📄️ Compute resource prerequisites
Suppose you have a Linux machine and would like to use it as a dendro compute resource. Here are the recommended steps to prepare it with the necessary software.
📄️ Host a Compute Resource
Each Dendro project comes equipped with a dedicated compute resource for executing analysis jobs. The default setting uses a compute resource provided by the authors with limitations on CPU, memory, and concurrent jobs, shared among all users. This public resource should only be used for testing with small jobs. Contact one of the authors if you would like to run more intensive processing or configure your own compute resources.
📄️ Worker nodes: AWS Batch
Here we provide instructions for setting up your compute resource to run Dendro jobs in AWS Batch. First you will need to provision the base AWS Batch infrastructure using CDK. This includes IAM roles, VPC, Security Group, EC2 Launch Templates, Batch Compute Environments and Batch Job Queues. Next you will need to configure your compute resource controller to submit jobs to AWS Batch. Finally, when you submit jobs from the web interface, you must select aws_batch as the run method.
📄️ Worker nodes: Slurm
WIP
📄️ Worker nodes: Kubernetes
WIP