Summary

The teardown phase of the OIDC auth tests fails with this error:

ERROR: CLUSTER_NAME not set

I believe it's coming from teardown-atlas-cluster.sh, which is called by auth_oidc/teardown.sh. The root cause seems to be that CLUSTER_NAME is not set. The Azure and GCP OIDC tests don't have the same problem, but they write CLUSTER_NAME to file in a different path (see auth_oidc/gcp/setup.sh vs auth_oidc/setup.sh).

The result is that the teardown phase fails consistently, possibly leaving Atlas resources running.

Confirmed this impacts at least the Rust Driver and Go Driver

Motivation

Who is the affected end user?

Drivers devs, possibly DBX leadership (if we're spending money on unterminated Atlas resources).

How does this affect the end user?

OIDC Evergreen tasks may fail. Maybe wastes money.

How likely is it that this problem or use case will occur?

It will always occur when using the standard OIDC test setup and teardown scripts.

If the problem does occur, what are the consequences and how severe are they?

Confusion about the failed OIDC Evergreen task, possibly blocked driver development and/or releases. Possible wasted money.

Is this issue urgent?

No.

Is this ticket required by a downstream team?

No.

Is this ticket only for tests?

Yes.

Acceptance Criteria

Update OIDC test teardown scripts to correctly tear down resources.