[GH-ISSUE #1885] [BUG]: When uploading a document getting network error #1226

New Issue

Originally created by @OllTee on GitHub (Jul 17, 2024).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/1885

How are you running AnythingLLM?

AnythingLLM desktop app

What happened?

When uploading a document getting network error

Are there known steps to reproduce?

Load Model for me it's Llama3 8B
Upload a file to the workspace
And when attempting to send a request i get this error:
Could not respond to message.
An error occurred while streaming response. network error

Originally created by @OllTee on GitHub (Jul 17, 2024). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/1885 ### How are you running AnythingLLM? AnythingLLM desktop app ### What happened? When uploading a document getting network error ### Are there known steps to reproduce? Load Model for me it's Llama3 8B Upload a file to the workspace And when attempting to send a request i get this error: Could not respond to message. An error occurred while streaming response. network error

yindo added the needs info / can't replicate Desktop investigating labels 2026-02-22 18:23:48 -05:00

yindo closed this issue

yindo commented

@timothycarambat commented on GitHub (Jul 17, 2024):

And when attempting to send a request i get this error:

Do you mean as a chat?

@timothycarambat commented on GitHub (Jul 17, 2024): > And when attempting to send a request i get this error: Do you mean as a chat?

yindo commented

@OllTee commented on GitHub (Jul 18, 2024):

I ask a question to it who is MattvsJapan and it anwsers wrongfully:
I import the yoututbers channel into AnythingLLM and the .html file gets imported:

3.When asking the same question again it give me this error this also happens with every other local file:

@OllTee commented on GitHub (Jul 18, 2024): 1. I ask a question to it who is MattvsJapan and it anwsers wrongfully: ![image](https://github.com/user-attachments/assets/768f5e9e-7fda-47be-a90b-8c5fff52f6d7) 2. I import the yoututbers channel into AnythingLLM and the .html file gets imported: ![image](https://github.com/user-attachments/assets/e2e5b97c-10d5-4194-9b53-804491d51aed) 3.When asking the same question again it give me this error this also happens with every other local file: ![image](https://github.com/user-attachments/assets/a9cdef28-03c8-4b10-8b77-967701c980fe)

yindo commented

@timothycarambat commented on GitHub (Jul 18, 2024):

What Embedder are you using?
What LLM are you using?
When you open the workspace's settings -> Vector database -> does "Vector count" have a non-zero number?

@timothycarambat commented on GitHub (Jul 18, 2024): - What Embedder are you using? - What LLM are you using? - When you open the workspace's settings -> Vector database -> does "Vector count" have a non-zero number?

yindo commented

@OllTee commented on GitHub (Jul 18, 2024):

Anything LLM Embedder.
Llama3 8B.
i have LanceDB and there is "There is no configuration needed for LanceDB." written on there.

@OllTee commented on GitHub (Jul 18, 2024): 1. Anything LLM Embedder. 2. Llama3 8B. 3. i have LanceDB and there is "There is no configuration needed for LanceDB." written on there.

yindo commented

@timothycarambat commented on GitHub (Jul 18, 2024):

I meant workspace settings, not the system settings - if you hover over the workspace on the sidebar there is a "gear" icon that will show - click that. That is the Workspace settings

And for the LLM, i mean the provider not the LLM model itself - sorry that was unclear

@timothycarambat commented on GitHub (Jul 18, 2024): I meant workspace settings, not the system settings - if you hover over the workspace on the sidebar there is a "gear" icon that will show - click that. That is the Workspace settings And for the LLM, i mean the provider not the LLM model itself - sorry that was unclear

yindo commented

@OllTee commented on GitHub (Jul 19, 2024):

Workspace Provider: System Default
Workspace Agent: None for now
Vector Count: 2

@OllTee commented on GitHub (Jul 19, 2024): 1. Workspace Provider: System Default 2. Workspace Agent: None for now 3. Vector Count: 2

yindo commented

@OllTee commented on GitHub (Jul 19, 2024):

and when is choose the Agent Provider as AnythingLLM it's the same error

@OllTee commented on GitHub (Jul 19, 2024): and when is choose the Agent Provider as AnythingLLM it's the same error

yindo commented

@timothycarambat commented on GitHub (Jul 19, 2024):

I still dont have the information for what LLM provider you are using. Please go to exactly this screen and tell me which provider is selected on screen.

