Gen3 Cloud Automation Recommendations

Hi,

I've just started working through the CSOC free commons tutorial (cloud-automation/csoc-free-commons-steps.md at master · uc-cdis/cloud-automation · GitHub) and had some questions about general details:

  • Is there a general structure regarding domain names for gen3 instances? (eg. gen3.DOMAINNAME,org or DOMAINNAME.datacommons,org, etc..)
  • Is there a recommended EC2 instance for an initial test, which can later be scaled or upgraded as necessary?
  • Is a Linux/Ubuntu VM ideal for the tutorial?

Thanks in advance!

Hi @ktaouk1,

thank you for your questions!

  1. There is no general structure, but we generally have to work with subdomains and own the actual domain, so if you own datacommons.io, you can setup gen3.datacommons.io, for example.
  2. Our largest container we're running is ETL, so as long as you run (for example) t2/3.xlarge nodes you can run everything, if not you will need to modify some code to request less resources because the nodes won't fit the containers. However, if you referred to an adminvm then t2.micro should be fine.
  3. Our VM runs on Ubuntu 18.04 so Ubuntu should be your best option.
1 Like