OpenZFS Developer Summit Introduction
- 4. Why OpenZFS?
● Raise awareness
○ open-source ZFS is awesome!
● Ensure continued improvements to ZFS are
available on all platforms
● Encourage collaboration
○ lots of smart people are working on ZFS
Avoid fragmentation, duplicated effort, stagnation.
- 5. State of OpenZFS
In the past 3 months:
● Website launch
○
○
○
○
>50,000 unique visitors
social & traditional media coverage
Edited by dozens of people
Wikipedia article, 1st page of google(“zfs”)
● Office Hours
○ >20 live participants; >800 subsequent views
● Talks at 5 conferences by 4 different people
○ reaching total of >300 people
● Developer Summit
- 6. Schedule
Monday
● Introductions
● Discussion sessions & presentations (~6 hrs)
○ Lunch; morning & afternoon coffee breaks
● (5:45pm) Beer bash & dinner (on-site)
Tuesday
● (9:30am) Hackathon
● (5:00pm) Hackathon presentations
● (6:16pm) Optional dinner nearby
- 7. Hackathon
Work in small teams (2-4 people)
Work with people you don’t usually
Work on subject area you don’t usually
Start a project and have something, however
small, to show at the end of the day
● Several ideas on DevSummit page
● Day will begin with idea presentations + group
formation
●
●
●
●
- 8. Discussion Priorities
● OpenZFS Code Repository
○ what code, how to manage, how to get there
● Report work in progress, get input
● Gather requirements for future projects
- 9. Topics (1 of 2)
● OpenZFS code repository (Matt Ahrens)
○ porting reports
● Fragmented performance (George Wilson)
● Portable FMA work-alike (Brian Behlendorf)
○ & spares in general
●
●
●
●
Community organization (Karyn Ritter)
Storage tiering (Boris Protopopov)
Performance investigation (Adam Leventhal)
Channel Programs (Chris S. & Max G.)
- 10. Topics (2 of 2)
●
●
●
●
●
Testing & testrunner suite (John Kennedy)
Scalability issues (Kirill Davydychev)
What we’re doing with ZFS (several people)
Multi-tennant ZFS (Rob M)
Interactions w/VM subsystem (Jorgen, Brian)
● Examining on-disk format w/mdb (Max B.)
● xattr performance (Prakash S)