So You Want to Contribute…

For general information on contributing to OpenStack, please check out the contributor guide to get started. It covers all the basics that are common to all OpenStack projects: the accounts you need, the basics of interacting with our Gerrit review system, how we communicate as a community, etc.

Below will cover the more project specific information you need to get started with the Cinder project, which is responsible for the following OpenStack deliverables:

cinder
os-brick
Shared library for managing local volume attaches.
python-cinderclient
Python client library for the OpenStack Block Storage API; includes a CLI shell.
python-brick-cinderclient-ext
Extends the python-cinderclient library so that it can handle local volume attaches.
Launchpad: (doesn’t have its own space, uses python-cinderclient’s)
cinderlib
Library that allows direct usage of Cinder backend drivers without cinder services.
rbd-iscsi-client
Library that provides a REST client that talks to ceph-isci’s rbd-target-api to export rbd images/volumes to an iSCSI initiator.
cinder-tempest-plugin
Contains additional Cinder tempest-based tests beyond those in the main OpenStack Integration Test Suite (tempest).

See the CONTRIBUTING.rst file in each code repository for more information about contributing to that specific deliverable. Additionally, you should look over the docs links above; most components have helpful developer information specific to that deliverable. (The main cinder documentation is especially thorough in this regard and you should read through it, particularly Background Concepts for Cinder and Programming HowTos and Tutorials.)

Communication

IRC

We use IRC a lot. You will, too. You can find infomation about what IRC network OpenStack uses for communication (and tips for using IRC) in the Setup IRC section of the main OpenStack Contributor Guide.

People working on the Cinder project may be found in the #openstack-cinder IRC channel during working hours in their timezone. The channel is logged, so if you ask a question when no one is around, you can check the log to see if it’s been answered: http://eavesdrop.openstack.org/irclogs/%23openstack-cinder/

weekly meeting

Wednesdays at 14:00 UTC in the #openstack-meeting-alt IRC channel. Meetings are logged: http://eavesdrop.openstack.org/meetings/cinder/

More information (including some pointers on meeting etiquette and an ICS file to put the meeting on your calendar) can be found at: http://eavesdrop.openstack.org/#Cinder_Team_Meeting

The meeting agenda for a particular development cycle is kept on an etherpad. You can find a link to the current agenda from the Cinder Meetings wiki page: https://wiki.openstack.org/wiki/CinderMeetings

The last meeting of each month is held simultaneously in videoconference and IRC. Connection information is posted on the meeting agenda.

weekly bug squad meeting

This is a half-hour meeting on Wednesdays at 15:00 UTC (right after the Cinder weekly meeting) in the #openstack-cinder IRC channel. At this meeting, led by the Cinder Bug Deputy, we discuss new bugs that have been filed against Cinder project deliverables (and, if there’s time, discuss the relevance of old bugs that haven’t seen any action recently). Info about the meeting is here: http://eavesdrop.openstack.org/#Cinder_Bug_Squad_Meeting

mailing list

We use the openstack-discuss@lists.openstack.org mailing list for asynchronous discussions or to communicate with other OpenStack teams. Use the prefix [cinder] in your subject line (it’s a high-volume list, so most people use email filters).

More information about the mailing list, including how to subscribe and read the archives, can be found at: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-discuss

virtual meet-ups

From time to time, the Cinder project will have video meetings to address topics not easily covered by the above methods. These are announced well in advance at the weekly meeting and on the mailing list.

Additionally, the Cinder project has been holding two virtual mid-cycle meetings during each development cycle, roughly at weeks R-18 and R-9. These are used to discuss follow-up issues from the PTG before the spec freeze, and to assess the development status of features and priorities roughly one month before the feature freeze. The exact dates of these are announced at the weekly meeting and on the mailing list.

cinder festival of XS reviews

This is a standing video meeting held the third Friday of each month from 14:00-16:00 UTC in meetpad to review very small patches that haven’t yet been merged. It’s held in video so we can quickly discuss issues and hand reviews back and forth. It is not recorded. Info about the meeting is here: http://eavesdrop.openstack.org/#Cinder_Festival_of_XS_Reviews

physical meet-ups

The Cinder project usually has a presence at the OpenDev/OpenStack Project Team Gathering that takes place at the beginning of each development cycle. Planning happens on an etherpad whose URL is announced at the weekly meetings and on the mailing list.

Contacting the Core Team

The cinder-core team is an active group of contributors who are responsible for directing and maintaining the Cinder project. As a new contributor, your interaction with this group will be mostly through code reviews, because only members of cinder-core can approve a code change to be merged into the code repository.

You can learn more about the role of core reviewers in the OpenStack governance documentation: https://docs.openstack.org/contributors/common/governance.html#core-reviewer

The membership list of cinder-core is maintained in gerrit: https://review.opendev.org/#/admin/groups/83,members

You can also find the members of the cinder-core team at the Cinder weekly meetings.

Project Team Lead

For each development cycle, Cinder project Active Technical Contributors (ATCs) elect a Project Team Lead who is responsible for running the weekly meetings, midcycles, and Cinder sessions at the Project Team Gathering for that cycle (and who is also ultimately responsible for everything else the project does).

  • You automatically become an ATC by making a commit to one of the cinder deliverables. Other people who haven’t made a commit, but have contributed to the project in other ways (for example, making good bug reports) may be recognized as “extra-ATCs” and obtain voting privileges. If you are such a person, contact the current PTL before the “Extra-ATC freeze” indicated on the current development cycle schedule (which you can find from the OpenStack Releases homepage .

The current Cinder project Project Team Lead (PTL) is listed in the Cinder project reference maintained by the OpenStack Technical Committee.

All common PTL duties are enumerated in the PTL guide.

