GithubHelp home page GithubHelp logo

buildout's Introduction

The Starzel Buildout

image

This is a standard Plone-buildout of the company Starzel.de.

Main features

  • It extends to config- and version-files on github shared by all projects that use the same version of Plone.
  • It allows to update a project simply by changing the version it extends.
  • It allows to update all projects of one version by changing remote files (very useful for HotFixes).
  • It is minimal work to setup a new project.
  • It has presets for development, testing, staging and production.
  • It has all the nice development-helpers we use.

Quickstart

git clone https://github.com/starzel/buildout SOME_PROJECT
cd SOME_PROJECT

Remove all files that are not needed for a project but are only used for the buildout itself.

rm -rf linkto README.rst README.txt .travis.yml VERSION.txt local_coredev.cfg CHANGES.rst

If you're not developing the buildout itself you want a create a new git repo.

rm -rf .git && git init

Add a file that contains a passwort. Do not use admin as a password in production!

echo -e "[buildout]\nlogin = admin\npassword = admin" > secret.cfg

Symlink to the file that best fits you local environment. At first that is usually development. Later you can use production or test. This buildout only uses local.cfg and ignores all local_*.cfg.

ln -s local_develop.cfg local.cfg

Create a virtualenv in Python 2.7 or Python 3.7 (Plone 5.2 only).

virtualenv .  # for Python 2.7
python3.7 -m venv .  # for Python 3 (Plone 5.2 only)

Install and configure Plone

./bin/pip install -r requirements.txt
./bin/buildout

Install git pre-commit hooks

./bin/pre-commit install

Structure

buildout.cfg

This contains the project settings (name, addons, checkouts etc.).

local.cfg

For each environment (development, production, test) there is a separate local_*.cfg-file. You create a symlink called local.cfg to one of these files depending on your environment. Each of the files includes the remote base.cfg that is hosted on github like this:

extends = https://raw.githubusercontent.com/starzel/buildout/5.1/linkto/base.cfg

This example refers to the tag 5.1 of this buildout that uses Plone 5.1 To use a different Plone-version simply change that to point to a different tag.

base.cfg

This remote file conatains most of the commonly used logic used for prodcution. It also includes two version-files that are also hosted on github:

pinned_versions_project.cfg

Here you pinn versions to overwrite or extend the hosted pinned_versions.cfg. These eggs are usually pinned for a reason and are usually not safe to be upgraded.

floating_versions_project.cfg

Here you overwrite and extend the hosted floating_versions.cfg. These eggs should usually be safe to be upgraded. ./bin/checkversions floating_versions_project.cfg will check pypi if there are newer releases for your pinned eggs.

Versions

We support the following versions of Plone:

To use this buildout with a version of Plone that is currently in development (a.k.a. the Coredev) please use:

Please note that new features are not always introduced to old versions.

  • Tags for development-versions (alpha, beta and rc) will exists but will be removed after the final release of that version.

Configuring a project

buildout.cfg contains the general project settings. Here you configure the name of the project, the eggs, source-checkouts and languages Plone will use.

Use in development

Symlink to the development-config:

ln -s local_develop.cfg local.cfg

The development-setup will build a simple instance with some useful tools (see below). The setup assumes that zeo, varnish and loadbalancing are only configured on production.

Install git pre-commit hooks using the pre-commit tool that was installed via requirements.txt:

./bin/pre-commit install

Use in production

Symlink to the production-config:

ln -s local_production.cfg local.cfg

A average project could use this stack pipeline:

nginx > varnish > nginx (for load-balancing) > at least 2 zeoclients > zeoserver

In local_production.cfg select the parts you really need.

parts +=
    zeoserver
    ${buildout:zeoclients}
    zeoclient_debug
    ${buildout:cron-parts}
    varnish-config
    backup
    precompiler
    nginx-config
    site_unit

Note that site_unit is a part that creates a systemd service unit which pulls in service units for zeoserver and any zeoclients you set up (see below). Starting, restarting and stopping the site unit will start, restart and stop, resp., the zeoserver and zeoclients. The site unit is the only systemd unit in the deployment that needs to be enabled or disabled.

If you add any zeoclients, also add their partnames to buildout:zeoclients. This is a list used in several places, e.g. to populate the dependencies of the main systemd service for the site.

zeoclients +=
    zeoclient2

A note on the execution environment of the processes: In order to have zeoserver and zeoclients run in the environment of the service user they run as, the systemd services need to be executed by a shell that has those environments loaded. This is what the run.sh script is about; see the comments within the script for details.

Server stack

Frontend webserver (Nginx)

The first Nginx manages the virtualhost, does url rewrites if needed and terminates the SSL (needs to be done before varnish).

A full demo-config with ssl and redirects can be found in this repo in /templates/demo_nginx.conf.

A minimal config without ssl can be found in the demo.plone.de project.

More information can also be found in the PloneDocs

Cache (Varnish)

After nginx we use varnish to cache the site. You can activate it like this:

parts +=
    ...
    varnish-config
    ...

Take a look in linkto/base.cfg for the varnish-config part, there are several switches to configure.

It is best practice to install varnish from your distribution repository. If you need to build varnish (e.g. because your system does not ship with the version you need), see plone.recipe.varnish. The same recipe that we use to configure varnish can also be used to build it.

