Avoid random generator races in concurrent Storage tests

[JinmingHu (Jinming-Hu)]

Purpose

Avoid calling StorageTest random helpers from concurrent test workers. These helpers are explicitly not thread-safe and can generate conflicting local filenames or resource names under load.

Changes

• Generate concurrent download filenames before launching workers in Blobs, Data Lake, and File Shares tests.
• Generate structured-message blob names before launching Blob workers.
• Use deterministic per-worker names in the disabled Data Lake pagination stress test.

Validation

• Built azure-storage-blobs-test, azure-storage-files-datalake-test, and azure-storage-files-shares-test in Release.
• Passed 3 consecutive LIVE iterations of:
  • BlockBlobClientTest.ConcurrentDownload_LIVEONLY_
  • BlockBlobClientTest.StructuredMessageTest_Concurrent_LIVEONLY_
  • DataLakeFileClientTest.ConcurrentDownload_LIVEONLY_
  • FileShareFileClientTest.ConcurrentDownload_LIVEONLY_
• The disabled Data Lake pagination stress test was compile-validated only because it requires a special AAD app without RBAC permissions.

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).
9 pipeline(s) were filtered out due to trigger conditions.
There may be pipelines that require an authorized user to comment /azp run to run.

[JinmingHu (Jinming-Hu)]

/azp run cpp - storage

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).
9 pipeline(s) were filtered out due to trigger conditions.
There may be pipelines that require an authorized user to comment /azp run to run.

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[Copilot]

Pull request overview

This PR updates Azure Storage unit tests to avoid calling non-thread-safe StorageTest random helpers from concurrent worker threads, preventing filename/blob-name collisions and race conditions during stress/concurrency runs.

Changes:

• Pre-generate per-worker download-to-file temporary filenames on the main thread before launching std::async workers (Blobs, Data Lake, File Shares).
• Pre-generate structured-message blob names before launching concurrent blob workers.
• Replace per-thread RandomString() file naming with deterministic workerIndex-fileIndex naming in the disabled Data Lake pagination stress test.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated no comments.

File	Description
sdk/storage/azure-storage-files-shares/test/ut/share_file_client_test.cpp	Passes a pre-generated temp filename into the async download-to-file worker to avoid thread-unsafe random calls.
sdk/storage/azure-storage-files-datalake/test/ut/datalake_path_client_test.cpp	Uses deterministic per-worker/per-file names to avoid concurrent RandomString() usage in a disabled stress test.
sdk/storage/azure-storage-files-datalake/test/ut/datalake_file_client_test.cpp	Moves temp filename generation out of the async worker by passing it as an argument.
sdk/storage/azure-storage-blobs/test/ut/block_blob_client_test.cpp	Moves temp filename and structured-message blob-name generation out of concurrent workers to avoid random-helper races.

---

💡 Add a code-review agent skill for context-aware, tailored reviews. Learn more in the docs.