With this change, containers are no longer run as "root" but as unprivileged
users. This is necessary in some environments, notably some Kubernetes
clusters.
To make this possible, we need to manually fix bind-mounted volumes in
docker-compose. This is pretty much equivalent to the behaviour in Kubernetes,
where permissions are fixed at runtime if the volume owner is incorrect. Thus,
we have a consistent behaviour between docker-compose and Kubernetes.
We achieve this by bind-mounting some repos inside "*-permissions" services.
These services run as root user on docker-compose and will fix the required
permissions, as per build/permissions/setowner.sh These services simply do not
run on Kubernetes, where we don't rely on bind-mounted volumes. There, we make
use of Kubernete's built-in volume ownership feature.
With this change, we get rid of the "openedx-dev" Docker image, in the sense
that it no longer has its own Dockerfile. Instead, the dev image is now simply
a different target in the multi-layer openedx Docker image. This makes it much
faster to build the openedx-dev image.
Because we declare the APP_USER_ID in the dev/docker-compose.yml file, we need
to pass the user ID from the host there. The only way to achieve that is with a
tutor config variable. The downside of this approach is that the
dev/docker-compose.yml file is no longer portable from one machine to the next.
We consider that this is not such a big issue, as it affects the development
environment only.
We take this opportunity to replace the base image of the "forum" image. There
is now no need to re-install ruby inside the image. The total image size is
only decreased by 10%, but re-building the image is faster.
In order to run the smtp service as non-root, we switch from namshi/smtp to
devture/exim-relay. This change should be backward-compatible.
Note that the nginx container remains privileged. We could switch to
nginxinc/nginx-unprivileged, but it's probably not worth the effort, as we are
considering to get rid of the nginx container altogether.
Close #323.
- 💥[Improvement] Fail on incorrect image name argument in `images
build/pull/push/printtag` commands.
- [Bugfix] Remove trailing slashes in docker-compose files for
[compatibility with docker-compose v2 in
WSL](https://github.com/docker/compose/issues/8558).
- [Improvement] `settheme` now works with preview domain.
- [Feature] Allow specifying extra pip packages through config.yml.
Added OPENEDX_EXTRA_PIP_REQUIREMENTS setting, which allows to specify
extra pip packages that should be installed.
Moved "openedx-scorm-xblock" package from Dockerfile to the new setting
in the config.yml.
When running "tutor k8s" we can disable the forum deployment and
service by setting RUN_FORUM to false, but the corresponding job is
added to jobs.yml unconditionally.
Add a conditional to tutor/templates/k8s/jobs.yml so that the forum
job definition is only included when RUN_FORUM is true.
Fixes #525.
In some cases, plugins declare an entry point but cannot be loaded.
This is the case when they depend on a version of tutor that is not the
one that is currently installed. This use case is very frequent when
working on multiple versions at the same time (i.e: right now, while we
are working on the Maple release). In such cases, it's best just to
ignore the plugin entirely rather than having to re-install all plugins
in the virtualenv.
- [Improvement] Upgrade all services to open-release/lilac.3.
- [Feature] Make it possible to override job configuration in
development: if they exist, `dev/docker-compose.jobs.yml` and
`dev/docker-compose.jobs.override.yml` will be loaded when running jobs.
- [Improvement] Faster `tutor local start` by building only necessary
images.
Previously, job declarations were always loaded from local/docker-compose.yml
and local/docker-compose.jobs.yml. This meant that it was not possible to
override job declarations in dev mode. It is now the case, with
dev/docker-compose.jobs.yml and dev/docker-compose.jobs.override.yml. Neither
of these files exist yet... But who knows? we might need this feature one day.
In any case the code is much cleaner now.
Before, custom `docker_compose_func` arguments had to be passed to job runners.
This was not very elegant. Also, it prevented us from loading custom job files
in development.
Here, we adopt a better object-oriented approach, where context classes are
ordered hierarchically.
This paves the way for loading `dev/docker-compose.jobs.yml` files in `tutor
dev init` commands -- which will be necessary to fix permissions in dev/local
mode.
Previously, we were building all images every time we ran a "local start"
command. This was causing unnecessary rebuild. Here, instead, we make use of
the `docker-compose up --build`. This means that only the required images will
be rebuilt.
Limits the memory chek to the 'local quickstart' command, makes error
handling more accurate and adds warning messages for some conditions.
Also adds a mention of this in troubleshooting.rst.
Adds a check in the 'local' command group that requires at least
4 GB of RAM to be allocated to Docker when running any of the
local subcommands on macOS. This addresses a common issue where
Docker's default setting (2 GB) causes startup to crash with
misleading error messages.
- 💥[Improvement] Change the `settheme` command such that, by default, a custom theme is assigned to the LMS and the CMS, both in production and development mode.
In conversations with edX, we learned that the name "edge" had negative
undertones for historical reasons. Thus, we switch to "nightly", which means
pretty much the same thing.
Here, we make it possible to automatically append a suffix to the version and app
name (in the sense of appdirs). This guarantees that a tutor edge project will
not accidentally override another community release.
In addition, we take the opportunity to document the tutor versioning format.
(I've been meaning to do that for a long time)
It is unnecessary to point to CI, or to indicate the doc version. Instead, we
link directly to the source code. Also, we improve the icon colors and general
appearance.
This ensures that any warning generated from compiling the docs is treated as
an error. Also, building the docs is now one of the steps performed in CI.
<rant>I attempted to actually run Tutor with Podman and I was sorely disappointed.
The only reliable source of docs that I found concerning the integration with
docker-compose is this blog post:
https://www.redhat.com/sysadmin/podman-docker-compose
There are no other official docs 😓
1. The instructions given in the blog post don't work out of the box. Launching
the podman service failed altogether on Ubuntu 20.04 and 20.10. It worked on
CentOS 8, but some parameters need to changed, such as the docker socket path.
2. After I got the podman service working, I managed to get an Open edX
platform running with tutor, but with the root user. Then, containers
complained that they could not write data to the bind-mounted volumes. I
attempted to run as a non-root user, and discovered that the podman socket is
only readable by root. This should explain why all commands from that blog post
are prefixed by sudo.
Long story short, I was hoping to update the tutorial. Instead, I'm just moving
it for the sake of better organisation. For the life of me, I do not understand
why some people would want to run Podman instead of Docker. Bad documentation
is an immediate turn-off for me. From my perspective, podman is mostly an
overblown marketina stunt.</rant>
There is too much information in each of the local/k8s/dev docs pages. The
"guides" that are listed in each one of those pages are moved either to "common
tasks" or to a dedicated "tutorials" section. This paves the way for more
comprehensive tutorials, where we describe how to run the latest master
branches of Open edX.
I am well aware that, as they stand, the tutorials are of poor quality and
should be rewritten. This is a task for another day/commit. For now, we only
move the contents to a separate part of the docs.
Also, we should add a "reference" section to the docs, where we add the result
of `tutor <subcommand> --help`.
Previously, the list of domain names to which a theme was assigned had to be
specified manually. Now, the themes are automatically assigned to the LMS and
the CMS, both in development and production modes.
