fix: Automate init and pass Default Region for data-on-eks/streaming/kafka/install.sh #342
Conversation
default region not being taken from env vars. Default variable taking presidence.
consistent to install.sh passing var from aws config
automate terraform init
jakeskyaws
changed the title
| # Apply modules in sequence | ||
| for target in "${targets[@]}" | ||
| do | ||
| echo "Applying module $target..." | ||
| apply_output=$(terraform apply -target="$target" -auto-approve 2>&1 | tee /dev/tty) | ||
| apply_output=$(terraform apply -target="$target" -var=region=$(aws configure get region) -auto-approve 2>&1 | tee /dev/tty) |
There was a problem hiding this comment.
This could potentially create issues for users who intend to deploy the cluster in a specific region as specified in the variables file.
Our install.sh script depends on the region variables defined in the blueprint. Users must take care to update the region within the variables file to match their intended deployment region. There are situations where their local region might be configured as eu-west-1, but their deployment target is us-west-2 or us-east-1.
There was a problem hiding this comment.
this is also not "sound" behavior for a declarative IaC tool/framework - you declare the value in code, and apply that. using a dynamic lookup could lead to destroying an entire regions worth of resources
|
This PR has been automatically marked as stale because it has been open 30 days |
|
Pull request closed due to inactivity. |
In data-on-eks/streaming/kafka/install.sh
When using Terraform the AWS_DEFAULT_REGION is not recognised. We are fetching the default with the aws cli as it is a prerequisite. Then passing as a variable.
This caused the default variable to take presidence causing the cluster to be deployed to "us-west-2"
Added to cleanup.sh for consistency