Doodba stands for Docker Odoo Base, and it is a highly opinionated image ready to put Odoo inside it, but without Odoo.
Yes, the purpose of this is to serve as a base for you to build your own Odoo project, because most of them end up requiring a big amount of custom patches, merges, repositories, etc. With this image, you have a collection of good practices and tools to enable your team to have a standard Odoo project structure.
BTW, we use Debian. I hope you like that.
Because developing Odoo is hard. You need lots of customizations, dependencies, and if you want to move from one version to another, it's a pain.
Also because nobody wants Odoo as it comes from upstream, you most likely will need to add custom patches and addons, at least, so we need a way to put all together and make it work anywhere quickly.
You can start working with this straight away with our scaffolding.
Basically, every directory you have to worry about is found inside /opt/odoo
.
This is its structure:
custom/
entrypoint.d/
build.d/
conf.d/
ssh/
config
known_hosts
id_rsa
id_rsa.pub
dependencies/
apt_build.txt
apt.txt
gem.txt
npm.txt
pip.txt
src/
private/
odoo/
addons.yaml
repos.yaml
common/
entrypoint.sh
build.sh
entrypoint.d/
build.d/
conf.d/
auto
addons/
odoo.conf
Let's go one by one.
Here you will put everything related to your project.
Any executables found here will be run when you launch your container, before running the command you ask.
Executables here will be aggregated with those in /opt/odoo/common/build.d
.
The resulting set of executables will then be sorted alphabetically (ascending) and then subsequently run.
Files here will be environment-variable-expanded and concatenated in
/opt/odoo/auto/odoo.conf
at build time.
It must follow the same structure as a standard ~/.ssh
directory, including
config
and known_hosts
files. In fact, it is completely equivalent to
~root/.ssh
.
The config
file can contain IdentityFile
keys to represent the private
key that should be used for that host. Unless specified otherwise, this
defaults to identity[.pub]
, id_rsa[.pub]
or id_dsa[.pub]
files found in
this same directory.
This is very useful to use deployment keys that grant git access to your private repositories.
Example - a private key file in the ssh
folder named my_private_key
for
the host repo.example.com
would have a config
entry similar to the below:
Host repo.example.com
IdentityFile ~/.ssh/my_private_key
Or you could just drop the key in id_rsa
and id_rsa.pub
files and it should
work by default without the need of adding a config
file.
Host key checking is enabled by default, which means that you also need to
provide a known_hosts
file for any repos that you wish to access via SSH.
In order to disable host key checks for a repo, your config would look something like this:
Host repo.example.com
StrictHostKeyChecking no
For additional information regarding this directory, take a look at this Digital Ocean Article.
Here you will put the actual source code for your project.
When putting code here, you can either:
- Use
repos.yaml
, that will fill anything at build time. - Directly copy all there.
Recommendation: use repos.yaml
for everything except for private
,
and ignore in your .gitignore
and .dockerignore
files every folder here
except private
, with rules like these:
odoo/custom/src/*
!odoo/custom/src/private
!odoo/custom/src/*.*
REQUIRED. The source code for your odoo project.
You can choose your Odoo version, and even merge PRs from many of them using
repos.yaml
. Some versions you might consider:
-
Original Odoo, by Odoo S.A..
-
OCB (Odoo Community Backports), by OCA. The original + some features - some stability strictness.
-
OpenUpgrade, by OCA. The original, frozen at new version launch time + migration scripts.
REQUIRED. Folder with private addons for the project.
A git-aggregator configuration file.
One entry per repo and addon you want to activate in your project. Like this:
website:
- website_cookie_notice
- website_legal_page
web:
- web_responsive
Advanced features:
-
You can bundle several YAML documents if you want to logically group your addons and some repos are repeated among groups, by separating each document with
---
. -
Addons under
private
andodoo/addons
are linked automatically unless you specify them. -
You can use
ONLY
to supply a dictionary of environment variables and a list of possible values to enable that document in the matching environments. -
If an addon is found in several places at the same time, it will get linked according to this priority table:
- Addons in
private
. - Addons in other repositories (in case one is matched in several, it will be random, BEWARE!). Better have no duplicated names if possible.
- Core Odoo addons from
odoo/addons
.
- Addons in
-
If an addon is specified but not available at runtime, it will fail silently.
-
You can use any wildcards supported by Python's glob module.
