Documentation Contributor Guide

This section describes the guidelines for contributing to the StarlingX documentation.

Quickstart

The StarlingX documentation uses reStructuredText (RST) markup syntax with Sphinx extensions. It uses the same contribution setup and workflow as the OpenStack documentation.

Setup for contribution

Follow the OpenStack instructions for setting up for contribution.

Make a change

  1. Make changes following the OpenStack instructions for:

    1. Starting a change

    2. Committing a change.

      Note

      StarlingX requires the use of a Signed-off-by header. Use the -s option with git commit.

  2. When writing documentation, follow Writing style and RST conventions.

  3. Build the documentation locally to verify your changes before committing. Follow the OpenStack instructions for Building documentation.

  4. If needed, follow up with edits to your patch following the OpenStack instructions for Responding to requests.

Find tasks and help needed

If you are looking for work to complete:

If you make a contribution that has an the associated story, task, or bug in the comment, link to the related story or bug as described in the Code Submission Guidelines.

Docs organization

Documentation for StarlingX is organized into the following sections:

Introduction

Overview of the StarlingX project.

Installation Guides

Release-specific installation and deployment guides.

Configuration

Configuration references for post-installation StarlingX system configuration.

Operation Guides

System administration and maintenance guides.

REST API Reference

REST API references for the StarlingX project. For additional information about where REST API documentation is located, see API documentation.

CLI Reference

Reference for the StarlingX project command line interface (CLI).

Developer Resources

Resources for developers using or building StarlingX.

Release Notes

Release notes for all StarlingX releases.

Contributor Guides

Overview and guidelines for contributing to StarlingX documentation.

API documentation

The structure and location of the REST API documentation deserves extra explanation.

Most REST API content is generated from the StarlingX project associated with the API. For example, the documentation for the StarlingX metal REST API is generated from the metal repository.

API references for StarlingX extensions are part of the docs repository, located in the api-ref project:

  • StarlingX extensions to the OpenStack Block Storage API

  • StarlingX extensions to the OpenStack Compute API

  • StarlingX extensions to the OpenStack Image API

  • StarlingX extensions to the OpenStack Networking API

The api-ref project also contains index pages used by Sphinx to generate the final content tree. Note that the REST API landing page used to render content in the generated website is found in the doc project.

For additional information on the API documentation, refer to API Documentation Contributor Guide.

Spec documentation

Spec documentation is found in the Starlingx specs project.

The specs/2019.03 directory contains the documentation files for approved and implemented specs.

Writing style

StarlingX documentation follows many (but not all!) of the writing style guidelines described in the OpenStack documentation writing style guide. Differences between the StarlingX and OpenStack practices are highlighted below.

  • Use Title Case for page titles. For example:

    ===============================
Documentation Contributor Guide
===============================

  • Start section titles with an action verb. Do not use a gerund (word that ends with -ing). For example:

    ------------------
Configure endpoint
------------------

Create rST Files

Use the tox -e newfile command to create new rST files.

About this task

All rST files created in StarlingX documentation repositories must have the following characteristics:

  • They must have unique file names.

  • They must have rST labels at the beginning of the files that match the file names.

Important

These rules apply to index files as well as those containing user documentation.

A utility is available for use from within each documentation repository you have installed to generate uniquely named files for you.

Prerequisites

You must have uuidgen installed on your system. This program is included by default on most modern Linux distributions. If it is not installed, consult your distribution’s documentation for instructions.

Procedure

  1. Change to the directory where you wish to create a new topic.

    Typically, this will be below the doc/source directory of the repository.

  2. Run the following tox command.

    tox -e newfile

  3. When prompted, enter a title for the new topic.

     You are about to create a new reStructuredText file in

 /home/jdoe/starlingx/docs/doc/source/intro

 or a content fragment file in doc/source/_includes

 If this is not what you want, press CTL-C to quit and change to the directory
 you want to create the file in.

 Enter a title for the new topic. The file name and topic label used for
 linking will be based on this value.


