Thursday, 2013-04-25

*** amyt has joined #openstack-meeting-alt00:02
*** markwash has quit IRC00:05
*** cp16net is now known as cp16net|away00:16
*** cp16net|away is now known as cp16net00:16
*** sarob has quit IRC00:28
*** bdpayne has quit IRC00:41
*** kagan has quit IRC00:41
*** bdpayne has joined #openstack-meeting-alt00:43
*** bdpayne has quit IRC00:44
*** zzs has quit IRC01:22
*** cp16net is now known as cp16net|away01:32
*** cp16net|away is now known as cp16net01:35
*** cp16net is now known as cp16net|away01:35
*** markwash has joined #openstack-meeting-alt01:41
*** rmohan has quit IRC02:02
*** rmohan has joined #openstack-meeting-alt02:03
*** amyt has quit IRC02:05
*** cp16net|away is now known as cp16net02:14
*** kagan has joined #openstack-meeting-alt02:20
*** kagan has quit IRC02:37
*** kagan has joined #openstack-meeting-alt03:29
*** bdpayne has joined #openstack-meeting-alt03:32
*** kagan has quit IRC03:32
*** kagan has joined #openstack-meeting-alt03:35
*** cp16net is now known as cp16net|away03:37
*** cp16net|away is now known as cp16net04:01
*** aignatov3 has joined #openstack-meeting-alt04:15
*** SergeyLukjanov has joined #openstack-meeting-alt04:22
*** amyt has joined #openstack-meeting-alt04:27
*** SergeyLukjanov has quit IRC04:56
*** sacharya has quit IRC05:02
*** aignatov3 has quit IRC05:02
*** bdpayne has quit IRC05:17
*** cp16net is now known as cp16net|away05:38
*** kagan has quit IRC05:40
*** kagan has joined #openstack-meeting-alt05:42
*** kagan has quit IRC05:56
*** SergeyLukjanov has joined #openstack-meeting-alt05:58
*** bdpayne has joined #openstack-meeting-alt06:00
*** bdpayne has quit IRC06:17
*** hartsocks has joined #openstack-meeting-alt06:21
*** bdpayne has joined #openstack-meeting-alt06:25
*** cp16net|away is now known as cp16net06:37
*** IlyaE has quit IRC07:19
*** bdpayne has quit IRC07:26
*** kagan has joined #openstack-meeting-alt07:47
*** kagan has quit IRC07:49
*** kagan has joined #openstack-meeting-alt08:23
*** kagan has quit IRC08:25
*** SergeyLukjanov has quit IRC11:14
*** openstack` has joined #openstack-meeting-alt17:27
*** yidclare has joined #openstack-meeting-alt17:27
*** yidclare has quit IRC17:29
*** hartsocks has joined #openstack-meeting-alt17:29
*** yidclare has joined #openstack-meeting-alt17:31
*** openstack` has joined #openstack-meeting-alt17:38
*** NikitaKonovalov has quit IRC17:39
*** NikitaKonovalov has joined #openstack-meeting-alt17:40
*** SergeyLukjanov has joined #openstack-meeting-alt17:41
*** dmitryme has joined #openstack-meeting-alt17:42
*** akuznetsov has joined #openstack-meeting-alt17:45
SergeyLukjanovhi everybody, Savanna meeting will start in 10 minutes17:54
*** mattf has joined #openstack-meeting-alt17:57
*** mattf has joined #openstack-meeting-alt17:57
*** nicolae_ has joined #openstack-meeting-alt17:58
*** aignatov3 has joined #openstack-meeting-alt17:58
*** sreshetnyak has joined #openstack-meeting-alt18:02
*** aignatov3 has left #openstack-meeting-alt18:02
*** hartsocks has quit IRC18:02
SergeyLukjanovok, I think it's time to start18:02
*** aignatov has joined #openstack-meeting-alt18:03
*** ruhe has joined #openstack-meeting-alt18:03
aignatov#startmeeting savanna18:03
openstack`Meeting started Thu Apr 25 18:03:49 2013 UTC.  The chair is aignatov. Information about MeetBot at http://wiki.debian.org/MeetBot.18:03
openstack`Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.18:03
openstack`The meeting name has been set to 'savanna'18:03
aignatovLet's start our Savanna meeting18:04
*** cloudchimp has quit IRC18:04
aignatovAgenda:18:04
aignatov#link https://wiki.openstack.org/wiki/Meetings/SavannaAgenda18:05
SergeyLukjanovI also want to note that this page can be edited by anyone18:06
SergeyLukjanovso, feel free to add items to this page to build an agenda for the next meeting18:06
aignatovthx18:06
*** openstack has joined #openstack-meeting-alt18:07
*** ChanServ sets mode: +o openstack18:07
aignatovLet's start our talk with news from Savanna team side18:07
aignatov1. Savanna at Summit18:08
dmitrymeWe presented Savanna on the summit18:08
aignatov#info Savanna at Summit18:08
dmitrymeand we got some feedback18:09
dmitrymewe've seen many people interested in the project18:10
dmitrymeand we will continue working with all parties who wish to colaborate18:10
dmitrymefor those who didn't attend our presentation here is the link:18:11
