Archive for October, 2015

Lwood-20151025

Introduction

Welcome to Last week on OpenStack Dev (“Lwood”) for 25th October 2015. For more background on Lwood, please refer here.

A bit of a shorter Lwood this week as the Mitaka Summit in Tokyo seems to have reduced the amount of Lwood-relevant traffic.

Basic Stats for week 19th to 25th October 2015 :

  • ~499 Messages (down about 19% from last week)
  • ~186 Threads (up about 2% from last week)

Notable Discussions

User Survey Results are out

Tom Fifield writes that the results of the most recent OpenStack user survey are now available here and it is, truly, compelling reading.  No matter what your involvement with OpenStack this is worthy of your time to go through and read.  Your humble correspondent will be taking a second pass at it right after I click send on this Lwood :)

New Stackalytics goodness

Ilya Shakhat notes that Stackalytics 0.9 is now up and running, a mixture of some neat new functionality as well as compatibility with upcoming infra changes (such as working with Gerrit 2.9+) Head over and check it out :)

Release Freeze until after Summit

Doug Hellman made note of the fact that barring anything unforeseen, there will be a freeze on new releases until after tne Summit – November 2nd.

OpenStack-HA IRC Channel

After some discussion with the -infra team, Adam Spiers points out there is now an IRC channel specifically for High Availability related discussions.

A new project – Watcher

Antoine Cabot announced the Watcher project on Wednesday last week – a project that seemingly makes use of data streamed from Ceilometer (and so isn’t, as I understand it querying the database) and allows this information to be used to trigger events with the aim of optimising resource usage.  Seems pretty neat and follows a presentation made at the Vancouver summit.

Release Notifications

A new section for Lwood, will see how useful/practical it proves to be – a wrapup of the weeks release/version bump notifications…

Comments

Yamaha Pedal Wiring

I did a little rewiring of my keyboard rig on the weekend, in particular the pedal board I use to keep things organised and quick to set up.  In the process of this I loomed three of the foot switches up together which necessitated re-terminating them with new plugs having cut the old ones off.

For future reference then, here is the wiring schema/colour codes for my particular pedals.  If Google picks this page up, might save someone else a little hassle :)

Yamaha FC4 Footswitch

The FC4 is a conventional switch, Normally closed, contacts open when the pedal is pressed down

  • Tip – N/C Contact – Inner conductor in cable
  • Sleeve – N/C Contact – Shield in cable

Yamaha FC5 Footswitch

Electrically the same as the FC5, just a different physical form factor.

  • Tip – N/C Contact – White conductor in cable
  • Sleeve – N/C Contact – Black conductor in cable

Yamaha FC7 Footpedal

So with the FC7 footpedal, when it’s pressed all the way down, the minimum resistance (100 ohms or so) is between the ring and tip, and maximum resistance (50k ohm approx) between ring and sleeve.  When the pedal is all the way up this reverses – the minimum resistance is between the ring and sleeve and maximum resistance between the ring and tip.

  • Tip – “Down” end of 50k ohm pot – White conductor in cable
  • Ring – Wiper of 50k ohm pot – Red conductor in cable
  • Sleeve – “Up” end of 50k ohm pot – Shield of cable

 

Comments

Lwood-20151018

Introduction

Welcome to Last week on OpenStack Dev (“Lwood”) for 18th October 2015. For more background on Lwood, please refer here.

Basic Stats for week 12th to 18th October 2015 :

  • ~619 Messages (up about 4% from last week)
  • ~182 Threads (up about 20% from last week)

Notable Discussions

Glance related security notice (OSSN 0057)

This post outlines the contents of OSSN 0057 – a DoS attack on Glance that can cause database resource depletion by continually requesting the creation of zero sized images.

Release communications for Mitaka

Doug Hellman sent an email to the PTLs and openstack-dev outlining plans to ensure more methodical and consistent release management for the Mitaka cycle.  In particular he implores PTLs, project release liaison people and other relevant folk to ensure they have email handling set up such that anything with the [release] topic gets flagged. Further details on the process will follow.

Tweaks to release request process

A later post from Doug outlines the first of these changes – some tweaks around what is expected for library versioning and constraint updates and a reminder that these new policies will be enforced starting this week.  More here.

Establishing Release and Documentation Liaisons for Mitaka

Rounding out our trifecta of Mitaka release related discussions, a few more words from Doug Hellman asking projects to find people to act as Release Liaisons as described here.

Lana Brindley chimed in with a timely reminder to the projects to consider who can be their Documentation Liaison as well.

Etherpads for Mitaka Design Summit

Thierry Carrez noted that there is now a page on the Wiki to collect the etherpads for the Summit.  Please check yours is listed if appropriate!

