Deployment Cancelled

Troubleshooting and help
, ,
1

I am unable get my docker finish build, and it seems like the health check or a similar build service is cancelling over timeout though docker is still clearly building after timeout is decided. I was using X-large instance, and the container is tested and compiles fine locally though takes time to build, a previous version without new libraries did build and deploy so the issue is mostly around compiling some python libraries taking its course. Any advice on pushing it over the fence?

Deployment ID 014c1610

image
image2224×546 169 KB

#8 1686.4 Building wheel for blis (pyproject.toml): still running…
#8 1746.8 Building wheel for blis (pyproject.toml): still running…
Instance is stopping
#8 1807.6 Building wheel for blis (pyproject.toml): still running…
canceled
#8 CANCELED
Instance stopped

2

hi @igal_nassima, What the app does during the start? Is it some kind of docker in docker configuration? Maybe using docker compose?

3

Uses Docker Compose:
Have a flask python web app, and PG in docker compose,

The failures started when I added spacy python library for NLP processing, which takes a while to compile, otherwise it started ok. I kept moving the server type up to more and more memory, but no luck yet.

My koyeb:

FROM koyeb/docker-compose
RUN apk add --no-cache git
COPY . /app/
CMD [“docker-compose”, “-f”, “./docker-compose.yml”, “up”]

my compose:

services:
python-app:
build:
context: .
dockerfile: Dockerfile
ports:
- “8000:8000”
volumes:
- .:/app:sb
depends_on:
db:
condition: service_healthy
links:
- db
db:
image: postgres
container_name: db
restart: always
ports:
- 5123:5432
env_file:
- .db.env
volumes:
- ./pData:/var/lib/postgresql/data
healthcheck:
test: [“CMD-SHELL”, “pg_isready -U $${POSTGRES_USER} -d $${POSTGRES_DB} -t 1”]
interval: 10s
timeout: 10s
retries: 10
start_period: 10s

4

Hi… Is timeout a dead end here? If we cannot up the server memory (paid account) and still cannot get a docker to not time out, we are a bit stuck.

5

You might need to go for a eSmall or eMedium instance. You might be using the entire memory of your instance.