Notice: this Wiki will be going read only early in 2024 and edits will no longer be possible. Please see: https://gitlab.eclipse.org/eclipsefdn/helpdesk/-/wikis/Wiki-shutdown-plan for the plan.
Difference between revisions of "Architecture Council/Meetings/August 11 2016"
| Line 26: | Line 26: | ||
* Last meeting: [[Architecture Council/Meetings/June 9 2016]] -- open actions see [[#Action_Items]] | * Last meeting: [[Architecture Council/Meetings/June 9 2016]] -- open actions see [[#Action_Items]] | ||
| − | * | + | === Jonas: {Bug|499399} Eclipse Infrastructure - Uptime and speed === |
| − | * | + | * Not meant as an offence, but there is a general perception (from clients) that the Eclipse.org infra is unstable |
| + | ** Build servers, Gerrit, Wiki, Eclipsecon.org ... even PUBLIC services like Eclipsecon.org were very slow and then down for 2 days ! | ||
| + | ** Why is Eclipsecon.org run internally and not on the Cloud ? | ||
| + | ** No suggestion for a solution, but the perception causes harm to the Community ! | ||
| + | ** Some people already moved away, or are considering to go away ("...have to wait 3 days until a build is done...") | ||
| + | ** Just a perception or is it a real problem (around uptime and speed) ? | ||
| + | ** Is the infra too complex to host with a given budget ? | ||
| + | |||
| + | * Dani: Agrees with the perception - issues started some time last year, already raised on the Board meeting | ||
| + | ** Please point to specific issues seen, if possible ... would like to stick to Eclipse.org, others might be worse in respect to response time | ||
| + | |||
| + | * Wayne: Please CC webmaster@eclipse.org on the bug | ||
| + | ** What are the core services that need to be at Eclipse.org, which others could be offloaded ? | ||
| + | ** For example, some projects want to use 3rd party services, which can't be hosted at Eclipse.org anyways | ||
| + | ** Webmaster been playing with Cloud services managed by the Eclipse Foundation (though surveys indicate hosting ourselves is more cost effective ... but then can't dynamically scale up) | ||
| + | ** Maybe we're at a crossing point where Cloud should be considered for some services at least | ||
| + | |||
| + | * Doug: In general happy with how things are going, just noticed some issues around the yearly release | ||
| + | |||
| + | * '''ACTION:''' Please collect data on issues / outages that led to a bad perception ? | ||
| + | * '''ACTION:''' What services need to be on Eclipse.org , which services could reasonably be outside ? | ||
| + | ** Put this data on {{bug|499399}} | ||
| + | |||
| + | === Tyler Jewell: EDP and distributing via Docker images === | ||
| + | * Wayne: Most projects produce some sort of a download to be consumed ("traditional model") | ||
| + | * This doesn't make much sense for Che - Docker makes much more sense there ("docker run eclipse/che") | ||
| + | ** Wayne secured the "eclipse" organization on dockerhub, investigating how to manage that with the Community | ||
| + | ** Already connected with the Mosquito project for input | ||
| + | ** Docker had some changes recently (Windows vs Linux etc) now pushing forward | ||
| + | |||
| + | * What does this mean with respect to the IP policy ? | ||
| + | ** It does make sense for Che to be consumable via docker for easy consumption | ||
| + | ** For consumers building their customized projects, it still makes sense for Docker to be "buildable" and to maintain an IP Log | ||
| + | ** Martin: Not so much concerned about the mechanics of downloads ... more concerned about the meaning of "exempt pre-req" (related to docker pre-req?) as well as download scanning software which checks downloads against the IP Log | ||
| + | ** Maybe if consumers are happy with whatever comes from dockerhub, it doesn't make sense for the EF to place additional constraints / restrictions | ||
| + | |||
| + | === Wayne Changes to IP Policy === | ||
| + | EF is working on a change to the IP Policy as [https://mmilinkov.wordpress.com/2016/06/29/overhauling-ip-management-at-the-eclipse-foundation/ blogged by Mike recently] | ||
| + | * Introducing a new, lighter-weight type of due diligence (license check on contained code only - no provenance) | ||
| + | ** Only check what a project "claims" for Type A releases, but not check if it's actually true | ||
| + | * Projects could choose to be "Type A" or "Type B" per release | ||
