CCBoot

Diskless render farm

Every render node on the same image, the same plugin versions, every night

A render farm is a cluster of PCs built to render computer-generated imagery. CCBoot boots all of them from one diskless image: install 3ds Max, V-Ray or your renderer once, and every node runs exactly that — no drifting plugin versions, no local disks to fail mid-render.

Start Your Free Trial See Pricing
days of rendering, split over 15 nodes
15 → 1
image for every node
1
local disks in the farm
0
CCBoot diskless render farm topology: designer PC, queue manager, NAS and diskless render nodes
CCBoot client and image management for render nodes

Why render farms go diskless

There is a huge amount of computation when 3D animation designers turn pictures and clips into animation. A queue manager (QM) distributes the processes across many computers: if one PC would need 15 days, 15 PCs working together finish in one.

Because these computers are used only for computing, giving each one a hard disk is a waste of money and resources — and if someone works on a node, it affects the computations. HDD nodes also fail: viruses, trojans, corrupted installs. That is where the CCBoot diskless boot solution comes to the 3D animation field.

  • Identical renderer and plugin versions on every node, guaranteed
  • Add a node: plug in, boot from network, it joins the queue
  • Reboot wipes any accidental change; nothing to reinstall
  • Spare CPU in the office at night can join the farm as diskless nodes

Deployment environment

The 3D animation studio deployment this solution was documented from.

Edition
CCBoot v3.0 Enterprise Edition, 40 PCs, June 2013
Network
Intel load balance; 2 × 26-port switches
Server
Windows Server 2008 R2, 16 GB RAM
Clients
Windows 7 x64, 16–24 GB RAM per node
Storage
NAS stores video resources and finished products
Software
3ds Max and V-Ray (distributed rendering plugin); After Effects plugin
CCBoot diskless render farm topology: designer PC, queue manager, NAS and diskless render nodes
Current farms run Windows 10/11 nodes with 64 GB+ RAM on 10G networks from the same CCBoot server — see server requirements.

Work principle

How the designer PC, queue manager, NAS and diskless render nodes fit together.

  1. 1

    Install the server side

    Install CCBoot server, 3ds Max and the queue manager (QM) on the server.

  2. 2

    Build the render node image

    Install CCBoot client, 3ds Max and V-Ray (the distributed rendering plugin) on a master node and upload the image. After the nodes diskless boot, V-Ray automatically monitors for and receives render tasks.

  3. 3

    Store assets on the NAS

    The NAS holds the video resources and finished products of the designer PC and the render nodes.

  4. 4

    Submit and distribute

    The designer works on a PC with an HDD and 3ds Max + V-Ray. When the animation is finished, the designer PC submits it to the QM; the QM finds the V-Ray instances and distributes the render tasks to the diskless nodes.

  5. 5

    Render and collect

    The render nodes process their tasks and save the finished product to the NAS.

Note: every software package can work with CCBoot like this, as long as it can do network rendering and/or distributed rendering with CPU power (not GPU) rendering.

Get Your FREE Trial Today

Boot two render nodes from one image on the trial and watch them pick up jobs from your queue manager.

Start Your FREE Trial Today

What the farm gains

The CCBoot features that matter most for compute nodes.

  • No More Buying Hard Drives

    Client PCs don’t need a hard drive… this alone could save you THOUSANDS of dollars.

  • Easy Updates

    When the “master” PC is updated, the rest of the PCs on the LAN will update with a simple reboot.

  • No More Viruses, Trojans, and Spyware

    The system will be restored after every reboot, leaving your PCs in a clean state at all times.

  • Easy Restore Points

    Set restore points such as create recovery, merge to last, and restore to last. This makes it easier to fix problems and maximize performance.

  • Efficient Cache Equipped

    Equipped with good cache performance, physical memory, and SSD cache support.

  • Energy Conserving

    Up to 40% energy could be saved, which is environmental and energy-saving.

Documentation

Step-by-step guides from the CCBoot wiki.

Browse the full wiki

Common questions

Which renderers work on a CCBoot diskless render farm?

Any renderer that supports network or distributed rendering on the CPU: V-Ray, Corona, Arnold CPU, Blender Cycles CPU, Cinema 4D Team Render, After Effects render engine and so on. Install it once on the master node image.

Does GPU rendering work?

The nodes boot a full Windows with their own GPUs, so GPU renderers run — but the documented deployment and the queue manager setup above target CPU distributed rendering. Test your GPU renderer’s licensing and node discovery on the trial.

How much RAM do render nodes need?

Render nodes are memory-heavy: 16–24 GB was typical in the 2013 deployment, 64 GB or more today. Set the CCBoot client RAM cache small so the renderer keeps the memory.

Where do scenes, textures and output live?

On a NAS or file server, not on the nodes. The diskless image holds only Windows and the renderer; assets are read from and written to shared storage over the LAN.

How do we update the renderer or a plugin on 40 nodes?

Boot one node in super-client mode, install the update, create a restore point and save. Every node runs the new version after its next reboot. If renders break, roll back to the previous restore point.

Related pages

Are You Ready To Give CCBoot A Try?

It only takes a few minutes to give CCBoot a try… (You’ll be glad you did)

Start Your FREE Trial TodayTalk to us