A thread on the Scheduler continues…

Flagged in last week’s Lwood at the better part of 64 messages for that week, the thread Ed Leafe kicked off is still going strong with about another 45 posts.  Ed provides some interesting comments here that somewhat sums up the conversation thus far.

The thread has been covering all sorts of things; Cassandra, numpy, scalability, how many nodes are a lot etc. etc. so it has moved beyond the original premise which was suggesting a way to allow experimentation with the Nova scheduler.  Still worth a read :)

OpenStack Infra moving to Gerrit 2.11

Zaro posted on behalf of the Infra team to announce plans to upgrade to Gerrit 2.11 (currently 2.8) shortly after the Mitaka summit.  Judging by the thread that followed, a very popular plan!  More details in the original post.

Magnum PTL announced

Somehow I missed the announcement in the week before last that the Magnum PTL election was concluded with Adrian Otto re-elected to the role.  My apologies for the oversight and congrats to Adrian!

More on the informal performance/large deployments group

As mentioned back in Lwood-20150927, Dina Belova kicked off a discussion on setting up an informal team to work specifically on OpenStack performance issues in large deployments.  It’s been confirmed that a session has been set aside for this group at the Tokyo Summit.

Comments

Lwood-20151011

Introduction

Welcome to Last week on OpenStack Dev (“Lwood”) for 11th October 2015. For more background on Lwood, please refer here.

Basic Stats for week 5th to 11th October 2015 :

  • ~596 Messages (up about 4% from last week)
  • ~174 Threads (down about 12% from last week)

Notable Discussions

TC Elections

Voting in the TC Elections wrapped up during the week past and the results were announced shortly thereafter.

The successful candidates were Anne Gentle (Rackspace), Doug Hellman (HP), Monty Taylor (IBM, OpenStack Foundation), Sean Dague (also of HP), Russell Bryant (Red Hat) and Kyle Mestery (also HP, I think).

A reminder you can read a little more about the candidates on this OpenStack Wiki page.

A little Launchpad housekeeping

The recent shifts on/off daylight savings time in many parts of the world make this a timely (sorry) request to please consider updating your timezone information in Launchpad to make scheduling meetings a bit easier.

Reviews requested for openstack-ansible-security

Major Hayden put out a request for interested and able folk to review the work thus far.  He points out that many are fairly easy to do – a small task and some doco and so an excellent place to start learning to do reviews.

Two new API Guidelines ready for Cross Project Review

…and they’ll be merged come October 16.  From the original post the titles are “Adds an API documentation guideline document” and “Add http400 for nonexistent resource”

Recording little everyday OpenStack successes

In this post Thierry Carrez announces a nice little initative to capture little moments of joy and success – the things seemingly not worth a blog post or mailing list thread, but cool none the less.  The post and subsequent thread has a bit more background but in essence the idea is to use the tag #success in the OpenStack channel of your choice and a bot will take it and record it on this Wiki page.  Neato!

A thread on the Scheduler

At the better part of 64 messages and counting, the thread Ed Leafe kicked off is one of the longer on the list for quite some time.  An interesting read, it follows up a post he made back in July (and mentioned in Lwood 20150719) suggesting a way to allow experimentation with the Nova scheduler by doing the work out of tree temporarily.

If Nova or general compute architectural discussions are your area of interest, a good one to read.  Also a nice example of (predominantly) postive collaboration by a wide range of OpenStack developers :)

Mitaka Design Summit Schedule up

Thierry concludes a thread from two weeks back in this post noting that the Mitaka Design Summit schedule is now up on sched.org

Keeping track of Ceilometer extensions

Gord Chung noted that there is now a specific section in the Ceilometer Wiki for tracking extensions to Ceilometer (such as plugins to support telemetry on different systems for example) that aren’t part of the main project itself.

Some interesting “orphaned” code – Attribute Mapping GUI for Horizon

An unusual if not unprecedented circumstance – University student creates a Useful Thing™ then gets a job that prevents them from finishing it…  In this case some GUI code for creating/editing attributes for Keystone.  After a little discourse on list the code was put up on review.openstack.org as a work in progress and review sought, or even better, someone to carry on from where the good seeming work has left off.

Refactoring the Glance image import process

Just a single post that refers back to a conversation started in September on priorities.  The resultant spec is here and the call has been made for feedback, particularly from folk that use rather than contribute to glance.

Recurring bug squash day for keystone

A well received call to have a weekly bug squash day for Keystone.  If you have the skills, please pitch in :)

PTL Election Analysis

As reported previously, the PTL elections took place at the end of September.  I did a little analysis of the employers of the various PTLs – this based on their company attribution in Stackalytics.  Note at the time of writing the Magnum PTL election is still underway.

