Best Practices of using DR software Optimally

Deadline
1

Hi

I know this is a long note, but I would like to know what would be the best practice to use a distributed rendering software on the systems while you are simultaneously working on a scene file. I have been looking at various options on this but have not hit upon any one good solution so far.

My current working setup is as follows:

  1. I have one system which I use for Modeling, Texturing and Lighting - (Lets call this Working Unit)
  2. I have another system in which I have stored all the Proxies, Models and their materials - (Lets call this Server)
  3. I have 3 other systems which I use for Renders (Nodes)

My current working process is as follows:

  1. I work on my Working Unit and copy the general material to the Working Unit.

  2. When I need some proxies, I import them to the scene but link the files to the Server unit (Proxy files/Materials etc)

  3. Render test renders on Working Unit and when satisfied,
    a. Archive the file,
    b. Copy to Server
    c. re-set paths manually
    d. Start Distributed render on Deadline on Server + 3 Nodes (I have license for 4 nodes) and continue to work on my Working Unit.

    While this system works fine, I have following issues:

  4. I am not able to optimally harness the power of the distributed renders for test renders. Sometimes, there might be a need to render the test scenes at decent resolution which takes up time. I currently render these on my Working Unit itself, as the above mentioned 4 steps are time consuming. This means 4 systems are lying idle along with ME.

  5. Even if I follow the above 4 steps and set the renders on the other systems (Server + 3 nodes), the network stalls my work on the Working Unit (especially when I work on the Materials) as it tries to connect to the Server which is under Render.

Furthermore, this is completely a manual process (all the above 4 steps), which means I cannot be using this process every time (which I would like to - given the fact that I have the systems to do this)

My question is:

How can I use the other systems (Server and Nodes) simultaneously when I am working on my Working Unit on a regular basis say even for test renders. Is there a way, I can automate the replication process on another node/server and also completely cut off my system from the rendering process?

Would be curious to know how others handle this.

Best Regards,

GP

2

how about this

Map S:\ to your server path with your data you arent going to copy locally to every machine - eg

\server\server only data\data

then Map T:\ on your worker machine to local data

\localmachine\data\

and Map T:\on the nodes to the server data

\server\data

run rsync.exe or the equivalent to keep \server\data and \localmachine\data synchronized [so they are the same, same paths etc so that T:\data\image.jpg is the same from the node and your workstation, even though they are pointing to different locations]

and then when you render the nodes will pull data from S as well your machine, but your local machine pulls the other data locally and the nodes remotely.

this assumes you are running windows…i’m sure something similar exists on OSX

cb

3

Thanks for the suggestion. Yes I am using Win 7 on all the systems. Let me understand this better and revert to you.

Best Regards

GP