forked from ciao-project/ciao
-
Notifications
You must be signed in to change notification settings - Fork 0
Weekly Meeting 2016 06 30
markdryan edited this page Jul 1, 2016
·
9 revisions
- opens (5 minutes)
- bugs (10 minutes): pre-seed a list of new or priority up/down candidates into the agenda for meeting focus
- prioritize
- scrub
- update/assign
- OpenStack Big Tent: to apply or not to apply? Happenings in June regarding the golang in OpenStack tend to preclude us at this point as an official project. The door is open to be an non-official project though still. Weigh the merits of striving to remain close to OpenStack versus letting ourselves be pushed away.
The meeting bot had an issue. I've pasted the full IRC logs at below and tried to manually extract the meeting summary. Several sections are missing though. There's no action items by person for example. Normal service should resume next week.
- Opens
- Salavador will be able to join ciao again for 50% of his time from next week
- ACTION Salvador will work on integrating existing BAT tests into the Single Machine Setup
- tcpepper provides link to testutil's readme
- ACTION: fuentess to work on integrating single machine tests into travis
- ACTION: fuentess to open github issues for two tasks
- We need to integrate single machine with BAT and then get single machine workingon travis
- obedmr_ will provide support to fuentess on dnsmasq
- leoswaldo has prioritized the list of rally tasks
- tests can be skipped in rally
- There are two interesting rally tests that we need to disable for now
- resize-server and boot-and-migrate
- it would be nice to have these tests
- AGREED: resize is out of scope for now. Not clear how much it is needed
- ACTION: Whole team. Come up with a migration plan, which includes evacuation
- CNCI is not currently migrateable
- obedmr_ is working on docker containers for ciao
- AGREED: that this is a good idea
- ACTION: obedmr_ to send his plan for dockerizing ciao
- Salavador will be able to join ciao again for 50% of his time from next week
- Bugs
- Openstack Big Tent
- Situation summarized by tcpepper
- Open stack does not work like normal open source projects. Ciao team have no rel plan for how to engage
- The ciao team are also not best place to contribute to open stack. They're usedto different languages architectures.
- And the problem is that we need to contribute for 2-3 years before being able toinfluence anything
- There doesn't seem to be any clear guidance on whether a project like ciao couldbe accepted as an official project
- So what do we do?
- Not bother at all or try to engage anyway even if success is not guaranteed?
- is it a lost cause that we'd be viewed as helpful?
- Our openstack talk getting accepted might be a sign that it's worth pursuing BigTent and that we're not locked out
- Note it seems that you can be part of open stack without being part of Big Tent, but we're not quite sure what this means
- ACTION: markusry put openstack back on the agenda for next week to discuss further.
- Situation summarized by tcpepper
- leoswaldo
- tcpepper
- fuentess
- markusry
- mrkz
- carlosag
- _erick0zcr
- obedmr-
- mcastelino
- sameo
<markusry> #startmeeting
<ciaomtgbot> Meeting started Thu Jun 30 16:00:32 2016 UTC. The chair is markusry. Information abut MeetBot at http://wiki.debian.org/MeetBot.
<ciaomtgbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
* _erick0zcr ([email protected]) has joined #ciao-project
<markusry> #topic opens
<mcastelino1> o/
* carlosag (c037362c@gateway/web/freenode/ip.192.55.54.44) has joined #ciao-project
* ciaomt ([email protected]) has joined #ciao-project
<leoswaldo> o/
<tcpepper> o/
<fuentess> o/
* ciaomt has quit (Client Quit)
<markusry> o/
* mcastelino1 is now known as mcastelino
<mrkz> o/
<carlosag> o/
<_erick0zcr> o/
<obedmr-> O/
<mcastelino> o/
<albertom> 0/
<markusry> Yes, forgot the roll call. Great start
* obedmr- is now known as obedmr_
* tcpepper has no opens this week
<markusry> Should I do topic again?
<fuentess> Hi, I have an open, I will be able to return to ciao next week with around 50% of my tme, there are still some tasks I need to complete on clear containers
<leoswaldo> markusry: it should be ok to continue, no tags called, so only comments
<fuentess> I am kind of lost with all new updates on ciao, so I will ramp up again next week and lease, if there is something you would like me to focus on, please let me know
<markusry> #info Salavador will be able to join ciao again for 50% of his time from next week
* peyote (c0373729@gateway/web/freenode/ip.192.55.55.41) has joined #ciao-project
<markusry> fuentess: What were you working on before you got repurposed?
