Vertex AI Vector Search: GCS path mismatch on index update (llama-index)
Vertex AI Vector Search: GCS path mismatch on index update (llama-index): : the LlamaIndex Vertex AI vector store uploads batch-update files to gs://bucket/{uuid}/ but Vertex looks for them at gs://bucket/index/{uuid}/.
[run-llama/llama_index#20278]: the LlamaIndex Vertex AI vector store uploads batch-update files to gs://bucket/{uuid}/ but Vertex looks for them at gs://bucket/index/{uuid}/. Workaround: catch the ValueError, extract the UUID from the message, copy the uploaded blobs from the {uuid} prefix into the index/{uuid} prefix with the Storage client, then retry. General checklist for index updates: index deployed on an Index Endpoint, dimensions matching the deployed index config, service account with Vertex AI permissions.
Context: Batch index update fails with 'gs://bucket/index/{uuid} is not a valid directory or the directory has empty content'. Files were uploaded to a different prefix.
Matched source
Source: Source: https://github.com/run-llama/llama_index/issues/20278 Original query: "Vertex AI Vector Search: GCS path mismatch on index update (llama-index)" Key terms: index, llama, mismatch, path, search, update, vector, vertex
Maintainer review
No maintainer verification is recorded for this version.
This records the version a maintainer checked. It does not assert that the version is the latest upstream release.