 Topic title:

    1. Review the directory (an example is highlighted above) that the utility will create the new file in.

      Note

      This does not apply if you choose to create a content fragment using the f option when prompted. In that case, the file will be saved to doc/source/_includes regardless of your current working directory.

    2. If this is not correct, press CTL-C to quit, change to the correct directory, and run the command again; otherwise, type the topic title and press ENTER.

  4. When prompted, select the type of rST stub file you want to create.

    Thanks. Now choose a topic type. Enter one of the following characters:

t) A task topic. Will contain the outline of a procedure.
i) An index.
r) A reference topic. Will contain a minimal list-table definition.
g) A minimal generic topic.
f) A content fragment included in an rST file. Will be saved to doc/source/_includes.
Topic type:

    Each option creates a stub file with different templated content useful for getting started. Press the corresponding key.

Results

The new rST file is created.

The title used in the new rST file matches what you typed exactly. However, some changes have been made to the file name and topic label.

For example, if you entered Architectural Considerations! as a title, listing the directory will show a file similar to the following:

$ ls

architectural-considerations--d9dd4c105700.rst

The following changes were made.

  • All alphabetical characters were converted to lower case.

  • Not shown The characters +, -, @, and & are replaced with plus, minus, at, and and respectively.

  • All spaces and other special characters, such as the ! were replaced by dashes.

  • A final dash and 12 digit random string were appended to the file name.

  • The extension .rst was added for all options except f, in which case the extension .rest was added.

  • If you chose to create an index file by selecting i when prompted, index- was prepended to the file name.

Examining the file reveals that the label matches the file name, while the title is preserved as typed. No label was added if you selected f.

cat architectural-considerations--d9dd4c105700.rst

.. _architectural-considerations--d9dd4c105700:

=============================
Architectural Considerations!
=============================

.. content here

When you reference this file in toctree and ref directives, use the file name/label string like this: architectural-considerations--d9dd4c105700

Automated quality checks

Several automated checks are available to help improve and maintain the quality of your documentation.

Some of these checks are run every time you perform a build and are intended to catch errors before they are submitted for review. Others are invoked independently of regular builds and are intended to identify problems prior to a release.

Formatting checks

You can build the HTML documentation locally using the tox -e docs command. After every successful build, several quality checks are performed against the build HTML output.

Checking for "grey bar" formatting errors in output ...
Found 2 HTML file(s) with greybar formatting issues:
./dist_cloud/kubernetes/reinstalling-a-subcloud-with-redfish-platform-management-service.html
./dist_cloud/kubernetes/installing-a-subcloud-without-redfish-platform-management-service.html
Using a browser, locate vertical grey bars in the left margin of the above file(s), then correct the issue(s) in the corresponding rST file(s).
Checking for ".. include::" errors in output ...
Checking for unexpanded substitution errors in output ...
Found 1 HTML file(s) that may have unexpanded substitution(s):

./node_management/kubernetes/hardware_acceleration_devices/enabling-mount-bryce-hw-accelerator-for-hosted-vram-containerized-workloads.html:| 1d02      | |SATA| controller               | Intel Corporation   |

Correct the issue(s) in the corresponding rST file(s).

This sample shows three problems.

Post-check issues and remedies

Test

Explanation

Remedy

Grey bars

Scans the output for evidence of rST vertical grey bars inserted into the output next to formatting errors and reports which files they were found in.​

  1. Open the file doc/build/html/index.html in a browser and navigate to the page reported in the output.

  2. Locate the grey bars.

    Tip

    Grey bars can be hard to find in some locations, such as notes, where they are obscured by a background fill. Look for other evidence of a problem such as an oversized font, text that appears to be randomly bolded, or senseless line breaks.

  3. Open the corresponding .rst file and find the location matching the grey bars in the output.

  4. Correct the issue.

Hint

Grey bars are often caused by indentation errors.

Include errors

Scans the output for malformed .. include:: statements that result in rST code and unintended content being exposed and reports which files they were found in.​

As above, find the problem in the appropriate .rst file by examining the .html file reported. Look for code fragments associated with .. include:: directives such as :start-after: and :end-before: that have been exposed in the final output.