| + | ** Expecting that Vertex would move to Type A ... others to do some releases Type A, and at some point do Type B | ||
| + | ** Sounds very similar to "parallel IP process" -- '''how to mark up what is what? How to deal with aggregates as being Type A or Type B ?''' | ||
| + | ** Wayne: Mature projects couldn't be Type B - only for Incubating ones. Helps them work out which software / licenses they actually need. | ||
| + | |||
| + | * Jonas: Has there been any survey on what Eclipse.org consumers actually want ? | ||
| + | ** There's a feeling that Eclipse.org actually does more than most companies do ... consumers like that, but is it actually necessary ? For example if node.js is used... | ||
| + | |||
| + | === Wayne: New AC Members === | ||
| + | * '''ACTION''' Wayne propose candidate | ||
| + | |||
| + | === Martin: UUID / Call Home Policy === | ||
| + | * Got in touch with Ian Skerret: No current plans on resurrecting UUID / data gathering. If there ever is, Ian would reach out to the AC. | ||
| + | * Jonas: Any conclusion on cleaning up existing UUIDs ? | ||
| + | ** Currently not discussed as an issue, since UUIDs were only generated in milestone builds. Feel free to reopen if this is considered an issue. | ||
| + | |||
| + | === Wayne: Inactive Projects === | ||
| + | * Projects that are not actively maintained but still considered useful | ||
| + | ** RSE, some modeling projects, ... | ||
| + | ** Discussion on {{bug|}} | ||
| + | |||
| + | * Projects that get funding initially, but stop when the funding ends - dysfunctional projects | ||
| + | ** Once funding returns, new committers restart - EDP has this as an exception, but not as a standard practice | ||
| + | ** This basically sacrifices meritocracy - how big of a deal is this ? Concerned about projects that lack continuity | ||
| + | ** Problem if EF has to constantly teach the committers the process ... if they went to Github, they'd have a hard time handing the keys from one to the next | ||
| + | ** Discussion on {{bug|}}, answer might be we don't care | ||
<!-- | <!-- | ||
| Line 112: | Line 177: | ||
* '''Regrets:''' | * '''Regrets:''' | ||
| − | * '''In attendance:''' Jonas Helming, Dani Megert, Martin O, | + | * '''In attendance:''' Wayne Beaton, Jonas Helming, Dani Megert, Martin O, Doug Schaefer, Michael Scharf, |
<!-- | <!-- | ||
Revision as of 11:52, 11 August 2016
| Meeting Title: | Architecture Council Monthly Meeting |
| Date & Time: | Thursday August 11, 2016 at 1100 Ottawa  HTML |  iCal
|
| Dial-in: | Let's use the Foundation's Asterisk setup for this call:
Participant conference extension: 701 then enter pin: 51968
|
Contents
Agenda / Notes
- Feel free to edit, but not during the call!
- Last meeting: Architecture Council/Meetings/June 9 2016 -- open actions see #Action_Items
Jonas: {Bug|499399} Eclipse Infrastructure - Uptime and speed
- Not meant as an offence, but there is a general perception (from clients) that the Eclipse.org infra is unstable
- Build servers, Gerrit, Wiki, Eclipsecon.org ... even PUBLIC services like Eclipsecon.org were very slow and then down for 2 days !
- Why is Eclipsecon.org run internally and not on the Cloud ?
- No suggestion for a solution, but the perception causes harm to the Community !
- Some people already moved away, or are considering to go away ("...have to wait 3 days until a build is done...")
- Just a perception or is it a real problem (around uptime and speed) ?
- Is the infra too complex to host with a given budget ?
- Dani: Agrees with the perception - issues started some time last year, already raised on the Board meeting
- Please point to specific issues seen, if possible ... would like to stick to Eclipse.org, others might be worse in respect to response time
- Wayne: Please CC webmaster@eclipse.org on the bug
- What are the core services that need to be at Eclipse.org, which others could be offloaded ?
- For example, some projects want to use 3rd party services, which can't be hosted at Eclipse.org anyways
- Webmaster been playing with Cloud services managed by the Eclipse Foundation (though surveys indicate hosting ourselves is more cost effective ... but then can't dynamically scale up)
- Maybe we're at a crossing point where Cloud should be considered for some services at least
- Doug: In general happy with how things are going, just noticed some issues around the yearly release
- ACTION: Please collect data on issues / outages that led to a bad perception ?