dmitryme#link http://www.slideshare.net/mirantis/savanna-hadoop-on-openstack18:11
aignatov#info Work on Savanna release 0.1.118:12
aignatovWe are continue working on the next minor release of Savanna18:12
aignatovIt will contain several fixes and improvements18:13
mattfthere was also a hortonworks presentation about savanna at summit18:13
mattf#link http://openstacksummitapril2013.sched.org/event/de4b42621ef876d086a000c31c0bdb3818:13
aignatovHere are:18:13
aignatov#link https://launchpad.net/savanna/+milestone/0.1.118:13
dmitrymeMatt, that was session held by Rackspace + Hortownworks18:14
mattf#link https://www.openstack.org/summit/portland-2013/session-videos/presentation/apache-hadoop-on-openstack18:14
*** cp16net is now known as cp16net|away18:14
dmitrymeguys presented their view on future of Savanna18:14
mattfit was, they turned into into a savanna presentation18:14
dmitrymeand in general we are aligned18:15
SergeyLukjanovcurrently we are working on finalizing 0.1.1 version18:15
*** sreshetnyak has quit IRC18:15
mattfalso related to summit -- hadoop came up in half a dozen presentations in one form or another -> definite interest in the topic18:15
SergeyLukjanovmattf, yes, hadoop is the hot topic at the summit18:16
SergeyLukjanovaignatov, let's continue on agenda18:17
aignatovok18:17
mattfas for savanna itself, a common topic was its relation to heat. the impression i took away was that there will be opportunities to coordinate w/ heat further down the roadmap18:18
mattfok, enough from me re summit18:18
aignatov#info architecture of provisioning pluggable mechanism18:18
SergeyLukjanovwe are working hard on cleaning our vision of architecture18:19
aignatovThere are several blupritnts related to this feature:18:19
aignatov#link https://blueprints.launchpad.net/savanna/0.218:20
SergeyLukjanovand trying to compile many different internal documents to the public pages with full description of pluggable mechanism and some other changes in version 0.218:20
SergeyLukjanovthe main goal of 0.2 version is to implement pluggable mechanism and several plugins for 3rd party tooling18:21
SergeyLukjanovadditionally we are looking on custom cluster topologies, file system placements and etc.18:21
aignatov#link https://blueprints.launchpad.net/savanna/+spec/pluggable-cluster-provisioning18:22
aignatov#link https://blueprints.launchpad.net/savanna/+spec/hierarchical-templates18:22
SergeyLukjanov"pluggable-cluster-provisioning" blueprint will be extended with workflow diagram and some additional info in a few days18:23
* mattf anxiously awaits that18:23
*** kagan has quit IRC18:24
*** devcamcar has joined #openstack-meeting-alt18:24
aignatovlet's move to the next topic18:25
*** hartsocks has joined #openstack-meeting-alt18:26
SergeyLukjanovif there are any questions or ideas about pluggable mechanism after reading blueprints, please, send email to savanna-all@lists.launchpad.net18:26
aignatov#diskimage-builder Hadoop elements for creating images working with Savanna18:26
aignatovSorry18:26
*** mestery_ is now known as mestery18:26
aignatov#info diskimage-builder Hadoop elements for creating images working with Savanna18:26
aignatovwe have the first version of Hadoop elements using diskimage-builder18:27
lifelesscool!18:28
aignatovWe will update our documentstion soon how to create own images witth different Hadoop version inside18:28
*** ErikB has joined #openstack-meeting-alt18:29
*** mordred has joined #openstack-meeting-alt18:30
aignatov#info integration tests for Savanna18:31
aignatovWe have started working on building CI process, basic integration tests are already in the repo18:32
mordredaignatov: I'd love for that stuff to start targetting the openstack ci systems - are you guys aware that the CI jobs are all in a git repo and available for you guys to add to?18:32
SergeyLukjanovwe already configured jenkins to be triggered for each new patch set and to use Launchpad SSO18:32
mattfhow do you invoke the integration tests?18:33
SergeyLukjanovmordred, yes, we are using OS CI to run pep8 and unit tests18:34
mordredgreat!18:34
mordredjust wanted to make sure - we're never sure if we communicate that clearly18:34
SergeyLukjanovbut we want to run very specific integration tests that will start multinode hadoop cluster18:35
SergeyLukjanovwhich requires 3-5 vms at OS18:35