Correct the issues by making the code comply with the documentation at:

https://docutils.sourceforge.io/docs/ref/rst/directives.html#include

Substitution errors

Scans the output for potential unexpanded substitutions such as |prod| and reports which files they were found in, along with the offending lines of HTML.

Note

This check cannot distinguish between a substitution and an ascii output table where cells are not properly padded. In either case, the problem needs to be fixed.

As above, find the problem in the appropriate .rst file by examining the .html file reported. Look for |<text>| code exposed in the output. In the corresponding .rst, find and correct the issue.

Hint

Substitutions are not allowed in code blocks, :ref:, :doc:, or within rST markup such as **, *`, `, and so on.

Substitutions cannot be used in ASCII “picture” style tables. If you need a substitution in a table, use the .. list-table:: format instead.

Spell checks

Spell checks are not performed as part of regular documentation builds. They are intended to be run periodically and prior to a release.

You can invoke the Sphinx link checker with the following command:

$ tox -e spellcheck

Sphinx will perform a temporary build and then check the output against a US English dictionary. Results are reported on the console and logged for future use.

Console output

Console output shows the path and name of the file an error was found in, the line number, the misspelled term and the full line to provide context.

doc/source/storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst:41: Spell check: aditional: used as aditional disk volumes for VMs booted from images.

Logs

Spell check logs are stored under doc/build/spelling in *.spelling files located and named for their rst counterparts.

For example, errors found in the file:

doc/source/storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst

are logged in the file:

doc/build/spelling/storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.spelling

Log files itemize one issue per line. For example:

storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst:41: (aditional)  used as aditional disk volumes for VMs booted from images
storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst:68: (num)  For more information on how placement group numbers, (pg_num) can be set
storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst:72: (num)  group numbers (pg_num) required based on pg_calc algorithm, estimates on
storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst:116: (num)  To list all the pools with their pg_num values, use the following command,
storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst:119: (num)  To get only the pg_num / pgp_num value, use the following command,
storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst:119: (num)  To get only the pg_num / pgp_num value, use the following command,
storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst:142: (num)  Increasing pg_num of a pool has to be done in increments of 64/
storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst:142: (num)  pg_num number, retry and wait for the cluster to be
storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst:149: (num)  pg_num of that pool, using the following commands:
storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst:162: (num)  pgp_num should be equal to pg_num.
storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst:162: (num)  pgp_num should be equal to pg_num.
storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst:203: (num)  pg_num, pgp_num, crush_rule.
storage/openstack/config-and-management-ceph-placement-group-number-dimensioning-for-storage-cluster.rst:203: (num)  pg_num, pgp_num, crush_rule

Note that the spell check in this example matched on the substring num several times in contexts such as pgp_num. Cases such as this may call for additional spell check customization.

Adding words

StarlingX documentation makes use of many technical terms that are not known to the default dictionary.

You can add these to the file doc/source/spelling_wordlist.txt.

This file contains one term per line.

Note

  • Care should be taken when adding terms to a custom dictionary to avoid errors not being reported. For example, “fs” may be correct in a code block but a typo in some other context. As a general rule, it is better to have the spell checker over-report than under-report.

  • It is important that spelling_wordlist.txt be kept in alphabetical order.

  • spelling_wordlist.txt is under git management and changes must be submitted for review and merge via a gerrit review.

RST conventions

StarlingX documentation follows many (but not all!) of the RST conventions described in the OpenStack documentation RST conventions guide. If RST markup is not listed in this section’s quick reference, refer to the OpenStack guide.

For detailed information about RST and Sphinx extensions, refer to the following documents:

Content reuse

Content can be reused in multiple places using the Sphinx .. include:: directive.

For example:

.. include:: /_includes/my_reusable_content.rest

“Include” files must not end in .rst. The StarlingX convention is to use .rest.

You can store multiple fragements of content in a single include file and use them in various places. To do this, use rst comments to deliniate where each begins and ends:

.. begin-fragement-1

   This content will be inserted using ``.. include::`` example 1, below.

