[Question]What's the difference between using tini and a shell script in a container about tini HOT 3 CLOSED

Using the shell script to start your node application will do that your application doesn't receive signals. So for example, if you run your container and then try to docker stop it, it will take 10 seconds, since Docker sends a SIGTERM to stop your container, but your container doesn't handle signals (because shell is the main process), so after 10 seconds Docker gives up and sends a SIGKILL to kill the container. Receiving a SIGKILL could lead to unexpected behaviour if you are running for example an HTTP server. Normally we stop servers using graceful shutdown, instead of just killing them.

It'll also happen if you try to ctrl+c your container. It won't stop.

This won't happen with Tini.

from tini.

@zjleon the answer depends on what your shell script does; and whether it uses exec when starting your NodeJS server.

Can you post your shell script? I'd be able to provide a more accurate answer then :)

from tini.

@krallin thx for reply,
this is my shell script to start the node server

// container.dev.web_static.initial.sh
npm install --prefix ./src/web_static
npm run dev:web_static

in the dockerfile, I just copy it to the container, then execute it

ADD ./scripts/container.dev.web_static.initial.sh ./initial.sh
CMD ["bash", "./initial.sh"]

the shell script right now is real simple, I can't see too much benefit I can get from using tini(from what I learn from the README), Is that right?

from tini.