Author Topic: Having issues when using distributed rendering with Substance2  (Read 2540 times)

I need help solving a problem that I'm sure many people have encountered when using distributed rendering with server nodes.

Basically, when I use distributed rendering, the buckets from the workstation CPUs read and render a substance2 material perfectly fine, but buckets from the render NODES resolve black. I don't get any errors in my dialogue about the plugin, so I'm thinking it's some kind of GPU issue.

I have two render pros from BOXX that are running ATS2400 onboard video graphics cards. Each machine is running  3ds Max 2019, V-Ray Next, and has the Substance Plugin successfully installed; however, I cannot open 3ds Max on these machines to change the Substance Settings from GPU to CPU. I changed the substance settings on my workstation to CPU to see if that would fix it, but the nodes still have their bucket issues.

Could someone shed some light on what is causing this problem and provide a solution? Is it a GPU issue? Is there a way to change the substance settings on these render nodes without opening 3ds Max?
 
Last Edit: June 06, 2019, 12:05:48 am

Hi Mark,

The settings on your local workstation will not affect the render nodes, they contain separate configuration files. I think the settings are stored in a .ini file, I'll have to see where that is saved to tomorrow when I get back to a Windows machine. That can be edited if you can access the machines remotely.

Can you provide more information on your distributed rendering setup? Is it rendering through 3ds Max, or are they rendering with vray standalone?

Thanks,
Galen
Software Engineer, Integrations
Maya, 3ds Max and Core Libraries

Hi Galen,

Thanks for your reply!

We are rendering through 3ds Max/Backburner. We have a local set up that uses 4 workstation PCs, and 2 render nodes. We can access any of the files on these machines, but the nodes cannot open 3ds Max directly. We dedicate a single workstation to be the manager and server that receives jobs and stores them in the backburner queue. When a job is started with distributed rendering ticked on, it connects to all PCs/Nodes that have their Distributed Rendering Spawner turned on. These DR hosts don't launch 3ds Max directly, but provided extra cores to process the job.

Here are the configuration settings taken from the Enu folder of a node and a workstation
https://kohlerftp.kohlerco.com/main.html?download&weblink=c375697e7b1526ca3675bb1319ea2f51&realfilename=3ds$20Max$20Settings.zip


When the job starts, a V-Ray Frame buffer launches, the scene data is transferred to the DR Hosts, and DR Hosts begin processing the job along with the server that launches it. DR Hosts can connect to any licensed machine that is starting a job on our network using DR, so this issue isn't exclusive to a backburner submission.

I apologize if you already know how DR works, but I just want to tick all the boxes. I think it's important to reiterate that there is no bucket issues when another Workstation connects as a DR Host.  The problem occurs when a render node connects as a DR Host. These nodes have powerful CPUs and the same software/plugins/licensing that the workstations do, but do not have GPUs that are capable of DirectX Features.

Is there any other information I can provide to help?

Last Edit: June 05, 2019, 05:04:43 pm