.. end-fragment-1

.. begin-fragement-2

   This content will be inserted using ``.. include::`` example 2, below.

.. end-fragment-2

To use one or the other in an rst topic, use the start after and end before parameters as shown below:

Example 1

========
My Topic
========

Integer sed tortor nisi. Vivamus feugiat, urna in posuere gravida, ligula
nunc hendrerit magna, nec tristique ex tortor non lorem.

...

.. include:: /_includes/my_reusable_content.rest
   :start-after: begin-fragement-1
   :end-before: end-fragement-1

Example 2

==============
My Other Topic
==============

Lorem ipsum dolor sit amet, consectetur adipiscing elit.

...

.. include:: /_includes/my_reusable_content.rest
   :start-after: begin-fragement-2
   :end-before: end-fragement-2

Content reuse in steps

Inserting reusable or conditional steps in a numbered list context causes the list to restart. For example:

An include file, mysteps.rest, containing:

.. begin-step

#. Blue

.. end-step

used in an rst file as follows:

#. Orange

#. Red

.. include:: mysteps.rest
   :start-after: begin-step
   :end-before: end-step

#. Green

Results in:

  1. Orange

  2. Red

  1. Blue

  2. Green

To avoid this, use substitute .. pre-include:: for .. include::

#. Orange

#. Red

.. pre-include:: mysteps.rest
   :start-after: begin-step
   :end-before: end-step

#. Green

This will result in the expected numbering sequence:

  1. Orange

  2. Red

  3. Blue

  4. Green

Note

  • Only the start-after and end-before paramters work with .. pre-include::.

  • Indentation within the rest file being referenced must match the calling context.

    • If the list is not indented, the additional step(s) to be inserted should not be indented.

    • If the list is indented (a sublist or in a nested block context), the same indentation must be applied to the additional steps.

Content conditionalization

Conditionalize across builds

If you need to conditionalize some content to be used in a specific build context, such as StarlingX or a 3rd party build that reuses StarlingX content, you can use the .. only:: directive.

For example:

.. only:: starlingx

   Integer sed tortor nisi. Vivamus feugiat, urna in posuere gravida, ligula
   nunc hendrerit magna, nec tristique ex tortor non lorem.

Three build contexts are available:

starlingx

Content will be included in the standard StarlingX documentation build.

partner

Content will be excluded from the StarlingX documentation build. To reuse this content in a 3rd party build, ensure that the partner tag is based to the builder.

openstack

Content will be included in an OpenStack documentation context. This content is included in the StarlingX documentation build.

Conditionalize across pages

If you want to reuse a block of content in multiple rst files from the same build using .. include:: directives, but need to exclude specific strings from one of those locations, you can use the hideable role and substitution.

You can hide both strings and blocks (paragraphs etc.).

Hiding strings

To hide a string, use the hideable role. For example:

.. start-prepare-servers-common

Prior to starting the |prod| installation, ensure that the |bare-metal|
servers are in the following state:

...

-   BIOS configured with Intel Virtualization (VTD, VTX)

    -  Disabled for controller-only servers and storage servers.

    -  Enabled for :hideable:`controller+worker (All-in-one) servers and` worker servers.

-   The servers are powered off.

.. end-prepare-servers-common

In the rst file where you want to include the text marked up with the :hideable: role, simply insert the content using the include directive:

.. _aio_duplex_install_kubernetes_r7:

================================================
Install Kubernetes Platform on All-in-one Duplex
================================================

...

--------------------------------
Prepare Servers for Installation
--------------------------------

.. include:: /shared/_includes/prepare-servers-for-installation-91baad307173.rest
   :start-after: start-prepare-servers-common
   :end-before: end-prepare-servers-common

...

The bullet:

-  Enabled for :hideable:`controller+worker (All-in-one) servers and` worker servers.

will render as:

  • Enabled for controller+worker (All-in-one) servers and worker servers.

In the rst file where you want to suppress the text marked up with the :hideable: role, do the same, but add the |hideable| substitution at the top of the file:

|hideable|

.. _aio_duplex_install_kubernetes_r7:

================================================
Install Kubernetes Platform on All-in-one Duplex
================================================

...