dmitrymeMatt, the integration tests are located there:18:36
dmitryme#link https://github.com/stackforge/savanna/tree/master/savanna/tests/integration18:36
dmitrymethe read me inside clarifies how to run them18:36
SergeyLukjanovso, we want to run integration tests in our internal OS lab18:36
*** dkehn has joined #openstack-meeting-alt18:36
*** dkehn has left #openstack-meeting-alt18:37
SergeyLukjanovbut logs will be public accessible and jenkins will publish +/- 1 verified to change requests18:37
mordredsweet. that's exciting18:37
SergeyLukjanovin the future we would like to be fully integrated into the OpenStack infrastructure18:38
ErikBDoes that require us to be incubated before that's possible?18:39
dmitrymeNo, OS nicely provides their infrastructure to any OS related project18:39
*** nicolae_ has quit IRC18:40
SergeyLukjanovErikB, I think it'll depends on which resources we would need to run itests18:40
dmitrymeI mean project does not need to be integrated to fully use OpenStack infra18:40
aignatovSo, that's all with agenda18:42
SergeyLukjanovI think that's all on agenda18:42
aignatov#info Q&A18:42
*** sreshetnyak has joined #openstack-meeting-alt18:42
*** plomakin has joined #openstack-meeting-alt18:43
mattfi've been following http://savanna.mirantis.com/index.html, and mentally translating instructions to work with fedora/rhel & rdo18:43
mattfhow can i contribute to the quick start etc guides?18:43
ErikBfyi - we should be close to socializing a light weight blueprint for HDP provider perhaps tomorrow18:43
SergeyLukjanovErikB, you asked in #savanna channel about which OS release we are aiming18:44
SergeyLukjanovErikB, we are trying to support any F+ releases18:44
SergeyLukjanovmattf, savanna.mirantis.com is an auto generated site from docs folder in our repo18:45
*** jmaron has joined #openstack-meeting-alt18:45
SergeyLukjanovhere is the quickstart - https://github.com/stackforge/savanna/blob/master/doc/source/quickstart.rst18:45
*** akuznetsov has quit IRC18:45
SergeyLukjanovit'll be cool if you can add info to quickstart in places which should be os-specific18:46
ErikB@Sergey, makes sense18:46
mattfhttps://launchpad.net/savanna points to "Quickstart: http://savanna.mirantis.com/quickstart.html" - can we change that to "Quickstart: https://github.com/stackforge/savanna/blob/master/doc/source/quickstart.rst" ?18:47
SergeyLukjanovmattf, just send the change request to savanna repo18:48
dmitrymeMatt, the savanna.mirantis.com is generated from https://github.com/stackforge/savanna/blob/master/doc/source18:48
SergeyLukjanovand when it'll merged into the master branch, savanna.mirantis.com will be auto generated18:48
SergeyLukjanovand will contain your changes18:48
ErikBI noticed that there were some differences between the controller api on mirantis.com vs. launchpad. I assume launchpad is the source of truth?18:48
mattflikewise for https://github.com/stackforge/savanna/blob/master/doc/source/restapi/v02.rst (from "API draft: http://savanna.mirantis.com/restapi/v02.html")?18:49
mattfErikB, that's the central point, need for an authoritative source18:49
dmitrymeMatt, exactly18:49
SergeyLukjanovErikB, could you please share direct links to pages?18:50
*** yidclare1 has joined #openstack-meeting-alt18:52
*** yidclare has quit IRC18:52
mattfi propose linking into the source repo. it's the actual authoritative source. it still gets nice formatting from github.18:52
mattfit also makes it clear that changes to the doc can be handled the same as changes to the code18:52
SergeyLukjanovmattf, changes to the doc are handled the same as changes to the code18:53
SergeyLukjanovjust send change request to the https://github.com/stackforge/savanna/blob/master/doc/source/restapi/v02.rst18:53
SergeyLukjanovand http://savanna.mirantis.com/restapi/v02.html would be updated automaticly18:54
*** yamahata has quit IRC18:54
ErikBI'll post a Q once I find the pages.18:54
SergeyLukjanovwe are using savanna.mirantis.com to provide additional features18:54
dmitrymeSergey, don't you need to add "DocImpact: blablabla" for that?18:54
SergeyLukjanovfor example, sire search18:54
dmitrymeto your commit message18:54
*** ErikB has quit IRC18:55
SergeyLukjanovdmitryme, partially correct, you need to add "DocBuild" to the commit message18:55
SergeyLukjanovbut if someone will send CR w/o it, I'll remind about it :)18:56
