tfgrid/circle_engineering
22
0
Fork 0
Code Issues 61 Pull Requests Packages Projects 1 Releases Wiki Activity
main
circle_engineering/meetings_notes/engineering_meeting_24-11-18.md

3.8 KiB
Raw Permalink Blame History

Engineering Circle Meeting 2024-11-18

Table of Contents

Attendees

  • Lee
  • Sabrina
  • Thabet
  • Mik
  • Scott

TODO

  • verification column
    • confirm all closed
  • in progress
    • confirm OK to move to 3.16
  • blocked
  • release issue, ok to close
  • 3.16
    • present timeline issue
    • check all issues
      • define clearly
        • specs
        • assignee
        • ETA

Main Content

  • possible hot fix
    • found two problems
      • rolling upgrdde part was communicating with substrate in a non-safe way
        • some leaking connections already fixed
      • way tfchain is implemented
        • active PR pending review
    • this together all issues with RPC should be mitigated
    • currently fix with ops team should work fine regardless
      • no worries in that regard
  • urgency
    • investigate statement
      • no workloads were affected
        • is this true?
        • bert says it is not true that no workloads were affected
    • workloads on GE node, e.g. ethereum
    • have to check
      • are the contracts still online
      • if contracts online, but not reachable
        • it gives more info on the issue
      • TODO make a ticket on this
        • Ashra will look into it
  • thursday issues on mycelium
    • timing out of mycelium
    • look as error if try to connect on the node but it can't as node is set on command line
      • this means user explicitly wants to connect to the node but it can't
      • dashboard belgium nodes, lots of reconnection
    • can you kick off peers that we can't use
      • e.g. 2 out of 5 nodes resetting connection
      • can mycelium automatically put the 2 nodes in the backlog
        • Lee: connection time out is 5 seconds, a bit short actually
          • will increase time-out to 1 min or more
            • should reduce spam
            • should reduce load on public node for reconnection
  • mycelium should be 1.0 in this release
    • we said this for 3.15 grid release
    • mycelium trying to keep semantic versioning
    • Kristof wants routing decision to be taken centrally
      • if we do this starting from 1.0 then we need to call it 2.0, which is too early
    • should go to 0.6 mycelium
    • might need to rewrite the change log
    • change the change to 0.6, and 0.57
  • mycelium gui
    • waiting on android
    • windows is there
  • local
    • ashra following up with

notes

  • qsfs
    • can be completed as soon as changes are released
    • we move it to 3.15
    • can be finished in the following

UI

  • overall stakeholders decision
    • between webteam, communication, ops, devs
    • if we decide something as developers, won't be communicated enough to other stakeholders
    • Sam can present that to company stakeholders
      • on wednesday
  • before there was a clean cut between grid.tf and threefold.io
  • threefold cloud instead of threefold dashboard
  • all urls, do we go with threefold.io
  • simple consensus
    • v3 stay as is
    • we can check how to go with v4

notes for 3.16 guidelines

  • add
    • Make sure not to add stories in the middle of the sprint
      • If there is a need, it needs to be agreed upon

website

  • check the website threefold.io with stakeholders on Wednesday
  • staging website should use www2.threefold.io, production www.threefold.io
  • can send website to internal group
  • full paragraph are not written either
  • send to internal group
  • wait until stakeholders call

Result

  • Moved qsfs back to 3.15
  • moved network filter to 3.16
  • closed mycelium issue
  • ongoing
    • mycelium GUI
      • almost there, in verification, need to set on manufacturer's store
    • local network
      • ongoing, asked Jan for feedback
    • threefold connect