Mintplex-Labs/anything-llm
1
0
Fork 0
mirror of https://github.com/Mintplex-Labs/anything-llm.git synced 2026-07-19 14:13:35 -04:00
Code Issues 328 Packages Projects Releases 10 Wiki Activity

[GH-ISSUE #4962] [BUG]: Illegal Instruction Error Causes Container Crash with Generic OpenAI API LLM Provider #3113

New Issue
Closed
opened 2026-02-22 18:32:39 -05:00 by yindo · 2 comments
yindo commented 2026-02-22 18:32:39 -05:00
Owner
Copy Link
Copy Source

Originally created by @siver1152 on GitHub (Feb 3, 2026).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/4962

How are you running AnythingLLM?

Docker (remote machine)

What happened?

Critical Bug Report: Generic OpenAI API Connection Causes Container Crash When Using Aphrodite Engine Inference Server

Issue Summary

I'm experiencing a critical bug when trying to connect to a generic OpenAI API endpoint causing a complete crash of the container itself.

Infrastructure Setup

Currently I have a server running Docker that has AnythingLLM running on it. On the same system I'm running my inference server which is outside of Docker running via Python virtual environment. I'm connecting to AnythingLLM via my desktop which is on the same subnet as the server. I can reach Aphrodite Engine from all 3 locations including from within the Docker container of AnythingLLM and a different random other Docker container using the info below in the report.

I'm running Aphrodite Engine, which exposes an OpenAI-compatible API. I configured it using the following Base URLs (I tested all variations):

  • http://<local IP, Docker IP, localhost>:2242/v1
  • http://<local IP, Docker IP, localhost>:2242/v1/
  • http://<local IP, Docker IP, localhost>:2242
  • http://<local IP, Docker IP, localhost>:2242/v1/chat/completions

I've tried using a random value for the API key as well as leaving the API key blank. I also attempted multiple string types for the API key along with variations in key length, but this did not change the behavior.

Configuration Details

For the chat model name, I tested:

  • Qwen/Qwen3-0.6B
  • Qwen3-0.6B

Model context window: 40,960
Max tokens: 48,832

I also tested these same settings with LM Studio and LocalAI. I specifically tried connecting the Aphrodite Engine OpenAI-compatible API to those other LLM providers to see how they would behave with the same endpoint. I tested them using the same settings and variations of settings, and I also applied any configuration changes that were allowed or required by those providers.

Evidence that the API is Reachable

When I add the generic OpenAI API to LM Studio and LocalAI, I can see in the logs that AnythingLLM is successfully making requests to the inference server. Here are some example log entries:

INFO 19:16:36 172.17.0.2:60508 - "GET /v1/models HTTP/1.1" 200
INFO 19:17:11 172.17.0.2:54448 - "GET /api/v0/models HTTP/1.1" 404
INFO 19:17:15 172.17.0.2:34108 - "GET /v1/models HTTP/1.1" 200
INFO 19:17:52 172.17.0.2:45374 - "GET /v1/models HTTP/1.1" 200
INFO 19:19:50 172.17.0.2:56500 - "GET /v1/models HTTP/1.1" 200

So the system is successfully detecting the model running in Aphrodite.

The configurations save successfully in all three providers. However, when testing the two other LLM providers (LM Studio and LocalAI), sending a prompt resulted in the same network streaming error, and the same container crash behavior occurred there as well.

Steps to Reproduce

  1. Setup a clean instance of AnythingLLM which allows for remote endpoints to connect
  2. setup and build Aphrodite Engine (link to the repo at the very bottom of this report)
  3. Configure the Generic OpenAI API provider using any one of the Base URLs listed above
    (you can also try and select LLM Studio or Local-AI for the LLM provider to see that the container and AnythingLLM can hit and reach the inference server via Generic OpenAI API)
  4. Save the configuration (this succeeds)
  5. Go to chat
  6. Send a prompt such as: ```
    tell me a joke

Expected Behavior

The system should stream and return a valid completion response from the Aphrodite OpenAI-compatible API.

Actual Behavior

I receive:

Could not respond to message.
An error occurred while streaming response. Network error.

Immediately after that, the entire Docker container crashes and must be restarted.

Network Connectivity Verification

I can confirm that I'm able to reach the generic API from:

  • The local system itself using localhost, the Docker IP and port, and the local network IP from outside the Docker container
  • A remote device using the local IP address of the host machine
  • From within the Docker container using local, remote, and Docker IP addresses, as well as the Docker hostname

Troubleshooting Performed

I have completely reset and restarted Docker multiple times. No other configurations were changed beforehand. The container crashes regardless of where or how I configure it. If I configure it during the initial setup, it also causes the container to crash.

_____________________________________________________________

Docker Container Logs

docker logs anythingllm
[collector] info: [TikTokenTokenizer] Initialized new TikTokenTokenizer instance.
[collector] info: Collector hot directory and tmp storage wiped!
[collector] info: Document processor app listening on port 8888
Environment variables loaded from .env
Prisma schema loaded from prisma/schema.prisma

✔ Generated Prisma Client (v5.3.1) to ./node_modules/@prisma/client in 668ms

Start using Prisma Client in Node.js (See: https://pris.ly/d/client)

import { PrismaClient } from '@prisma/client'
const prisma = new PrismaClient()

or start using Prisma Client at the edge (See: https://pris.ly/d/accelerate)

import { PrismaClient } from '@prisma/client/edge'
const prisma = new PrismaClient()


See other ways of importing Prisma Client: http://pris.ly/d/importing-client

Environment variables loaded from .env
Prisma schema loaded from prisma/schema.prisma
Datasource "db": SQLite database "anythingllm.db" at "file:../storage/anythingllm.db"

SQLite database anythingllm.db created at file:../storage/anythingllm.db

34 migrations found in prisma/migrations

Applying migration `20230921191814_init`
Applying migration `20231101001441_init`
Applying migration `20231101195421_init`
Applying migration `20231129012019_add`
Applying migration `20240113013409_init`
Applying migration `20240118201333_init`
Applying migration `20240202002020_init`
Applying migration `20240206181106_init`
Applying migration `20240206211916_init`
Applying migration `20240208224848_init`
Applying migration `20240210004405_init`
Applying migration `20240216214639_init`
Applying migration `20240219211018_init`
Applying migration `20240301002308_init`
Applying migration `20240326231053_init`
Applying migration `20240405015034_init`
Applying migration `20240412183346_init`
Applying migration `20240425004220_init`
Applying migration `20240430230707_init`
Applying migration `20240510032311_init`
Applying migration `20240618224346_init`
Applying migration `20240821215625_init`
Applying migration `20240824005054_init`
Applying migration `20241003192954_init`
Applying migration `20241029203722_init`
Applying migration `20241029233509_init`
Applying migration `20250102204948_init`
Applying migration `20250226005538_init`
Applying migration `20250318154720_init`
Applying migration `20250506214129_init`
Applying migration `20250709230835_init`
Applying migration `20250725194841_init`
Applying migration `20250808171557_init`
Applying migration `20260130040204_init`

The following migrations have been applied:

migrations/
  └─ 20230921191814_init/
    └─ migration.sql
  └─ 20231101001441_init/
    └─ migration.sql
  └─ 20231101195421_init/
    └─ migration.sql
  └─ 20231129012019_add/
    └─ migration.sql
  └─ 20240113013409_init/
    └─ migration.sql
  └─ 20240118201333_init/
    └─ migration.sql
  └─ 20240202002020_init/
    └─ migration.sql
  └─ 20240206181106_init/
    └─ migration.sql
  └─ 20240206211916_init/
    └─ migration.sql
  └─ 20240208224848_init/
    └─ migration.sql
  └─ 20240210004405_init/
    └─ migration.sql
  └─ 20240216214639_init/
    └─ migration.sql
  └─ 20240219211018_init/
    └─ migration.sql
  └─ 20240301002308_init/
    └─ migration.sql
  └─ 20240326231053_init/
    └─ migration.sql
  └─ 20240405015034_init/
    └─ migration.sql
  └─ 20240412183346_init/
    └─ migration.sql
  └─ 20240425004220_init/
    └─ migration.sql
  └─ 20240430230707_init/
    └─ migration.sql
  └─ 20240510032311_init/
    └─ migration.sql
  └─ 20240618224346_init/
    └─ migration.sql
  └─ 20240821215625_init/
    └─ migration.sql
  └─ 20240824005054_init/
    └─ migration.sql
  └─ 20241003192954_init/
    └─ migration.sql
  └─ 20241029203722_init/
    └─ migration.sql
  └─ 20241029233509_init/
    └─ migration.sql
  └─ 20250102204948_init/
    └─ migration.sql
  └─ 20250226005538_init/
    └─ migration.sql
  └─ 20250318154720_init/
    └─ migration.sql
  └─ 20250506214129_init/
    └─ migration.sql
  └─ 20250709230835_init/
    └─ migration.sql
  └─ 20250725194841_init/
    └─ migration.sql
  └─ 20250808171557_init/
    └─ migration.sql
  └─ 20260130040204_init/
    └─ migration.sql
      
