E2B port not open error

[quancore]

I am trying to build e2b from template

# Custom E2B Sandbox Template with S3 Integration
# Base image is required - only Debian-based images are supported
FROM e2bdev/code-interpreter:latest

# Accept build arguments for AWS credentials
ARG AWS_ACCESS_KEY_ID
ARG AWS_SECRET_ACCESS_KEY
ARG AWS_REGION=us-west-2
ARG S3_BUCKET_NAME

# Set environment variables from build args
ENV AWS_ACCESS_KEY_ID=${AWS_ACCESS_KEY_ID}
ENV AWS_SECRET_ACCESS_KEY=${AWS_SECRET_ACCESS_KEY}
ENV AWS_REGION=${AWS_REGION}
ENV S3_BUCKET_NAME=${S3_BUCKET_NAME}

# Update package list
RUN apt-get update

# Install S3 filesystem tools and utilities
RUN apt-get install -y \
    s3fs \
    awscli \
    curl \
    wget \
    fuse \
    jq \
    git \
    vim \
    htop \
    && rm -rf /var/lib/apt/lists/*

# Install goofys for high-performance S3 mounting
RUN curl -L https://github.com/kahing/goofys/releases/latest/download/goofys -o /usr/local/bin/goofys \
    && chmod +x /usr/local/bin/goofys

# Install only additional packages not pre-installed in E2B code interpreter
RUN pip install --no-cache-dir \
    boto3 \
    python-dotenv \
    pyarrow \
    fastparquet

# Create necessary directories
RUN mkdir -p /s3-data /workspace /tmp/artifacts

# Set up proper permissions for FUSE
RUN echo "user_allow_other" >> /etc/fuse.conf

# Copy startup script to a location accessible by user
COPY startup.sh /usr/local/bin/startup.sh

# Make script executable for all users
RUN chmod +x /usr/local/bin/startup.sh

# Note: E2B will run startup.sh via the start_cmd in e2b.toml
# No systemd setup needed - E2B handles container initialization differently

# Set workspace as working directory  
WORKDIR /workspace

# This is a config for E2B sandbox template.
# You can use template ID (5efyoklxeppejs2prlu4) or template name (sentient-e2b-s3) to create a sandbox:

# Python SDK
# from e2b import Sandbox, AsyncSandbox
# sandbox = Sandbox("sentient-e2b-s3") # Sync sandbox
# sandbox = await AsyncSandbox.create("sentient-e2b-s3") # Async sandbox

# JS SDK
# import { Sandbox } from 'e2b'
# const sandbox = await Sandbox.create('sentient-e2b-s3')

team_id = "10c72e9b-7c76-4dce-b9cf-815a783fcf8e"
start_cmd = "startup.sh"
dockerfile = "e2b.Dockerfile"
template_name = "sentient-e2b-s3"
template_id = "5efyoklxeppejs2prlu4"

Startup script etc running properly but still I receive

Error executing code: {"sandboxId":"idqbbkxbbanwbejb1jdi1","message":"The sandbox is 
  running but port is not open","port":49999,"code":502}: This error is likely due to sandbox timeout. You can modify the 
  sandbox timeout by passing 'timeout' when starting the sandbox or calling '.set_timeout' on the sandbox with the desired
   timeout.

[linear]

E2B-2914 E2B port not open error

[mishushakov]

Hey, the cause is most likely that the start command isn't starting the Jupyter Server.

From our documentation:

e2b template build -c "/root/.jupyter/start-up.sh"

https://e2b.dev/docs/sandbox-template#how-to-create-custom-sandbox

[quancore]

@mishushakov I did not understand, I do not use jupyter, why I need that? I have another start command I have indicated in the TOML file. Basically my startup script is mounting some dirs via goofys during e2b build and then when sandbox created, it would start with mounting @mlejva

[mishushakov]

Hey, you would need to start Jupyter and Code Interpreter Server in order to run code in the sandbox using .run_code command

[quancore]

With code interpreter image, basically we always need to use this default startup script and cannot use custom one? What is your recommandation if I want to use a custom startup to mount dirs?

[mishushakov]

You would have to customize the default startup script:
https://github.com/e2b-dev/code-interpreter/blob/main/template/start-up.sh

[quancore]

I have modified the starup script and started jupyter kernel properly and it works