<fuentess> markusry: on the BAT stuff
<leoswaldo> *I also have one, let me know when I can start with it
<sameo> o/
<fuentess> I saw that Kristen already updated some BAT Tests, I can take a look at them and contnue working on what is missing
<markusry> fuentess: Okay. So kristen has created a set scripts for BAT that integrate with Jenkns
<markusry> One of the things we were talking about was getting the BAT scripts working in the Sinle Machine Setup
<markusry> I think.
<mcastelino> agree. that would be good first thing to do
<fuentess> markusry, mcastelino: good, I can start working on that
<markusry> great.
<markusry> #action Salvador will work on integrating existing BAT tests into the Single Machine Stup
<mcastelino> also based on the work obedmr_ did... i.e. running dnsmasq on the macvtap, we shouldalso be able to run single VM on Travis
<tcpepper> fuentess: also check out https://github.com/01org/ciao/blob/master/testutil/README.md
<markusry> Oh, now that would be fancy
<tcpepper> for some info on various test options we've been growing
<markusry> #info tcpepper provides link to testutils readme.
<fuentess> thanks tcpepper, taking a look
<markusry> #link https://github.com/01org/ciao/blob/master/testutil/README.md
* mrkz has quit (Remote host closed the connection)
<ciaoslackbridge> <> Pssst! I didn’t unfurl <https://github.com/01org/ciao/blob/master/testutil/RADME.md> because it was already shared in this channel quite recently (within the last hour) and didn’t want to clutter things up.
<markusry> mcastelino: Is that something fuentess could work on as well, or is there lots of tric networking in there.
<markusry> I'm referirng to singlemachine on Travis
<mcastelino> markusry: I think he can.. basically take obedmr_ work (not the container stuff) butthe dnsmasq part
<mcastelino> i can help if we get stuck
<markusry> fuentess: Is that okay with you?
<fuentess> sure :)
<fuentess> so 2 tasks on my side for next week
<markusry> #action fuentess to work on integrating signle machine tests into travis
<mcastelino> markusry: I see two steps
<markusry> #action fuentess to open github issues for two tasks
<mcastelino> integrate single machien with BAT
<mcastelino> then single machine with BAT on Travis
<markusry> sounds good. I think that's what we've already captured with our actions
<markusry> Let me make it explicit
<obedmr_> fuentess: please let me know if need help on dnsmasq stuff
<markusry> #info We need to integrate single machine with BAT and then get single machine workingon travis
<fuentess> thanks obedmr_, will do
<markusry> #info obedmr_ will provide support to fuentess on dnsmasq
<markusry> Shall we move on to leoswaldo's open?
<fuentess> markusry: I think so
* mrkz ([email protected]) has joined #ciao-project
<markusry> fuentess: Welcome back BTW
<fuentess> thanks markusry
<markusry> leoswaldo:What's your issue?
<leoswaldo> I worked on making a prioritization list for next rally task on ciao. basically its hre (I'll move it to ciao's github wiki)
<leoswaldo> http://paste.openstack.org/show/523834/
<leoswaldo> there are two task that might be good to talk about
<leoswaldo> resize-server
<leoswaldo> and boot-and-migrate
<markusry> #info leoswaldo has prioritized the list of rally tasks
<leoswaldo> these tasks for what I know have not been started, or planned yet
<tcpepper> does rally have knobs? can we tell it we don't want to run a boot-and-migrate test fo now?
<markusry> #link http://paste.openstack.org/show/523834/
<leoswaldo> yes
<tcpepper> ok
<leoswaldo> we run each scenario indepently
<tcpepper> we do need to have a session to talk about our "evacuate" ideas and make an implementaion plan
<markusry> #info tests can be skipped in rally
<leoswaldo> but these two tasks would be very good to have them in CIAO
<leoswaldo> would it be good to start talking about migration of machines or wait some months forthat ?
<markusry> #info There are two interesting rally tests that we need to disable for now
<markusry> #info resize-server and boot-and-migrate
<markusry> #info it would be nice to have these tests
<markusry> leoswaldo: I think we can discuss in the next few weeks.
<markusry> I guess we can't have migration until we have storage.
<tcpepper> some time back we'd talked about resize and weren't quite sure whether it was truly a ommon use case
<leoswaldo> right
<markusry> resize means change the resources assigned to a workload?
