letrain

when i have the igpu-vf passed through it vm works fine via remote desktop. but when i change it back to VNC in the unraid GUI it never initializes the display. i've had these issues before after passing through a nvidia gpu as well. not sure why i can't go back to vnc after using passthrough on a prior vm boot.

@zhtengwi'll be sad if this plugin dies. i've seen some forks on github, but not updated. i see its just some scripts, a page for sr-iov, and just updates the module for the kernel (i915.ko)?. i wonder if it would be possible to fork and just update the module, but i'm not familar with plugin creation, or if simply updated the module for 6.1.49 would be the only needed fix. added: it does appear that the i915.ko driver exists already. so would this plugin only need to install the conf file, add the page to emhttp, and patch libvirt.php? it appears the kernel module is the failure point, causing it to remove the conf files, and libvirt edit, but not emhttp edit. but this page is useless as it also removes the i915-sriov folder so the config will not stay. i'm not familar enough with scripting but could i just removed the kernel module part from the plugin script? or does this plugin include a different i915.ko driver? got this far.... its late an i'm past my knowledge. i915 0000:00:02.0: driver does not support SR-IOV configuration via sysfs

don't bind to vfio. it creates the VF's then just add those on your vm template page, read a few posts prior this exact question has been asked. sr-iov is not VFIO. also until this is updated for 6.1.49 it will not create vf's

thats what it did for me. looking at docker hub tag 27.0.1 and latest appear identical.

How do you apply this?

So what the final solution here? Everything in the list needed? Custom kernel? Rom resize script? Everything works bare metal. But it won't work in the vm. Unraid 6.12.3 here, is 6.12.1 needed? Was there regression? Host BIOS Enable ReBAR support - Host BIOS Enable 4G Decoding - Enable & Boot Custom Kernel syslinux configuration (near beginning of this thread) - Boot Unraid in UEFI Mode - VM must use UEFI BIOS - VM must have the top line of XML from <domain type='kvm'> to: <domain type='kvm' xmlns:qemu='http://libvirt.org/schemas/domain/qemu/1.0'> - VM must have added the following (after the </device> line, before the </domain> line): <qemu:commandline> <qemu:arg value='-fw_cfg'/> <qemu:arg value='opt/ovmf/X-PciMmio64Mb,string=65536'/> </qemu:commandline>

ok what did i do wrong? its enabled in bios. asrock calls it Clever access memory. it is on in the bios. the xml is edited. everything is on. i ran the user script.... and the vbios is up to date for the 3090.. @KptnKMan it works in bare metal..but i just can't get it to work in the VM.

the script doesn't work anymore.... it is already bound to vfio. passthrough is working fine, but it keeps saying tat rombar is not enabled in bios so i figured getting the vbios might help it work better. oh well. "Okay dumping vbios file named GeForce RTX 3090 to the location /mnt/user/isos/vbios/ cat: rom: Input/output error Um.... somethings gone wrong and I couldn't dump the vbios for some reason Sometimes when this happens all we need to do to fix this is 'stub' or 'bind to the vfio' the gpu and reboot the server This can be done in Unraid 6.8.3 with the use of the vfio config plugin or if you are on Unraid 6.9 or above it can be done directly from the gui in Tools/System Devices .....So please do this and run the script again"

how do i prevent remux from converting all my mkv files to mp4 first, is there a way to limit this to avi files only?

hasn't hit CA yet. i'll check the github

thanks for the reply. it was worth a shot. i'll patiently wait. i've lived without it before, i can again. was using gvt-g, and now had sr-iov on new setup. i dont shut down often but i just upgraded my hardware and thought that might of been causing my unmounting during shutdown problem. or my boot usb was starting to go. tried several things and the update fixed my shutdown issue. its not often but reboots and shutdowns do occur.

