lupusmichaelis / lupus-dev-toolkit Goto Github PK

When accessing /folder through a reverse proxy $virtualhost, and no index file's present, will redirect to http://$container_ip/folder/ instead of http://$virtualhost/folder/.

https://aquasecurity.github.io/trivy/v0.18.3/installation/

In base script build/igor/library.sh, “environement” should be “environment”.

This typo could be a huge headache for other users of the images.

Need to bump up minor revision on each image once it's done, building up everything.

Instead of incrementing base tag image in Dockerfile, let's make it an argument we can control from the build process. This could allow more generic building as the base tag could mention the OS, and other relevant information.

Package managers have their quirks and gotcha than are not easy to be kept in mind at all time. It would be a fine idea to provide wrapping helper which would take care of the details.

Package names might be tricky to be distribution agnostic, we might be wise to keep that for a later issue.

https://www.thegeekstuff.com/2016/04/mysql-optimize-table/

The base docker-entrypoint is quite naive, and only allow for executing the command that's provided to it. We should be smarter and find a way for the image author to specify the base command of the container (or even retrieve it from the CMD of the last dockerfile).

https://jpetazzo.github.io/2014/06/17/transparent-squid-proxy-docker/
https://gist.github.com/hvrauhal/f98d7811f19ad1792210

Not always easy to know how to properly set a specific feature in a specific file. Lets provide generic helpers for the task, that will call the distribution specific and image base specific tools to set up properly.

Maybe Salt or Ansible are already doing the job and we should pick one of them.

Provide helpers for easier sed, merge, diff and other format specific manipulation.

See if we can build leaner images by copying binary and configuration, rather than carrying over the whole OS.

A possible implementation would be to have leaf images, so the branch images can still be specialized by subsequent authors.

We don't necessary need to carry over all helper scripts that allow us to build image and configure them. We need a mechanism to clean up that mess as soon as they're not requested anymore.

It is too easy to discard valuable data with docker compose down -v, let's build some script that will avoid the pain of data loss. This is to compensate the lack of official support: moby/moby#33782