[LMCache][Example] Align the PYTHONHASHSEED for prefillers and decoders for KV chunks hashing

[zejunchen-zejun]

Add and align the PYTHONHASHSEED for prefillers and decoders for hashing the chunks.
For now the LMCache uses the hash seed from the vllm, so it is needed to align the hash seed for all prefillers and decoders, so that the decoders can identify and retrieve the kv cache chunks saved by the prefillers.
https://github.com/LMCache/LMCache/blob/dev/lmcache/v1/token_database.py#L106-L107

[gemini-code-assist]

Code Review

This pull request aligns the PYTHONHASHSEED for LMCache examples to ensure compatibility between prefillers and decoders. However, it introduces a security vulnerability by hardcoding the seed to a fixed value. My review includes a critical comment to address this by making the seed configurable and adding a security warning, which is crucial even for an example script.

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[zejunchen-zejun]

Hi, @ApostaC @KuntaiDu

For the issue LMCache/LMCache#1095, we add the flag to align the hash seed in the lmcache example code in vllm, could you help review?

Thank you. :)

[ApostaC]

Hi @zejunchen-zejun , this looks good and thanks for the contribution! Can you fix the DCO issue by sign off the commits? Please follow the instructions here: https://docs.vllm.ai/en/latest/contributing/#dco-and-signed-off-by

Also, you may want to setup and run pre-commit locally to lint the code. It can also help you sign off the commits automatically.