Why do we need a dedicated video out

[landS]

Given the excellent state of IPMI, SSH, and plugins like Shell-In-A-Box, why the heck does unraid need a dedicated GPU out?

 

Not trolling on this one, truly curious.

[JonathanM]

Given the excellent state of IPMI, SSH, and plugins like Shell-In-A-Box, why the heck does unraid need a dedicated GPU out?

 

Not trolling on this one, truly curious.

Very few consumer boards have IPMI, but mostly I think it's a KVM limitation, not something that unraid has direct control over. However, how would you troubleshoot a network connection failure with no local console? I suppose you could redirect the console to a serial port like many other network appliances, but then managing unraid turns into a game of trying to connect a second machine just to get an output screen.

[jonp]

Given the excellent state of IPMI, SSH, and plugins like Shell-In-A-Box, why the heck does unraid need a dedicated GPU out?

 

Not trolling on this one, truly curious.

 

You don't need a dedicated GPU out if you're not doing GPU pass through with a VM.  And even then, you don't need it if the GPU is an AMD-based GPU.  It is only if you are doing GPU pass through and the GPU is NVIDIA.  This is a GPU-vendor specific issue right now, and we plan to look into this more in the future.

[landS]

@jonp - that is great news!  This means I will not need another gpu on planned server (MBD-X10SRA-F-O (no GPU), E5 1650V3, AMD R9 280 win 10 passthrough)

 

@jonathanm - I was going to say, if not a kvm limitation, make option switch driven with note to re-enable by booting stick from another PC should network troubleshooting be required.

[tayshun12]

Given the excellent state of IPMI, SSH, and plugins like Shell-In-A-Box, why the heck does unraid need a dedicated GPU out?

 

Not trolling on this one, truly curious.

 

You don't need a dedicated GPU out if you're not doing GPU pass through with a VM.  And even then, you don't need it if the GPU is an AMD-based GPU.  It is only if you are doing GPU pass through and the GPU is NVIDIA.  This is a GPU-vendor specific issue right now, and we plan to look into this more in the future.

 

Hey Jonp! Just wondering what you mean when you say you need it for Nvidia GPUs? Im doing a build right now with an ATI R9 290 and a Nvidia 560ti (re-purpose of an old card) and I was hoping to use my integrated graphics as the hosts Graphics card. Would this suffice?

[saarg]

Given the excellent state of IPMI, SSH, and plugins like Shell-In-A-Box, why the heck does unraid need a dedicated GPU out?

 

Not trolling on this one, truly curious.

 

You don't need a dedicated GPU out if you're not doing GPU pass through with a VM.  And even then, you don't need it if the GPU is an AMD-based GPU.  It is only if you are doing GPU pass through and the GPU is NVIDIA.  This is a GPU-vendor specific issue right now, and we plan to look into this more in the future.

 

Hey Jonp! Just wondering what you mean when you say you need it for Nvidia GPUs? Im doing a build right now with an ATI R9 290 and a Nvidia 560ti (re-purpose of an old card) and I was hoping to use my integrated graphics as the hosts Graphics card. Would this suffice?

Using your integrated graphics would be no problem. Just be sure you set it to use the integrated graphics as the initial graphics device. Then you should be able to use both the AMD and nvidia card for VM's.

[tayshun12]

Great thanks for the info!

[tim8v]

old thread, but... i am in the middle of a version 6 unraid build. my board is an Asus X99-Deluxe, with no onboard GPU. I have a GTX970 in the first slot which i have used to install unRAID. I want to use that card for a windows VM, is that possible or will i need another card?

[saarg]

old thread, but... i am in the middle of a version 6 unraid build. my board is an Asus X99-Deluxe, with no onboard GPU. I have a GTX970 in the first slot which i have used to install unRAID. I want to use that card for a windows VM, is that possible or will i need another card?

Most likely you'll need another card for unraid to grab and use for the console. I think there might have been someone that got it working using the rom file from the GPU, but I migh also remember wrong.

Be aware that the X99 platform have been troublesome to use for virtualization because of  bad support for separating the devices into separate IOMMU groups. The ACS override to separate the devices into groups also fails on most motherboards. 

[bungee91]

Most likely you'll need another card for unraid to grab and use for the console. I think there might have been someone that got it working using the rom file from the GPU, but I migh also remember wrong.

Be aware that the X99 platform have been troublesome to use for virtualization because of  bad support for separating the devices into separate IOMMU groups. The ACS override to separate the devices into groups also fails on most motherboards.

 

No hating on X99  , we have proper ACS and no IOMMU grouping issues (I think your thinking Skylake and Z/H170 chipset).

The thread you're mentioning for the rom, etc. is here (I have not attempted this) http://lime-technology.com/forum/index.php?topic=43644.msg452464#msg452464

[saarg]

For some reason I thought X99 was skylake. My bad  I'll go stand  in the corner now... 

[veritas2884]

First, thanks for the amazing work on this!

 

I am think of doing a new build with a Asus ROG Crosshair VIII Impact Mini DTX AM4 Motherboard and a AMD Ryzen 9 3900X 12-core.

I am going to move my P2000 to the new build.

 

Here is my question, if I start unraid in terminal mode will I still be able to use my P2000 for Plex transcoding duties since this build will have no spot for a second GPU and no on-board GPU?

 

Thanks!

[tjb_altf4]

If Plex is running in a container (and not a VM) and you utilise the NVIDIA build of unraid, then it can be shared.