222twotwotwo opened a new issue, #140: URL: https://github.com/apache/dubbo-go-pixiu-samples/issues/140
<!-- Please use this template while reporting a bug and provide as much info as possible. Not doing so may result in your bug not being addressed in a timely manner. Thanks! --> **What happened**: When Pixiu probes readiness in the run.sh script of the ai/kvcache/mock sample, it uses the same prompt as the formal test request, which pre-warms the token cache in Pixiu's memory. The subsequent reset only clears the status on the mock side but does not clear Pixiu's token cache. As a result, formal test requests hit the cache and skip the /tokenize call, causing tokenize_calls to be 0 and the assertion to fail. **What you expected to happen**: The run.sh script should pass all assertions correctly without being affected by the probe request. **How to reproduce it (as minimally and precisely as possible)**: 1. Configure the PIXIU_SOURCE environment variable. 2. Run bash run.sh in the ai/kvcache/mock directory. 3. Observe the assertion failure: tokenize was not called on engine-a. **Anything else we need to know?**: The KVCache routing function works normally. The issue is caused by the probe request polluting the token cache, not a functional bug of Pixiu. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