i'm was using the sr-iov plugin in unraid 6.12.2. i upgraded to 6.12.3 because i did have the docker unclean shutdown problem. no sr-iov doesn't work.. i found this in the logs. "Jul 15 12:17:33 Tower kernel: i915 0000:00:02.0: driver does not support SR-IOV configuration via sysfs" wondering if its a driver issue. thanks

unraid 6.12.3 was just released. it broke sr-iov. worked fine in 6.12.2. 6.12.3 - "Docker This release resolves an issue where Docker does not properly stop when the array Stops, which can result in an unclean shutdown. If Docker containers have issues starting after a while, and you are running Plex, go to your Plex Docker container settings, switch to advanced view, and add this to the Extra Params: --no-healthcheck" I had the this problem with dockers and unclean shutdown, and disks not unmounting so upgraded. after upgrading sr-iov was missing. after trying to reinstall it this is the dialog. it appears in settings again, but as soon as i try to save config, and enable it doesn't work. thanks EDIT: found this in the logs "Jul 15 12:17:33 Tower kernel: i915 0000:00:02.0: driver does not support SR-IOV configuration via sysfs" not sure what happened between 6.12.2 and 6.12.3. its weird that sr-iov is in settings, but not on the plugin page. and "apps" still gives me the option to install, but fails.

any luck getting time to work correctly?? its right on the time stamp but the images are way off. they are being written as 7pm being 00-00-00! when i change this in the config the docker won't even start.

second this. i'm very interested in this setup as opposed to an HA setup with one wanip address.

works fine for me. did you ever figure it out.

any luck figuring this out? it just started for me.

Docker-wyze-bridge brings me here. Should I go to GitHub? It is a template question but I don't see the template listed here.

its been a while. I'm not back to wanting to mine in a nice t-rex docker with 1 gpu... only now i'm getting this error. 20220501 15:15:29 TREX: Can't find nonce with device [ID=0, GPU #0], cuda exception: CUDA_ERROR_OUT_OF_MEMORY its is a p2000. i mined just fine before, and have been happily mining with it in windows up until now. its worked fine for ETH mining. its right at the GB limit. i did have to run the special script in windows so its was "headless" and would be 100% free to mine. but i've never had this issue in t-rex and have no idea how to fix. EDIT: UnRaid Version: 6.10.0-rc5 so... went to "ptrfrll/nv-docker-trex:cuda10-latest" and it seems to be working. i was on "latest" because nvidia-smi = NVIDIA-SMI 495.46 Driver Version: 495.46 CUDA Version: 11.5 so i was using latest t-rex because cuda 11. is this going to be an issue going forward?

nice. thank you for all your efforts. i really do think usb-ip is underrated.

i'm on rc3.... i uploaded my diagnostics just in case. i'm not sure whats going on. i thought it might of been dynamix plugin of some kind. i recently installed his file manager and wasn't sure if something with the gui messed up... i've used your usb plugin before. i was trying out virtualhere so i removed all the usb stuff. now i've readded. is there anything else i can send over that could help? it could just be my server setup? i followed the wiki. i added them to usbip-bind@hardware:id.service instead of 1-1.2 this time. not sure if thats the issue. i can see them just clicking attach nothing happens. thanks! edit : just saw fix common problems plugin is throwing an error at me. tower-diagnostics-20220313-1417.zip

not sure if its because of resent updates on the RC branch but the clicking the "Attach" button does nothing. it sees my server, and my devices, but clicking attach does nothing.

Search the topic for "stereo" you will find the custom repo from last year. Add it and search for clone. I just found it. Might have to restart the docker, or refresh page for it come up.

Went to rc2. Didn't realize it was out yet. I saw nvidia did officially support. I saw on the forums here somewhere unraid didn't yet or something. It's working fine without rom. Moved to rc2. Still didn't work... So changed up the pcie slots for everything. Then I had to turn on downstream in vm settings as the top two pcie are tied together and the p2000 occupied one it them. Well that worked for booting dual gpus in one vm. Awesome. Then Nvidia plugin and reboot still working. Somewhere along the way it works now. Rc2, downstream, no roms for either card, amd gpu reset plugin (honestly the VM should never reboot unless the server does), Nvidia plugin and my p2000 is happily mining eth with only 5gb of ram thanks for help. I still not sure why the original issue occured. I tried the same solutions. Only difference is rc2 and reset bug plugin. All those combos were tried previously. BIos is up to date. 4g decode is on, resize bar is not on. When I researched i had seen it wasn't "supported yet". And as far as I know latest bios. Haven't checked in a while. With windows 11 out I'm sure there was an update recently. As everything is working I don't want to bork it. as it seems every system is particular.

anyone else get any variables working? i tried separating out gpus and it wouldn't do that even though i added the gpu-id's tried the arguments. tried just amd mining and arguments didn't work. yes privileged is on.