This project consists of a set of Ansible playbooks that allow you to deploy a host running Docker with a set of dependencies for running pan-cancer workflows.
See recommended parameters at example_params.json
Also see an example inventory file for two hosts:
[master]
i-346b76d2 ansible_ssh_host=54.246.12.180 ansible_ssh_user=ubuntu ansible_ssh_private_key_file=/home/ubuntu/.ssh/green_snake.pem
i-3b6b76dd ansible_ssh_host=54.216.128.201 ansible_ssh_user=ubuntu ansible_ssh_private_key_file=/home/ubuntu/.ssh/green_snake.pem
[all_groups:children]
master
If you are going to run the DKFZ/EMBL workflow, you will need to run the DEWrapperWorkflow playbook. To do this, ensure that you specify DEWrapper in your json variables:
"workflow_name":"DEWrapper",
"workflows":["Workflow_Bundle_DEWrapperWorkflow_1.0.2_SeqWare_1.1.0","Workflow_Bundle_HelloWorld_1.0-SNAPSHOT_SeqWare_1.1.0"]
Alternatively, if using Bindle config files to specify variables, use the following
install_workflow = True
workflow_name = DEWrapper
workflows=Workflow_Bundle_DEWrapperWorkflow_1.0.2_SeqWare_1.1.0,Workflow_Bundle_HelloWorld_1.0-SNAPSHOT_SeqWare_1.1.0
Note that for the DKFZ/EMBL workflow, a large batch of dependency files needs to be in the /datastore directory. However, this directory may be ephemeral (for example when using lvm). This means that the Amazon image process will not retain this information. In practice, this probably just means re-running this playbook via youxia deployment and most steps will be skipped.