langgenius/dify
1
0
Fork 0
mirror of https://github.com/langgenius/dify.git synced 2026-07-21 20:15:31 -04:00
Code Issues 775 Packages Projects Releases 10 Wiki Activity

The full-text search recall test has been yielding no results, and the VECTOR_STORE setting is not working #20923

New Issue
Closed
opened 2026-02-21 20:09:48 -05:00 by yindo · 1 comment
yindo commented 2026-02-21 20:09:48 -05:00
Owner
Copy Link
Copy Source

Originally created by @xiaoqi951109 on GitHub (Dec 10, 2025).

Self Checks

  • I have read the Contributing Guide and Language Policy.
  • This is only for bug report, if you would like to ask a question, please head to Discussions.
  • I have searched for existing issues search for existing issues, including closed ones.
  • I confirm that I am using English to submit this report, otherwise it will be closed.
  • 【中文用户 & Non English User】请使用英语提交，否则会被关闭 ：）
  • Please do not modify this template :) and fill in all the required fields.

Dify version

1.10.1

Cloud or Self Hosted

Self Hosted (Docker)

Steps to reproduce

When I used the main branch code 1.10.1 for Docker Compose self-hosted deployment, I found that the full-text index of the knowledge base did not work. No matter how I conducted recall tests, there were no results. However, the same documents and models worked normally in the cloud.
At the same time, I found that even if the vector database is set to something like elasticsearch or qdrant, the recall test for the [VECTOR_STORE=elasticsearch] document would still report an error, indicating that it cannot connect to Weaviate.
[Could not connect to Weaviate:Connection to Weaviate failed. Details: .]
When I tried to download version 1.9.2 and followed the same steps as mentioned above, everything worked fine.

✔️ Expected Behavior

full-text index of the knowledge base did work by weaviate and elasticsearch

❌ Actual Behavior

No response

Originally created by @xiaoqi951109 on GitHub (Dec 10, 2025). ### Self Checks - [x] I have read the [Contributing Guide](https://github.com/langgenius/dify/blob/main/CONTRIBUTING.md) and [Language Policy](https://github.com/langgenius/dify/issues/1542). - [x] This is only for bug report, if you would like to ask a question, please head to [Discussions](https://github.com/langgenius/dify/discussions/categories/general). - [x] I have searched for existing issues [search for existing issues](https://github.com/langgenius/dify/issues), including closed ones. - [x] I confirm that I am using English to submit this report, otherwise it will be closed. - [x] 【中文用户 & Non English User】请使用英语提交，否则会被关闭 ：） - [x] Please do not modify this template :) and fill in all the required fields. ### Dify version 1.10.1 ### Cloud or Self Hosted Self Hosted (Docker) ### Steps to reproduce When I used the main branch code 1.10.1 for Docker Compose self-hosted deployment, I found that the full-text index of the knowledge base did not work. No matter how I conducted recall tests, there were no results. However, the same documents and models worked normally in the cloud. At the same time, I found that even if the vector database is set to something like elasticsearch or qdrant, the recall test for the [VECTOR_STORE=elasticsearch] document would still report an error, indicating that it cannot connect to Weaviate. [Could not connect to Weaviate:Connection to Weaviate failed. Details: .] When I tried to download version 1.9.2 and followed the same steps as mentioned above, everything worked fine. ### ✔️ Expected Behavior full-text index of the knowledge base did work by weaviate and elasticsearch ### ❌ Actual Behavior _No response_
yindo added the 👻 feat:rag label 2026-02-21 20:09:48 -05:00
yindo closed this issue 2026-02-21 20:09:48 -05:00
yindo commented 2026-02-21 20:09:49 -05:00
Author
Owner
Copy Link
Copy Source

@dosubot[bot] commented on GitHub (Dec 10, 2025):

Hi @xiaoqi951109! I'm Dosu and I’m helping the dify team.

The main cause is a change in Docker Compose profiles in Dify v1.10.1: the Weaviate vector database is now optional and not started by default, so if the Weaviate container isn’t running, recall tests and full-text indexing will fail—even if you set VECTOR_STORE to elasticsearch or qdrant. You need to explicitly start the correct vector store container using the profile, for example:
docker compose --profile weaviate -p dify up -d
details.

