[Store] Structure FilereadWorkerPoolConfig environment settings

[bitborne]

Description

Part of #3809. Extract MC_FILEREAD_WORKERS from the transfer-task source into a private, client-owned FilereadWorkerPoolConfig and the common environment catalog. The worker pool still takes one process-wide snapshot on first construction, defaults to 10 workers for absent/empty/invalid values, and emits the existing warning with the raw input for a nonempty invalid value. The unrelated NoF settings and the public worker-pool API stay unchanged.

The shared typed integer parser is used for nonempty input. This intentionally accepts surrounding ASCII whitespace (for example, "2 "), which the previous strtol path rejected; the supported positive integer behavior and first-use caching are covered by direct config and worker-count tests. This is a single-owner extraction following merged PRs #3872 and #4036.

Open PR #3424 also edits transfer_task.cpp while refactoring NoF internals, including the adjacent #ifdef USE_NOF boundary. It does not migrate MC_FILEREAD_WORKERS; if it merges before this PR, the NoF guard overlap needs resolving at integration time. This PR does not rebase or modify NoF behavior preemptively.

Module

• Mooncake Store (mooncake-store)
• Common (mooncake-common)

Type of Change

• Refactor

How Has This Been Tested?

Test commands:

cmake --build build --parallel 232 --target fileread_worker_pool_config_test fileread_worker_pool_test transfer_task_test mooncake_store_client_objects
ctest --test-dir build -R '^(fileread_worker_pool_config_test|transfer_task_test)$' --output-on-failure
GTEST_FILTER='*-TransferTaskTest.BatchGetOffloadObjectCopiesPinnedHostToGpu' ctest --test-dir build -R '^(fileread_worker_pool_config_test|fileread_worker_pool_test|transfer_task_test)$' --output-on-failure
ctest --test-dir build -R '^fileread_worker_pool_test$' --repeat until-fail:10 --output-on-failure
./scripts/code_format.sh --check --staged
pre-commit run --files mooncake-common/include/environment_variables.h mooncake-store/src/CMakeLists.txt mooncake-store/src/config/fileread_worker_pool_config.cpp mooncake-store/src/config/fileread_worker_pool_config.h mooncake-store/src/transfer_task.cpp mooncake-store/tests/CMakeLists.txt mooncake-store/tests/fileread_worker_pool_config_test.cpp mooncake-store/tests/fileread_worker_pool_test.cpp

Test results:

• Unit tests pass for the new config, isolated worker-pool lifecycle, and all non-GPU baseline-failing transfer-task cases (3/3 focused CTest targets with that one Google Test case filtered).
• Integration tests pass (if applicable)
• Manual testing done (describe below)

The unfiltered target has one pre-existing GPU pinned-host allocation failure: TransferTaskTest.BatchGetOffloadObjectCopiesPinnedHostToGpu gets a null buffer. The same test fails at the same assertion in the independently built branch with unmodified fileread worker code, and fails when run alone; it is not caused by this extraction. Before the production change, the new whitespace/worker-count test failed against the old parser; after the change its isolated binary passes 10 repetitions. The client object target builds, and touched-file pre-commit and staged-line clang-format 20 checks pass.

Checklist

• I have performed a self-review of my own code
• I have formatted my code using ./scripts/code_format.sh
• I have run pre-commit on the files changed in this PR and all hooks pass
• I have updated the documentation (if applicable)
• I have added tests to prove my changes are effective
• For changes >500 LOC: I have filed an RFC issue

No documentation update: the public API and positive-integer contract remain the same; whitespace acceptance is described above.

AI Assistance Disclosure

• No AI tools were used
• AI tools were used (specify below)

Codex assisted with ownership analysis, implementation, RED-to-GREEN focused tests, build/format/hook verification, and review. The human submitter will review every changed line and is responsible for the change end to end.