This is an incubator for the scripts, plans, and playbooks needed to run Galaxy benchmarking experiments.
This is a work in progress and will need a massive reorganization and clean up before it is ready for prime time. Much of this documentation is aspirational in nature and describes what we want to do eventually, and not what is currently being done.
The idea is to be able to cd to one of the subdirectories and:
- provision the infrastructure
- install Galaxy
- configure Galaxy (create user, upload data and workflows)
- run benchmarks
- save/copy metrics data
- destroy the infrastructure
Each experiment should include a provision.sh script that is responsible for creating the actual infrastructure that Galaxy will run on. The provision.sh script can use whatever means necessary to accomplish this task, whether it be Terraform plans, Ansible playbooks, or running eksctl/gcloud/et al.
This should essentiall be:
helm install galaxy galaxy/galaxy -n galaxy -f values.yml
Where the values.yml file contains the specific customizations need for installation on this cluster.
TBD
- Configure
abm
- Derive the Galaxy URL with
kubectl - Create an admin user and get their API key
- Upload the workflows
- Upload the data
TBD
TBD
TBD