[BUG] error in upsert config: Error: documentStoreServices.insertIntoVectorStore - Error: documentStoreServices._insertIntoVectorStoreWorkerThread - TypeError: fetch failed

[bodzebod]

Describe the bug
Hi,
Using Flowise AI 2.1.5.

I try to set up a document store with a file loader. The upsert config is made of:

• ollama embedding model
• postgresql with pgvector as vector store
• the same postgresql as record manager

When clicking "upsert", I get the error:

Status: 500
Error: documentStoreServices.insertIntoVectorStore - Error: documentStoreServices._insertIntoVectorStoreWorkerThread - TypeError: fetch failed

I don't see any output in flowiseAI docker log. Is it possible to increase the level of verbosity for debugging?

To Reproduce
Steps to reproduce the behavior:

1. Create a document store with a file loader
2. Reproduce the same upsert config as in the screenshot
3. Click upsert button
4. See error

Expected behavior
I expect a successful upsert.

Screenshots
If applicable, add screenshots to help explain your problem.

Flow

Setup

• Installation: Docker
• Flowise Version: 2.1.5
• OS: Linux
• Browser: Edge

Additional context
Many thanks for your help!

[bodzebod]

I found the issue. For embedding, I'm using ollama serving from a remote server.
I updated it recently, and forgot to update the file /etc/systemd/system/ollama.service to allow remote access, i.e:
Environment="OLLAMA_HOST=0.0.0.0"

After updating the service (sudo systemctl daemon-reload) and restarting (sudo systemctl restart ollama), I no longer have the error. The error was occuring because the embedding model was not accessible.

[laurentbekaert593-rgb]

its not working?

[laurentbekaert593-rgb]

i have this error

Status: 500
Error: documentStoreServices.insertIntoVectorStoreMiddleware - Error: documentStoreServices.insertIntoVectorStore - Error: documentStoreServices._insertIntoVectorStoreWorkerThread - Error: getaddrinfo ENOTFOUND postgres