JonMikelV

Wow - your patience with idiots like me is incredible! Turns out I've spent HOURS (days?) trying to find (or create) the "extra parameters" field in "Show more settings..." when I really just needed to toggle the "Basic View" / "Advanced View" switch in the upper right of the page. D'Oh! I wonder if I was confusing myself with the Docker overview page "Advanced View" (that shows the CPU & memory) thinking it was the the Docker *EDIT* page "Advanced View" that I needed. Either way, all resolved - thanks!

Hmm... I'm sure I'm not confusing the CPU percentage shown above the line with the memory usage shown below it. I guess trying to explain what I was thinking of just made my post more confusing, not less. 😞 I must be incorrectly using the --memory parameter from: I've tried this test as EXTRA_PARAMETERS and standalone Variable values with & without double-dashes but still see max memory in the usage column: Resulting in: Sorry for turning this into a standard Docker question - am I able to move the discussion out of Feature Requests?

On the Docker tab, the "Advanced View" helpfully shows memory load as "currently used" / "system memory". But if this is truly memory LOAD should it reflect usage of the "--memory" parameter? For example, if I have a container using 8G of memory on a 32G server it shows as "8.00 GiB / 32.00 GiB" which suggests 25% usage, but if the container were run with a "--memory=10G" parameter then the container would actually be at 80% usage. On a related note, it might be nice to include a total docker numbers - like the total CPU & memory used by ALL dockers.

I've played with a few VMs created from scratch without (much) issue, but in testing migrating a physical box to a VM I'm running into super slow disk IO (like 100-200 KB/s max). I'm wondering if because the .img file was created outside of my no-COW "domains" folder, maybe it "carried in" the C attribute when I moved it under "domains"? I tried lsattr on the file but end up getting an "lsattr: Inappropriate ioctl for device while reading flags" error. Does anybody have any suggestions on how to check Copy On Write attribute for a file in Unraid or know if it is removed when a file is moved (or copied) into a no-COW folder?

@Zudnic, did you create the `/mnt/user/ubuntudata` folders before running the `mount` command?

@pmeth, intuerrupts is a good thought - are you passing through any hardware (even USB ports) that might be getting grabbed by the first machine then appear unavaiable to the second?

I've recently started having issues where No VNC connections to my VMs randomly lock up after anywhere from 10 min. to 10 days. The VM page shows them still running and I can ping the machines, but any No VNC connection (not matter from what 'client') just connects but shows the VM screen from the time of the lockup (based on the clock). I have been getting around it by force-quitting and restarting the VM but would prefer to figure out how to diagnose and resolve this. The only VMs I use long enough to notice the issue are both Ubuntu based, if that is likely to matter.

I'm no good at reading the diagnostics, but does the issue go away if you allocate specific CPUs to specific VMs? I've run into similar issues VM crashing issues if I run multiple VMs where the sum of the allocated memory is too close to actual system memory.

Have you created other VMs successfully? Are you able to ping the VM once it's started?

it looks to me like you CAN create the VMs and they do boot, but the installs "fail out" when trying to do network access during the installation? If that's the case, sounds like one of your listed changes (probably not Letsenrypt) might be the culprit. Is your Jitsi IP range in the 192.168.86.xx address space seen in the second image? I'm wondering if your VMs are connecting to the IP range and not finding a DHCP server to give them an IP address.

I assume you're not attempting any video card pass through? As @Vannie suggested, ports might be an issue - does the built in Unraid VM page "VNC Remote" feature give the same issues as standalone viewers? When you get the blank screen are you still able to ping the VMs (or even connect to any shares that might exist)?

