GHunter

the GT 710, 730, and 1030 are all low priced cards and work well. They are about as cheap as you can get and you can get them in PCIE 1x connections and single slot designs.

I'm having this same problem. Just happened 20 minutes ago or so. filesvr-diagnostics-20190719-1020.zip

Not a single diagnostics posted in this thread either.

As it turns out, I didn't have to bind anything. My VM's just take control of the GPU's when they are started and returns then back to unRaid when they are shutdown. Smooth upgrade.

Thanks. It's gonna be a fun ride as I upgraded to server class components as shown in my signature.

I'm installing a new motherboard today that does not have an onboard GPU. It's a Supermicro X10DAC. Do i need to hide all of my GPU's so I can run a headless system and use all of my GPU's for VM's using the new vfio-bind method in unRaid 6.7.0? example: BIND=<device> <device> Where <device> is a Domain:Bus:Device.Function string, for example, BIND=01:00.0 02:00.0 save as /config/vfio-pci.cfg Thanks in advance!

Shutting down the array has always shutdown the VM's and most other services. Wish it didn't work this way but it does.

you should make a feature request for this. I think it would be a fairly simple fix as the logic already handles started VM's that are running normally. Basically resume all paused VM's first, then continue on with the stopping process as normal or something similar to this.

@queueizCan you be more specific? What did you try exaclty to get this working? Thanks!

Good idea. I'd like to see this added as well.

Or you can buy a 2 port HBA card and pass that through.

I used a nVidia GTX 750 for mine. I use all nVidia GPUs in my builds as you can see in my signature.

I haven't had any luck getting a decent price on 2 Intel Xeon E5-2687W v3 processors so I had to go with 2 Intel Xeon E5-2667 v3 processors instead. They are 8 core CPU's so I lost a few cores over my original plan but on the bright side, they were cheap and they run a little faster at a base clock speed of 3.2 GHz. For the CPU coolers, I picked up 2 Cooler Master Hyper 212 EVOs. I've seen other people use them in their server builds and they should fit in my 4U case. Now I need to order my memory and a few miscellaneous items and plan on getting this built in 2 to 3 weeks! Luckily memory prices have dropped and I found 64 gigs of ECC memory for $290 at Newegg.

I have it installed and working fine using Q35 and OVMF with a dedicated GPU. Didn't have any problems installing it as it was pretty straight forward. I haven't used it in ages though.

Glad you've made progress on this. I have 4 of the 50ft cables myself. 1 of them went bad and monoprice replaced it for me under warranty.

Thanks for posting this. I've used notify in some of my scripts as well and this will certainly add more flexibility. I appreciate it!

Me too! I've got quite a few VMs and could really take advantage of this feature when added.

Ended up purchasing a Supermicro X10DAC instead of the X10DAX because it has an onboard LSI HBA. Still looking on ebay for the CPU's and memory and hope to find and purchase them this week or early next week.

This is for the guest VM.

FYI. @jonp mentioned in another post about dropping support for Q35 and Windows based VM's. More info here: So keep that in mind before switching machine types for your Windows VM's.

Intel is the only platform for iGPU passthrough support and it is still fairly new so YMMV with Intel support as well.

I would pull your flash drive and check the file system on it from another PC. Also, try booting in safe mode. I think there may have been a similar problem reported earlier in the RC series where a plugin caused the issue. Either way, can't hurt to try.

You should change your memory to use 4 sticks instead of 2 as your board supports quad channel memory for faster IO speed.

No. To take advantage of this patch, you'll have to continue to edit the XML manually.

Will do. I'll post back if I have questions or run into a problem. Thanks for your help!!