<tcpepper> or if it might be sufficient to stop the workload, edit the workload definition, and sart it fresh
<tcpepper> markusry: yes
<markusry> RIght, so this isn't supported. It's supported in payloads but nowhere else
<tcpepper> it's something fancy hypervisors tend to support. so then there's pressure for orchesration to support it.
<leoswaldo> ok, so I
<tcpepper> but....do users really do it much?
<leoswaldo> resize no that much
<markusry> Need to see how this would work with docker as well.
<chamings> for pets they do
<chamings> when they undersize the original
<markusry> Seems some scoping work is needed if we do want to support this
<mcastelino> also resize what... RAM and CPU or more than that
<markusry> rootfs?
<chamings> yes
<tcpepper> though pets do tend to have maintenance windows
<chamings> all of those
<chamings> tcpepper: yes
<markusry> Personally, I think there are higher priority things to implement first
<chamings> and actually it turns out to generally be easier and more reliable in practice to use ata volumes and migrate between VM's anyway
<chamings> not arguing hte low prio, just pointing out that it does get used. carry on
<tcpepper> I'd want to see data on how well the pet OS's truly support hotplug of resources. It' just not that common imho which means its not likely to be reliable.
<tcpepper> so let's just declare right now that resize is out of our scope until we have demand t make it in scope
<markusry> Sounds good to me
<markusry> #agreed resize is out of scope for now. Not clear how much it is needed
<leoswaldo> ok, I think thats it from my opens
<leoswaldo> thks
<markusry> Great. So we will discuss migration soon. Is that the consensus
<markusry> ?
<tcpepper> yes we need to
<markusry> #action Whole team. Scope migration
<mcastelino> FYI: the CNCI today is not yet migrateable
<markusry> #info CNCI is not currently migrateable
<tcpepper> markusry: more broad though...scope our "evacuation" plan, which is both more and lessthan migration
* Patifa (textual@nat/intel/x-lamhgnftwmbrcrei) has joined #ciao-project
<markusry> #action Whole team. Come up with a migration plan, which includes evacuation
<markusry> So any more opens?
<obedmr_> yes,
<obedmr_> just a quick one
<obedmr_> regarding the Ciao on the top of docker containers for development
<obedmr_> what do you think?
<obedmr_> is it ok to create a CIAO org in the dockerhub for putting it?
<mcastelino> I think we should
<leoswaldo> I think its really cool for developers, I already tested and works pretty good :)
<tcpepper> this is another area where we need a more cohesive plan
<markusry> Does this mean that launcher runs in a container?
<obedmr_> yes
<tcpepper> packaging in various forms...dockerhub, debian, fedora
<obedmr_> yep
<markusry> Does all the networking stuff still work?
<obedmr_> markusry: yes, it's working
<mcastelino> obedmr_: did you address the DHCP question
<tcpepper> with "package" for each component or node's needs so you can make a control node, net ode, and lots of comp nodes easy
<obedmr_> mcastelino: still working in the dummy interface thing
<tcpepper> and the all-in-one to easily just play with it in one go
<markusry> #info obedmr_ is working on docker containers for ciao
<markusry> #agreed that this is a good idea
<obedmr_> I will send more information about it and what's the plan
<obedmr_> I'm proposing
<markusry> #action obedmr_ to send his plan for dockerizing ciao
<markusry> Shall we move on?
<obedmr_> yep
<markusry> tcpepper: Do you want to talk about OS briefly, before we do some bugs
<markusry> ?
<tcpepper> lets do bugs first
<tcpepper> if we don't get to OS we can talk about it later
<tcpepper> ie: it can bump to next week easily enough
<markusry> Okay, we only have 4 P2s
<markusry> Shall we do these and then do OS Big Tent
<markusry> Not sure how important it is to go through the P3s.
<markusry> #topic Bugs
<markusry> #link P2 bugs: https://github.com/01org/ciao/issues?q=is%3Aopen+is%3Aissue+label%3Abuglabel%3AP2
<tcpepper> we can at least glance through the P3's since the list is really long and propose escaating any that need it?
<markusry> #link https://github.com/01org/ciao/issues/267
<markusry> One of mine. It's a bug that needs to be fixed otherwise vendoring messes up your gopth
<markusry> easy fix though.
<markusry> #action markusry to fix for Sprint2
<markusry> #link https://github.com/01org/ciao/issues/98
<markusry> leoswaldo: Looks like you're working on this
<leoswaldo> pull request sent, I only need to rework from sameo's feedback
<leoswaldo> #link https://github.com/01org/ciao/pull/318
<leoswaldo> sending it later today
<markusry> #info Pr issued for #267
<markusry> #link https://github.com/01org/ciao/issues/17
* mrkz has quit (Ping timeout: 244 seconds)
<markusry> Me again. It a docker resource issue. It's not clear we're computing statistics for ontainers properly.