- ACTION: What services need to be on Eclipse.org , which services could reasonably be outside ?
- Put this data on bug 499399
Tyler Jewell: EDP and distributing via Docker images
- Wayne: Most projects produce some sort of a download to be consumed ("traditional model")
- This doesn't make much sense for Che - Docker makes much more sense there ("docker run eclipse/che")
- Wayne secured the "eclipse" organization on dockerhub, investigating how to manage that with the Community
- Already connected with the Mosquito project for input
- Docker had some changes recently (Windows vs Linux etc) now pushing forward
- What does this mean with respect to the IP policy ?
- It does make sense for Che to be consumable via docker for easy consumption
- For consumers building their customized projects, it still makes sense for Docker to be "buildable" and to maintain an IP Log
- Martin: Not so much concerned about the mechanics of downloads ... more concerned about the meaning of "exempt pre-req" (related to docker pre-req?) as well as download scanning software which checks downloads against the IP Log
- Maybe if consumers are happy with whatever comes from dockerhub, it doesn't make sense for the EF to place additional constraints / restrictions
Wayne Changes to IP Policy
EF is working on a change to the IP Policy as blogged by Mike recently
- Introducing a new, lighter-weight type of due diligence (license check on contained code only - no provenance)
- Only check what a project "claims" for Type A releases, but not check if it's actually true
- Projects could choose to be "Type A" or "Type B" per release
- Expecting that Vertex would move to Type A ... others to do some releases Type A, and at some point do Type B
- Sounds very similar to "parallel IP process" -- how to mark up what is what? How to deal with aggregates as being Type A or Type B ?
- Wayne: Mature projects couldn't be Type B - only for Incubating ones. Helps them work out which software / licenses they actually need.
- Jonas: Has there been any survey on what Eclipse.org consumers actually want ?
- There's a feeling that Eclipse.org actually does more than most companies do ... consumers like that, but is it actually necessary ? For example if node.js is used...
Wayne: New AC Members
- ACTION Wayne propose candidate
Martin: UUID / Call Home Policy
- Got in touch with Ian Skerret: No current plans on resurrecting UUID / data gathering. If there ever is, Ian would reach out to the AC.
- Jonas: Any conclusion on cleaning up existing UUIDs ?
- Currently not discussed as an issue, since UUIDs were only generated in milestone builds. Feel free to reopen if this is considered an issue.
Wayne: Inactive Projects
- Projects that are not actively maintained but still considered useful
- RSE, some modeling projects, ...
- Discussion on bug
- Projects that get funding initially, but stop when the funding ends - dysfunctional projects
- Once funding returns, new committers restart - EDP has this as an exception, but not as a standard practice
- This basically sacrifices meritocracy - how big of a deal is this ? Concerned about projects that lack continuity
- Problem if EF has to constantly teach the committers the process ... if they went to Github, they'd have a hard time handing the keys from one to the next
- Discussion on bug , answer might be we don't care
Attendees
All AC Members are invited.
- PMC Reps please confirm attendance or list your delegate below. Every PMC is required to name a primary and backup delegate, and to ensure that one delegate attends the meeting.
| BIRT: | Wenfeng Li | Wenbin He |
| DTP: | Brian Payton | Linda Chan |
| Eclipse: | Dani Megert | Mike Wilson |
| Modeling: | Ed Merks | Cédric Brun Eike Stepper |
| Mylyn: | Steffen Pingel | Mik Kersten |
| RT: | Christian Campo | Tom Watson Doug Clarke Ian Bull |
| SOA: | Adrian Mos | Marc Dutoo |
| Technology: | Gunnar Wagenknecht | Wayne Beaton |
| Tools: | Doug Schaefer | |
| WTP: | Chuck Bridgham | Neil Hauge |
| LocationTech: | Jim Hughes | |
| IoT: | Julien Vermillard |
All Attendees
- Regrets:
- In attendance: Wayne Beaton, Jonas Helming, Dani Megert, Martin O, Doug Schaefer, Michael Scharf,