Also, once a dataset is created, its vector store type is saved in its metadata. Changing VECTOR_STORE in your environment only affects new datasets; existing ones will keep using the original backend (often Weaviate), so recall tests may still try to connect to Weaviate unless you delete/recreate the dataset or migrate its index_struct source.

To fix:

  1. Make sure the correct vector store container is running (Weaviate, Elasticsearch, or Qdrant) and healthy.
  2. Double-check your .env and middleware.env files for VECTOR_STORE, DB_TYPE, and COMPOSE_PROFILES settings.
  3. Restart your containers after making changes.
  4. For existing datasets, consider deleting/recreating them if you want to switch vector stores.

If you still see connection errors, check container logs and network settings. If you run into UUID or MySQL issues, upgrading to a newer Dify version is recommended reference.

To reply, just mention @dosu.

How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other

Chat with Dosu Open in Cursor Join Discord Share on X

@dosubot[bot] commented on GitHub (Dec 10, 2025): <!-- Greeting --> Hi @xiaoqi951109! I'm [Dosu](https://go.dosu.dev/dosubot) and I’m helping the dify team. <!-- Answer --> The main cause is a change in Docker Compose profiles in Dify v1.10.1: the Weaviate vector database is now optional and not started by default, so if the Weaviate container isn’t running, recall tests and full-text indexing will fail—even if you set VECTOR_STORE to elasticsearch or qdrant. You need to explicitly start the correct vector store container using the profile, for example: `docker compose --profile weaviate -p dify up -d` [details](https://github.com/langgenius/dify/issues/29135). Also, once a dataset is created, its vector store type is saved in its metadata. Changing VECTOR_STORE in your environment only affects new datasets; existing ones will keep using the original backend (often Weaviate), so recall tests may still try to connect to Weaviate unless you delete/recreate the dataset or migrate its index_struct [source](https://github.com/langgenius/dify/blob/88b20bc6d00cc668d733c69e0316af8c8f71e734/api/core/rag/datasource/vdb/vector_factory.py#L38-L327). To fix: 1. Make sure the correct vector store container is running (Weaviate, Elasticsearch, or Qdrant) and healthy. 2. Double-check your `.env` and `middleware.env` files for VECTOR_STORE, DB_TYPE, and COMPOSE_PROFILES settings. 3. Restart your containers after making changes. 4. For existing datasets, consider deleting/recreating them if you want to switch vector stores. If you still see connection errors, check container logs and network settings. If you run into UUID or MySQL issues, upgrading to a newer Dify version is recommended [reference](https://github.com/langgenius/dify/issues/29181). <!-- Dosu Comment Footer --> *To reply, just mention [@dosu](https://go.dosu.dev/dosubot).* --- <sup>How did I do? [Good](https://app.dosu.dev/response-feedback/afa69576-5093-4be9-be1a-81eed01c38f0?feedback_type=great_response) | [Irrelevant](https://app.dosu.dev/response-feedback/afa69576-5093-4be9-be1a-81eed01c38f0?feedback_type=irrelevant_answer) | [Incorrect](https://app.dosu.dev/response-feedback/afa69576-5093-4be9-be1a-81eed01c38f0?feedback_type=incorrect_sources) | [Verbose](https://app.dosu.dev/response-feedback/afa69576-5093-4be9-be1a-81eed01c38f0?feedback_type=too_verbose) | [Hallucination](https://app.dosu.dev/response-feedback/afa69576-5093-4be9-be1a-81eed01c38f0?feedback_type=hallucination) | [Report 