All-Hands-AI
Error: connect ECONNREFUSED 127.0.0.1:3000
Describe the bug
on the frontend, I am getting this error
ERROR: Failed connection to server. Please ensure the server is reachable at ws://
at the backend, I am getting the below error Running the app...
[email protected] start vite --port 3001 --host 0.0.0.0
VITE v5.2.7 ready in 578 ms
➜ Local: http://localhost:3001/
➜ Network: http://
Setup and configuration
Current version:
commit c37124d74022fc73f32a576d88dd59db2ebfe2ac (HEAD -> main, origin/main, origin/HEAD) Author: Xingyao Wang [email protected] Date: Tue Apr 2 14:58:28 2024 +0800
My config.toml and environment vars (be sure to redact API keys):
cat config.toml
LLM_API_KEY="sk-xxxxxxxxxxxxxxxxxxxxxxxxxx"
LLM_MODEL="open-devin-preview-gpt4"
WORKSPACE_DIR="./workspace"
My model and agent (you can see these settings in the UI): Initializing agent (may take up to 10 seconds)... nothing after that
Commands I ran to install and run OpenDevin:
Steps to Reproduce:
- the only change I made apart from the instructions provided is in the makefile @cd frontend && npm run start -- --port $(FRONTEND_PORT) --host "0.0.0.0" <- added 0.0.0.0 to be accessible over the network.
Logs, error messages, and screenshots:
Additional Context
Error: connect ECONNREFUSED 127.0.0.1:3000
Looks like you forget to start the backend or mock server at 3000
Open file opendevin/frontend/vite.config and there corect these lines of code "const BACKEND_HOST = process.env.BACKEND_HOST || "127.0.0.1:3000";
// check BACKEND_HOST is something like "example.com"
if (!BACKEND_HOST.match(/^([\w\d-]+(.[\w\d-]+)+(:\d+)?)/)) {
throw new Error(
Invalid BACKEND_HOST ${BACKEND_HOST}, example BACKEND_HOST 127.0.0.1:3000,
);
}" with those corected ones "const BACKEND_HOST = process.env.BACKEND_HOST || "127.0.0.1:3001";
// check BACKEND_HOST is something like "example.com"
if (!BACKEND_HOST.match(/^([\w\d-]+(.[\w\d-]+)+(:\d+)?)/)) {
throw new Error(
Invalid BACKEND_HOST ${BACKEND_HOST}, example BACKEND_HOST 127.0.0.1:3001,
);
}" - the error is that frontend try to comunicate on 127.0.0.1:3000 and the backend try to communicate on 127.0.0.1:3001 . Afer that save and rerun . It will work ok now .
the only change I made apart from the instructions provided is in the makefile @cd frontend && npm run start -- --port $(FRONTEND_PORT) --host "0.0.0.0" <- added 0.0.0.0 to be accessible over the network.
Why do you need the backend to be accessible over the network? Are you running backend and front end on different boxes? I suspect your setup will run fine if you dropped --host "0.0.0.0"
If you need to keep it, you need to set the environment variable BACKEND_HOST to your backend server's IP address or DNS.
#576
@JustinLin610 - Yes it is the issue
When I tried wget without sudo, I got
Cannot write to ‘/tmp/llama_index/models--BAAI--bge-small-en-v1.5/snapshots/5c38ec7c405ec4b44b94cc5a9bb96e735b38267a/1_Pooling/config.json’ (Success)
the fact that it prints (Success) at the end could be misleading
with sudo
2024-04-02 21:46:29 (2.78 MB/s) - ‘/tmp/llama_index/models--BAAI--bge-small-en-v1.5/snapshots/5c38ec7c405ec4b44b94cc5a9bb96e735b38267a/1_Pooling/config.json’ saved [190/190]
Success Starting backend... INFO: Started server process [12999] INFO: Waiting for application startup. INFO: Application startup complete. INFO: Uvicorn running on http://127.0.0.1:3000 (Press CTRL+C to quit)
have a question though, do I have to start the back end separately? If so, where can I have a nohup added?
The newest makefile has a run target that runs both backend and frontend for you.
have a question though, do I have to start the back end separately? If so, where can I have a nohup added?
@lijom80 You don't need to, now make run will let you start both the front and back end at the same time and see their outputs simultaneously.
i've got the same problem with the new make run command, giving me this error:
5:39:57 PM [vite] http proxy error: /litellm-models
Error: connect ECONNREFUSED 127.0.0.1:3000
at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1602:16)
5:39:57 PM [vite] http proxy error: /litellm-agents
Error: connect ECONNREFUSED 127.0.0.1:3000
at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1602:16)
5:39:57 PM [vite] http proxy error: /litellm-models
Error: connect ECONNREFUSED 127.0.0.1:3000
at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1602:16)
5:39:57 PM [vite] http proxy error: /litellm-agents
Error: connect ECONNREFUSED 127.0.0.1:3000
at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1602:16)
5:39:58 PM [vite] ws proxy error:
Error: connect ECONNREFUSED 127.0.0.1:3000
at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1602:16)
5:40:03 PM [vite] http proxy error: /litellm-models
Error: connect ECONNREFUSED 127.0.0.1:3000
at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1602:16)
5:40:03 PM [vite] http proxy error: /litellm-agents
Error: connect ECONNREFUSED 127.0.0.1:3000
at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1602:16)
5:40:03 PM [vite] http proxy error: /litellm-models
Error: connect ECONNREFUSED 127.0.0.1:3000
at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1602:16)
5:40:03 PM [vite] http proxy error: /litellm-agents
Error: connect ECONNREFUSED 127.0.0.1:3000
at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1602:16)
5:40:03 PM [vite] ws proxy error:
Error: connect ECONNREFUSED 127.0.0.1:3000
at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1602:16)
Backend always take some time to start, it there any output logs from backend?
no thanks, for me the issue was some cuda library not properly installed on the wsl. I solved it.
My services are running as well. I realised after reboot the node, and running backend again, it installed few more components.
Also, the new make run does not run the backend automatically for some reason.
I have a question, what is the ideal infra requirements for make it run smoothly?
On Wed, 3 Apr 2024, 22:02 Francesco Dotta, @.***> wrote:
no thanks, for me the issue was some cuda library not properly installed on the wsl. I solved it.
— Reply to this email directly, view it on GitHub https://github.com/OpenDevin/OpenDevin/issues/568#issuecomment-2035070098, or unsubscribe https://github.com/notifications/unsubscribe-auth/BF3SFTXL2WJLBENCJ2EGGO3Y3QVLLAVCNFSM6AAAAABFTF6E7CVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMZVGA3TAMBZHA . You are receiving this because you were mentioned.Message ID: @.***>
My services are running as well. I realised after reboot the node, and running backend again, it installed few more components. Also, the new make run does not run the backend automatically for some reason. I have a question, what is the ideal infra requirements for make it run smoothly?
make build will install dependences to run smoothly.
Seems user can run opendevin now. But maye still keep an eye on backend launch slowly problem.
