回复: Fixing memory initialization logic and query timeout overflows

[冯 庆新 <qi...@hotmail.com>]

Well done！
Thank you for your careful work on this PR. These enhancements will definitely help improve the robustness of IoTDB.

Best regards,

Qingxin Feng
________________________________
发件人: jackietien97@gmail.com <ja...@gmail.com>
发送时间: 2026年1月23日 14:57
收件人: dev@iotdb.apache.org <de...@iotdb.apache.org>
主题: Fixing memory initialization logic and query timeout overflows

Hi all,

I've submitted a pull request to fix a few issues regarding memory allocation and timeout handling in the DataNode module.

First, regarding memory configuration, I realized that the memory limit per fragment instance wasn't being distributed fairly among query threads. I've updated the logic to divide `maxBytesPerFragmentInstance` by `queryThreadCount` to prevent potential over-allocation. I also adjusted the initialization sequence to ensure the thread count is loaded before dependent memory calculations run.

Second, I fixed a bug in `DriverTask` where calculating deadlines with large timeouts could result in an integer overflow. This was causing some issues with infinite loops during schema fetches, so I added a check to cap the value at `Long.MAX_VALUE` if an overflow occurs.

The PR link is: https://github.com/apache/iotdb/pull/17071

Best regards,
--------------------
Yuan Tian