All-Things-Docker-and-Kubernetes

Status Codes and Restart Policies

It is helpful to know status codes especially when you’re troubleshooting errors.

For container issues, we can specify restart policies inside our docker-compose file. The default restart policy is “no”, which means container won’t be restarted if it stops or crashes.

“No”

Here’s an example of a docker-compose file. Notice that we defined “no” enclosed in qoutes because without the quotes(no), YAML formatting will interpret it as a boolean false.

version: '3'
services:

  node-app:
    image: node
    restart: "no"

always vs. on-failure

If we use always, the container will automatically attempt to restart if it crashes. On the other-hand, on-failure would only restart the container if it returns an error status code (non-zero status codes). We also don’t need to enclosed them in quotes.

version: '3'
services:

  node-app:
    image: node
    restart: "no"

  web-server:
    image: redis
    restart: always

  worker:
    image: ubuntu
    restart: on-failure

Use always for containers that needs to be up 100% of the time, like web servers.

Use on-failure for containers that runs batch jobs or files processes.

Check out the labs for Docker Compose in this repository to see how restart policies are used.

Back to first page

This site is open source. Improve this page.