--------------------------------
Prepare Servers for Installation
--------------------------------

.. include:: /shared/_includes/prepare-servers-for-installation-91baad307173.rest
   :start-after: start-prepare-servers-common
   :end-before: end-prepare-servers-common

...

The bullet:

-  Enabled for :hideable:`controller+worker (All-in-one) servers and` worker servers.

will render as:

  • Enabled for worker servers.

Hiding blocks

To hide a block, wrap it in a container directive with the argument hideable

For example, create an include file install-status.rest with the following contents:

The **deploy status** field has the following values:

.. container:: hideable

   ``Pre-Install``
       This status indicates that the ISO for the subcloud is being updated by
       the Central Cloud with the boot menu parameters, and kickstart
       configuration as specified in the ``install-values.yaml`` file.

   ``Installing``
       This status indicates that the subcloud's ISO is being installed from
       the Central Cloud to the subcloud using the Redfish Virtual Media
       service on the subcloud's |BMC|.

.. container::

   ``Bootstrapping``
       This status indicates that the Ansible bootstrap of |prod-long|
       software on the subcloud's controller-0 is in progress.

and load it in two different contexts:

a.rst
.. include:: install-status.rest
b.rst
|hideable|

.. include:: install-status.rest

The output from a.rst will include all three definitions. The output from b.rst will include only the bootstrap definition.

RST quick reference

Acronyms

Define acronym at first instance on page. After definition, use acronym only.

Input:

:abbr:`CPU (Central Processing Unit)`

Output:

CPU

Code samples

Format code snippets as standalone literal blocks.

Input:

::

  ping 8.8.8.8

Output:

ping 8.8.8.8

Commands

Format commands using the Sphinx command role.

Input:

:command:`system help`

Output:

Use the system help command for the full list of options.

Cross-references

Cross-reference to arbitrary locations in a document using the ref role and a named target. Named targets must precede a section heading. For more information on references, see Internal Hyperlink Targets.

Input:

.. _my_named_target:

**********
My section
**********

This is the section we want to reference.

...

This is the reference to :ref:`my_named_target`.

Output:

My section

This is the section we want to reference.

…

This is the reference to My section.

Information blocks

Emphasize information using notices (an admonition in Sphinx). Different types of notices exist to emphasize degrees of information importance.

Input:

.. note::

   Use a ``note`` for a generic message.

.. seealso::

   Use ``seealso`` for extra but helpful information.

.. important::

   Use ``important`` for details that can be easily missed, but should not be
   ignored by a user and are valuable before proceeding.

.. warning::

   Use ``warning`` to call out information the user must understand
   to avoid negative consequences.

Output:

Note

Use a note for a generic message.

See also

Use seealso for extra but helpful information.

Important

Use important for details that can be easily missed, but should not be ignored by a user and are valuable before proceeding.

Warning

Use warning to call out information the user must understand to avoid negative consequences.

Inline elements

Format most inline elements such as filenames and paths, code fragments, parameters, or options with double back ticks.

Input:

``/path/to/file.name``
``--option``

Output:

Open the /path/to/file.name file.

Optionally pass the --option with the command.

Refer to the OpenStack Inline elements guide for markup for other inline elements.

Lists

Use a bulleted list for a sequence of items whose order does not matter, such as a list of features.

Input:

* Banana
* Apple
* Orange

Output:

  • Banana

  • Apple

  • Orange

Use an enumerated list for a sequence of items whose order matters, such as in an ordered sequence of installation steps.

Input:

#. Wash apple.
#. Peel apple.
#. Eat apple.

Output:

  1. Wash apple.

  2. Peel apple.

  3. Eat apple.

Use a definition list for an unordered list where each item has a short definition, such as term/definition pairs.

Input:

Command A
  Description of command A.

Command B
  Description of command B.

Output:

Command A

Description of command A.

Command B

Description of command B.

