fix: add FlexKVConnectorV1 to PdConnector allowed first connector types

[linhu-nv]

Fix PdConnector TypeError when launching disaggregated FlexKV serving (disagg_flexkv.sh).

PdConnector.__init__ only allowed DynamoConnector and LMCacheConnectorV1 as the first connector. When FlexKVConnectorV1 is configured as the first connector (KV offloading
role), it raises:

TypeError: Expected first connector to be DynamoConnector or LMCacheConnectorV1, got FlexKVConnectorV1

This PR adds FlexKVConnectorV1 to the allowed first connector types using the same optional-import pattern established for LMCacheConnectorV1 in PR #4319.

Fixes: NVBug 6088159

Changes

• Add optional import for FlexKVConnectorV1 from vllm.distributed.kv_transfer.kv_connector.v1.flexkv_connector (soft dependency, no error if FlexKV is not installed)
• Extend allowed_first_types and error message to include FlexKVConnectorV1

Test plan

• Launch examples/backends/vllm/launch/disagg_flexkv.sh — confirm prefill worker starts without TypeError
• Launch existing disagg_kvbm.sh / disagg_lmcache.sh — confirm no regression for DynamoConnector and LMCacheConnectorV1 paths
• Verify FlexKV-not-installed case: import falls back gracefully, PdConnector still works with DynamoConnector/LMCacheConnectorV1

---

Summary by CodeRabbit

• New Features

  • Added FlexKV as an alternative KV offloading connector option.

• Documentation

  • Updated documentation to reflect the new connector option availability.

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[github-actions]

🌿 Fern Docs Preview: https://nvidia-preview-a7f28d4c-8770-4b1b-8394-cb481f3fa2c1.docs.buildwithfern.com/dynamo/dev

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: b58a33b6-1650-4f19-a701-acf7aa2bb29d

📥 Commits

Reviewing files that changed from the base of the PR and between ee3389e and 5e40097.

📒 Files selected for processing (1)

• lib/bindings/kvbm/python/kvbm/vllm_integration/connector/pd_connector.py

---

Walkthrough

Adds optional FlexKV support to PdConnector as an alternative first KV offloading connector. Introduces guarded import for FlexKVConnectorV1, conditionally updates connector type validation, and updates class documentation to reflect the additional option.

Changes

Cohort / File(s)	Summary
FlexKV Integration lib/bindings/kvbm/python/kvbm/vllm_integration/connector/pd_connector.py	Adds optional FlexKV support with guarded import for FlexKVConnectorV1. Conditionally extends the type validation set for the first connector to accept FlexKV alongside KVBM and LMCache. Updates error message and class documentation to reflect the new option.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~8 minutes

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The pull request title accurately and concisely summarizes the main change: adding FlexKVConnectorV1 to the allowed first connector types in PdConnector.
Description check	✅ Passed	The pull request description is comprehensive and well-structured, covering the problem statement, solution approach, specific changes made, and a detailed test plan with checkboxes.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[dagil-nvidia]

/ok to test 5e40097

[dagil-nvidia]

/ok to test c23b9b7

[dagil-nvidia]

/ok to test c23b9b7

[dagil-nvidia]

/ok to test f5fa63c