This example shows these advanced features:
# Spanish Localization
l10n-spain:
- l10n_es # Overrides built-in l10n_es under odoo/addons
server-tools:
- "*date*" # All modules that contain "date" in their name
- module_auto_update # Makes `autoupdate` script actually autoupdate addons
web:
- "*" # All web addons
---
# Different YAML document to separate SEO Tools
website:
- website_blog_excertp_img
server-tools: # Here we repeat server-tools, but no problem because it's a
# different document
- html_image_url_extractor
- html_text
---
# Enable demo ribbon only for devel and test environments
ONLY:
PGDATABASE: # This environment variable must exist and be in the list
- devel
- test
web:
- web_environment_ribbon
---
# Enable special authentication methods only in production environment
ONLY:
PGDATABASE:
- prod
server-tools:
- auth_*
Files to indicate dependencies of your subimage, one for each of the supported package managers:
apt_build.txt
: build-time dependencies, installed before any others and removed after all the others too. Usually these would include Debian packages such asbuild-essential
orpython-dev
.apt.txt
: run-time dependencies installed by apt.gem.txt
: run-time dependencies installed by gem.npm.txt
: run-time dependencies installed by npm.pip.txt
: a normal piprequirements.txt
file, for run-time dependencies too. It will get executed with--update
flag, just in case you want to overwrite any of the pre-bundled dependencies.
This folder is full of magic. I'll document it some day. For now, just look at the code.
Only some notes:
-
Will compile your code with
PYTHONOPTIMIZE=1
by default. -
Will remove all code not used from the image by default (not listed in
/opt/odoo/custom/src/addons.yaml
), to keep it thin.
This directory will have things that are automatically generated at build time.
It will be full of symlinks to the addons you selected in addons.yaml
.
It will have the result of merging all configurations under
/opt/odoo/{common,custom}/conf.d/
, in that order.
I will document all build arguments and environment variables some day, but for now keep this in mind:
-
This is just a base image, full of tools. You need to build your project subimage from this one, even if your project's
Dockerfile
only contains these 2 lines:FROM tecnativa/doodba MAINTAINER Me <me@example.com>
-
The above sentence becomes true because we have a lot of
ONBUILD
sentences here, so at least your project must have a./custom
folder along with itsDockerfile
for it to work. -
All should be magic if you adhere to our opinions here. Just put the code where it should go, and relax.
There is a good collections of tools available in the image that help dealing with Odoo and its peculiarities:
A handy CLI tool to automate addon management based on the current environment.
It allows you to install, update, test and/or list private, extra and/or core
addons available to current container, based on current addons.yaml
configuration.
Call addons --help
for usage instructions.
The CLI text editor we all know, just in case you need to inspect some bug in hot deployments.
Just a little shell script that you can use to add logs to your build or entrypoint scripts:
log INFO I'm informing
Little shell shortcut for exporting a translation template from any addon(s). Usage:
pot my_addon,my_other_addon
Little shortcut to make your odoo shell
scripts executable.
For example, create this file in your scaffolding-based project:
odoo/custom/shell-scripts/whoami.py
. Fill it with:
#!/usr/local/bin/python-odoo-shell
from __future__ import print_function
print(env.user.name)
print(env.context)
Now run it:
$ chmod a+x odoo/custom/shell-scripts/whoami.py # Make it executable
$ docker-compose build --pull # Rebuild the image, unless in devel
$ docker-compose run --rm odoo custom/shell-scripts/whoami.py
Another little shell script, useful for debugging. Just run it like this and Odoo will execute unit tests in its default database:
unittest my_addon,my_other_addon
Note that the addon must be installed for it to work. Otherwise, you should run it as:
unittest my_addon,my_other_addon -i my_addon,my_other_addon
Environment variables are there so that if you need to connect with the database, you just need to execute:
docker exec -it your_container psql
The same is true for any other Postgres client applications.
VSCode debugger. If you use this editor with its python module, you will find it useful.
To debug, add this Python code somewhere:
import ptvsd
ptvsd.enable_attach("my_secret", address=("0.0.0.0", 6899))
print("ptvsd waiting...")
ptvsd.wait_for_attach()
Of course, you need to have properly configured your VSCode. To do so, make
sure in your project there is a .vscode/launch.json
file with these minimal
contents:
{
"version": "0.2.0",
"configurations": [
{
"name": "Attach to debug in devel.yaml",
"type": "python",
"request": "attach",
"localRoot": "${workspaceRoot}/odoo",
"remoteRoot": "/opt/odoo",
"port": 6899,
"secret": "my_secret",
"host": "localhost"
}
]
}
Then, execute that configuration as usual.