Section headings

Use up to three levels of headings in one file using the following characters:

  • Heading 1 (Page Title in Title Case) - underline and overline with equal signs;

    • Heading 2 (Major page sections in Sentence case) - underline and overline with dashes;

      • Heading 3 (subsections in Sentence case) - underline and overline with asterisks.

Example RST:

==============
Document Title
==============

Introduce the topic using 1-2 concise sentences. It should tell the user what
info can be found on this page.

.. contents::  // Use a local TOC to aid user navigation in the page
   :local:
   :depth: 1

---------------
Section heading
---------------

Lorem ipsum dolor sit amet, consectetur adipiscing elit.

******************
Subsection heading
******************

Integer sed tortor nisi. Vivamus feugiat, urna in posuere gravida, ligula
nunc hendrerit magna, nec tristique ex tortor non lorem.

Release activities

Several activies must be completed in preparation for each release.

Release notes updates

Various sections in the StarlingX release notes must be updated on the master branch before the release branch for the upcoming release (for example r/stx.9.0) is created.

The Release notes can be found in the directory doc/source/releasenotes.

Version number updates

Product and related software version numbers must be updated throughout the source rst files.

  • Some of these changes must me made manually by searching for and replacing the appropriate values.

  • More common version values are implemented as Sphinx substitutions: simple, reusable placeholders.

    • For the current version of StarlingX, edit the value of this-ver in the file doc/source/_this.txt.

      .. |this-ver| replace:: r9.0

    • For other values, consult the file doc/source/_vendor/vendor_strings.txt.

Version menu updates

A menu is available on each page of StarlingX documentation to allow easy switching between documentation versions. This menu must be updated on master and each release branch.

To do so:

  1. On a working branch, open the file doc/source/conf.py.

  2. If you are updating on a new branch, change the configuration assignment starlingxdocs_plus_this_version to the name of the new branch.

    For example:

    starlingxdocs_plus_this_version="Version 9.0"

    Otherwise, confirm that the value is correct for the branch you are updating.

  3. Locate the configuration assignment starlingxdocs_plus_other_versions.

    This assignment takes a key,value pair for each branch other than the branch you are configuring.

    • Each key is the name of the branch that will appear in the menu.

    • Each value is the name of the branch.

    For example, if you are editing conf.py on the branch r/stx.7.0, then this assignment would container a key pair for each of

    r/stx.6.0, r/stx.8.0, <newer-release> …, and master.

    These pairs are separated by commas and must be listed from oldest to most recent, always ending with the master branch. The key for the master branch is Latest.

    For example, for release 9.0, the r/stx.9.0 version of this file would have this assignment:

    starlingxdocs_plus_other_versions = [("Version 6.0","r/stx.6.0"),("Version 7.0","r/stx.7.0"),("Version 8.0","r/stx.8.0"),("Latest","")]

    Note

    During a documentation build, Sphinx prepends a forward slash to each value in starlingxdocs_plus_other_versions to construct the correct URLs, such as https://docs.starlingx.io/r/stx.8.0/. Since the master documentation is at the base URL docs.startlingx.io, this means that the value must be an empty string, as show above.

Adding the promte job

A promote job must be created for merged reviews on a new release branch to be published to docs.starlingx.io.

Prerequisites

You must have the OpenStack project-config repo available.

Procedure

  1. In the project-config repo, open the file zuul.d/starlingx-jobs.yaml.

  2. Locate the promote-stx-tox-docs-site job.

  3. Add the new branch to the list of branches that will be published.

    For example:

    - job:
    name: promote-stx-tox-docs-site
    parent: opendev-promote-docs-base
    description: |
      Promote content from openstack-tox-docs job for
      starlingx/doc repository only.

      Publish the results of a sphinx build to
      /afs/.openstack.org/project/starlingx.io/www

    final: true
    vars:
      download_artifact_job: openstack-tox-docs
    # add older release branches as desired
    branches:
      - master
      - r/stx.5.0
      - r/stx.6.0
      - r/stx.7.0
      - r/stx.8.0
      - r/stx.9.0

  4. Save the file and commit a gerrit review.

See https://review.opendev.org/c/openstack/project-config/+/873266 for a sample promote job update with reviewers.