Thanks for sharing! I may give PrimeOS a try one of these days, but a while ago I gave in and picked up a cheap Chromebook at a holiday sale to play with. Figured it was cheaper than all the time I was spending fighting with the VM. :-)

Why the bashing? Just because I don't know Proxmox doesn't mean I don't know VMs. (Plus you confirmed it's not supported - seems a solid suggestion to me.) One person's "bullshit" is another's avoidance of technical details about drive exposure until sure the audience is ready for it. You shouldn't feel the need to apologise for your opinion, it's just as valid as anybody else's! Plus, thanks to you I know more about Proxmox use that I did before. 🙂

Installing Unraid would replace Win 10. If you don't need Win 10 functionality you could run Plex in Docket container in Unraid. This is likely the most effective way to do it. If you still want Win 10 on the box then you'd have to create A VM and install Windows in that VM. At that point you could run Plex in the Win 10 VM instead of (or in addition to) an Unraid Docker container. If using a VM then performance will vary depending on what physical hardware you pass through (dedicate) to the VM. For example, a default VM see a generic video card which means if you wanted something running in the VM to use you the GPU on your physical video card you would have to pass it through to the VM. Does that answer your questions or did I make things worse? :-)

If you're planning to play with Unraid under Proxmox that should be fine but I wouldn't recommend it for long term use. I know little of Proxmox but I would guess that Unraid wouldn't have the right kind of drive access to safely "do it's thing" if running in a Proxmox VM.

Welcome @BrazMan! 1. Not having used torguard myself I can't say for sure, but I suspect you can keep using torguard VPN but exactly how it's done will depend on how you decide to run Plex. Will it be in a Docker container or a VM (running Linux or Windows 10)? 2. I see ToGruard Docker containers at DockerHub so it's likely there is one for Unraid, but I haven't checked yet. 3. See answer #1 (how to set up remote Plex access will depend on whether Plex is running in a Docker or VM but consider watching Spaceinvader One's excellent video on containers and VPNs). 4. See #1 (setting up direct GPU use depends on if Plex is running in Docker or a VM). Probably the "easiest" for you to relate to would be to put Windows 10 in a VM passing your video card through for direct VM use, then just do everything the same way you have been. However, that would not necessarily be taking the most advantage of using Unraid. To do that you might want to consider running Plex in a Docker.

Here's what ultimately worked for me. ~~~ VM Config: Unraid Share: /mnt/user Unraid Mount tag: user ~~~ Addition to /etc/fstab: #2020-04-14 Sample auto-mount via "Unraid Mount tag" user /mnt/user 9p trans=virtio,version=9p2000.L,_netdev,rw 0 0 I've read that 9p is actually pretty bad for performance but I've never been bothered enough to figure out an alternative. 🙂 Note that I got a pretty big speed boost by adding msize=262144: user /mnt/user 9p msize=262144,trans=virtio,version=9p2000.L,_netdev,rw 0 0 Of course I only tested with a simple dd command, so YMMV: dd if=/dev/zero of=/mnt/user/testfile bs=1G count=1 oflag=direct; rm ~/mnt/user/testfile

I'm a glutton for punishment so looked anyway - I can't get past "Booting 'local image A'".

Thanks for the link @sharpfork. I had read through that as well and decided it sounded like a lot more work than it probably needed to be - at least here 2 years later. 🙂 What I've tried so far is: - download the CloudReady image from https://www.neverware.com/freedownload - unzip the cloudready-free-74.4.57-32-bit.bin (or cloudready-free-74.4.58-64-bit.bin) file into my ISOs folder - Create a VM with: + Logical CPUs: 2 + Initial & Max Memoru: 2G [min. spec. for CloudReady] + Machine: Q35-2.1 + BIOS: OVMF + USB Controller: 3.0 (qemu XHCI) [boots faster than USB 2] + OS Install ISO: path to cloudready-free-74.4.58-64-bit.bin + OS Install CDRom Bus: USB + Primary vDisk Location: Auto [created as 16G drive - again min. spec. for CloudReady] + Primary vDisk Bus: SATA + Graphics Card: VNC + VNC Video Driver: Cirrus [If you left the default of QXL (best) I didn't see any of the boot screens mentioned below] + VNC Password: <blank> + VNC Keyboard: English-United States (en-us) + Sound Card: None + Network MAC: <whatever defaulted> + Network Bridget: br0 + USB Devices <none> + Other PIC Devices: <none> With those settings, I am able boot the VM and see: - the "TianoCore" logo screen for about 3 seconds - a blank screen for about 8 seconds - the CloudReady boot logo for about 1 second - the animated "Your system is repairing itself. Please wait." page for about 3 seconds - the whole process start over Eventually I have to force-stop the VM. Per the Neverware support pages, this sort of boot loop is usually indicative of unsupported hardware of some sort so I've decided there may be something about the specific hardware my Unraid is running on that CloudReady doesn't like. I might be wrong, but I haven't made time to play around with anything else (like the 27 partition process in the link you provided).