This is another great debugger that includes remote debugging via telnet, which can be useful for some cases, or for people that prefer it over wdb.
To use it, inject this in any Python script:
import pudb.remote
pudb.remote.set_trace(term_size=(80, 24))
Then open a telnet connection to it (running in 0.0.0.0:6899
by default).
It is safe to use in production environments if you know what you are doing and do not expose the debugging port to attackers. Usage:
docker-compose exec odoo telnet localhost 6899
We found this one to be the most useful tool for downlading code, merging it and placing it somewhere.
This little script wraps git-aggregator
to make it work fine and
automatically with this image. Used in the scaffolding's setup-devel.yaml
step.
Example repos.yaml
file
This example merges several sources:
./odoo:
defaults:
# Shallow repositores are faster & thinner. You better use
# $DEPTH_DEFAULT here when you need no merges.
depth: $DEPTH_MERGE
remotes:
ocb: https://github.com/OCA/OCB.git
odoo: https://github.com/odoo/odoo.git
target:
ocb $ODOO_VERSION
merges:
- ocb $ODOO_VERSION
- odoo refs/pull/13635/head
We set an $OPENERP_SERVER
environment variable pointing to the autogenerated
configuration file so you don't have to worry about
it. Just execute odoo
and it will work fine.
Note that version 9.0 has an odoo
binary to provide forward compatibility
(but it has the odoo.py
one too).
Get up and running quickly with the provided scaffolding.
I will assume you know how to use Git, Docker and Docker Compose.
git clone https://github.com/Tecnativa/doodba-scaffolding.git myproject
cd myproject
ln -s devel.yaml docker-compose.yml
chown -R $USER:1000 odoo/auto
chmod -R ug+rwX odoo/auto
export UID GID="$(id -g $USER)" UMASK="$(umask)"
docker-compose build --pull
docker-compose -f setup-devel.yaml run --rm odoo
docker-compose up
And if you don't want to have a chance to do a git merge
and get possible
future scaffolding updates merged in your project's git log
:
rm -Rf .git
git init
The scaffolding provides you a boilerplate-ready project to start developing Odoo in no time.
This scaffolding comes with some environment configurations, ready for you to extend them. Each of them is a Docker Compose file almost ready to work out of the box (or almost), but that will assume that you understand it and will modify it.
After you clone the scaffolding, search for XXX
comments, they will help
you on making it work.
Set it up with:
export UID GID="$(id -g $USER)" UMASK="$(umask)"
docker-compose -f setup-devel.yaml run --rm odoo
Once finished, you can start using Odoo with:
docker-compose -f devel.yaml up --build
This allows you to track only what Git needs to track and provides faster Docker builds.
You might consider adding this line to your ~/.bashrc
:
export UID GID="$(id -g $USER)" UMASK="$(umask)"
To browse Odoo go to http://localhost:${ODOO_MAJOR}069
(i.e. for Odoo 11.0 this would be http://localhost:11069
).
This environment has several special features:
This is one of the greatest Python debugger available, and even more for Docker-based development, so here you have it preinstalled.
I told you, this image is opinionated. 😉
To use it, write this in any Python script:
import wdb
wdb.set_trace()
It's available by default on the development environment, where you can browse http://localhost:1984 to use it.
DO NOT USE IT IN PRODUCTION ENVIRONMENTS. (I had to say it).
It provides a fake SMTP server that intercepts all mail sent by Odoo and displays a simple interface that lets you see and debug all that mail comfortably, including headers sent, attachments, etc.
- For development, it's in http://localhost:8025
- For testing, it's in http://$DOMAIN_TEST/smtpfake/
- For production, it's not used.
All environments are configured by default to use the bundled SMTP relay. They are configured by these environment variables:
SMTP_SERVER
SMTP_PORT
SMTP_USER
SMTP_PASSWORD
SMTP_SSL
EMAIL_FROM
For them to be useful, you need to remove any ir.mail_server
records in your
database.
The Docker network is in --internal
mode, which means that it has
no access to the Internet. This feature protects you in cases where
a production database is restored and Odoo tries to connect to
SMTP/IMAP/POP3 servers to send or receive emails. Also when you are
using connectors,
mail trackers
or any API sync/calls.
If you still need to have public access, set internal: false
in the
environment file, detach all containers from that network, remove the network,
reatach all containers to it, and possibly restart them. You can also just do:
docker-compose down
docker-compose up -d
Usually a better option is whitelisting.
