Following this post set of posts on running Nextflow with AWS Batch:

• Nextflow blog: Scaling with AWS Bach
• This blog: Running CAW with AWS Batch
• Alexander Peltzer’s blog: Running nf-co.re pipelines with AWSBatch
• Anthony Underwood’s blog: Running Nextflow on AWS batch

I’m writing a small guide on how to make a custom AMI with a big enough size to run some actual data with.

All the instructions are in fact listed in the custom AMI Nextflow docs and this solution is there, but as I struggled once already, I prefer to document it for everyone and especially future me.

• On the AWS EC2 Dashboard, launch a new instance
• 1. Choose AMI: Search for and choose Amazon ECS-Optimized Amazon Linux AMI in the AWS MArketplace
• 2. Choose Instance Type: Choose a t2.micro
• 3. Configure Instance: Leave default configuration
• 4. Add Storage: Input 500 GiB or any other size, but 22 GiB is definitively not big enough for actual real life data.
• 5. Add Tags: Do as you want
• 6. Configure Security Group: Just make sure you can connect to the EC2 Instance.
• 7. Review Instance Launch: Procure a key pair that you have so that you can access your Instance
• Launch the newly created Instance (Actions -> Instance State -> Start), it should be easy to find
• Connect providing the key pair specified earlier.
• The ssh command provided by AWS is wrong, you need to replace root by ec2-user
• sudo yum update to update the system
• Install awscli using conda, so install conda first, and actually install wget before

sudo yum install wget
wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86\_64.sh
bash Miniconda3-latest-Linux-x86\_64.sh -p /home/ec2-user/miniconda
# You can either source .bashrc or specify the path to conda
source .bashrc
conda install -c conda-forge awscli

• Edit this file: /etc/sysconfig/docker-storage
• Add this option --storage-opt dm.basesize=500GB (or specify any other size you prefer)
• sudo service docker restart to restart the docker daemon
• Check that you actually have the size you wanted with docker info | grep -i base and docker info | grep -i data

For all their support, thanks to:

• alneberg
• apeltzer
• KochTobi
• t-neumann