MeetBSD 2018 Trip Report: Rodney Grimes12/04/2018
I wish to extend my thanks to the FreeBSD Foundation for the travel grant which allowed me to attend MeetBSD 2018 as a representative of the Release Engineering team. I would also like to thank Michael Dexter and iXsystems for advancing the funds, and making travel and hotel arrangements to accommodate my stay in Santa Clara for the event. Furthermore I would like thank Sabina Anja, Sean Chittenden, Brad Davis, Michael Dexter, and Michael Lucas who paid for various meals throughout the conference.
I traveled and shared a room with Michael Dexter, which is becoming the norm for us as we both are from Portland, Oregon. Hopefully someday soon some of you can come to our fine city for one of these conferences.
We arrived early on Wednesday due to the economics of flights and short distance traveled from Portland. In the early evening a small group of us went to dinner at Pedros and participated in various discussions.
Thursday was the developers summit, I attended the following talks:
LWPMFS: LightWeight Persistent Memory Filesystem:
Interesting topic, I had not thought much about optimizing the file system for NVDIMM use.
Evaluating GIT for FreeBSD:
I learned several things about the GIT work flow, things it does well, and things it comes short of for our current work flow. Unfortunately the post talk Q&A session did not occur since the talk had run into lunch time.
NUMA by Mark Johnson
This greatly expanded my understanding of what is going on under the hood in the NUMA code base.
Have / Need / Want by everyone
I am very happy with how this went, and am glad that the method used at BSDCan, for the 12.0 release target items, was reused so that we have a real document of our results.
I did not attend any additional sessions during the MeetBSD conference, see below:
I was involved in several hallway tracks, including a fairly in-depth conversation with Rick Miller of Verisign about how our release model would affect his work flow, what changes could make it better for them, what types of changes would be very hard for them to work with, etc.
A work session with Peter Wemm on some long standing “requests” for updates to the svn repository seed files for base, ports and docs was completed, this has been an action item for almost 2 years. In addition there is now a seed file for the GSOC repository! Thank you Peter.
During a late evening in the hackers lounge I had a discussion with Chuck Tuffli about his first hand work experience with how git has worked or failed to work. This continued into a discussion of how one might implement in svn the git model of git commit and git push. We called our hypothetical model svn stage, and that it would just modify and store state info in the .svn directory. The next evening, while having a discussion with Peter Wemm this topic came up, and as I explained our “model” to him, he replied that it is called svn shelf/unshelf and is already available in svn. This feature may be useful to help the Security Officer with the issue of not being able to make a publicly visible commit until embargo is lifted.