All migrations have been successfully applied.
[backend] info: [EncryptionManager] Loaded existing key & salt for encrypting arbitrary data.
[backend] info: [TokenManager] Initialized new TokenManager instance for model: gpt-3.5-turbo
[backend] info: [ContextWindowFinder] Pulling remote model map...
[backend] info: [TokenManager] Returning existing instance for model: gpt-3.5-turbo
[backend] info: prisma:info Starting a sqlite pool with 65 connections.
[backend] info: [ONBOARDING PATCH] Legacy instance is already onboarded, marking onboarding as complete. You will not see this message again.
[backend] info: [TELEMETRY DISABLED] Telemetry is marked as disabled - no events will send. Telemetry helps Mintplex Labs Inc improve AnythingLLM.
[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: [PushNotifications] Generating new VAPID keys...
[backend] info: [PushNotifications] New VAPID keys generated!
[backend] info: [BackgroundWorkerService] Starting...
[backend] info: [BackgroundWorkerService] Service started with 1 jobs ["cleanup-orphan-documents"]
[backend] info: [PushNotifications] Loading single user mode subscriptions...
[backend] info: Primary server in HTTP mode listening on port 3001
[backend] info: [ContextWindowFinder] zai:titan-embed-text-v2 - context window is not a positive number. Got 0.
[backend] info: [ContextWindowFinder] zai:embed-v4.0 - context window is not a positive number. Got 0.
[backend] info: [ContextWindowFinder] zai:gemini-embedding-001 - context window is not a positive number. Got 0.
[backend] info: [ContextWindowFinder] zai:text-embedding-005 - context window is not a positive number. Got 0.
[backend] info: [ContextWindowFinder] zai:text-multilingual-embedding-002 - context window is not a positive number. Got 0.
[backend] info: [ContextWindowFinder] zai:codestral-embed - context window is not a positive number. Got 0.
[backend] info: [ContextWindowFinder] zai:mistral-embed - context window is not a positive number. Got 0.
[backend] info: [ContextWindowFinder] zai:text-embedding-3-large - context window is not a positive number. Got 0.
[backend] info: [ContextWindowFinder] zai:text-embedding-3-small - context window is not a positive number. Got 0.
[backend] info: [ContextWindowFinder] zai:text-embedding-ada-002 - context window is not a positive number. Got 0.
[backend] info: [ContextWindowFinder] Remote model map synced and cached
prisma:info Starting a sqlite pool with 65 connections.
[backend] info: [220:273]: No direct uploads path found - exiting.
[bg-worker][cleanup-orphan-documents] info: [220:273]: No direct uploads path found - exiting.
[backend] warn: Child process exited with code 0 and signal null
[backend] info: Worker for job "cleanup-orphan-documents" exited with code 0
[backend] info: [MetaGenerator] fetching custom meta tag settings...
[backend] error: LMStudio:listModels Connection error.
[backend] error: LMStudio:listModels Connection error.
[backend] error: LMStudio:listModels Connection error.
[backend] error: LMStudio:listModels Connection error.
[backend] info: [Event Logged] - update_llm_provider
[backend] info: [Event Logged] - workspace_created
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [NativeEmbedder] Initialized Xenova/all-MiniLM-L6-v2
[backend] info: [LMStudio] initialized with model: Qwen/Qwen3-0.6B
/usr/local/bin/docker-entrypoint.sh: line 26:   220 Illegal instruction     (core dumped) node /app/server/index.js
[collector] info: [TikTokenTokenizer] Initialized new TikTokenTokenizer instance.
[collector] info: Collector hot directory and tmp storage wiped!
[collector] info: Document processor app listening on port 8888
Environment variables loaded from .env
Prisma schema loaded from prisma/schema.prisma

✔ Generated Prisma Client (v5.3.1) to ./node_modules/@prisma/client in 622ms

Start using Prisma Client in Node.js (See: https://pris.ly/d/client)

import { PrismaClient } from '@prisma/client'
const prisma = new PrismaClient()

or start using Prisma Client at the edge (See: https://pris.ly/d/accelerate)

import { PrismaClient } from '@prisma/client/edge'
const prisma = new PrismaClient()


See other ways of importing Prisma Client: http://pris.ly/d/importing-client

Environment variables loaded from .env
Prisma schema loaded from prisma/schema.prisma
Datasource "db": SQLite database "anythingllm.db" at "file:../storage/anythingllm.db"

34 migrations found in prisma/migrations


No pending migrations to apply.
[backend] info: [EncryptionManager] Loaded existing key & salt for encrypting arbitrary data.
[backend] info: [TokenManager] Initialized new TokenManager instance for model: gpt-3.5-turbo
[backend] info: [TokenManager] Returning existing instance for model: gpt-3.5-turbo
[backend] info: prisma:info Starting a sqlite pool with 65 connections.
[backend] info: [TELEMETRY DISABLED] Telemetry is marked as disabled - no events will send. Telemetry helps Mintplex Labs Inc improve AnythingLLM.
[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: [BackgroundWorkerService] Starting...
[backend] info: [BackgroundWorkerService] Service started with 1 jobs ["cleanup-orphan-documents"]
[backend] info: [LMStudio] Error caching context windows Error: LMStudio:cacheContextWindows - Not Found
    at /app/server/utils/AiProviders/lmStudio/index.js:92:19
    at process.processTicksAndRejections (node:internal/process/task_queues:95:5)
    at async LMStudioLLM.cacheContextWindows (/app/server/utils/AiProviders/lmStudio/index.js:84:7)
    at async eagerLoadContextWindows (/app/server/utils/boot/eagerLoadContextWindows.js:19:7)
    at async Server.<anonymous> (/app/server/utils/boot/index.js:70:7)
[backend] info: [LMStudio] Context windows cached for all models!
[backend] info: ⚡Pre-cached context windows for LMStudio
[backend] info: [PushNotifications] Loaded existing VAPID keys!
[backend] info: [PushNotifications] Loading single user mode subscriptions...
[backend] info: Primary server in HTTP mode listening on port 3001
[backend] info: [MetaGenerator] fetching custom meta tag settings...
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [Event Logged] - update_llm_provider
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [NativeEmbedder] Initialized Xenova/all-MiniLM-L6-v2
[backend] info: [LMStudio] initialized with model: Qwen/Qwen3-0.6B
/usr/local/bin/docker-entrypoint.sh: line 26:   153 Illegal instruction     (core dumped) node /app/server/index.js
[collector] info: [TikTokenTokenizer] Initialized new TikTokenTokenizer instance.
[collector] info: Collector hot directory and tmp storage wiped!
[collector] info: Document processor app listening on port 8888
Environment variables loaded from .env
Prisma schema loaded from prisma/schema.prisma

✔ Generated Prisma Client (v5.3.1) to ./node_modules/@prisma/client in 611ms

Start using Prisma Client in Node.js (See: https://pris.ly/d/client)

import { PrismaClient } from '@prisma/client'
const prisma = new PrismaClient()

or start using Prisma Client at the edge (See: https://pris.ly/d/accelerate)

import { PrismaClient } from '@prisma/client/edge'
const prisma = new PrismaClient()


