What should be noted when using VectorDBBench to test Milvus in a local area network (LAN) environment?

[jkc132454-blip]

I have downloaded train.parquet, shuffle_train.parquet, neighbors.parquet, and scalar_labels.parquet offline and placed them in the /tmp/vectordb_bench/dataset/openai/openai_small_50k/ directory. However, when I run vectordbbench batchcli --batch-config-file vectordb_bench/config-files/batch.yml, I get an error as shown in the screenshot. How should I resolve this?

[alwayslove2013]

@jkc132454-blip, I understand you want to run vdbbench in Local mode without it accessing the network for validation.

To bypass the integrity check that compares local files against AWS S3 or Aliyun OSS, you can simply modify the source code to make the validation functions return true directly.

This will allow the tool to use your locally downloaded dataset without attempting any online verification.

VectorDBBench/vectordb_bench/backend/data_source.py

Lines 60 to 62 in 631831b

	def validate_file(self, remote: pathlib.Path, local: pathlib.Path) -> bool:
	info = self.bucket.get_object_meta(remote.as_posix())

VectorDBBench/vectordb_bench/backend/data_source.py

Lines 147 to 149 in 631831b

	def validate_file(self, remote: pathlib.Path, local: pathlib.Path) -> bool:
	# info() uses ls() inside, maybe we only need to ls once
	info = self.fs.info(remote)