How to use F5 image generator tool to prep BIG-IP images for AWS and Azure


You can use the F5 image generator ( ) to create BIG-IP images for use in public Clouds. This method creates a private machine image and is therefore not limited to the official BIG-IP versions released via the marketplace.

The F5 image generator comes as a multi-component install, or as a docker container.

This document shows the steps required to generate a functional private image with the docker method within both AWS and Azure.


Common Prerequisite



You can provision a Ubuntu (all major Linux distros should work with the container method) host using i3.metal instance type – this instance type is the cheapest with hardware virtualisation support. The presence of hardware virtualisation speeds up the image creation process, but it is not essential.

Ensure the root device has enough disk space (e.g., 100 GB) – this is needed to hold all files the image creation process requires

You can also use an existing host, provided it has docker installed and has enough disk space.

Once the host is ready, go through the following steps as a reference.

cd /home/ubuntu

Get ovftool onto the host (need to download from VMware first). Below is an example of getting it via curl.

curl -OJ

Install ovftool and copy extracted files to local directory  

chmod +x VMware-ovftool-4.4.3-18663434-lin.x86_64.bundle
sudo ./VMware-ovftool-4.4.3-18663434-lin.x86_64.bundle --eulas-agreed
sudo cp -r /usr/lib/vmware-ovftool /home/ubuntu/vmware-ovftool

Get BIG-IP ISO and optional EHF ISO

curl -OJ

Run below to obtain/run the container, enable hardware virtualisation support and mount local directory into the container

sudo docker run -it --device="/dev/kvm" -v "/home/ubuntu:/mnt" f5devcentral/f5-bigip-image-generator:latest

Inside the container, run the following

cp -r /mnt/vmware-ovftool /usr/lib/vmware-ovftool/; sudo chmod +x /usr/lib/vmware-ovftool/ovftool /usr/lib/vmware-ovftool/ovftool.bin;


which ovftool

Create config.yml file for the image generator. Check for all options

cat << EOF > config.yaml
AWS_BUCKET: "cz-abgmbh" 
AWS_REGION: "us-east-1"
MODULES: "all" 
REUSE: "Yes"
ISO: "/mnt/BIGIP-"

The image generator creates the image from the ISO and then uploads it to an existing s3 bucket, after which the image is converted into a snapshot, which is eventually registered as an AMI.

The AWS key must carry the appropriate permissions. See this ( ) for detail.

As well as a service role named ‘vmimport’ be created. See this ( ) for detail.

Once all the permissions and the role are ready, run the build process as below.

build-image -c config.yaml

The process will generate the image, upload it to the s3 bucket, convert it to a snapshot and print out an AWS AMI ID.



To generate the image for Azure, the process is that the image generator creates the image, uploads it to an existing storage container, it will then convert that image to an Azure image and place it within the same Azure resource group where the existing storage container is.

A service principal is required with ‘Contributor’ role assignment.

Refer to this ( ) for detail.

Below is a sample config.yaml file.

AZURE_REGION: "Australia East"
MODULES: "all"
PLATFORM: "azure"
REUSE: "Yes"
ISO: "/mnt/BIGIP-"


Hope the processes described in this article make it easier for people getting the image up and running in AWS and Azure. GCP should be very similar, please leave a comment if you are having issues with GCP and I will take a look as well. 


Updated Mar 23, 2022
Version 2.0

Was this article helpful?

1 Comment