Ph9214

thanks, just wanted to be sure, i'm already mounting with the Unassigned Devices plugin, the share drive feature works so thanks for the tip

can I use the symlink command https://www.cyberciti.biz/faq/creating-soft-link-or-symbolic-link/ to let a network share access and pretend to serve a folder on an "external" drive that is not part of the array, or will that mess up the parity operations or permissions making it cause serious errors or be unaccessable in the share?

I have a 120GB cache drive and 6x2TB drives 1 for parity with a 250GB ssd and another 120GB (identical to the cache drive) ssd my cache drive keeps filling up to 100% and then resetting to normal operation (according to the alert messages) I have 5 separate shares on each of the 5 non parity 2TB drives with just a vdisk in each (so I can do software raid 0 in windows vm), the images are currently 941 GB each in size the shares are apparently using lots of space (~40GB) on the cache drive and I was wondering why they wrote the vdisks to the cache since I haven't written anything to them yet? I was expecting it to just write the changes to the cache drive and not the whole file! Are they just trying to clone the files to the cache drive completely? Also my Shares say that they are unprotected, is that because of them being partially on the cache drive

also, it is to pass through the co-processor, I wanted to run multiple vms on the Phi with visualized cores (so windows won't freak out because of 600+ cpu cores)

Doesn't look easy. https://software.intel.com/en-us/articles/getting-kernel-based-virtual-machine-kvm-to-work-with-intel-xeon-phi-coprocessors Admittedly that article is a little dated, but I couldn't find anything more recent with a quick search. does unraid run on customized redhat? Nope. Heavily customized slackware. But the virtual machine software unraid uses is KVM. Well the guide uses rpm packages and kernel patches, including uninstalling and reinstalling the packages used to manage the kvm environments, is there a way to install unraid as a suit on top of red hat?

Doesn't look easy. https://software.intel.com/en-us/articles/getting-kernel-based-virtual-machine-kvm-to-work-with-intel-xeon-phi-coprocessors Admittedly that article is a little dated, but I couldn't find anything more recent with a quick search. does unraid run on customized redhat?

I was wondering if you could assign cores from a Xeon Phi pcie card to a vm in unraid? Maybe running parallel virtualization processes on the built in micro OS on the phi to emulate 4 cpu cores? Seems promising... Any ideas?

It does work with usb storage devices from what I have tested and I don't know why it shouldn't as it just uses virsh. well, I will try it with more usb storage devices, I think the problem was that (not sure but probably) the storage device was mounted in the host.

I very little (no) experience with extensions (making them) (the installation part seems easy enough) any quick tips for a learning side project? Maybe listing the pcie properties (of lspci) of a selected device? I heave no idea where to start so any help is appreciated.

thanks for the tip I will try that

what I meant to say was streaming of the desktop (and more specificaly the output of the gpu) from a virtual machine not actual video/movie files

Yes I do this to, but, I want to be able to also connect to that same vm over vnc and get the power of the gpu and use it for video stream encoding to get much better performance when connecting remotely

Can I pass through individual vgpus of a nvidia k1 or k2 server gpu accelerator in unraid? can I get any benefit from a k1 or k2 card over a standard gpu in unraid?

I want to make/have made for me a plugin that automatically attaches hotpluged usb devices to a selected vm unless they are already assigned in the xml file of another vm in which case it assigns the usb device to the vm which has it in the xml (I don't know why unraid doesn't allow you to disconnect and reconnect usb devices that already are assigned to a vm) this plugin would basically be a fork of this https://lime-technology.com/forum/index.php?topic=47240.0 but it would automatically do it for a pre specified vm (unless it is already assigned to another vm in it's xml)

I am new to this whole unraid vnc thing, but basically what I want/need to to is to encode the video stream with h264 for improved performance, and I also want to accelerate the vm with a dedicated gpu installed in the unraid server so I can run graphics intense applications in the unraid vm, It seems that the vnc option should be separate from the assigning a gpu option, is there a way I can connect to a vm with a dedicated gpu assigned?

it doesn't work with usb storage devices also, can you make one that automatically attaches anything you plug in to a specified vm

ya exact same issue, wish I had known that fix a year ago I wonder why that works? also, If you can, use the igpu instead of a 3rd card

which motherboard? mine was a gigabyte z97x-ud3h-bk

A while ago when I was trying to configure a VM on my system the VM using a GPU in a PCIE slot that went through the south bridge there were extreme graphical glitches (anything 3d wouldn't show up and some windows had strange white lines through them) is this a known problem (gpus on the south bridge not working properly) I don't have any xml files (the server has since been uninstalled) just wondering if there was a fix I didn't know about.

you may need to enable MSISupport for those devices (had a different issue (audio stuttering), but it might help)

the disks im cloning from are the destination (moving from normal pc to unraid), also I need them to be local to the vm

I have 5x 2 TB drives in a windows software raid 0 array that I want to move to a windows vm inside unraid, what is the best way to do that? I was thinking that I could shrink them to their minimum size inside windows by resizing the partition (they only have about 5TB total on them) and then using the dd command to clone them individually to img/iso (idk which type unraid uses for virtual disks) to a 5TB drive and then formatting them in unraid (and adding a parity drive (unrelated)) and coppying the img/iso files to the disks themselves, increasing the size in the web console using this https://lime-technology.com/forum/index.php?topic=39731.0 and assinging them all to a win 10 vm. would this work? or would I lose all my data? thx in advance

No I just wanted to pass through the gpu, I have no ther free slots, it's not to big of a deal if I can't get it to work, I would just like to know why such a feature hasn't been implemented in unraid to split up groups, is there a problem with ACS override or can the groups simply not be split because of the shared bandwidth and if so why? this is a limitation of the combination of the BIOS in your motherboard and the Linux kernel. It is not something that can be resolved at the unRAID level which runs above these. is it possible to make a patch for my bios and or the linux kernel? Not possible I'm afraid. Your motherboard has 4 PCIe slots. Why not just swap around the hardware so this card is in one of the slots that will isolate in an iommu group (unless all other devices in the other 3 slots, are being passed through already and this is the 4th PCI device you need to pass though) ya, I already have 3 other vms not a big problem.

No I just wanted to pass through the gpu, I have no ther free slots, it's not to big of a deal if I can't get it to work, I would just like to know why such a feature hasn't been implemented in unraid to split up groups, is there a problem with ACS override or can the groups simply not be split because of the shared bandwidth and if so why? this is a limitation of the combination of the BIOS in your motherboard and the Linux kernel. It is not something that can be resolved at the unRAID level which runs above these. is it possible to make a patch for my bios and or the linux kernel?

No I just wanted to pass through the gpu, I have no ther free slots, it's not to big of a deal if I can't get it to work, I would just like to know why such a feature hasn't been implemented in unraid to split up groups, is there a problem with ACS override or can the groups simply not be split because of the shared bandwidth and if so why?