Skip to content

Latest commit

 

History

History
139 lines (93 loc) · 4.81 KB

2021-03-09-Community-Meeting.md

File metadata and controls

139 lines (93 loc) · 4.81 KB

Gluster Community Meeting - 09/03/2021

Previous Meeting details:

Date/Time: Check the community calendar

Bridge

Attendance

Name (#gluster-dev alias) - company

  • Nikhil - Red Hat
  • Pranith Kumar K - PhonePe
  • Nishith Vihar S - Red Hat
  • Sunil Kumar - Red Hat
  • Amar Tumballi - kadalu.io
  • Shwetha Acharya - Red Hat
  • SajuMohammed - Red Hat
  • Srijan Sivakumar - Red Hat
  • Rinku Kothiya - Red Hat
  • Aravinda - kadalu.io

User stories

  • rebalance:
    • when we had 'readdir + stat' was way slower (~200x), compared to only readdir.
    • fsync cost - some fsyncs took upto 25second in few cases! Fails meet SLA. Can we do O_DIRECT writes, so we can avoid fsync?
    • rebalance start & rebalance stop was costly (doesn't serve any purpose!)
    • data movement - layout overlap is a costly operation!
    • Work around was to get the filelist in separate input files, and pass it to rebalance script, so we avoid crawling.

Community

  • Project metrics:
Metrics Value
Coverity 64
Clang Scan 89
Test coverage 70.6 %
Gluster User Queries in last 14 days 8
Total Github issues 358

  • Any release updates?

    • Request team to tag the patches that need to go in Release-9.1

  • Blocker issues across the project?

    • None

  • Notable thread from mailing list

    • None

Conferences / Meetups

  • SYSTOR 21

    • June 14th - June 16th

  • Talks related to gluster:

    • None.

  • SDC India 2021

    • Announcements regarding plans and dates for SDC India 2021 will be made in March.

GlusterFS - v10 and beyond

  • Gluster X Backlog - Under planning, a Github tracker will soon be added to track the issues.

Developer focus

  • Can we have a real-time check for coverity scan on each PR as a Github action?

  • Any design specs to discuss?

    • readdir through file #2197
    • Option dependency #2019
    • Simple-Quota is being deployed by kadalu.io team, and is working smoothly. Would be great to get it merged by GlusterFS X #1750

  • How do we get to the internal faster? for a developer?

    • How can we help onboarding faster?
    • No Design docs, no overviews etc
    • Makes hard for external developer!
    • Transfer of knowledge is key for project's progress!
    • README.md in each directory!
    • The flow of how to go about the code understanding!
    • Videos are key!
    • Very very important for project's success!

  • Why one Self-Heal daemon per node? instead of 1 per volume? 1 per brick etc?

    • Ideally we thought number of self-healing required is less, so one per node, for all volumes would be good enough.

Component status

  • Arbiter - Syntax update for volume creation #2207
  • AFR - Stable
  • DHT - Stable
  • EC - Stable
  • DOC - Stable
  • Geo Replication - Stable
  • Glusterd - Stable
  • thin-arbiter - Stable

Recent Blog posts / Document updates

  • No new blog posts in the recent times.

Host

  • Who will host next meeting? - Srijan
    • Host will need to send out the agenda 24hr - 12hrs in advance to mailing list, and also make sure to send the meeting minutes.
    • Host will need to reach out to one user at least who can talk about their usecase, their experience, and their needs.
    • Host needs to send meeting minutes as PR to http://github.com/gluster/community
    • Host has to send the meeting minutes as a mail to the gluster-devel and gluster-users list.

Notetaker

  • Who will take notes from the next meeting? - Srijan

RoundTable

  • Blog posts which have been written post Dec 2020 not reflecting in planet.gluster.org.
    • Infra issue : 118
  • Plan to have video sessions on alternate tuesdays to discuss more on the in-depth working of different features/components of Gluster

Action Items on host

  • Check-in Minutes of the previous meeting