See other ways of importing Prisma Client: http://pris.ly/d/importing-client

Environment variables loaded from .env
Prisma schema loaded from prisma/schema.prisma
Datasource "db": SQLite database "anythingllm.db" at "file:../storage/anythingllm.db"

34 migrations found in prisma/migrations


No pending migrations to apply.
[backend] info: [EncryptionManager] Loaded existing key & salt for encrypting arbitrary data.
[backend] info: [TokenManager] Initialized new TokenManager instance for model: gpt-3.5-turbo
[backend] info: [TokenManager] Returning existing instance for model: gpt-3.5-turbo
[backend] info: prisma:info Starting a sqlite pool with 65 connections.
[backend] info: [TELEMETRY DISABLED] Telemetry is marked as disabled - no events will send. Telemetry helps Mintplex Labs Inc improve AnythingLLM.
[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: [BackgroundWorkerService] Starting...
[backend] info: [BackgroundWorkerService] Service started with 1 jobs ["cleanup-orphan-documents"]
[backend] info: [LMStudio] Error caching context windows Error: LMStudio:cacheContextWindows - Not Found
    at /app/server/utils/AiProviders/lmStudio/index.js:92:19
    at process.processTicksAndRejections (node:internal/process/task_queues:95:5)
    at async LMStudioLLM.cacheContextWindows (/app/server/utils/AiProviders/lmStudio/index.js:84:7)
    at async eagerLoadContextWindows (/app/server/utils/boot/eagerLoadContextWindows.js:19:7)
    at async Server.<anonymous> (/app/server/utils/boot/index.js:70:7)
[backend] info: [LMStudio] Context windows cached for all models!
[backend] info: ⚡Pre-cached context windows for LMStudio
[backend] info: [PushNotifications] Loaded existing VAPID keys!
[backend] info: [PushNotifications] Loading single user mode subscriptions...
[backend] info: Primary server in HTTP mode listening on port 3001
[backend] info: [MetaGenerator] fetching custom meta tag settings...
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported.
[backend] error: Error: The OPENAI_API_KEY environment variable is missing or empty; either provide it, or instantiate the OpenAI client with an apiKey option, like new OpenAI({ apiKey: 'My API Key' }).
    at new OpenAI (/app/server/node_modules/openai/index.js:72:19)
    at openAiModels (/app/server/utils/helpers/customModels.js:136:18)
    at getCustomModels (/app/server/utils/helpers/customModels.js:62:20)
    at /app/server/endpoints/system.js:1090:41
    at Layer.handle [as handle_request] (/app/server/node_modules/express/lib/router/layer.js:95:5)
    at next (/app/server/node_modules/express/lib/router/route.js:149:13)
    at /app/server/utils/middleware/multiUserProtected.js:60:7
    at Layer.handle [as handle_request] (/app/server/node_modules/express/lib/router/layer.js:95:5)
    at next (/app/server/node_modules/express/lib/router/route.js:149:13)
    at validatedRequest (/app/server/utils/middleware/validatedRequest.js:20:5)
[backend] error: LocalAI:listModels 404 <!DOCTYPE html>
<html class="no-js theme-auto center-alignment-no" lang="en-EN" >
<head>
  <meta charset="UTF-8">
  <meta name="endpoint" content="None">
  <meta name="description" content="SearXNG — a privacy-respecting, open metasearch engine">
  <meta name="keywords" content="SearXNG, search, search engine, metasearch, meta search">
  <meta name="generator" content="searxng/2026.1.30+ad42b553b">
  <meta name="referrer" content="no-referrer">
  <meta name="robots" content="noarchive">
  <meta name="viewport" content="width=device-width, initial-scale=1">
  <title>SearXNG</title>
  <script type="module" src="/static/themes/simple/sxng-core.min.js" client_settings="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"></script>
  <link rel="stylesheet" href="/static/themes/simple/sxng-ltr.min.css" type="text/css" media="screen">
  <link title="SearXNG" type="application/opensearchdescription+xml" rel="search" href="/opensearch.xml?method=POST&amp;autocomplete=">
  <link rel="icon" href="/static/themes/simple/img/favicon.png" sizes="any">
  <link rel="icon" href="/static/themes/simple/img/favicon.svg" type="image/svg+xml">
  <link rel="apple-touch-icon" href="/static/themes/simple/img/favicon.png">
</head>
<body class="None_endpoint" >
  <main id="main_404" class="">

    <nav id="links_on_top"><a href="/info/en/about" class="link_on_top_about"><svg viewBox="0 0 512 512" aria-hidden="true" class="sxng-icon-set-big"><path d="M248 64C146.39 64 64 146.39 64 248s82.39 184 184 184 184-82.39 184-184S349.61 64 248 64z" fill="none" stroke="currentColor" stroke-miterlimit="10" stroke-width="32"/><path fill="none" stroke="currentColor" stroke-linecap="round" stroke-linejoin="round" stroke-width="32" d="M220 220h32v116"/><path fill="none" stroke="currentColor" stroke-linecap="round" stroke-miterlimit="10" stroke-width="32" d="M208 340h88"/><path d="M248 130a26 26 0 1026 26 26 26 0 00-26-26z" fill="currentColor" stroke="currentColor" stroke-miterlimit="10" stroke-width="1"/></svg><span>About</span></a><a href="/preferences" class="link_on_top_preferences"><svg viewBox="0 0 512 512" class="ionicon sxng-icon-set-big" aria-hidden="true"><path d="M262.29 192.31a64 64 0 1 0 57.4 57.4 64.13 64.13 0 0 0-57.4-57.4M416.39 256a154 154 0 0 1-1.53 20.79l45.21 35.46a10.81 10.81 0 0 1 2.45 13.75l-42.77 74a10.81 10.81 0 0 1-13.14 4.59l-44.9-18.08a16.11 16.11 0 0 0-15.17 1.75A164.5 164.5 0 0 1 325 400.8a15.94 15.94 0 0 0-8.82 12.14l-6.73 47.89a11.08 11.08 0 0 1-10.68 9.17h-85.54a11.11 11.11 0 0 1-10.69-8.87l-6.72-47.82a16.07 16.07 0 0 0-9-12.22 155 155 0 0 1-21.46-12.57 16 16 0 0 0-15.11-1.71l-44.89 18.07a10.81 10.81 0 0 1-13.14-4.58l-42.77-74a10.8 10.8 0 0 1 2.45-13.75l38.21-30a16.05 16.05 0 0 0 6-14.08c-.36-4.17-.58-8.33-.58-12.5s.21-8.27.58-12.35a16 16 0 0 0-6.07-13.94l-38.19-30A10.81 10.81 0 0 1 49.48 186l42.77-74a10.81 10.81 0 0 1 13.14-4.59l44.9 18.08a16.11 16.11 0 0 0 15.17-1.75A164.5 164.5 0 0 1 187 111.2a15.94 15.94 0 0 0 8.82-12.14l6.73-47.89A11.08 11.08 0 0 1 213.23 42h85.54a11.11 11.11 0 0 1 10.69 8.87l6.72 47.82a16.07 16.07 0 0 0 9 12.22 155 155 0 0 1 21.46 12.57 16 16 0 0 0 15.11 1.71l44.89-18.07a10.81 10.81 0 0 1 13.14 4.58l42.77 74a10.8 10.8 0 0 1-2.45 13.75l-38.21 30a16.05 16.05 0 0 0-6.05 14.08c.33 4.14.55 8.3.55 12.47" fill="none" stroke="currentColor" stroke-linecap="round" stroke-linejoin="round" stroke-width="32px"/></svg><span>Preferences</span></a></nav>
<div class="center">
    <h1>Page not found</h1>
    <p>Go to <a href="/">search page</a>.</p>
</div>
  </main>
  <footer>
    <p>
    Powered by <a href="/info/en/about">SearXNG</a> - 2026.1.30+ad42b553b — a privacy-respecting, open metasearch engine<br>
        <a href="https://github.com/searxng/searxng">Source code</a>
        | <a href="https://github.com/searxng/searxng/issues">Issue tracker</a>
| <a href="/stats">Engine stats</a>        | <a href="https://searx.space">Public instances</a>
    </p>
  </footer>
</body>
</html>
[backend] error: LocalAI:listModels Connection error.
[backend] error: LocalAI:listModels Connection error.
[backend] error: LocalAI:listModels Connection error.
prisma:info Starting a sqlite pool with 65 connections.
[backend] info: [153:206]: No direct uploads path found - exiting.
[bg-worker][cleanup-orphan-documents] info: [153:206]: No direct uploads path found - exiting.
[backend] warn: Child process exited with code 0 and signal null
[backend] info: Worker for job "cleanup-orphan-documents" exited with code 0
[backend] info: [Event Logged] - update_llm_provider
[backend] info: [NativeEmbedder] Initialized Xenova/all-MiniLM-L6-v2
/usr/local/bin/docker-entrypoint.sh: line 26:   153 Illegal instruction     (core dumped) node /app/server/index.js
[collector] info: [TikTokenTokenizer] Initialized new TikTokenTokenizer instance.
[collector] info: Collector hot directory and tmp storage wiped!
[collector] info: Document processor app listening on port 8888
Environment variables loaded from .env
Prisma schema loaded from prisma/schema.prisma