What we can determine at least is that the connection issue is not the vector database

@timothycarambat commented on GitHub (Jul 19, 2024): I still dont have the information for _what_ LLM provider you are using. Please go to *exactly* this screen and tell me which provider is selected on screen. What we can determine at least is that the connection issue is not the vector database <img width="1775" alt="Screenshot 2024-07-19 at 3 20 23 PM" src="https://github.com/user-attachments/assets/be8cba7e-f702-4b28-89b2-4ec297ed0a9f">

yindo commented

@OllTee commented on GitHub (Jul 19, 2024):

@OllTee commented on GitHub (Jul 19, 2024): ![image](https://github.com/user-attachments/assets/7c9f97c6-db41-4527-b0d1-a293e34847b3)

yindo commented

@timothycarambat commented on GitHub (Jul 19, 2024):

Okay, are you on a machine with a GPU? If so, what is in it? It seems like the internal ollama process is not attaching or is failing to use the GPU- which since you are on Windows would not be out of the realm of possibilities.

The next best step is to get logs from debug mode
https://docs.useanything.com/debug

If you open powershell and drag the desktop icon into the terminal and remove the double quotes and press enter it will show all the logs of the app. Then you would send a chat and it would tell us what the error is.

If you swapped to another LLM this problem would "go away",

@timothycarambat commented on GitHub (Jul 19, 2024): Okay, are you on a machine with a GPU? If so, what is in it? It seems like the internal ollama process is not attaching or is failing to use the GPU- which since you are on Windows would not be out of the realm of possibilities. The next best step is to get logs from debug mode https://docs.useanything.com/debug If you open powershell and drag the desktop icon into the terminal and remove the double quotes and press enter it will show all the logs of the app. Then you would send a chat and it would tell us what the error is. If you swapped to another LLM this problem would "go away",

yindo commented

@OllTee commented on GitHub (Jul 20, 2024):

GTX 970

[Preferences] preference config stored at C:\Users\ot\AppData\Roaming\anythingllm-desktop\config.json
Prisma schema loaded from prisma\schema.prisma
Datasource "db": SQLite database "anythingllm.db" at "file:C:/Users/ot/AppData/Roaming/anythingllm-desktop/storage/anythingllm.db"

Already in sync, no schema change or pending migration was found.

Running generate... (Use --skip-generate to skip the generators)
Error: EPERM: operation not permitted, copyfile 'C:\Users\ot\AppData\Roaming\Prisma\master\61e140623197a131c2a6189271ffee05a7aa9a59\windows\libquery-engine' -> 'C:\Program Files\AnythingLLM\resources\backend\node_modules\prisma\query_engine-windows.dll.node'
Prisma schema loaded from prisma\schema.prisma
Error: EPERM: operation not permitted, copyfile 'C:\Users\ot\AppData\Roaming\Prisma\master\61e140623197a131c2a6189271ffee05a7aa9a59\windows\libquery-engine' -> 'C:\Program Files\AnythingLLM\resources\backend\node_modules\prisma\query_engine-windows.dll.node'
[OllamaProcessManager] Ollama will bind on port 11434 when booted.
[Preferences] Will load window with last know bounds.
[collector] info: Collector hot directory and tmp storage wiped!
[collector] info: [production] AnythingLLM Standalone Document processor listening on port 8888
[backend] info: [TELEMETRY ENABLED] Anonymous Telemetry enabled. Telemetry helps Mintplex Labs Inc improve AnythingLLM.
[backend] info: prisma:info
[backend] info: [TELEMETRY SENT]
[backend] info: Hot loading of AnythingLLMOllama - LLM_PROVIDER is anythingllm_ollama.
[backend] info: [NativeEmbedder] Initialized
[backend] info: [CommunicationKey] RSA key pair generated for signed payloads within AnythingLLM services.
[backend] info: [EncryptionManager] Loaded existing key & salt for encrypting arbitrary data.
[backend] info: [production] AnythingLLM Standalone Backend listening on port 3001
[OllamaProcessManager] SINGLETON LOCK: Using existing OllamaProcessManager.
[OllamaProcessManager] [windows] Ollama subprocess running. Port 11434. PID 6496.
[backend] info: OllamaAPI offline - retrying. 1/3
[backend] info: [BackgroundWorkerService] Feature is not enabled and will not be started.
[backend] info: [NativeEmbedder] Initialized
[backend] info: [NativeEmbedder] Embedded Chunk 1 of 1