🐛](https://app.dosu.dev/response-feedback/afa69576-5093-4be9-be1a-81eed01c38f0?feedback_type=bug_report) | [Other](https://app.dosu.dev/response-feedback/afa69576-5093-4be9-be1a-81eed01c38f0?feedback_type=other)</sup> [![Chat with Dosu](https://dosu.dev/dosu-chat-badge.svg)](https://app.dosu.dev/6088a4f4-073a-43f3-b3d4-4f06fb0c05a0/ask?utm_source=github)&nbsp;[![Open in Cursor](https://dosu.dev/dosu-in-cursor.svg)](https://cursor.com/link/prompt?text=The%20main%20cause%20is%20a%20change%20in%20Docker%20Compose%20profiles%20in%20Dify%20v1.10.1%3A%20the%20Weaviate%20vector%20database%20is%20now%20optional%20and%20not%20started%20by%20default%2C%20so%20if%20the%20Weaviate%20container%20isn%E2%80%99t%20running%2C%20recall%20tests%20and%20full-text%20indexing%20will%20fail%E2%80%94even%20if%20you%20set%20VECTOR_STORE%20to%20elasticsearch%20or%20qdrant.%20You%20need%20to%20explicitly%20start%20the%20correct%20vector%20store%20container%20using%20the%20profile%2C%20for%20example%3A%20%20%0A%60docker%20compose%20--profile%20weaviate%20-p%20dify%20up%20-d%60%20%20%0A%5Bdetails%5D%28https%3A//github.com/langgenius/dify/issues/29135%29.%0A%0AAlso%2C%20once%20a%20dataset%20is%20created%2C%20its%20vector%20store%20type%20is%20saved%20in%20its%20metadata.%20Changing%20VECTOR_STORE%20in%20your%20environment%20only%20affects%20new%20datasets%3B%20existing%20ones%20will%20keep%20using%20the%20original%20backend%20%28often%20Weaviate%29%2C%20so%20recall%20tests%20may%20still%20try%20to%20connect%20to%20Weaviate%20unless%20you%20delete/recreate%20the%20dataset%20or%20migrate%20its%20index_struct%20%5Bsource%5D%28https%3A//github.com/langgenius/dify/blob/88b20bc6d00cc668d733c69e0316af8c8f71e734/api/core/rag/datasource/vdb/vector_factory.py%23L38-L327%29.%0A%0ATo%20fix%3A%20%20%0A1.%20Make%20sure%20the%20correct%20vector%20store%20container%20is%20running%20%28Weaviate%2C%20Elasticsearch%2C%20or%20Qdrant%29%20and%20healthy.%20%20%0A2.%20Double-check%20your%20%60.env%60%20and%20%60middleware.env%60%20files%20for%20VECTOR_STORE%2C%20DB_TYPE%2C%20and%20COMPOSE_PROFILES%20settings.%20%20%0A3.%20Restart%20your%20containers%20after%20making%20changes.%20%20%0A4.%20For%20existing%20datasets%2C%20consider%20deleting/recreating%20them%20if%20you%20want%20to%20switch%20vector%20stores.%0A%0AIf%20you%20still%20see%20connection%20errors%2C%20check%20container%20logs%20and%20network%20settings.%20If%20you%20run%20into%20UUID%20or%20MySQL%20issues%2C%20upgrading%20to%20a%20newer%20Dify%20version%20is%20recommended%20%5Breference%5D%28https%3A//github.com/langgenius/dify/issues/29181%29.)&nbsp;[![Join Discord](https://img.shields.io/badge/join-5865F2?logo=discord&logoColor=white&label=)](https://go.dosu.dev/discord-bot)&nbsp;[![Share on X](https://img.shields.io/badge/X-share-black)](https://twitter.com/intent/tweet?text=%40dosu_ai%20helped%20me%20solve%20this%20issue!&url=https%3A//github.com/langgenius/dify/issues/29406)
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
Knowledge Pipeline
Proposal
WIP
cloud
dependencies
feat:webapp
good first issue
outdated
performance
pull-request
Mirrored from GitHub Pull Request
 python:uv
refactor
size:L
stale
status: accepting prs
status: blocked
status: in discussion
stt & tts
vulnerablity
web
wont-fix
🌊 feat:workflow
🌚 invalid
🐞 bug
👻 feat:rag
💪 enhancement
📚 documentation
🔨 feat:tools
🙋‍♂️ question
🙏 help wanted
🤔 cant-reproduce
🤖 feat:agent
No Label 👻 feat:rag
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langgenius/dify#20923
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?