✔ Generated Prisma Client (v5.3.1) to ./node_modules/@prisma/client in 617ms

Start using Prisma Client in Node.js (See: https://pris.ly/d/client)

import { PrismaClient } from '@prisma/client'
const prisma = new PrismaClient()

or start using Prisma Client at the edge (See: https://pris.ly/d/accelerate)

import { PrismaClient } from '@prisma/client/edge'
const prisma = new PrismaClient()


See other ways of importing Prisma Client: http://pris.ly/d/importing-client

Environment variables loaded from .env
Prisma schema loaded from prisma/schema.prisma
Datasource "db": SQLite database "anythingllm.db" at "file:../storage/anythingllm.db"

34 migrations found in prisma/migrations


No pending migrations to apply.
[backend] info: [EncryptionManager] Loaded existing key & salt for encrypting arbitrary data.
[backend] info: [TokenManager] Initialized new TokenManager instance for model: gpt-3.5-turbo
[backend] info: [TokenManager] Returning existing instance for model: gpt-3.5-turbo
[backend] info: prisma:info Starting a sqlite pool with 65 connections.
[backend] info: [TELEMETRY DISABLED] Telemetry is marked as disabled - no events will send. Telemetry helps Mintplex Labs Inc improve AnythingLLM.
[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: [PushNotifications] Loaded existing VAPID keys!
[backend] info: [BackgroundWorkerService] Starting...
[backend] info: [BackgroundWorkerService] Service started with 1 jobs ["cleanup-orphan-documents"]
[backend] info: [PushNotifications] Loading single user mode subscriptions...
[backend] info: Primary server in HTTP mode listening on port 3001
prisma:info Starting a sqlite pool with 65 connections.
[backend] info: [154:207]: No direct uploads path found - exiting.
[bg-worker][cleanup-orphan-documents] info: [154:207]: No direct uploads path found - exiting.
[backend] warn: Child process exited with code 0 and signal null
[backend] info: Worker for job "cleanup-orphan-documents" exited with code 0
[backend] info: [Event Logged] - update_llm_provider
[backend] info: [Event Logged] - update_llm_provider
[backend] info: [MetaGenerator] fetching custom meta tag settings...
[backend] info: [NativeEmbedder] Initialized Xenova/all-MiniLM-L6-v2
[backend] info: [GenericOpenAiLLM] Inference API: http://172.25.50.36:2242/v1 Model: Qwen/Qwen3-0.6B
/usr/local/bin/docker-entrypoint.sh: line 26:   154 Illegal instruction     (core dumped) node /app/server/index.js
[collector] info: [TikTokenTokenizer] Initialized new TikTokenTokenizer instance.
[collector] info: Collector hot directory and tmp storage wiped!
[collector] info: Document processor app listening on port 8888
Environment variables loaded from .env
Prisma schema loaded from prisma/schema.prisma

✔ Generated Prisma Client (v5.3.1) to ./node_modules/@prisma/client in 603ms

Start using Prisma Client in Node.js (See: https://pris.ly/d/client)

import { PrismaClient } from '@prisma/client'
const prisma = new PrismaClient()

or start using Prisma Client at the edge (See: https://pris.ly/d/accelerate)

import { PrismaClient } from '@prisma/client/edge'
const prisma = new PrismaClient()


See other ways of importing Prisma Client: http://pris.ly/d/importing-client

Environment variables loaded from .env
Prisma schema loaded from prisma/schema.prisma
Datasource "db": SQLite database "anythingllm.db" at "file:../storage/anythingllm.db"

34 migrations found in prisma/migrations


No pending migrations to apply.
[backend] info: [EncryptionManager] Loaded existing key & salt for encrypting arbitrary data.
[backend] info: [TokenManager] Initialized new TokenManager instance for model: gpt-3.5-turbo
[backend] info: [TokenManager] Returning existing instance for model: gpt-3.5-turbo
[backend] info: prisma:info Starting a sqlite pool with 65 connections.
[backend] info: [TELEMETRY DISABLED] Telemetry is marked as disabled - no events will send. Telemetry helps Mintplex Labs Inc improve AnythingLLM.
[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: [PushNotifications] Loaded existing VAPID keys!
[backend] info: [BackgroundWorkerService] Starting...
[backend] info: [BackgroundWorkerService] Service started with 1 jobs ["cleanup-orphan-documents"]
[backend] info: [PushNotifications] Loading single user mode subscriptions...
[backend] info: Primary server in HTTP mode listening on port 3001
prisma:info Starting a sqlite pool with 65 connections.
[backend] info: [153:207]: No direct uploads path found - exiting.
[bg-worker][cleanup-orphan-documents] info: [153:207]: No direct uploads path found - exiting.
[backend] warn: Child process exited with code 0 and signal null
[backend] info: Worker for job "cleanup-orphan-documents" exited with code 0
[backend] info: [MetaGenerator] fetching custom meta tag settings...
[backend] info: [NativeEmbedder] Initialized Xenova/all-MiniLM-L6-v2
[backend] info: [GenericOpenAiLLM] Inference API: http://172.25.50.36:2242/v1 Model: Qwen/Qwen3-0.6B
/usr/local/bin/docker-entrypoint.sh: line 26:   153 Illegal instruction     (core dumped) node /app/server/index.js

_____________________________________________________________

Aphrodite Engine Repository

GitHub: https://github.com/aphrodite-engine/aphrodite-engine

Are there known steps to reproduce?

Steps to Reproduce

  1. Setup a clean instance of AnythingLLM which allows for remote endpoints to connect
  2. setup and build Aphrodite Engine (link to the repo at the very bottom of this report)
  3. Configure the Generic OpenAI API provider using any one of the Base URLs listed above
    (you can also try and select LLM Studio or Local-AI for the LLM provider to see that the container and AnythingLLM can hit and reach the inference server via Generic OpenAI API)
  4. Save the configuration (this succeeds)
  5. Go to chat
  6. Send a prompt such as:
  7. tell me a joke

