first writing data to the fast cache drive and then later moving the data to the array. My problem with this solution is that the mover actually removes the data from the cache drive. As far as I understand there is no option for the mover to keep the data on the cache drive and the array at the same time. Is this correct or did I miss something? I will take a look at CA Backup to see if it can do what I want. Best, Michael

Hey, I just setup an Unraid server which is mainly going to be used for running VMs. I would like to have fast access to each VM's vdisk. I also have some data which I need fast access to. Of course I would also like to have this data parity protected. How do I do this with Unraid? As I understand adding a cache drive will only improve my write speed (assuming I have the share set to "prefer"). So what I would like to have is a cache pool which is periodically backed up on the array. Does this exist in Unraid? Is there another way of achieving what I want? Note, I know I could add a second cache drive and set it to raid 1 but this seems wasteful since I already have an array which is parity protected. Best, Michael

Hey, I'm trying to get VMs working on Unraid. I have a server with 2 GPUs (2070 Super & 2080 Ti). I created a new Ubuntu VM and got it running without a GPU first. This worked perfectly. Then I passed the 2070 Super to the VM and installed all the Nvidia drivers. At this point everything still worked perfectly even after one or two reboots. So the next thing I wanted to test was to pass the 2080 Ti instead of the 2070 Super. At this point the VM refused to boot. Now it does not boot even if I only pass the 2070 Super or no GPU at all. Here is a list of things I tried to fix this issue but non of them worked. 1. Pass GPU with or without the sound card 2. Manually edit the XML to pass all four functions of the GPU (I tried setting multifunction='on') 3. Rebooting the server Does anyone have an idea how to fix this problem? Any suggestions would be greatly appreciated. Best, Michael

Thank you very much for the quick reply! Having the data only on the cache drive and then making a backup from time to time seems like a reasonable idea. Do you have any idea how I managed to get such good disk performance in passmark? The VM I tested on is only 20GB and completely fits in RAM. Could it be that Unraid somehow moves the whole VM into RAM?

Hi everyone, I have an Unraid server on which I'm running a bunch of VMs. My array consists of 2 of HDDs (no parity yet) and I have a fast SSD around that I'm currently not using. I ran passmark on a Windows 10 VM and got a result that I do not understand. According to the results my disk performance is in the 99th percentile. How can this be? My 2 HDDs should not be fast at all, right? Now on to my question about cache. My idea was to have files that I want to access fast on the SSD (cache) and also on the array (for backup). That would mean that the VMs only ever interact with the SSD and the mover writes back the updates to the array when it has time. As far as I can see Unraid does not really have that option. Both the "prefer" and "only" mode would result in the data being stored on the cache drive only while the "yes" and "no" modes would result in the data only being stored on the array only, right? Did I get something completely wrong here? Best, Michael

Thank you very much for the replies! Daedalus asked why I do not want to use the Linux subsystem for programming. I do have experience using this and it works quite well. However, the whole reason why I'm looking into unraid is that I would like to separate things as much as possible. That means that I do not want to use the same OS for gaming as I do for work. About question 4. If I install a VM directly onto an SSD do I need to give it it's own fixed size partition or is unraid able to give a VM more size when needed? So could I just install a couple of SSDs and then put as many VMs on them as I want and unraid will handle the exact allocation for me? From what I have heard so far it does sound like unraid might be the thing I am looking for. Since I'd be buying a whole new setup to use with unraid I wanted to ask if there is anything I need to pay attention to when buying hardware? I know that for the PCIe passthrough my CPU and my mainboard need to support IOMMU passthrough. I think most modern CPUs (AMD and Intel) and mainboards support this, right? Is there anything else I need to look out for?

Hi everyone, I am trying to figure out whether or not unraid is the right tool to fulfill my requirements. Currently I use a natively installed windows 10 which I use for leisure activities (gaming, browsing, etc). I also have virtual box installed with a bunch of linux VMs which I use for work (mostly programming). There are two reasons I use linux VMs for work. The first being that linux is nicer for programming, the second being that I like to keep things separated as much as possible. However, VM performance kind of sucks which makes heavy tasks such as machine learning a pain. My solution for that is to use windows for such tasks but then work and other stuff starts mixing again which I would like to avoid. So my idea for an unraid system would be to have a CPU with a high core count and two GPUs (a high end one and a mid range one). The idea would then be to have the ability to flexibly assign cores and GPUs to different VMs depending on what I currently need. So lets say I want to train some neural network on the high end GPU and play some (CPU heavy) game while waiting. I would then start the linux VM with the NN with the high end GPU and just a few cores and also start a windows VM with the mid range GPU and the rest of the CPU. This should in theory let me do both things simultaneously without too much interference, right? Now that you hopefully understand how I plan on using the system on to some more concrete questions. 1. Do you have an easier solution for my requirements? 2. Will unraid let me easily assign different resources to a VM before it is started? 3. Will linux / windows complain if they get assigned different GPUs (or no GPU at all) every time they are started? Let's assume I only have Nvidia GPUs. 4. Is it possible to save the data for each VM on a large (and redundant) HDD and then load it onto an SSD once it is started? In other words basically use the SSD as a cache? 5. How good is remote desktop performance between two VMs? In case it is bad is there any other easy way to use multiple VMs with the same peripherals at the same time? If you read through all of the above thank you for your time!