<markusry> Haven't started looking at this.
<markusry> But will take a look to see what kubernetes does.
<markusry> It seems to use a library for this
<markusry> #info issue 17 is not started. markusry will fix this for sprint2 and check out how Kbernetes handles this.
<markusry> #linke https://github.com/01org/ciao/issues/16
<markusry> me again
* AmyLee7 ([email protected]) has joined #ciao-project
<markusry> It's related to the previous bug and also to our discussion about resizing.
<leoswaldo> #link https://github.com/01org/ciao/issues/16
<ciaoslackbridge> <> Pssst! I didn’t unfurl <https://github.com/01org/ciao/issues/16> because it as already shared in this channel quite recently (within the last hour) and I didn’t want to cluter things up.
<markusry> leoswaldo: Thanks
<markusry> #info issue 16. markusry will fix for sprint 2. Hasn't started yet
<markusry> Okay, so onto P3s?
<markusry> Here's the list
<markusry> https://github.com/01org/ciao/issues?q=is%3Aopen+is%3Aissue+label%3Abug+label%3AP3
<markusry> How should we handle this? Do we want to go through each individually
<markusry> ?
<markusry> There are 12
<tcpepper> maybe everybody just glance throughit and holler if you think soemthing should be highr
<tcpepper> take a minute or two for that?
<markusry> Yep. Sounds good
<tcpepper> and we have some non-P'd label:bug issues too. I'm not sure how to query those easily
<markusry> I'll see if I can find them
<tcpepper> I think https://github.com/01org/ciao/issues/320 is the only one currently
<markusry> mcastelino: This one is yours.
<mcastelino> yes... it is a P3 for now I think
<markusry> #link https://github.com/01org/ciao/issues/320
<ciaoslackbridge> <> Pssst! I didn’t unfurl <https://github.com/01org/ciao/issues/320> because itwas already shared in this channel quite recently (within the last hour) and I didn’t want to cluter things up.
<markusry> #info Issue 320 should be a P3.
<markusry> #info Issue 320 now is a P3
<markusry> Does anyone have any thoughts on those other P3 bugs?
<tcpepper> I know the line between bug and enhancement is subjective, but...issues #2, #4, #6 fee more like enhancements than P3 bugs to me
<markusry> That's fine with me. Does anyone have any objection to this?
<markusry> #6 is sort of storage related as well.
<tcpepper> no further comments on the P3's from me
<markusry> #info Issues 2, 4 and 6 are no longer bugs. They're now enhancements
<markusry> Me neither
<markusry> Shall we move onto Big Tent
<tcpepper> sure
<markusry> #topic Openstack Big Tent
<markusry> tcpepper: You have the floor
<tcpepper> sigh
<tcpepper> :)
<tcpepper> as humans, developers, the caio team's mostly new to OpenStack so we've been trying tofigure out what it means to be more engaged with OpenStack the past few months
<tcpepper> I feel like we've assumed it's a normal open source project in the way of ones we've sen before
<tcpepper> it is not
<tcpepper> it has a unique governance form
<tcpepper> but there's not a clear document describing how to be a part of the group
<tcpepper> at the summit we heard "commit to fixing bugs for 2-3 years, get known in the communit, then try to influence future decisions"
<tcpepper> that's a foreign way of doing open developing for us
<markusry> And also we don't know python
<markusry> At least I don't
<tcpepper> so that's a good segway into the other issue
<markusry> tcpepper: So what is the first issue?
<tcpepper> we chose a different language, thinking that a good tool for a job would be acceptable
<tcpepper> the first issue is we don't quite know what we're doing, what we're aiming for
<markusry> Is it just that you want to raise the point that openstack does not work how we expectd
<markusry> Okay.
<tcpepper> second would be that we're different
<tcpepper> in language, in architecture
<markusry> #info Open stack does not work like normal open source projects. Ciao team have no rel plan for how to engage
<tcpepper> and also in experience, whether in the language chosen or how we've grown to attack prblems and collaborate in the open
<tcpepper> so we're a fish out of water
<tcpepper> we'd like guidance from the project, but what is out there formally is actually ambiguus
<markusry> #info The ciao team are also not best place to contribute to open stack. They're usedto different languages architectures.
