Run GNS3 on a remote server

Split the load on multiple machine

Requirements

Limitations

Setup

Add your remote servers

Setup the appliance templates

Last updated
June 15, 2017

Contributors
Julien Duponchelle
Edit this Article
Last updated
June 15, 2017

Contributors
Julien Duponchelle
Edit this Article

Scale GNS3

GNS3 itself doesn’t have limitation on the number of nodes you can run into it. The limitation will be your hardware.

Run GNS3 on a remote server

If your local machine is not enough you can run GNS3 on a bigger machine locally or on a cloud Provider like Packet.

See Install on a remote server

Split the load on multiple machine

In order to cross the limitation of your hardware you can use multiple GNS3 server for running the same topology.

An example is Jason C. Neumann running GNS3 in a Raspberry Pi cluster:

https://www.gns3.com/discussions/the-worlds-first-gns3-beowulf-cl

NOTEThe PI use an ARM processor it’s not adapted to run x86 code on Qemu. Also since this forum post the interfaces has change.

Requirements

  • You need the same version of the GNS3 server on all machines
  • All server need to be on the same network and nothing should prevent them to communicate with each others

Limitations

  • A node need to be explicitly created on a server. Their is no smart allocation
  • You need to create an appliance template by server
  • You can’t move a node from a server to another
  • Portable project are not supported

NOTEThis is not definitive. GNS3 is evolving and contribution are welcomes

Setup

Add your remote servers

First you need to add your servers in the preferences.

And will see the resources usage:

Setup the appliance templates

When you create your appliance template choose a remote server:

At the node name step choose a name helping to remember on which server you will run the node:

Create a second template for the other server. You have now two node that allow you to manually dispatch the load between servers.

You can use the server summary in order to know which server as free resources