This environment is just a template. It is not production-ready. You must change many things inside it, it's just a guideline.
It includes pluggable smtp
and backup
services.
Once you fixed everything needed, run it with:
docker-compose -f prod.yaml up --build --remove-orphans
Remember that you will want to backup the filestore in /var/lib/odoo
volume.
For production and [test][] templates to work fine, you need to have a working Traefik inverse proxy in each node.
To have it, use this inverseproxy.yaml
file:
version: "2.1"
services:
proxy:
image: traefik:1.6-alpine
networks:
shared:
private:
public:
volumes:
- acme:/etc/traefik/acme:rw,Z
ports:
- "80:80"
- "443:443"
depends_on:
- dockersocket
restart: unless-stopped
privileged: true
tty: true
command:
- --ACME.ACMELogging
- --ACME.Email=you@example.com
- --ACME.EntryPoint=https
- --ACME.HTTPChallenge.entryPoint=http
- --ACME.OnHostRule
- --ACME.Storage=/etc/traefik/acme/acme.json
- --DefaultEntryPoints=http,https
- --EntryPoints=Name:http Address::80 Redirect.EntryPoint:https
- --EntryPoints=Name:https Address::443 TLS
- --LogLevel=INFO
- --Docker
- --Docker.EndPoint=http://dockersocket:2375
- --Docker.ExposedByDefault=false
- --Docker.Watch
dockersocket:
image: tecnativa/docker-socket-proxy
privileged: true
networks:
private:
volumes:
- /var/run/docker.sock:/var/run/docker.sock
environment:
CONTAINERS: 1
NETWORKS: 1
SERVICES: 1
SWARM: 1
TASKS: 1
restart: unless-stopped
networks:
shared:
internal: true
driver_opts:
encrypted: 1
private:
internal: true
driver_opts:
encrypted: 1
public:
volumes:
acme:
Then boot it up with:
docker-compose -p inverseproxy -f inverseproxy.yaml up -d
This will intercept all requests coming from port 80 (http
) and redirect them
to port 443 (https
), it will download and install required SSL certificates
from Let's Encrypt whenever you boot a new production instance, add the
required proxy headers to the request, and then redirect all traffic to/from
odoo automatically.
It includes a security-enhaced proxy to reduce attack surface when listening to the Docker socket.
This allows you to:
- Have multiple domains for each Odoo instance.
- Have multiple Odoo instances in each node.
- Add an SSL layer automatically and for free.
A good rule of thumb is test in testing before uploading to production, so this environment tries to imitate the production one in everything, but removing possible pollution points:
Test it in your machine with:
docker-compose -f test.yaml up --build
This environment also needs a global inverse proxy.
In examples below I will skip the -f <environment>.yaml
part and assume you
know which environment you want to use.
Also, we recommend to use run
subcommand to create a new container with same
settings and volumes. Sometimes you may prefer to use exec
instead, to
execute an arbitrary command in a running container.
docker-compose run --rm odoo psql
You will need to restart it whenever any Python code changes, so to do that:
docker-compose restart -t0 odoo
In production:
docker-compose restart odoo https
docker-compose run --rm odoo odoo --stop-after-init --init addon1,addon2
docker-compose run --rm odoo unittest addon1,addon2
For all services in the environment:
docker-compose logs -f --tail 10
Only Odoo's:
docker-compose logs -f --tail 10 odoo
docker-compose run --rm odoo odoo -i addon1,addon2 --stop-after-init
docker-compose run --rm odoo odoo -u addon1,addon2 --stop-after-init
Add module_auto_update
from https://github.com/OCA/server-tools to your
installation following the standard methods of repos.yaml
+ addons.yaml
.
Now we will install the addon:
docker-compose up -d
docker-compose run --rm odoo --stop-after-init -u base
docker-compose run --rm odoo --stop-after-init -i module_auto_update
docker-compose restart odoo
It will automatically update addons that got updated every night. To force that automatic update now in a separate container:
docker-compose up -d
docker-compose run --rm odoo autoupdate
docker-compose restart odoo
docker-compose run --rm odoo pot addon1[,addon2]
Now copy the relevant parts to your addon1.pot
file.
docker-compose run --rm odoo odoo shell
docker-compose run --rm -p 127.0.0.1:$SomeFreePort:8069 odoo
Then open http://localhost:$SomeFreePort
.
This image is production-ready, but it is constantly evolving too, so some new features can break some old ones, or conflict with them, and some old features might get deprecated and removed at some point.