Additional responsibilities for the Cinder PTL can be found by reading through the Managing the Development Cycle section of the Cinder documentation.

New Feature Planning

The Cinder project uses both “specs” and “blueprints” to track new features. Here’s a quick rundown of what they are and how the Cinder project uses them.

specs
Exist in the cinder-specs repository. Each spec must have a Launchpad blueprint (see below) associated with it for tracking purposes.
A spec is required for any new Cinder core feature, anything that changes the Block Storage API, or anything that entails a mass change to existing drivers.
It contains a README.rst file explaining how to file a spec.
You can read rendered specs docs at:
blueprints
Exist in Launchpad, where they can be targeted to release milestones.
Examples of changes that can be covered by a blueprint only are:
  • adding a new volume, backup, or target driver; or

  • adding support for a defined capability that already exists in the base volume, backup, or target drivers

Feel free to ask in #openstack-cinder or at the weekly meeting if you have an idea you want to develop and you’re not sure whether it requires a blueprint and a spec or simply a blueprint.

The Cinder project observes the following deadlines. For the current development cycle, the dates of each (and a more detailed description) may be found on the release schedule, which you can find from: https://releases.openstack.org/

  • spec freeze (all specs must be approved by this date)

  • new driver merge deadline

  • new target driver merge deadline

  • new feature status checkpoint

  • driver features declaration

  • third-party CI compliance checkpoint

Additionally, the Cinder project observes the OpenStack-wide deadlines, for example, final release of non-client libraries (os-brick), final release for client libraries (python-cinderclient), feature freeze, etc. These are also noted and explained on the release schedule for the current development cycle.

Task Tracking

We track our tasks in Launchpad. See the top of the page for the URL of each Cinder project deliverable.

If you’re looking for some smaller, easier work item to pick up and get started on, search for the ‘low-hanging-fruit’ tag in the Bugs section.

When you start working on a bug, make sure you assign it to yourself. Otherwise someone else may also start working on it, and we don’t want to duplicate efforts. Also, if you find a bug in the code and want to post a fix, make sure you file a bug (and assign it to yourself!) just in case someone else comes across the problem in the meantime.

Reporting a Bug

You found an issue and want to make sure we are aware of it? You can do so in the Launchpad space for the affected deliverable:

Getting Your Patch Merged

Before your patch can be merged, it must be reviewed and approved.

The Cinder project policy is that a patch must have two +2s before it can be merged. (Exceptions are documentation changes, which require only a single +2, and specs, for which the PTL may require more than two +2s, depending on the complexity of the proposal.) Only members of the cinder-core team can vote +2 (or -2) on a patch, or approve it.

Note

Although your contribution will require reviews by members of cinder-core, these aren’t the only people whose reviews matter. Anyone with a gerrit account can post reviews, so you can ask other developers you know to review your code … and you can review theirs. (A good way to learn your way around the codebase is to review other people’s patches.)

If you’re thinking, “I’m new at this, how can I possibly provide a helpful review?”, take a look at How to Review Changes the OpenStack Way.

There are also some Cinder project specific reviewing guidelines in the Code Reviews section of the Cinder Contributor Guide.

Patches lacking unit tests are unlikely to be approved. Check out the Testing section of the Cinder Contributors Guide for a discussion of the kinds of testing we do with cinder.

In addition, some changes may require a release note. Any patch that changes functionality, adds functionality, or addresses a significant bug should have a release note. You can find more information about how to write a release note in the Release notes section of the Cinder Contributors Guide.

Keep in mind that the best way to make sure your patches are reviewed in a timely manner is to review other people’s patches. We’re engaged in a cooperative enterprise here.

If your patch has a -1 from Zuul, you should fix it right away, because people are unlikely to review a patch that is failing the CI system.

  • If it’s a pep8 issue, the job leaves sufficient information for you to fix the problems yourself.

  • If you are failing unit or functional tests, you should look at the failures carefully. These tests guard against regressions, so if your patch causing failures, you need to figure out exactly what is going on.

  • The unit, functional, and pep8 tests can all be run locally before you submit your patch for review. By doing so, you can help conserve gate resources.

  • Other test failures: we also run integration tests in the gate that run your changes in the context of an OpenStack deployment, where cinder and os-brick interact with users, admins, and other services. Sometimes these tests will fail, and it may not obviously be your patch’s fault. Keep in mind, however, that the failure could still be a cinder issue, for which the cinder project (which includes you, as a contributor) is responsible. So please take a few minutes to look over the logs from the failing test job to see if you can identify the issue.

    • If you’re not sure how to do this, ask in the #openstack-cinder channel (or during open discussion at the weekly cinder meeting), and someone will walk you through the basic process.

    • You can tell Zuul to do a recheck, but first:

      • Make sure you look at the job’s build history, because if the job is failing consistently, it’s probably due to some particular issue that must be fixed before the job will start passing again. So a recheck in this situation will just waste resources. Check the mailing list or ask in IRC or look at the comments on your patch (sometimes a reviewer will leave a note saying not to recheck until some other patch has merged).

      • When you think a recheck is appropriate, make sure you follow the OpenStack community guidelines for How to Handle Test Failures.

How long it may take for your review to get attention will depend on the current project priorities. For example, the feature freeze is at the third milestone of each development cycle, so feature patches have the highest priority just before M-3. Likewise, once the new driver freeze is in effect, new driver patches are unlikely to receive timely reviews until after the stable branch has been cut (this happens three weeks before release). Similarly, os-brick patches have review priority before the nonclient library release deadline, and cinderclient patches have priority before the client library release each cycle. These dates are clearly noted on the release schedule for the current release, which you can find from https://releases.openstack.org/

You can see who’s been doing what with Cinder recently in Stackalytics: https://www.stackalytics.io/report/activity?module=cinder-group