If you use the system-varnish only need the [varnish-config] part, it will generate the config (vcl) for you. In /etc/varnish/default.vcl include the generated vcl:

vcl 4.0;

include "<path to your buildout>/etc/varnish.vcl";

A systemctl restart varnish should activate the new config. To use one varnish installation with serveral vhosts, see the Varnish with multiple sites section below.

Loadbalancer (Nginx)

Another Nginx spreads the requests to several Zeoclients, here is a minimal config. In production you can look at the demo.plone.de project

# starzel (zeoclients)
upstream starzel_zeoclients {
    ip_hash;
    server 127.0.0.1:8082;
    server 127.0.0.1:8083;
    server 127.0.0.1:8084;
    }

The ip_hash option is needed for multiple Zeoclients, more information can be found in this issue

The ip and port has to be the same as the settings for the zeoclients in then part [bindips] and [ports].

Varnish with multiple sites

The generated varnish config works with a single vhost, for multiple sites/domains we need a custom varnish config. This configuration is not yet build into the buildout script/template, we need to do the changes manually in a copy of a varnish config file (just copy the varnis4.vcl over to /etc/varnish and include it in default.vcl).

In the varnish.vcl we need to add the additional backend, note the different loadbalancer port.

backend 001 {
   .host = "localhost";
   .port = "8091";
   .connect_timeout = 0.4s;
   .first_byte_timeout = 300s;
   .between_bytes_timeout  = 60s;
}

backend 002 {
   .host = "localhost";
   .port = "8081";
   .connect_timeout = 0.4s;
   .first_byte_timeout = 300s;
   .between_bytes_timeout  = 60s;
}

In sub vcl_recv we remove the backend (set req.backend_hint = backend_000;) and add this switch:

if (req.http.host == "my_host") {
    set req.backend_hint = 001;
}
else {
    set req.backend_hint = 002;
}

This does the vhost routing to the different backends. "my_host" is the upstream name of the cache, see the config of demo.plone.de project. The Varnish config can be tested with this command: varnishd -C -f /etc/varnish/default.vcl

Use for test-instances

Create a copy of local_production.cfg called local_test.cfg and modify it according to your needs.

Warning

If test runs on the same server as production:

In this case you need a different name for the project on test. Otherwise one will overwrite the database of the other. Because of this the name of the project must not be set in buildout.cfg but in the local_*.cfg-files.

Features

Debugging and Development

packages

All eggs of your buildout will be symlinked to in parts/packages.

zopepy

Run ./bin/zopepy to have a python-prompt with all eggs of your buildout in its python-path.

checkversions

Run ./bin/checkversions floating_versions_project.cfg to check if your pinned eggs are up-to-date.

stacktrace

The part stacktrace-script adds a bash-script ./bin/stack.sh that will print the current stacktrace to stdout. Useful to find out what Plone is doing when it's busy.

This was removed in 5.2.2 because it only works with ZServer (i.e. in Python 2). Use https://pypi.org/project/py-spy/ instead.

pre-commit

This installs a pre-commit-hook that runs several code analysis tests including black.

mrbob

This part adds bobtemplates.plone to simplify the creation of new addons.

test

Run tests for your test-eggs

coverage-test

Generate coverage-reports for your test-eggs in parts/test/.

i18n

Restrict loaded languages

By default only german ('de') is loaded on startup. In your buildout.cfg you can override the loaded languages using language = de en fr. This setting also affects the languages used in the i18nize-xxx part. (see http://maurits.vanrees.org/weblog/archive/2010/10/i18n-plone-4#restrict-the-loaded-languages)

i18nize-xxx

Modify the commented-out part i18nize-xxx to get a script that runs i18ndude fro an egg. Here is an example for the egg dynajet.site adding a script ./bin/i18nize-site.

[i18nize-site]
recipe = collective.recipe.template
input = ${buildout:directory}/i18nize.in
output = ${buildout:bin-directory}/i18nize-site
mode = 775
dollar = $
domain = dynajet.site
packagepath = ${buildout:directory}/src/dynajet.site/src/dynajet/site
languages = ${buildout:languages}
i18nize-all

This runs all i18nize commands for a package.

Testing

Setup for gitlab-ci

The config local_ci.cfg can be used by your ci-system to run the buildout.

Deployment

monitoring

Change the settings for maxram to have memmon restart an instance when it uses up to much memory.

Sentry logging

Configure zeoclients to send tracebacks to Sentry in local_production.cfg by uncommenting it and adding a dsn. You also need to enable the egg raven. Repeat for each zeoclient.

Hotfixes

This buildout automatically includes the correct Hotfixes for the version of Plone you use. E.g. the extends-file for Plone 5.0.6 https://raw.githubusercontent.com/starzel/buildout/5.0.6/linkto/base.cfg pulls in the file https://raw.githubusercontent.com/starzel/buildout/master/linkto/hotfixes/5.0.6.cfg which in turn contains the pinns and eggs for all HotFixes for that version.

By having the hotfixes-files in the master-branch we can easily update Hotfixes for each version without having to move any tags. The same day a Hotfix is published the corresponding extends-files will be updated. You simply have to rerun buildout and restart your site to include them.

Notes

local.cfg and secret.cfg must never be versioned. The file .gitignore in this buildout already prevent this.

It might feel weird that buildout.cfg loads local.cfg, but this avoids some weird behavior of buildouts extends-feature.

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.