On deploying the PD Disaggregation example I found multiple issues:

• deprecated objects being created like described in Deprecate InferenceModel #121
• current Inference scheduler for PD disaggregation is using ghcr.io/llm-d/llm-d-inference-scheduler:v0.2.1 and the latest version available is ghcr.io/llm-d/llm-d-inference-scheduler:v0.3.1
• the config file for the inferencescheduler in the example uses the default config and not the default-pd-config.yaml
• i found an issue where the vLLM baked in the "ghcr.io/llm-d/llm-d:v0.2.0" image does not have available the proper NixlConnector