Bug: sam delete doesn't seem to respect MFA session duration, or doesn't know where to find cached session

[rhbecker]

Description:

Every sam delete invocation requires entry of a new token. This does not occur with sam deploy using the same profile.

Maybe there's some small chance this issue shares a root cause with what's being addressed in PR #6551.

This issue seems similar to the resolved #1623, and perhaps the resolved #1682.

Steps to reproduce:

1. Configure an AWS CLI profile with MFA enabled.
2. Switch into that profile.
3. Execute a sam deploy.
4. Enter the requested code.
5. Try another sam deploy just to see you're not asked for the code again.
6. Execute a sam delete; you're prompted for a code regardless of whether still within the expected session duration.

Observed result:

The prompt for the code comes after answering the following question, which might indicate the location of the issue ...

Are you sure you want to delete the stack <SOME STACK NAME> in the region us-west-2 ? [y/N]

Expected result:

The sam delete command should know how to locate the session cache and should respect duration_seconds settings configured for the AWS CLI profile, just as sam deploy seems to.

Additional environment details (Ex: Windows, Mac, Amazon Linux etc)

The config for the profile I'm using ...

[profile example.mfa]
duration_seconds = 43200
mfa_serial = arn:aws:iam::<REDACTED>:mfa/testuser
output = yaml
region = us-west-2
role_arn = arn:aws:iam::<REDACTED>:role/<REDACTED>
role_session_name = example
source_profile = example

The output of sam --info ...

{
  "version": "1.107.0",
  "system": {
    "python": "3.11.3",
    "os": "Linux-5.15.90.1-microsoft-standard-WSL2-x86_64-with-glibc2.31"
  },
  "additional_dependencies": {
    "docker_engine": "Not available",
    "aws_cdk": "Not available",
    "terraform": "Not available"
  },
  "available_beta_feature_env_vars": [
    "SAM_CLI_BETA_FEATURES",
    "SAM_CLI_BETA_BUILD_PERFORMANCE",
    "SAM_CLI_BETA_TERRAFORM_SUPPORT",
    "SAM_CLI_BETA_RUST_CARGO_LAMBDA"
  ]
}

[lucashuy]

Thanks for raising this issue! I can reproduce the problem where sam delete doesn't seem to be using the same cache as sam deploy. We'll need to dig a little deeper into what delete does differently than deploy to find the root cause.