@OllTee commented on GitHub (Jul 20, 2024): GTX 970 [Preferences] preference config stored at C:\Users\ot\AppData\Roaming\anythingllm-desktop\config.json Prisma schema loaded from prisma\schema.prisma Datasource "db": SQLite database "anythingllm.db" at "file:C:/Users/ot/AppData/Roaming/anythingllm-desktop/storage/anythingllm.db" Already in sync, no schema change or pending migration was found. Running generate... (Use --skip-generate to skip the generators) Error: EPERM: operation not permitted, copyfile 'C:\Users\ot\AppData\Roaming\Prisma\master\61e140623197a131c2a6189271ffee05a7aa9a59\windows\libquery-engine' -> 'C:\Program Files\AnythingLLM\resources\backend\node_modules\prisma\query_engine-windows.dll.node' Prisma schema loaded from prisma\schema.prisma Error: EPERM: operation not permitted, copyfile 'C:\Users\ot\AppData\Roaming\Prisma\master\61e140623197a131c2a6189271ffee05a7aa9a59\windows\libquery-engine' -> 'C:\Program Files\AnythingLLM\resources\backend\node_modules\prisma\query_engine-windows.dll.node' [OllamaProcessManager] Ollama will bind on port 11434 when booted. [Preferences] Will load window with last know bounds. [collector] info: Collector hot directory and tmp storage wiped! [collector] info: [production] AnythingLLM Standalone Document processor listening on port 8888 [backend] info: [TELEMETRY ENABLED] Anonymous Telemetry enabled. Telemetry helps Mintplex Labs Inc improve AnythingLLM. [backend] info: prisma:info [backend] info: [TELEMETRY SENT] [backend] info: Hot loading of AnythingLLMOllama - LLM_PROVIDER is anythingllm_ollama. [backend] info: [NativeEmbedder] Initialized [backend] info: [CommunicationKey] RSA key pair generated for signed payloads within AnythingLLM services. [backend] info: [EncryptionManager] Loaded existing key & salt for encrypting arbitrary data. [backend] info: [production] AnythingLLM Standalone Backend listening on port 3001 [OllamaProcessManager] SINGLETON LOCK: Using existing OllamaProcessManager. [OllamaProcessManager] [windows] Ollama subprocess running. Port 11434. PID 6496. [backend] info: OllamaAPI offline - retrying. 1/3 [backend] info: [BackgroundWorkerService] Feature is not enabled and will not be started. [backend] info: [NativeEmbedder] Initialized [backend] info: [NativeEmbedder] Embedded Chunk 1 of 1

yindo commented

@timothycarambat commented on GitHub (Jul 20, 2024):

To confirm, the logs above are exactly what is shown when the error is reproduced, correct? Looks like the prompt is sent, embedded, and then ???

@timothycarambat commented on GitHub (Jul 20, 2024): To confirm, the logs above are exactly what is shown when the error is reproduced, correct? Looks like the prompt is sent, embedded, and then ???

yindo commented

https://github.com/user-attachments/assets/cea12804-ceac-44b4-b5b1-e3275c884ce0

@OllTee commented on GitHub (Jul 20, 2024):

@OllTee commented on GitHub (Jul 20, 2024): https://github.com/user-attachments/assets/cea12804-ceac-44b4-b5b1-e3275c884ce0

yindo commented

@timothycarambat commented on GitHub (Jul 30, 2024):

After reviewing this for some time this detail sticks out a lot

GTX 970

What are the specs of your machine? Your machine very well may be dramatically underpowered or possibly outdatated to do CPU embedding and may not even support AVX2 instructions.

What are the full specs of your machine?

@timothycarambat commented on GitHub (Jul 30, 2024): After reviewing this for some time this detail sticks out a lot > GTX 970 What are the specs of your machine? Your machine very well may be dramatically underpowered or possibly outdatated to do CPU embedding and may not even support AVX2 instructions. What are the full specs of your machine?

yindo commented

@OllTee commented on GitHub (Aug 2, 2024):

@OllTee commented on GitHub (Aug 2, 2024): ![image](https://github.com/user-attachments/assets/012489b0-154d-4eea-87c3-25c752337ca9)

yindo commented

@timothycarambat commented on GitHub (Aug 2, 2024):

Your computer does not support AVX2 instruction sets (AVX2 is like 10 years old now!!) so you cannot use the default lancedb. You can set up a cloud or local vector database via any of the other vector storage providers, but you can't use the no-setup default with a machine that old.

Alternatively, you can use a cloud-hosted version of AnythingLLM and avoid that entirely. That is why you have been getting such odd errors.

@timothycarambat commented on GitHub (Aug 2, 2024): <img width="926" alt="Screenshot 2024-08-02 at 1 50 16 PM" src="https://github.com/user-attachments/assets/e8db494c-d5a0-4871-8915-99fe54447d2d"> Your computer does not support AVX2 instruction sets (AVX2 is like 10 years old now!!) so you cannot use the default `lancedb`. You can set up a cloud or local vector database via any of the other vector storage providers, but you can't use the no-setup default with a machine that old. Alternatively, you can use a cloud-hosted version of AnythingLLM and avoid that entirely. That is why you have been getting such odd errors.

yindo commented