http://localhost:11434/api/embeddings": dial tcp [::1]:11434: connect: connection refused

[vbogdanro]

Description

Hi, nice application :)

I try to use it with local ollama & weaviate vector DB

Seems to be an default communication issue, verba try to connect on IPV6 tcp [::1] and receive this error

✘ {'errors': {'error': [{'message': 'send POST request: Post
"http://localhost:11434/api/embeddings": dial tcp [::1]:11434: connect:
connection refused'}]}, 'status': 'FAILED'}

IPV4 request
curl -X POST http://localhost:11434/api/embeddings -d '{
"model": "mxbai-embed-large",
"prompt": "Why is the sky blue?"
}'
{"embedding":[-0.2021758258342743....

IPV6 request
curl -X POST http://[::1]:11434/api/embeddings -d '{
"model": "mxbai-embed-large",
"prompt": "Why is the sky blue?"
}' -H "Content-Type: application/json"
{"embedding":[-0.2021758258342743,

cat /etc/redhat-release
Fedora release 40 (Forty)
uname -a
Linux fedora-lab 6.9.11-200.fc40.x86_64 #1 SMP PREEMPT_DYNAMIC Thu Jul 25 18:17:34 UTC 2024 x86_64 GNU/Linux

I tried to disable IPV6 from OS, the same issue

Is this a bug or a feature?

• Bug
• Feature

Steps to Reproduce

goldenverba==1.0.4

grep -i host /etc/systemd/system/ollama.service
Environment="OLLAMA_HOST=0.0.0.0:11434"

netstat -tulpn | grep 11434
tcp6 0 0 :::11434 :::* LISTEN 21062/ollama

---

cat .env
WEAVIATE_URL_VERBA=http://WEAVIATE_LOCAL_IP:8080

OLLAMA_URL=http://localhost:11434
OLLAMA_MODEL=codestral:22b-v0.1-q5_0
OLLAMA_EMBED_MODEL=mxbai-embed-large

---

verba start
INFO: Will watch for changes in these directories: ['/home/bogdan/codeVB/weaviate/verba']
INFO: Uvicorn running on http://localhost:8000 (Press CTRL+C to quit)
INFO: Started reloader process [23724] using WatchFiles
/home/bogdan/venvs/verba/lib64/python3.12/site-packages/pypdf/_crypt_providers/_cryptography.py:32: CryptographyDeprecationWarning: ARC4 has been moved to cryptography.hazmat.decrepit.ciphers.algorithms.ARC4 and will be removed from this module in 48.0.0.
from cryptography.hazmat.primitives.ciphers.algorithms import AES, ARC4
ℹ Setting up client
ℹ No Auth information provided
✔ Connected to Weaviate
✘ Schema initialization failed Create class! Unexpected status code:
422, with response body: {'error': [{'message': 'vectorizer: no module with name
"text2vec-openai" present'}]}.
✘ Schema initialization failed Create class! Unexpected status code:
422, with response body: {'error': [{'message': 'vectorizer: no module with name
"text2vec-cohere" present'}]}.
ℹ Setting up components
✔ Config Saved in Weaviate
✘ {'errors': {'error': [{'message': 'send POST request: Post
"http://localhost:11434/api/embeddings": dial tcp [::1]:11434: connect:
connection refused'}]}, 'status': 'FAILED'}
ℹ Setting READER to BasicReader
ℹ Setting CHUNKER to TokenChunker
ℹ Setting EMBEDDER to ADAEmbedder
ℹ Setting RETRIEVER to WindowRetriever
ℹ Setting GENERATOR to GPT3
INFO: Started server process [23726]
INFO: Waiting for application startup.
INFO: Application startup complete.

GUI side:

If i try to import a document i revive this errors
(INFO) Importing...
(INFO) Importing 1 files with BasicReader
(INFO) Importing cerere_vieriu_bogdan_serviciul_politiei_rutiere_bac_u-1.pdf
(SUCCESS) Loaded 1 documents in 0.5s
(INFO) Starting Chunking with TokenChunker
(SUCCESS) Chunking completed with 30 chunks in 0.0s
(INFO) Starting Embedding with OllamaEmbedder
(ERROR) Embedding not successful: Document 4429dadd-7701-4865-adea-a009451e8ddd not found None
(ERROR) Unexpected error: Document 4429dadd-7701-4865-adea-a009451e8ddd not found None

server side:

✔ Loaded 1 documents in 0.5s
Chunking documents: 100%|████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 1014.59it/s]
✔ Chunking completed with 30 chunks in 0.0s
Vectorizing Chunks: 100%|████████████████████████████████████████████████████████| 30/30 [00:03<00:00, 8.70it/s]
Vectorizing document chunks: 100%|█████████████████████████████████████████████████| 1/1 [00:03<00:00, 3.45s/it]
ℹ (1/1) Importing document
xyz.pdf with 1 batches
✘ {'errors': {'error': [{'message': 'send POST request: Post
"http://localhost:11434/api/embeddings": dial tcp [::1]:11434: connect:
connection refused'}]}, 'status': 'FAILED'}
Importing batches: 100%|███████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 28.31it/s]
INFO: 127.0.0.1:39730 - "POST /api/import HTTP/1.1" 200 OK

Additional context

[thomashacker]

Thanks for the issue! Can you communicate with Ollama through the terminal? We just released the new v2 version, maybe upgrading might fix the issue

[geeman11]

Im having the same issue, even after updating to v2!

[thomashacker]

Are you receiving the same error?