In order of number of PTLs we have HP (9), Mirantis (7), Rackspace (6), IBM (4), Redhat (3) and Intel (2).  The remaining 11 PTLs are all from different companies.  

I didn’t have ready access to historical data, but will be able to prepare a perspective after the next PTL election.  Below you’ll find the results tabulated and, in a first for Lwood, a graph!

OpenStackPTLsOct2015

Project Name Employer
Barbican Douglas Mendizabal Rackspace
Ceilometer Gordon Chung Huawei
ChefOpenstack Jan Klare X-ion
Cinder Sean Mcginnis Dell
Community App Catalog Christopher Aedo IBM
Congress Tim Hinrichs Styra
Cue Vipul Sabhaya HP
Designate Graham Hayes HP
Documentation Lana Brindley Rackspace
Glance Flavio Percoco Red Hat
Heat Sergey Kraynev Mirantis
Horizon David Lyle Intel
I18n Ying Chun Guo IBM
Infrastructure Jeremy Stanley OpenStack Foundation
Ironic Jim Rollenhagen Rackspace
Keystone Steve Martinelli IBM
Kolla Steven Dake Cisco
Magnum PTL Election in progress TBC
Manila Ben Swartzlander NetApp
Mistral Renat Akhmerov Mirantis
Murano Serg Melikyan Mirantis
Neutron Armando Migliaccio HP
Nova John Garbutt Rackspace
OpenStack UX Piet Kruithof HP
OpenStackAnsible Jesse Pretorius Rackspace
OpenStackClient Dean Troyer Intel
Oslo Davanum Srinivas Mirantis
Packaging-deb Thomas Goirand Mirantis
PuppetOpenStack Emilien Macchi Red Hat
Quality Assurance Matthew Treinish HP
Rally Boris Pavlovic Mirantis
RefStack Catherine Diep IBM
Release cycle management Doug Hellmann HP
RpmPackaging Dirk Mueller SUSE
Sahara Sergey Lukjanov Mirantis
Searchlight Travis Tripp HP
Security Robert Clark HP
Solum Devdatta Kulkarni Rackspace
Swift John Dickinson SwiftStack
TripleO Dan Prince Red Hat
Trove Craig Vyvial HP
Zaqar Fei Long Wang Catalyst IT

 

Comments

Lwood-20151004

Introduction

Welcome to Last week on OpenStack Dev (“Lwood”) for 4th October 2015. For more background on Lwood, please refer here.

Basic Stats for week 28th September to 4th October 2015 :

  • ~573 Messages (up about 2% from last week)
  • ~174 Threads (up about 13% from last week)

Notable Discussions

TC Elections

The TC Elections for six positions are underway, closing date for voting for those eligible is 9th October 2015 23:59:59 – UTC time and date.  If you’re able to vote you should already have received an email with a voting link from Tristan Cacqueray (on/around 3rd October if mine is anything to go by). You can read Tristan’s announcement post here and a some thought provoking observations about being an ATC from Flavio Percoco.

A quick scan suggests that all 19 candidates have posted supporting information to the dev mailing list or you can refer to the OpenStack Wiki page that has all the relevant information.

Discussion about new testing for OpenStack Compatible mark

This post was flagged in last week’s Lwood to little fanfare.  A bit of discussion started up about it in the week past including a post about related discussions among the Neutron team.

To recap the overall thread discusses new requirements that come into force in November whereby the OpenStack Foundation will require vendors requesting “OpenStack Compatible” storage driver licenses to start passing the Cinder third-party integration tests.  This all part of a broader initiative to better use the OpenStack brand to enforce better interoperability.

Asserting that a project follows the base deprecation policy

A timely reminder that following the most recent TC meeting there is now a definition for a standard base deprecation policy for OpenStack projects. PTLs are requested to review the definition and if appropriate propose a change adding that tag to their project.  Full details in the OP including a reminder that it’s ok, particularly for “younger” projects to not make such an attestation.

Proposed Mitaka release schedule

From Tierry Carrez a post noting that the proposed schedule for Mitaka is now available here.  By all accounts no surprises – similar to previous releases and even an acknowledgement of some time off over the Christmas/New Year period.

Infra needs Gerrit developers

A call out to anyone able to help out the Infra team with work on Gerrit. As it is rather wryly noted, Gerrit forms a key part of the CI infrastructure for OpenStack so it’s in everyone’s best interests to have Gerrit be as awesome and reliable as possible.  Please consider!

Tokyo Summit Track/Room/Time Allocations

For anyone attending the Tokyo Summit, a further post on track, room and time allocations. Feedback requested if there are any glaring issues.

A request for feedback from the OpenStack UX team

Most PTLs would already be across this but it’s none the less encouraging to see the UX team reaching out requesting feedback on personas to facilitate creating user stories and the like.  Original post is here.

Comments