Knowledge created in versions prior to 1.9.1 is not usable after upgrading to 1.9.2

TL;DR

Reporter-confirmed workaround from xiaoyesoso: change the knowledge base's embedding model, which triggers an automatic index rebuild and restores the knowledge base.

Steps

1. Contributor DhruvGorasiya identified the root cause: Weaviate 1.27.0+ requires the new named-vectors schema format (vectorConfig); existing knowledge data stored by Weaviate 1.19.0 uses the old format, so queries with weaviate-client v4 fail. The team published a migration guide and migration scripts preserving embeddings (see docs/weaviate/WEAVIATEMIGRATIONGUIDE). Reporter-confirmed workaround from xiaoyesoso: change the knowledge base's embedding model, which triggers an automatic index rebuild and restores the knowledge base.

When to use

You are seeing this: Contributor DhruvGorasiya identified the root cause: Weaviate 1.27.0+ requires the new named-vectors schema format (vectorConfig); existing knowledge data stored by Weaviate 1.19.0 uses the old format, so queries with weaviate-client v4 fail. Use this skill when you run into "Knowledge created in versions prior to 1.9.1 is not usable after upgrading to 1.9.2".

When not to use

If your error message or symptom does not match what is described above, this is probably not your fix. Search for your exact error text instead of forcing this one to fit.

Versions

Versions mentioned in the source: Weaviate 1.27.0, Weaviate 1.19.0. If you are on something much newer or older, the details may have shifted.

Why this happens

The original report does not dig into a root cause. It documents the symptom and the fix that resolved it.

Sources

• https://github.com/langgenius/dify/blob/main/CONTRIBUTING.md
• https://github.com/langgenius/dify/issues/1542
• https://github.com/langgenius/dify/discussions/categories/general
• https://github.com/langgenius/dify/issues

In gh:langgenius/dify#27291 (https://github.com/langgenius/dify/issues/27291), someone reported: Issue langgenius/dify#27291 (closed, 113 comments): ### Self Checks - [x] I have read the Contributing Guide and Language Policy. - [x] This is only for bug report, if you would like to ask a question, please head to Discussions. - [x] I have searched for existing issues search for existing issues, including closed ones. - [x] I confirm that I am using English to submit this report, otherwise it will be closed. - [x] 【中文 The verified resolution from the thread: Contributor DhruvGorasiya identified the root cause: Weaviate 1.27.0+ requires the new named-vectors schema format (vectorConfig); existing knowledge data stored by Weaviate 1.19.0 uses the old format, so queries with weaviate-client v4 fail. The team published a migration guide and migration scripts preserving embeddings (see docs/weaviate/WEAVIATEMIGRATIONGUIDE). Reporter-confirmed workaround from xiaoyesoso: change the knowledge base's embedding model, which triggers an automatic index rebuild and restores the knowledge base.