I attended my first PTG (Project Teams Gathering https://www.openstack.org/ptg) this year in Dublin Ireland from February 24 to March 2.  I’ve been working on OpenStack code (and code that uses it or packages and installs it) for a few years now, and recently was made a Monasca Core Reviewer. I attended the OpenStack Summit in Vancouver, which was quite a different experience (should blog about that some time – I have lots of notes I could dig up).

The PTG is different from the Summits.  This was only the third one that has been held, and the first I’d attended.  I had a bit of a sense of what to expect from the meetings after attending some of the developer sessions at the Vancouver Summit, and after the virtual mid-cycle meeting we had for Monasca last year.  But the format was still a bit challenging.

Read below for my take on the first two days, and look for other posts about the rest of the week and weather/travel.

The first two days

The first two days of the PTG are intended to be more open to cross-project collaborations and to give attendees a chance to stretch out and make connections.  That is a great idea, but in practice I found it challenging.  My focus and reason for attending was monitoring (https://wiki.openstack.org/wiki/Monasca), but I’ve touched a few other OpenStack projects along the way.  But the sessions scheduled for Monday didn’t have anything that jumped out at me as being something I needed to be part of.  I know a bit about Keystone, but we already had excellent representation there from my team. The best fit I found on the schedule was the Security SIG (Security recently went from being a project to a SIG), and it was great for the first hour. [1]  Then the discussion among the small team that was there died out (several things were planned for later when interested engineers would be available) and the meeting broke up so the team members could go to other sessions they were more interested in.  Which left me trying to find another session.

Looking back at the Monday schedule, I realized I didn’t know much of anything about Watcher or Vitrage.  A little googling told me they might have some relation to monitoring.  One of my coworkers was interested in Vitrage, so I wandered to the Watcher room.  I slipped in the back and listened, but the discussion was all around some deep specific use cases and without any context it was difficult to gain anything useful out of it.  After another hour and a half, they broke the team meeting and I was again left without a plan.

Jet lag was setting in, so the rest of the morning wasn’t very productive.  After lunch I sat in on the Edge discussion, which was a much larger group than Security or Watcher (25+ people vs 5ish).  But again, without context it was difficult to gain anything.

Tuesday wasn’t much better.  I spent the morning doing some work, then the afternoon in the Release Cycle discussion.  I had intended to go to the Vitrage session (some of my coworkers had arranged some time to try to coordinate with them), but jet lag and insomnia meant I slept in until 10. Release Cycle was again a large session with a lot of interested parties, but the discussion was not coming to any sort of conclusion. The later conclusion to keep the current schedule but work on longer maintenance and “skip level” or “fast forward” upgrades is probably a good compromise.

Overall, I think what was missing from the first two days were some general interest sessions and some advertising.  There needed to be some descriptions about what sessions were available and why you, a developer on X, would be interested in attending a session about Y.  And there needed to be some catch-all sessions for developers who aren’t interested in the inner workings of some companies pet project.  As it was, I could have skipped Monday and possibly Tuesday with no big loss (and saved a few hundred dollars of travel costs).

Rest of the week, Travel and weather

Look for follow on posts to discuss the meatier part of the week and the record breaking weather we encountered (#SnowpenStack).

 

[1] I did learn one good tidbit out of the Security meeting. Bandit is moving out of OpenStack to an Apache project that will make it more accessible as a general Python security scanning tool.  I have a rant about how that is much better than the Gnocchi story, but I’ll save that for another blog.