Aphrodite Engine Repository

https://github.com/aphrodite-engine/aphrodite-engine
Originally created by @siver1152 on GitHub (Feb 3, 2026). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/4962 ### How are you running AnythingLLM? Docker (remote machine) ### What happened? # Critical Bug Report: Generic OpenAI API Connection Causes Container Crash When Using Aphrodite Engine Inference Server ## Issue Summary I'm experiencing a critical bug when trying to connect to a generic OpenAI API endpoint causing a complete crash of the container itself. ## Infrastructure Setup Currently I have a server running Docker that has AnythingLLM running on it. On the same system I'm running my inference server which is outside of Docker running via Python virtual environment. I'm connecting to AnythingLLM via my desktop which is on the same subnet as the server. I can reach Aphrodite Engine from all 3 locations including from within the Docker container of AnythingLLM and a different random other Docker container using the info below in the report. I'm running Aphrodite Engine, which exposes an OpenAI-compatible API. I configured it using the following Base URLs (I tested all variations): - `http://<local IP, Docker IP, localhost>:2242/v1` - `http://<local IP, Docker IP, localhost>:2242/v1/` - `http://<local IP, Docker IP, localhost>:2242` - `http://<local IP, Docker IP, localhost>:2242/v1/chat/completions` I've tried using a random value for the API key as well as leaving the API key blank. I also attempted multiple string types for the API key along with variations in key length, but this did not change the behavior. ## Configuration Details For the chat model name, I tested: - `Qwen/Qwen3-0.6B` - `Qwen3-0.6B` **Model context window:** 40,960 **Max tokens:** 48,832 I also tested these same settings with LM Studio and LocalAI. I specifically tried connecting the Aphrodite Engine OpenAI-compatible API to those other LLM providers to see how they would behave with the same endpoint. I tested them using the same settings and variations of settings, and I also applied any configuration changes that were allowed or required by those providers. ## Evidence that the API is Reachable When I add the generic OpenAI API to LM Studio and LocalAI, I can see in the logs that AnythingLLM is successfully making requests to the inference server. Here are some example log entries: ``` INFO 19:16:36 172.17.0.2:60508 - "GET /v1/models HTTP/1.1" 200 INFO 19:17:11 172.17.0.2:54448 - "GET /api/v0/models HTTP/1.1" 404 INFO 19:17:15 172.17.0.2:34108 - "GET /v1/models HTTP/1.1" 200 INFO 19:17:52 172.17.0.2:45374 - "GET /v1/models HTTP/1.1" 200 INFO 19:19:50 172.17.0.2:56500 - "GET /v1/models HTTP/1.1" 200 ``` So the system is successfully detecting the model running in Aphrodite. The configurations save successfully in all three providers. However, when testing the two other LLM providers (LM Studio and LocalAI), sending a prompt resulted in the same network streaming error, and the same container crash behavior occurred there as well. ## Steps to Reproduce 1. Setup a clean instance of AnythingLLM which allows for remote endpoints to connect 2. setup and build Aphrodite Engine (link to the repo at the very bottom of this report) 3. Configure the Generic OpenAI API provider using any one of the Base URLs listed above (you can also try and select LLM Studio or Local-AI for the LLM provider to see that the container and AnythingLLM can hit and reach the inference server via Generic OpenAI API) 4. Save the configuration (this succeeds) 5. Go to chat 6. Send a prompt such as: ``` tell me a joke ``` ## Expected Behavior The system should stream and return a valid completion response from the Aphrodite OpenAI-compatible API. ## Actual Behavior I receive: ``` Could not respond to message. An error occurred while streaming response. Network error. ``` Immediately after that, the entire Docker container crashes and must be restarted. ## Network Connectivity Verification I can confirm that I'm able to reach the generic API from: - The local system itself using localhost, the Docker IP and port, and the local network IP from outside the Docker container - A remote device using the local IP address of the host machine - From within the Docker container using local, remote, and Docker IP addresses, as well as the Docker hostname ## Troubleshooting Performed I have completely reset and restarted Docker multiple times. No other configurations were changed beforehand. The container crashes regardless of where or how I configure it. If I configure it during the initial setup, it also causes the container to crash. # _____________________________________________________________ # Docker Container Logs ``` docker logs anythingllm [collector] info: [TikTokenTokenizer] Initialized new TikTokenTokenizer instance. [collector] info: Collector hot directory and tmp storage wiped! [collector] info: Document processor app listening on port 8888 Environment variables loaded from .env Prisma schema loaded from prisma/schema.prisma ✔ Generated Prisma Client (v5.3.1) to ./node_modules/@prisma/client in 668ms Start using Prisma Client in Node.js (See: https://pris.ly/d/client) ``` import { PrismaClient } from '@prisma/client' const prisma = new PrismaClient() ``` or start using Prisma Client at the edge (See: https://pris.ly/d/accelerate) ``` import { PrismaClient } from '@prisma/client/edge' const prisma = new PrismaClient() ``` See other ways of importing Prisma Client: http://pris.ly/d/importing-client Environment variables loaded from .env Prisma schema loaded from prisma/schema.prisma Datasource "db": SQLite database "anythingllm.db" at "file:../storage/anythingllm.db" SQLite database anythingllm.db created at file:../storage/anythingllm.db 34 migrations found in prisma/migrations Applying migration `20230921191814_init` Applying migration `20231101001441_init` Applying migration `20231101195421_init` Applying migration `20231129012019_add` Applying migration `20240113013409_init` Applying migration `20240118201333_init` Applying migration `20240202002020_init` Applying migration `20240206181106_init` Applying migration `20240206211916_init` Applying migration `20240208224848_init` Applying migration `20240210004405_init` Applying migration `20240216214639_init` Applying migration `20240219211018_init` Applying migration `20240301002308_init` Applying migration `20240326231053_init` Applying migration `20240405015034_init` Applying migration `20240412183346_init` Applying migration `20240425004220_init` Applying migration `20240430230707_init` Applying migration `20240510032311_init` Applying migration `20240618224346_init` Applying migration `20240821215625_init` Applying migration `20240824005054_init` Applying migration `20241003192954_init` Applying migration `20241029203722_init` Applying migration `20241029233509_init` Applying migration `20250102204948_init` Applying migration `20250226005538_init` Applying migration `20250318154720_init` Applying migration `20250506214129_init` Applying migration `20250709230835_init` Applying migration `20250725194841_init` Applying migration `20250808171557_init` Applying migration `20260130040204_init` The following migrations have been applied: migrations/ └─ 20230921191814_init/ └─ migration.sql └─ 20231101001441_init/ └─ migration.sql └─ 20231101195421_init/ └─ migration.sql └─ 20231129012019_add/ └─ migration.sql └─ 20240113013409_init/ └─ migration.sql └─ 20240118201333_init/ └─ migration.sql └─ 20240202002020_init/ └─ migration.sql └─ 20240206181106_init/ └─ migration.sql └─ 20240206211916_init/ └─ migration.sql └─ 20240208224848_init/ └─ migration.sql └─ 20240210004405_init/ └─ migration.sql └─ 20240216214639_init/ └─ migration.sql └─ 20240219211018_init/ └─ migration.sql └─ 20240301002308_init/ └─ migration.sql └─ 20240326231053_init/ └─ migration.sql └─ 20240405015034_init/ └─ migration.sql └─ 20240412183346_init/ └─ migration.sql └─ 20240425004220_init/ └─ migration.sql └─ 20240430230707_init/ └─ migration.sql └─ 20240510032311_init/ └─ migration.sql └─ 20240618224346_init/ └─ migration.sql └─ 20240821215625_init/ └─ migration.sql └─ 20240824005054_init/ └─ migration.sql └─ 20241003192954_init/ └─ migration.sql └─ 20241029203722_init/ └─ migration.sql └─ 20241029233509_init/ └─ migration.sql └─ 20250102204948_init/ └─ migration.sql └─ 20250226005538_init/ └─ migration.sql └─ 20250318154720_init/ └─ migration.sql └─ 20250506214129_init/ └─ migration.sql └─ 20250709230835_init/ └─ migration.sql └─ 20250725194841_init/ └─ migration.sql └─ 20250808171557_init/ └─ migration.sql └─ 20260130040204_init/ └─ migration.sql All migrations have been successfully applied. [backend] info: [EncryptionManager] Loaded existing key & salt for encrypting arbitrary data. [backend] info: [TokenManager] Initialized new TokenManager instance for model: gpt-3.5-turbo [backend] info: [ContextWindowFinder] Pulling remote model map... [backend] info: [TokenManager] Returning existing instance for model: gpt-3.5-turbo [backend] info: prisma:info Starting a sqlite pool with 65 connections. [backend] info: [ONBOARDING PATCH] Legacy instance is already onboarded, marking onboarding as complete. You will not see this message again. [backend] info: [TELEMETRY DISABLED] Telemetry is marked as disabled - no events will send. Telemetry helps Mintplex Labs Inc improve AnythingLLM. [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: [PushNotifications] Generating new VAPID keys... [backend] info: [PushNotifications] New VAPID keys generated! [backend] info: [BackgroundWorkerService] Starting... [backend] info: [BackgroundWorkerService] Service started with 1 jobs ["cleanup-orphan-documents"] [backend] info: [PushNotifications] Loading single user mode subscriptions... [backend] info: Primary server in HTTP mode listening on port 3001 [backend] info: [ContextWindowFinder] zai:titan-embed-text-v2 - context