steve1977

Thanks. Yes, running a few of them. Thought dockers are quite low in CPU need as long as I don't do any transcoding (i.e., I am not using Plex or others). So, 1C/2T may be plentyful?

Thanks for your reply. What about leaving one core for Unraid and triple-assigning the remaining five cores in parallel to the three VMs? I will rarely use all three VMs in parallel at full load. Any experience in double- or triple-assigning the same cores to several VMs?

Any thoughts?

I have been using sabnzbd docker for a while and it is working well. Big thanks for the development! I have assigned the sabnzbd a dedicated IP separate from the docker. I had anticipated that there may be some issues related to folder mappings to the unraid raid (which now is on a different IP) and also about the connectivity with Radarr/Sonarr. However, this does not appear to be the case and things are working. Do I still need to anticipate any issues or does this magically work? If so, I am curious why mapping and connectivity still work now?

That's great. Can you elaborate how to do it? Not just download and install on-top?

Thanks again @gridrunner for the video and all the other helpful Unraid videos. I gave it another try with the very latest clover file and now all appears to work for me. Fingers crossed this stays stable for me. One Q - is it helpful and even possible to update Clover to later versions over the next weeks and months. Would I need an USB image again or could do this right out of the VM? Right now, my VM starts a xmas Clover screen upon bootup.

Thanks, very clear. I posted in a separate thread regarding core allocations. Coming back to my other question: anyone else experiencing the issue that LE VM "freezes" frequently. I can still open the VM GUI, but can no longer move between the menus. It shows the Kodi GUI, but that's it and no action possible.

Is this just an issue for those with pass-through GPU or is the standard install broken now?

Great to see this sticky and would love to get advice how to set up the right CPU and Ram assignment. My System: Cpu: Intel i7 7800X: 6C/12T @ 3.5ghz Motherboard: Asus Prime X299-A Ram: 32GB DDR 3 Graphiccards: GTX1050 (host), GTX1050 (VM1) I'd like to run: 1) Unraid with dockers (sabnzbd, radarr, sonarr), no transcoding/Plex 2) VM1: Win10 for gaming 3) VM2: MacOS for Photos 4) VM3: LE Kodi 5) Optional: VM4: Win10 for either gaming or general use (TBD) Would love to seek advice how to assign CPU and Ram. Currently, I have it set up as follows: Unraid: no assignment (not used 0/6, 12GB) VM1: 4/10, 5/11, 8GB VM2: 2/8, 3/9, 8GB VM3: 1/7, 4GB No VM4 setup yet I am wondering whether my VM1 would perform better if I were to assign an additional core to it? Also curious whether I can set VM3 set up more efficiently and whether helpful to assign additional ram to any of the VMs? If I were to set up VM4 as additional gaming VM (or all-purpose VM), how would I need to change my config? Thanks in advance and happy new year!

Thanks for your immediate reply. I wasn't aware. I don't need to assign any core to VMs? I.e., not ticking any box when setting up the docker? I also do have one gaming VM. I noticed very visible performance differences when assigning 1, 2 or 4 cores. Would not assigning any core make things more flexible and actually better than assigning just 2 cores (i.e., using maximum number of cores if available)? Let me focus on LE VM (without passthrough) then instead of docker, which indeeed may be easier for me. Fingers crossed the LE VM image for Unraid will eventually see some update. It feels abundant now.

I am hesitating between trying to get the docker or the LE VM to work. Is @Squid's recommendation still unchanged to go with a VM? My issue with the VM is that it has not been seeing recent update for the LE headless image and that it freezes when running for some time. While I don't mind the RAM overhead, I do mind that I need to assign a CPU core. Curious why you didn't mention this?

I am facing the same issue. I have created two Windows 10 VMs a few months ago and worked flawlessly. Now, this no longer works and I only got the shell upon startup / creation. I need to use OVMF, so switching to Seabios is not a solution. Also, exactly the same computer with same files worked before flawlessly with this setup. The only difference I can think of is that I upgraded to the 6.4.0_rc15e. Could this be the issue? Any thoughts appreciated!

are things easier if not passing through a gpu? is the clover update important or can i just keep an “old” version? main use case of my mac vm is mac photos. does the photos app benefit from use of a passed through gpu? unfortunately, it’s quite slow / sluggish for me.

Great to hear. This reads complicated. I thought you could just upgrade and it works. Why do you need to reinstall the graphics driver again and what do you need to do with Clover? Anyone tried a fresh install recently and can confirm whether latest Clover driver fixes the "black screen" issue?

Any thoughts whether this is now working with Clover later than 4200?

I have some structural issues that prevent me from opening ports on my router. I have been thinking of work-arounds and having a VPS may be a good and safe solution? I got inspired by this article https://jordancrawford.kiwi/home-server-without-portforward/. Would love to get your thoughts whether that's possible with Nextcloud running in an Unraid Docker. If I read this correctly, I may not need letsencrypt in this case, but it would require me to tunnel the Unraid Nextcloud docker to my VPS through some form of VPN? Any thoughts and advice for me?

I don't perceive this at all as picking on me. Actually, I need to say that community members of this forum in general are among the most polite and helpful compared to all other forums I am engaging. Part of the reason what gets me interested to pursue more and more new Unraid related projects.

Got it thanks. Just browsed this threat and realized that a native implementation within the app is the most requested new feature. Is there still hope for native support?

Thanks. I have a VPN client installed on my router. What you are referring to is a VPN server in a docker, right? If I were to succeed to set this up, would this allow me to access my full Unraid even including all dockers and VMs? Would my routerVPN (client) conflict?

Has there been any update on remote access? if still “just” VPN, can you advice how this works? Sent from my iPhone using Tapatalk

Mmh... How do you access the three GPUs? Do you have screens attached or using some form of remote desktop? Does the issue only happen if you run all three VMs in parallel or also individually? I still face some problems when running two Win VMs in parallel, which I don't face when not passing through the GPU. If I may ask, what are you using three GPUs in paralle for and do you have experience with Gamestream?

I have also spent many hours (if not literally days). I still don't have it fully working, but at least made some meaningful improvements. Have you tried to use OVMF instead of Seabios? OVMF appears to work a lot better. Also, are you using RDC, which may be another source of issue?

Thanks. Sparklyballs mentioned three issues though. Do you suspect the closed ports causing the first two issues? Why is 445 "working", but 444 not.

Thanks. I had read through it and tried to follow it. I have no clue what is occupying port 444. I assume from this thread that 443 is taken by Unraid. I can see i the docker settings what other dockers are using, but port 444 is not among them. Is there some other software or plugin for Unraid that can tell me what ports are being used by what? I have signed up for duckdns amd included all respective information (following the how-to). Duckdns is running. Not sure what I am missing within Letsecrypt though?