FIX: Ignite CLI Docker container exits immediately after container restart #4852
Description
Context
I used Ignite CLI to run my Cosmos SDK chain inside a Docker container.
The container starts and the chain runs normally the first time.
However, after stopping the container and starting it again, the container exits immediately instead of keeping the chain running.
This issue blocks normal development workflow because the container cannot be restarted reliably after being stopped.
No major code changes were made between runs. The issue appears to be related to how Ignite CLI behaves inside Docker when restarting an existing container.
Describe the Bug
When running Ignite CLI inside a Docker container:
The container runs normally on first start
After stopping the container
Restarting the same container causes it to exit immediately
The chain does not stay running in the background
This behavior is unexpected, as the container should continue running the chain process after restart.
Steps to Reproduce
Create or use an existing Cosmos SDK chain with Ignite CLI
Run Ignite CLI inside a Docker container (e.g. using ignite chain serve)
Confirm that the container and chain are running normally
Stop the Docker container
docker stop <container_name>
Start the container again
docker start <container_name>
Observe that the container exits immediately
Expected Behavior
After restarting the Docker container:
The Ignite CLI process should continue running
The Cosmos chain should start and stay alive
The container should remain in a running state
Actual Behavior
The Docker container exits immediately after restart
The chain does not stay running
No interactive prompt is available
The container must be recreated to run successfully again
Environment Details
OS: Linux (Docker container)
docker-compose.yml
services:
ignite:
image: ignitehq/cli:v29.6.2
container_name: ingsha-ignite
tty: true
stdin_open: true
working_dir: /apps
volumes:
# Ignite CLI 的 home & cache
- ../sdh:/home/tendermint
- ../src:/apps
ports:
# - "1317:1317" # Cosmos REST API
- "26657:26657" # Tendermint RPC
# - "4500:4500" # gRPC
# environment:
# - HOME=/home/tendermint
# restart: always
command: bash
Ignite Version:
Ignite CLI version: ignitehq/cli:v29.6.2
Ignite CLI config version: ignitehq/cli:v29.6.2
Cosmos SDK version: v0.53.5
Go Version:
(Go is provided inside the Ignite Docker image)
Docker: Docker Engine (latest stable)
Notes
The issue only occurs after restarting an existing container
First-time container execution works correctly
This might be related to:
Ignite CLI running in interactive mode
Process lifecycle / TTY handling inside Docker
ignite chain serve not being suitable as PID 1 without a wrapper
Using docker exec to manually start the chain again
However, a reliable restart behavior is expected when using Ignite CLI in Docker-based workflows.