Args can be read out as $1, $2, $3, $4, $5 inside deploy.sh.

That being said, Environment variables should be accessible the way you’re using them, you’ll have to review Dockerfile and deploy.sh to identify the problem. You can start by specifying 

runs = “env”

with no arg to look at all the environment variables present.

Hi @wei 

I would think that is how it works, but the very top of the script reads in passed args.

PROJECT=$1
shift
SERVICE=$1
shift

…

The project would be $1 and it does not post correctly. However, if I put in my args like this

args = [“project1”, “app1”, “TOKEN”, “role1”, “PASSWORD”] (PASSWORD and TOKEN are secrets)

Then my log comes back better:

bad decrypt
You are in project:=============>project1
You are deploying to service:===>app1

 But I can tell the secrets still didn’t make it through correctly, and I’d rather not expose it to debug but I may have to.

Just create a new test secret and see if it comes back. Regardless, secrets are masked with [FILTERED] in the output anyways. Try my “env” test mentioned above if you still have troubles.

The runs env block, are you suggesting a block right before the script block? Such as

action “environment check” {
  needs = [“Previous step”]
  uses = “env”
}

Also, I echoed out all passed args to the script then cat’d that script. I also wanted to see if there is a difference from my test secret being with our without $.

My result is, they were passed but not interpolated.

My test secret was TESTSECRET with a value of: TESTTESTTEST

Args looked like this:

args = “projec1 app1 $TOKEN encrypted_role $PASS TESTSECRET $TESTSECRET”

My cat (from the deploy.sh script) showed this:

project1 app1 $TOKEN encrypted_role $PASS TESTSECRET $TESTSECRET

Try echo out env in the deploy.sh using echo $(env) to make sure it contains everything you want.

And you have to have $ before variables otherwise they are not interpolated.