window is not a positive number. Got 0. [backend] info: [ContextWindowFinder] zai:embed-v4.0 - context window is not a positive number. Got 0. [backend] info: [ContextWindowFinder] zai:gemini-embedding-001 - context window is not a positive number. Got 0. [backend] info: [ContextWindowFinder] zai:text-embedding-005 - context window is not a positive number. Got 0. [backend] info: [ContextWindowFinder] zai:text-multilingual-embedding-002 - context window is not a positive number. Got 0. [backend] info: [ContextWindowFinder] zai:codestral-embed - context window is not a positive number. Got 0. [backend] info: [ContextWindowFinder] zai:mistral-embed - context window is not a positive number. Got 0. [backend] info: [ContextWindowFinder] zai:text-embedding-3-large - context window is not a positive number. Got 0. [backend] info: [ContextWindowFinder] zai:text-embedding-3-small - context window is not a positive number. Got 0. [backend] info: [ContextWindowFinder] zai:text-embedding-ada-002 - context window is not a positive number. Got 0. [backend] info: [ContextWindowFinder] Remote model map synced and cached prisma:info Starting a sqlite pool with 65 connections. [backend] info: [220:273]: No direct uploads path found - exiting. [bg-worker][cleanup-orphan-documents] info: [220:273]: No direct uploads path found - exiting. [backend] warn: Child process exited with code 0 and signal null [backend] info: Worker for job "cleanup-orphan-documents" exited with code 0 [backend] info: [MetaGenerator] fetching custom meta tag settings... [backend] error: LMStudio:listModels Connection error. [backend] error: LMStudio:listModels Connection error. [backend] error: LMStudio:listModels Connection error. [backend] error: LMStudio:listModels Connection error. [backend] info: [Event Logged] - update_llm_provider [backend] info: [Event Logged] - workspace_created [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [NativeEmbedder] Initialized Xenova/all-MiniLM-L6-v2 [backend] info: [LMStudio] initialized with model: Qwen/Qwen3-0.6B /usr/local/bin/docker-entrypoint.sh: line 26: 220 Illegal instruction (core dumped) node /app/server/index.js [collector] info: [TikTokenTokenizer] Initialized new TikTokenTokenizer instance. [collector] info: Collector hot directory and tmp storage wiped! [collector] info: Document processor app listening on port 8888 Environment variables loaded from .env Prisma schema loaded from prisma/schema.prisma ✔ Generated Prisma Client (v5.3.1) to ./node_modules/@prisma/client in 622ms Start using Prisma Client in Node.js (See: https://pris.ly/d/client) ``` import { PrismaClient } from '@prisma/client' const prisma = new PrismaClient() ``` or start using Prisma Client at the edge (See: https://pris.ly/d/accelerate) ``` import { PrismaClient } from '@prisma/client/edge' const prisma = new PrismaClient() ``` See other ways of importing Prisma Client: http://pris.ly/d/importing-client Environment variables loaded from .env Prisma schema loaded from prisma/schema.prisma Datasource "db": SQLite database "anythingllm.db" at "file:../storage/anythingllm.db" 34 migrations found in prisma/migrations No pending migrations to apply. [backend] info: [EncryptionManager] Loaded existing key & salt for encrypting arbitrary data. [backend] info: [TokenManager] Initialized new TokenManager instance for model: gpt-3.5-turbo [backend] info: [TokenManager] Returning existing instance for model: gpt-3.5-turbo [backend] info: prisma:info Starting a sqlite pool with 65 connections. [backend] info: [TELEMETRY DISABLED] Telemetry is marked as disabled - no events will send. Telemetry helps Mintplex Labs Inc improve AnythingLLM. [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: [BackgroundWorkerService] Starting... [backend] info: [BackgroundWorkerService] Service started with 1 jobs ["cleanup-orphan-documents"] [backend] info: [LMStudio] Error caching context windows Error: LMStudio:cacheContextWindows - Not Found at /app/server/utils/AiProviders/lmStudio/index.js:92:19 at process.processTicksAndRejections (node:internal/process/task_queues:95:5) at async LMStudioLLM.cacheContextWindows (/app/server/utils/AiProviders/lmStudio/index.js:84:7) at async eagerLoadContextWindows (/app/server/utils/boot/eagerLoadContextWindows.js:19:7) at async Server.<anonymous> (/app/server/utils/boot/index.js:70:7) [backend] info: [LMStudio] Context windows cached for all models! [backend] info: ⚡Pre-cached context windows for LMStudio [backend] info: [PushNotifications] Loaded existing VAPID keys! [backend] info: [PushNotifications] Loading single user mode subscriptions... [backend] info: Primary server in HTTP mode listening on port 3001 [backend] info: [MetaGenerator] fetching custom meta tag settings... [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [Event Logged] - update_llm_provider [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [NativeEmbedder] Initialized Xenova/all-MiniLM-L6-v2 [backend] info: [LMStudio] initialized with model: Qwen/Qwen3-0.6B /usr/local/bin/docker-entrypoint.sh: line 26: 153 Illegal instruction (core dumped) node /app/server/index.js [collector] info: [TikTokenTokenizer] Initialized new TikTokenTokenizer instance. [collector] info: Collector hot directory and tmp storage wiped! [collector] info: Document processor app listening on port 8888 Environment variables loaded from .env Prisma schema loaded from prisma/schema.prisma ✔ Generated Prisma Client (v5.3.1) to ./node_modules/@prisma/client in 611ms Start using Prisma Client in Node.js (See: https://pris.ly/d/client) ``` import { PrismaClient } from '@prisma/client' const prisma = new PrismaClient() ``` or start using Prisma Client at the edge (See: https://pris.ly/d/accelerate) ``` import { PrismaClient } from '@prisma/client/edge' const prisma = new PrismaClient() ``` See other ways of importing Prisma Client: http://pris.ly/d/importing-client Environment variables loaded from .env Prisma schema loaded from prisma/schema.prisma Datasource "db": SQLite database "anythingllm.db" at "file:../storage/anythingllm.db" 34 migrations found in prisma/migrations No pending migrations to apply. [backend] info: [EncryptionManager] Loaded existing key & salt for encrypting arbitrary data. [backend] info: [TokenManager] Initialized new TokenManager instance for model: gpt-3.5-turbo [backend] info: [TokenManager] Returning existing instance for model: gpt-3.5-turbo [backend] info: prisma:info Starting a sqlite pool with 65 connections. [backend] info: [TELEMETRY DISABLED] Telemetry is marked as disabled - no events will send. Telemetry helps Mintplex Labs Inc improve AnythingLLM. [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: [BackgroundWorkerService] Starting... [backend] info: [BackgroundWorkerService] Service started with 1 jobs ["cleanup-orphan-documents"] [backend] info: [LMStudio] Error caching context windows Error: LMStudio:cacheContextWindows - Not Found at /app/server/utils/AiProviders/lmStudio/index.js:92:19 at process.processTicksAndRejections (node:internal/process/task_queues:95:5) at async LMStudioLLM.cacheContextWindows (/app/server/utils/AiProviders/lmStudio/index.js:84:7) at async eagerLoadContextWindows (/app/server/utils/boot/eagerLoadContextWindows.js:19:7) at async Server.<anonymous> (/app/server/utils/boot/index.js:70:7) [backend] info: [LMStudio] Context windows cached for all models! [backend] info: ⚡Pre-cached context windows for LMStudio [backend] info: [PushNotifications] Loaded existing VAPID keys! [backend] info: [PushNotifications] Loading single user mode subscriptions... [backend] info: Primary server in HTTP mode listening on port 3001 [backend] info: [MetaGenerator] fetching custom meta tag settings... [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] info: [LMStudio] No context windows cached - Context window may be inaccurately reported. [backend] error: Error: The OPENAI_API_KEY environment variable is missing or empty; either provide it, or instantiate the OpenAI client with an apiKey option, like new OpenAI({ apiKey: 'My API Key' }). at new OpenAI (/app/server/node_modules/openai/index.js:72:19) at openAiModels (/app/server/utils/helpers/customModels.js:136:18) at getCustomModels (/app/server/utils/helpers/customModels.js:62:20) at /app/server/endpoints/system.js:1090:41 at Layer.handle [as handle_request] (/app/server/node_modules/express/lib/router/layer.js:95:5) at next (/app/server/node_modules/express/lib/router/route.js:149:13) at /app/server/utils/middleware/multiUserProtected.js:60:7 at Layer.handle [as handle_request] (/app/server/node_modules/express/lib/router/layer.js:95:5) at next (/app/server/node_modules/express/lib/router/route.js:149:13) at validatedRequest (/app/server/utils/middleware/validatedRequest.js:20:5) [backend] error: LocalAI:listModels 404 <!DOCTYPE html> <html class="no-js theme-auto center-alignment-no" lang="en-EN" > <head> <meta charset="UTF-8"> <meta name="endpoint" content="None"> <meta name="description" content="SearXNG — a privacy-respecting, open metasearch engine"> <meta name="keywords" content="SearXNG, search, search engine, metasearch, meta search"> <meta name="generator" content="searxng/2026.1.30+ad42b553b"> <meta name="referrer" content="no-referrer"> <meta name="robots" content="noarchive"> <meta name="viewport" content="width=device-width, initial-scale=1"> <title>SearXNG</title> <script type="module" src="/static/themes/simple/sxng-core.min.js" client_settings="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"></script> <link rel="stylesheet" href="/static/themes/simple/sxng-ltr.min.css" type="text/css" media="screen"> <link title="SearXNG" type="application/opensearchdescription+xml" rel="search" href="/opensearch.xml?method=POST&amp;autocomplete="> <link rel="icon" href="/static/themes/simple/img/favicon.png" sizes="any"> <link rel="icon" href="/static/themes/simple/img/favicon.svg" type="image/svg+xml"> <link