CIJTF Meeting Tracker Issue

[bedeho]

CIJTF Meeeting no.1

Scheduling

TBD.

Logistics

Date: TBD.
Time: TBD.
Location: TBD.
Lead: TBD.
Note-taker: TBD.

Agenda

Title	Description	Duration	Action
Greetings	Everyone gives a quick intro.	5m	Informal
Review Backlog Highlights	There is a substantial list of enhancements that have been proposed, let's review the highlights as a basis for future work and priorities.	45m	Informal/Discussion
Q1 2023 Priorities	What should be focused on as a first step?.	20m	Discussion

Backlog

• Standalone Colossus repo: Standalone Colossus repo #4448
• Standalone Argus repo:Standalone Argus repo #4449
• Update Colossus README: Update Colossus README #4450
• Update Argus README: Update Argus README  #4451
• Authentication; Infrastructure authentication #4414
• Standardise Argus access policy signals in metadata: Standardise Argus access policy signals in metadata #4332
• Update reverse proxy recommendations for Argus and Colossus to support HTTP/2 : Update reverse proxy recommendations for Argus and Colossus to support HTTP/2 #3829
• Lead infrastructure (error) monitoring: Lead infrastructure (error) monitoring #4452
• Lighthouse: Lighthouse - An automated storage interrogation service #4270

Meeting Notes

TBD

[singulart]

Any actual meeting notes to leaf through?

[bedeho]

This meeting has not happened, this is a preliminary agenda at this point. Feel free to open an issue with a new agenda item.

[yasiryagi]

Here is the meeting minutes:

Time:

12:00 16/06/2023

Agenda:

• Current storage operation model.
• Current option for scaling? can worker run multiple servers?
  Finding:
• Currently one worker can run one node.
• Each buckets has a objects number and size (in bytes) limits.
• Currently workers running multiple nodes will only have an identical replica in all nodes.
• Currently experience showed that transactorAccoutId keys can only be the role key.
• Due to transactorAccoutId keys can only be the role key, operator running multiple buckets will run the same buckets on all servers in full replication.
• If operator can run with multiple transactorAccoutId keys, possibly operator can have different buckets assigned to different keys and end points.
• Command line only allows moving one bag/channels at a time from one bucket to another. It is mentioned that that there is an extrinsics that allows movement in patches.

Actions:

• Provide logs for transactorAccoutId keys can only be the role key.
• Find out why that transactorAccoutId keys can only be the role key. @mnaamani
• Explore options of having bag patch movement from one bucket to another. @mnaamani
• Explore the option of having the ability to assign bucket for the node. @mnaamani

Next session:

• Feedback on the previous actions.
• Replication limit
• Sharding
• Industry standards and options to scaling
• Visions of the end game of scaling