I doubt you'll hear much about actual ChromeOS. As far as I know Google doesn't provide any self-install type options (such as an ISO). That being said, ChromeOS is Google's customized version of the open source ChromiumOS (just like the Chrome browser is Google's customized version of the the Chromium browser). And just like how there are other customized versions of the Chromium browser (such as Brave, among others) there are other customized versions of ChromiumOS. This website has a nice comparison of the most common ones: - https://www.electromaker.io/blog/article/flint-os-vs-chromium-os-vs-cloudready-which-chrome-os-is-best To summarize it - as of Jan. 2019: - ChromiumOS - 'the original', open source, not necessarily the most polished or feature packed, - FydeOS (formerly FlintOS) - many variants (such as for RaspberryPi), supports Android apps, etc. - CloudReady (also formerly FlintOS) - more enterprise focused, supports CloudReady Linux apps, etc. So you might want to do some searching for ChromiumOS, FydeOS, or CloudReady and consider trying one or more of those as VMs on your Unraid box. If you, come on back and share your own experiences and thoughts! 🙂

I like shiny new things and haven't had any major issues with updates in the past so decided to try 6.6.7 -> 6.7.0 even though I'm not near the box right now.... Oops. 🙂 30 minutes later I'm still at the "System is rebooting" page and while the device replies to pings it won't accept (actively rejects?) connection requests to the GUI. I have SFTP access and one of my remotely visible auto-start Dockers is running but another isn't and my auto-start VMs aren't. I have remote access to other machines on that network and the Unraid SMB shares do seem to be visible. I'm fine waiting until I have physical access to the device but if people have suggestions for remotely (via SSH) debugging I could poke around a bit. 🙂 Edit: Thanks @moose, apparently it just got stuck shutting down - once I physically power cycled it everything (including the expected loss-of-power parity check) was fine - hello 6.7.0! 🙂

I'm curious how you figured out what plugin was causing the unexpected spinups.

I'm glad you figured it out! I hadn't answered because I've not used privoxy before. In case in run into the same issues you had, what was the variable you needed to add?

I've found a few posts about running Streambaby (a Java based TiVo video streaming app) on unRAID but they seem to be so old they're running it ON unRAID, not in a Docker. I checked Apps (and DockerHub) for Streambaby but found nothing so I'm curious if it's possible to run Java (and subsequently Streambaby) in a Docker container along with why I might NOT want to do so. Thanks. Jon-Mikel

Another apology for bringing this thing back from the dead, but I've poked around a bit and still am not finding anything that sounds like a "definitive answer". According to the UnRAID manual 6 section on "The Engine" it sounds like Copy-on-write being enabled (set to "Auto") is good for Docker containers because it lets them use the same base OS image until one of them needs to make a change to it, at which point a copy is made and the "customizing" container user the copy while all others continue to use the original I suspect it's DISABLED (set to "no") for the "domains" share (which is for holding saved VM instances) because of the overhead associated with making copies of the VM file (or changed blocks of the file) every time the guest OS makes changes to the disk image. But that's purely a guess.