rel="apple-touch-icon" href="/static/themes/simple/img/favicon.png"> </head> <body class="None_endpoint" > <main id="main_404" class=""> <nav id="links_on_top"><a href="/info/en/about" class="link_on_top_about"><svg viewBox="0 0 512 512" aria-hidden="true" class="sxng-icon-set-big"><path d="M248 64C146.39 64 64 146.39 64 248s82.39 184 184 184 184-82.39 184-184S349.61 64 248 64z" fill="none" stroke="currentColor" stroke-miterlimit="10" stroke-width="32"/><path fill="none" stroke="currentColor" stroke-linecap="round" stroke-linejoin="round" stroke-width="32" d="M220 220h32v116"/><path fill="none" stroke="currentColor" stroke-linecap="round" stroke-miterlimit="10" stroke-width="32" d="M208 340h88"/><path d="M248 130a26 26 0 1026 26 26 26 0 00-26-26z" fill="currentColor" stroke="currentColor" stroke-miterlimit="10" stroke-width="1"/></svg><span>About</span></a><a href="/preferences" class="link_on_top_preferences"><svg viewBox="0 0 512 512" class="ionicon sxng-icon-set-big" aria-hidden="true"><path d="M262.29 192.31a64 64 0 1 0 57.4 57.4 64.13 64.13 0 0 0-57.4-57.4M416.39 256a154 154 0 0 1-1.53 20.79l45.21 35.46a10.81 10.81 0 0 1 2.45 13.75l-42.77 74a10.81 10.81 0 0 1-13.14 4.59l-44.9-18.08a16.11 16.11 0 0 0-15.17 1.75A164.5 164.5 0 0 1 325 400.8a15.94 15.94 0 0 0-8.82 12.14l-6.73 47.89a11.08 11.08 0 0 1-10.68 9.17h-85.54a11.11 11.11 0 0 1-10.69-8.87l-6.72-47.82a16.07 16.07 0 0 0-9-12.22 155 155 0 0 1-21.46-12.57 16 16 0 0 0-15.11-1.71l-44.89 18.07a10.81 10.81 0 0 1-13.14-4.58l-42.77-74a10.8 10.8 0 0 1 2.45-13.75l38.21-30a16.05 16.05 0 0 0 6-14.08c-.36-4.17-.58-8.33-.58-12.5s.21-8.27.58-12.35a16 16 0 0 0-6.07-13.94l-38.19-30A10.81 10.81 0 0 1 49.48 186l42.77-74a10.81 10.81 0 0 1 13.14-4.59l44.9 18.08a16.11 16.11 0 0 0 15.17-1.75A164.5 164.5 0 0 1 187 111.2a15.94 15.94 0 0 0 8.82-12.14l6.73-47.89A11.08 11.08 0 0 1 213.23 42h85.54a11.11 11.11 0 0 1 10.69 8.87l6.72 47.82a16.07 16.07 0 0 0 9 12.22 155 155 0 0 1 21.46 12.57 16 16 0 0 0 15.11 1.71l44.89-18.07a10.81 10.81 0 0 1 13.14 4.58l42.77 74a10.8 10.8 0 0 1-2.45 13.75l-38.21 30a16.05 16.05 0 0 0-6.05 14.08c.33 4.14.55 8.3.55 12.47" fill="none" stroke="currentColor" stroke-linecap="round" stroke-linejoin="round" stroke-width="32px"/></svg><span>Preferences</span></a></nav> <div class="center"> <h1>Page not found</h1> <p>Go to <a href="/">search page</a>.</p> </div> </main> <footer> <p> Powered by <a href="/info/en/about">SearXNG</a> - 2026.1.30+ad42b553b — a privacy-respecting, open metasearch engine<br> <a href="https://github.com/searxng/searxng">Source code</a> | <a href="https://github.com/searxng/searxng/issues">Issue tracker</a> | <a href="/stats">Engine stats</a> | <a href="https://searx.space">Public instances</a> </p> </footer> </body> </html> [backend] error: LocalAI:listModels Connection error. [backend] error: LocalAI:listModels Connection error. [backend] error: LocalAI:listModels Connection error. prisma:info Starting a sqlite pool with 65 connections. [backend] info: [153:206]: No direct uploads path found - exiting. [bg-worker][cleanup-orphan-documents] info: [153:206]: No direct uploads path found - exiting. [backend] warn: Child process exited with code 0 and signal null [backend] info: Worker for job "cleanup-orphan-documents" exited with code 0 [backend] info: [Event Logged] - update_llm_provider [backend] info: [NativeEmbedder] Initialized Xenova/all-MiniLM-L6-v2 /usr/local/bin/docker-entrypoint.sh: line 26: 153 Illegal instruction (core dumped) node /app/server/index.js [collector] info: [TikTokenTokenizer] Initialized new TikTokenTokenizer instance. [collector] info: Collector hot directory and tmp storage wiped! [collector] info: Document processor app listening on port 8888 Environment variables loaded from .env Prisma schema loaded from prisma/schema.prisma ✔ Generated Prisma Client (v5.3.1) to ./node_modules/@prisma/client in 617ms Start using Prisma Client in Node.js (See: https://pris.ly/d/client) ``` import { PrismaClient } from '@prisma/client' const prisma = new PrismaClient() ``` or start using Prisma Client at the edge (See: https://pris.ly/d/accelerate) ``` import { PrismaClient } from '@prisma/client/edge' const prisma = new PrismaClient() ``` See other ways of importing Prisma Client: http://pris.ly/d/importing-client Environment variables loaded from .env Prisma schema loaded from prisma/schema.prisma Datasource "db": SQLite database "anythingllm.db" at "file:../storage/anythingllm.db" 34 migrations found in prisma/migrations No pending migrations to apply. [backend] info: [EncryptionManager] Loaded existing key & salt for encrypting arbitrary data. [backend] info: [TokenManager] Initialized new TokenManager instance for model: gpt-3.5-turbo [backend] info: [TokenManager] Returning existing instance for model: gpt-3.5-turbo [backend] info: prisma:info Starting a sqlite pool with 65 connections. [backend] info: [TELEMETRY DISABLED] Telemetry is marked as disabled - no events will send. Telemetry helps Mintplex Labs Inc improve AnythingLLM. [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: [PushNotifications] Loaded existing VAPID keys! [backend] info: [BackgroundWorkerService] Starting... [backend] info: [BackgroundWorkerService] Service started with 1 jobs ["cleanup-orphan-documents"] [backend] info: [PushNotifications] Loading single user mode subscriptions... [backend] info: Primary server in HTTP mode listening on port 3001 prisma:info Starting a sqlite pool with 65 connections. [backend] info: [154:207]: No direct uploads path found - exiting. [bg-worker][cleanup-orphan-documents] info: [154:207]: No direct uploads path found - exiting. [backend] warn: Child process exited with code 0 and signal null [backend] info: Worker for job "cleanup-orphan-documents" exited with code 0 [backend] info: [Event Logged] - update_llm_provider [backend] info: [Event Logged] - update_llm_provider [backend] info: [MetaGenerator] fetching custom meta tag settings... [backend] info: [NativeEmbedder] Initialized Xenova/all-MiniLM-L6-v2 [backend] info: [GenericOpenAiLLM] Inference API: http://172.25.50.36:2242/v1 Model: Qwen/Qwen3-0.6B /usr/local/bin/docker-entrypoint.sh: line 26: 154 Illegal instruction (core dumped) node /app/server/index.js [collector] info: [TikTokenTokenizer] Initialized new TikTokenTokenizer instance. [collector] info: Collector hot directory and tmp storage wiped! [collector] info: Document processor app listening on port 8888 Environment variables loaded from .env Prisma schema loaded from prisma/schema.prisma ✔ Generated Prisma Client (v5.3.1) to ./node_modules/@prisma/client in 603ms Start using Prisma Client in Node.js (See: https://pris.ly/d/client) ``` import { PrismaClient } from '@prisma/client' const prisma = new PrismaClient() ``` or start using Prisma Client at the edge (See: https://pris.ly/d/accelerate) ``` import { PrismaClient } from '@prisma/client/edge' const prisma = new PrismaClient() ``` See other ways of importing Prisma Client: http://pris.ly/d/importing-client Environment variables loaded from .env Prisma schema loaded from prisma/schema.prisma Datasource "db": SQLite database "anythingllm.db" at "file:../storage/anythingllm.db" 34 migrations found in prisma/migrations No pending migrations to apply. [backend] info: [EncryptionManager] Loaded existing key & salt for encrypting arbitrary data. [backend] info: [TokenManager] Initialized new TokenManager instance for model: gpt-3.5-turbo [backend] info: [TokenManager] Returning existing instance for model: gpt-3.5-turbo [backend] info: prisma:info Starting a sqlite pool with 65 connections. [backend] info: [TELEMETRY DISABLED] Telemetry is marked as disabled - no events will send. Telemetry helps Mintplex Labs Inc improve AnythingLLM. [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: [PushNotifications] Loaded existing VAPID keys! [backend] info: [BackgroundWorkerService] Starting... [backend] info: [BackgroundWorkerService] Service started with 1 jobs ["cleanup-orphan-documents"] [backend] info: [PushNotifications] Loading single user mode subscriptions... [backend] info: Primary server in HTTP mode listening on port 3001 prisma:info Starting a sqlite pool with 65 connections. [backend] info: [153:207]: No direct uploads path found - exiting. [bg-worker][cleanup-orphan-documents] info: [153:207]: No direct uploads path found - exiting. [backend] warn: Child process exited with code 0 and signal null [backend] info: Worker for job "cleanup-orphan-documents" exited with code 0 [backend] info: [MetaGenerator] fetching custom meta tag settings... [backend] info: [NativeEmbedder] Initialized Xenova/all-MiniLM-L6-v2 [backend] info: [GenericOpenAiLLM] Inference API: http://172.25.50.36:2242/v1 Model: Qwen/Qwen3-0.6B /usr/local/bin/docker-entrypoint.sh: line 26: 153 Illegal instruction (core dumped) node /app/server/index.js ``` # _____________________________________________________________ ## Aphrodite Engine Repository GitHub: https://github.com/aphrodite-engine/aphrodite-engine ### Are there known steps to reproduce? ## Steps to Reproduce 1. Setup a clean instance of AnythingLLM which allows for remote endpoints to connect 2. setup and build Aphrodite Engine (link to the repo at the very bottom of this report) 3. Configure the Generic OpenAI API provider using any one of the Base URLs listed above (you can also try and select LLM Studio or Local-AI for the LLM provider to see that the container and AnythingLLM can hit and reach the inference server via Generic OpenAI API) 4. Save the configuration (this succeeds) 5. Go to chat 6. Send a prompt such as: 7. ``` tell me a joke ``` ## Aphrodite Engine Repository ``` https://github.com/aphrodite-engine/aphrodite-engine ```
yindo added the possible bug label 2026-02-22 18:32:39 -05:00
yindo closed this issue 2026-02-22 18:32:39 -05:00
yindo commented 2026-02-22 18:32:40 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Feb 5, 2026):