The best you can do is to subscribe to the compatibility breakage announcements issue.
These instructions assume you use the official scaffolding.
To remove the www.
prefix, set these params in the .env
file:
DOMAIN_PROD=example.com
DOMAIN_PROD_ALT=www.example.com
To add the www.
prefix, it is almost the same:
DOMAIN_PROD=www.example.com
DOMAIN_PROD_ALT=example.com
Of course, both domains should point to the same machine before booting, or Let's Encrypt might ban your server for some time.
In .env
, set DOMAIN_PROD
to host1.com,host2.com,www.host1.com
, etc.
This image includes a hack that will set the initial language to load when Odoo creates its database for the first time. These conditions must match:
$PGDATABASE
is set.- That database does not yet exist.
$INITIAL_LANG
is set to any Odoo lang code. I.e.es_ES
.- Odoo is booted.
I use Fish, how to export needed variables?
Do:
set -x UID (id -u $USER)
set -x GID (id -g $USER)
set -x UMASK (umask)
You can make those variables universal (available in all terminals you open
from now on) by using set -Ux
instead of set -x
.
Most likely you are using versions 8.0
or 9.0
of the image. If so:
- Edit
devel.yaml
. - Search for the line that starts with
command:
in theodoo
service. - Change it for a command that actually works with your version:
odoo --workers 0
for Odoo 8.0.odoo --workers 0 --dev
for Odoo 9.0.
Of course. There's no guarantee that we will like it, but please do it. 😉
It runs triggers when doing the automatic build in the Docker Hub. Check this.
Can I have my own scaffolding?
You probably should, and rebase on our updates. However, if you are planning on a general update to it that you find interesting for the general-purpose one, please send us a pull request.
Let's suppose you want to use test.yaml
environment by default,
no matter where you clone the project:
ln -s test.yaml docker-compose.yaml
git add docker-compose.yaml
git commit
Let's suppose you only want to use devel.yaml
in your local development
machine by default:
ln -s devel.yaml docker-compose.yml
Notice the difference in the prefix (.yaml
vs. .yml
). Docker Compose will
use the .yml
one if both are found, so that's the one we considered you
should use in your local clones, and that's the one that will be git-ignored by
default by the scaffolding .gitignore
file.
As a design choice, the scaffolding defaults to being explicit.
Version-pinning is a good idea to keep your code from differing among image updates. It's the best way to ensure no updates got in between the last time you checked the image and the time you deploy it to production.
You can do it through its sha256 code.
Get any image's code through inspect, running from a computer where the correct image version is downloaded:
docker image inspect --format='{{.RepoDigests}}' tecnativa/doodba:10.0-onbuild
Alternatively, you can browse this image's builds, click on the one
you know it works fine for you, and search for the digest
word using your
browser's search in page system (Ctrl+F usually).
You will find lines similar to:
[...]
10.0: digest: sha256:fba69478f9b0616561aa3aba4d18e4bcc2f728c9568057946c98d5d3817699e1 size: 4508
[...]
8.0: digest: sha256:27a3dd3a32ce6c4c259b4a184d8db0c6d94415696bec6c2668caafe755c6445e size: 4508
[...]
9.0: digest: sha256:33a540eca6441b950d633d3edc77d2cc46586717410f03d51c054ce348b2e977 size: 4508
[...]
Once you find them, you can use that pinned version in your builds, using a Dockerfile similar to this one:
# Hash-pinned version of tecnativa/doodba:10.0-onbuild
FROM tecnativa/doodba@sha256:fba69478f9b0616561aa3aba4d18e4bcc2f728c9568057946c98d5d3817699e1
Make sure there's a ir.config_parameter
called report.url
with the value
http://localhost:8069
.
This can become useful when you have isolated environments
(like in devel.yaml
and test.yaml
by default) but you need to allow
some external API access for them. I.e., you could use
Google Fonts API for your customer's reports, and those reports
would take forever and end up rendering badly in staging environments.
In such case, we recommend using the tecnativa/whitelist image. Read its docs there.
Just head to our project and open an issue or pull request.
If you plan to open a pull request, remember that you will usually have to open two of them:
- Targeting the
master
branch, from which the main images are built. This pull request must include tests. - Targeting the
scaffolding
branch, which serves as the base for projects using this base image. This one is not always required.
If you need to add a feature or fix for scaffolding
, before merging that PR,
we need tests that ensure that backwards compatibility with previous
scaffolding versions is preserved.