-
Notifications
You must be signed in to change notification settings - Fork 336
Description
Description:
I can successfully run the all-in-one HyperDX image with:
docker run -p 8080:8080 -p 4317:4317 -p 4318:4318 docker.hyperdx.io/hyperdx/hyperdx-all-in-one
It works perfectly on Ubuntu 25.04 with Docker 29.0.3 (build 511dad6).
However, when using the dev compose setup:
git clone https://github.com/hyperdxio/hyperdx.git
cd hyperdx
docker-compose -f docker-compose.dev.yml up
The following happens:
- hdx-oss-dev-ch-server-1 (ClickHouse) starts successfully.
- The other containers (db-1, otel-collector-1, otel-collector-json-1) appear to start (Docker shows "Created"), but never become healthy.
- The process hangs indefinitely, and eventually Docker itself becomes unresponsive—commands like docker ps, docker stop, or docker-compose down hang or fail.
- After this, the only recovery is to restart the Docker engine.
- Even after restart, the stuck containers cannot be removed normally—they persist in a limbo state until a full systemctl restart docker and manual cleanup.
The only log output before failure is:
ch-server-1 | /entrypoint.sh: create new user 'default' instead 'default'
ch-server-1 | Processing configuration file '/etc/clickhouse-server/config.xml'.
ch-server-1 | Merging configuration file '/etc/clickhouse-server/config.d/docker_related_config.xml'.
dependency failed to start: container hdx-oss-dev-ch-server-1 is unhealthy
System Info:
OS: Ubuntu 25.04
Docker: 29.0.3, build 511dad6
CPU: 4 cores
RAM: 15.62 GiB
Compose file: docker-compose.dev.yml from main branch
Expected behavior:
All services start cleanly, or at least fail gracefully without destabilizing Docker.
Actual behavior:
Docker engine hangs or crashes; containers become undeletable; full Docker reset required.
Impact:
This makes local development using docker-compose.dev.yml currently unusable for me, despite the all-in-one image working fine.
Request:
Investigate possible race conditions or init loops in MongoDB or OTel collector containers
Happy to provide full logs if helpful!