The llm-d-inference-scheduler e2e tests currently fail when run using Podman.

Hardcoded docker Commands: Despite being able to set CONTAINER_TOOL to “podman” in the Makefile, the run_e2e.sh script, which is executed by the test suite, contains hardcoded docker commands (e.g., docker pull, docker images). This directly breaks the workflow for users who only have Podman installed, as the script fails when it tries to execute a command that doesn't exist. This inconsistency forces Podman users into complex workarounds or prevents them from running the tests altogether.

Example: https://github.com/llm-d/llm-d-inference-scheduler/blob/main/test/scripts/run_e2e.sh#L14 and https://github.com/llm-d/llm-d-inference-scheduler/blob/main/test/e2e/e2e_suite_test.go#L128

These issues block developers who primarily use Podman, increase setup friction, and hinder effective local testing and CI alignment.