[dyarza]

Hi everyone,

We just purchased a new system for M~R on our network, I could not be more excited (4 Xeon 5130 cores, etc) The speed increase is tremendous from my workstation.

I am noticing a potentially fatal problem with my setup, I hope that there is a workaround.

Here is the configuration:

-- The new machine (lets call it Rendermatic) lives on the network and it has M~R 1.1 running on Windows Server 2003.

-- The M~R manager and server components are running on it.

--Individual users create mxs files in their prefered way, SketchUp, 3D Max, Cinema 4D and even Studio.

--Using the mxcl -d each user connects to the manager or Rendermatic and can see the render queue and submit jobs.

It all seems great, but I noticed this afternoon while working on a rendering that Rendermatic is not using my updated mxs file.

-- I have a file: "S:\Scenes\Residence.mxs"

-- I render with Rendomatic using the above method and see that I need to make some changes.

-- I go back into (SketchUp in this case) and make some changes, I also update some mxm files.

-- I export the mxs again, overwriting the old one, so it is still Residence.mxs

-- I launch mxcl -d again, load the file and render over the network with Rendermatic.

The image has not changed.

If I log into Rendermatic and stop the server and manager and re-launch, then it renders the updated scene. This of course defeats the purpose of having this new machine shared among many users.

Sorry for the long post but I figure this needsd the detail.

Any insight would be great, maybe this is a known thing... or it needs to be fixed?

Thanks,

[Q2]

Hi dyaraza

I am having the same problem here. Looks like this is a bug. I usually give my corrected files a new name and start the whole process again, no fun but so far the only way this works out for me. Let us know if you find a solution.

Ciao

Q!

[dyarza]

Yes, giving the MXS a new name solves it, but it would sure be nice not to have to do it.

I hope that it is a bug that gets fixed, because I am really liking this workflow.

[Inaki]

Two things:

I dont think is correct to run the manager and the server on the same machine as I understand u are doing here.

I do recomend restar every node once every render is finished

[dyarza]

Two things:

I dont think is correct to run the manager and the server on the same machine as I understand u are doing here.

I do recomend restar every node once every render is finished

I have read about not running manager and server on the same machine before, it seems to work albeit with the problems described. I will try to run the manager on a different machine and see if it helps.

As for restarting each node after each render, it completely defeats the purpose of a renderfarm if you have to do that. In our case we just have the machine set up so that multiple users can use it and having to restart each time kills the workflow. If these are limitations of the software I would hope that they are considered bugs and beig looked at.

Having said that, giving the MXS a new name solves the problem and I am very happy with how the process is running. We have six interns who can all render to that machine without having to log into it and mess with it.