<tcpepper> over the past two months (and two years) there's been a lot of (rehashed) discussion i the openstack community and goverance bodies around "what is openstack"
<markusry> #info And the problem is that we need to contribute for 2-3 years before being able toinfluence anything
<tcpepper> and what is allowable for an "official" project in it, versus one that's "just part ofits ecosystem"
<tcpepper> there wasn't a clear decision within those discussions
<tcpepper> ie: one that would formally exclude us, or formally allow something as different as weare
<tcpepper> at the extremes we could assume we'd never get it, and not bother with trying, stay diferent and even allow ourselves to become more different
<markusry> #info There doesn't seem to be any clear guidance on whether a project like ciao couldbe accepted as an official project
<tcpepper> or assume there's value in being close and getting in and participating, and keep tryig at it
<markusry> #info so what do we do?
<tcpepper> We started from a point of wanting to help make a better OpenStack
<markusry> #info Not bother at all or try to engage anyway even if success is not guaranteed?
<tcpepper> is it a lost cause that we'd be viewed as helpful?
<markusry> good question
<markusry> #info is it a lost cause that we'd be viewed as helpful?
<markusry> Interesting points and good questions.
<tcpepper> it's hard to be motivated and engaged when that's starting to feel like a lost cause
<tcpepper> I don't know that we need to make a decision immediately, but over the coming months w need to be discussing this
<markusry> Okay, well we've got the main issues noted down I hope.
<markusry> Does the summary seem okay to you?
<tcpepper> one specific thing that could be a signpost is whether we get a talk accepted to the nxt summit or not
<tcpepper> markusry: yeah definitely
<tcpepper> I'm curious if you all see another other signpost pending which could influence a choie
<tcpepper> obviously the past month's choice on no golang as an allowable language is one
<markusry> #info Our openstack talk getting accepted might be a sign that it's worth pursuing BigTent and that we're not locked out
<carlosag> carlosag: afk
<tcpepper> or at least staying close to OpenStack
<tcpepper> one thing discussed this month by the TC is that Big Tent isn't a requirement to be a art of openstack
<tcpepper> there's some vague official versus non-official line where you're not in Big Tent but ou are openstack
* tcpepper is frankly confused by the TC's statements
<markusry> #info Note it seems that you can be part of open stack without being part of Big Tent,but we're not quite sure what this means
<markusry> I guess my main concern is about the amount of work required.
<markusry> Anyway, it's 18:03 now.
<markusry> Should we close the meeting, or at least try
<markusry> ?
<tcpepper> yeah we can keep discussing in next meetings
<tcpepper> I'm curious to hear others' thoughts on tradeoffs
<markusry> Let's put it back on the agenda for next meeting. We can think about it in the meantie
<markusry> Thanks for raising the issues
<tcpepper> sounds like a plan
<markusry> #action markusry put openstack back on the agenda for next week
<markusry> Okay, here it goes.
<markusry> Fingers crossed everyone
<markusry> #endmeeting
<tcpepper> :)
* mrkz (mesiment@nat/intel/x-ifdiozhsaycptreo) has joined #ciao-project
<fuentess> thanks markusry
<markusry> It didn't work did it?
<fuentess> seems ike not :(
<tcpepper> I don't think so
<markusry> #endmeeting
<markusry> Nooo
<tcpepper> #ENDMEEETING
<markusry> #please end meeting
<tcpepper> ciaomtgbot: #endmeeting
<ciaomtgbot> tcpepper: Error: "#endmeeting" is not a valid command.
<tcpepper> ciaomtgbot: endmeeting
<ciaomtgbot> tcpepper: Error: "endmeeting" is not a valid command.
<tcpepper> ciaomtgbot: help
<ciaomtgbot> tcpepper: (help [<plugin>] [<command>]) -- This command gives a useful description o what <command> does. <plugin> is only necessary if the command is in more than one plugin.
<markusry> help
<chamings> ciaomtgbot: help meetbot
<ciaomtgbot> chamings: Error: There is no command "meetbot".
<markusry> #ciaomtgbot help
<chamings> ciaomtgbot: help meetbot endmeeting
<ciaomtgbot> chamings: Error: There is no command "meetbot endmeeting".
<chamings> who is the Owner of ciaomtgbot
<chamings> ?
<markusry> kristen
<chamings> :(
Development
- Release Process
- QA
- Vendoring FAQ
- [Single VM Development Environment] (https://github.com/01org/ciao/wiki/Single-VM-Machine-Development-Environment)
Architecture
Usage