[BUG] Investigate Multi-Instance Deployment Issues (same account, different regions) about autonomous-driving-data-framework

[chamcca]

existing modules should be reviewed to ensure the Roles they create are unique for a given deployment. these could be named using hashes or include region identifier.

seed-farmer v2.0 will address this, but in the meantime ADDF modules could be patched

[dgraeber]

For the stack trace above, the use of the 'masterrole' suffix on the role is causing the issue (in this case). Searching for masterrole and where to be remediated by adding the hash suffix:

• modules/core/eks/stack.py
• modules/core/eks/modulestack.yaml
• modules/core/eks/deployspec.yaml
• modules/core/aws-batch/modulestack.yaml

The above files for eks need to make sure the role names are correlated to one another. The instance in aws-batch seems like a single instance

The method _execute_deploy should probably have another parameter for hash so it can be passed in to the modulestack.yaml as a parameter (??) so it can be postpended to the role name reference

[dgraeber]

The issue in core/aws-batch has been addressed in #35