You are using lancedb and are too old of a CPU that does not support AVX2 #1331

@timothycarambat commented on GitHub (Feb 5, 2026): You are using lancedb and are too old of a CPU that does not support AVX2 #1331
yindo commented 2026-02-22 18:32:40 -05:00
Author
Owner
Copy Link
Copy Source

@siver1152 commented on GitHub (Feb 5, 2026):

i haven't messed with, changed or played with LancerDB. im not trying to do embedding or anything. i'm just try to send a chat and get a reply.
its a fresh container install of AnythingLLM using the documented install guide for Linux. No configuration other then the LLM provider everything else is default.

im running Proxmox with a AMD EPYC 7002 Series processors (ROME arch)
also ive confirmed the system i have has AVX/AVX2 support and that its enabled on all my VM in my ecosystem.
The VM is running the latest Ubuntu serve.

im going to try and test this on AnythingLLM desktop and see if it will work there. i can also setup a ChromaDB if needed and re-test it with that enabled and setup.

See image below that showing AVX/AVX2 support.

One thing i should have noted is im pulling the container from Docker Hub.

Image
@siver1152 commented on GitHub (Feb 5, 2026): i haven't messed with, changed or played with LancerDB. im not trying to do embedding or anything. i'm just try to send a chat and get a reply. its a fresh container install of AnythingLLM using the documented install guide for Linux. No configuration other then the LLM provider everything else is default. im running Proxmox with a AMD EPYC 7002 Series processors (ROME arch) also ive confirmed the system i have has AVX/AVX2 support and that its enabled on all my VM in my ecosystem. The VM is running the latest Ubuntu serve. im going to try and test this on AnythingLLM desktop and see if it will work there. i can also setup a ChromaDB if needed and re-test it with that enabled and setup. See image below that showing AVX/AVX2 support. One thing i should have noted is im pulling the container from Docker Hub. <img width="1043" height="278" alt="Image" src="https://github.com/user-attachments/assets/83320a86-94a2-4a6b-9101-b8c675a1704e" />
yindo changed title from [BUG]: Illegal Instruction Error Causes Container Crash with Generic OpenAI API LLM Provider to [GH-ISSUE #4962] [BUG]: Illegal Instruction Error Causes Container Crash with Generic OpenAI API LLM Provider 2026-06-05 14:50:22 -04:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
Desktop
Docker
Integration Request
Integration Request
OS: Linux
OS: Mobile
OS: Windows
UI/UX
blocked
bug
bug
core-team-only
documentation
duplicate
embed-widget
enhancement
feature request
github_actions
good first issue
investigating
needs info / can't replicate
possible bug
pull-request
Mirrored from GitHub Pull Request
 question
stage: specifications
wontfix
No Label possible bug
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: Mintplex-Labs/anything-llm#3113
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?