dmitrymeok :-)18:56
SergeyLukjanovjfyi you can open routing table at our site - http://savanna.mirantis.com/http-routingtable.html :)18:57
*** jcru has quit IRC18:57
*** cp16net|away is now known as cp16net18:57
*** yamahata has joined #openstack-meeting-alt18:57
ruhealso I'd like to note that all the OS projects keep developer documentation on the project doc site (not on github). for example http://docs.openstack.org/developer/swift/18:58
mattfif you go to the project landing page https://launchpad.net/savanna and follow links to the documentation you're sent off to savanna.mirantis.com, which doesn't have an obvious linkage back to https://github.com/stackforge/savanna/blob/master/doc/source/18:58
SergeyLukjanovi think that our time is over18:58
dmitrymeMatt, but why do you need that linkage?18:59
mattfit sounds like the commit of a doc change may cause a regeneration of the savanna.mirantis.com docs, but only with a specific command in the commit message?18:59
dmitrymeOnly contributors need to know that?18:59
dmitryme* only contributors need to know that18:59
mattfi had to come here and ask how to propose changes to the docs. it's very obvious if the docs are served from github, or a wiki.18:59
*** jcru has joined #openstack-meeting-alt19:00
SergeyLukjanovmattf, we will add some info about how to contribute to the docs19:00
mattfruhe, do you have to be officially incubating to host on docs.os.o/developer ?19:01
ruhe@mattf that's what I'd like to know too19:02
mattfmordred, do you know? ^^19:02
mordredreading scrollback19:03
mattfmordred, do you have to be officially incubating to host on docs.os.o/developer ?19:03
mordredyes19:03
mordredalthough we do have readthedocs hooks you can install19:03
mattfthanks19:03
mordredI believe several of the other stackforge projects use that feature19:04
mattfneat19:04
SergeyLukjanovmordred, thank you, we'll take a look on it19:04
mordredcool. look at project gear in modules/openstack_project/files/jenkins_job_builder/config/projects.yaml for an example19:05
mattfthanks all, ttyl19:05
aignatovlet' finish this meeting19:06
aignatovthank you all19:06
aignatov#endmeeting19:06
*** sreshetnyak has quit IRC19:06
SergeyLukjanovooops :)19:07
aignatovhmm19:07
SergeyLukjanovlooks like meetingbot is asleep19:07
clarkbyeah it came back halfway through the meeting19:08
clarkbthere were problems with netsplit fallout19:09
*** flaper87 has joined #openstack-meeting-alt19:09
*** yidclare1 has quit IRC19:10
*** mattf has left #openstack-meeting-alt19:12
*** aignatov has left #openstack-meeting-alt19:16
*** yidclare has joined #openstack-meeting-alt19:17
*** cloudchimp has joined #openstack-meeting-alt19:18
*** lindj_ has joined #openstack-meeting-alt19:20
*** ruhe has left #openstack-meeting-alt19:22
*** ErikB has joined #openstack-meeting-alt19:23
*** aignatov has joined #openstack-meeting-alt19:29
*** lindj_ has left #openstack-meeting-alt19:30
*** lindj_ has joined #openstack-meeting-alt19:30
*** NikitaKonovalov has quit IRC19:32
*** aignatov has left #openstack-meeting-alt19:37
*** IlyaE has quit IRC19:47
*** devcamcar has quit IRC19:47
*** devcamcar has joined #openstack-meeting-alt19:47
*** jbresnah has joined #openstack-meeting-alt19:51
*** plomakin has quit IRC19:53
*** lindj_ has left #openstack-meeting-alt20:00
markwashHello, glance folks!20:00
rosmaitahowdy!20:00
markwashlets get going20:01
flaper87\o/20:01
markwash#startmeeting20:01
openstackmarkwash: Error: A meeting name is required, e.g., '#startmeeting Marketing Committee'20:01
markwash:-(20:01
*** eglynn has joined #openstack-meeting-alt20:01
markwash#startmeeting Glance20:01
openstackMeeting started Thu Apr 25 20:01:28 2013 UTC.  The chair is markwash. Information about MeetBot at http://wiki.debian.org/MeetBot.20:01
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.20:01
*** openstack changes topic to " (Meeting topic: Glance)"20:01
openstackThe meeting name has been set to 'glance'20:01
markwashWe've got a lot of blueprints to talk about today!20:01
*** bdpayne has quit IRC20:01
markwashI posted a new standing agenda here https://wiki.openstack.org/wiki/Meetings/Glance20:02
markwashbut we probably will only get into the new blueprints part20:02
markwashI want to leave time for open discussion at the end though, so please yell at me if we go over 50 minutes20:02
markwashsigh20:02
markwash#topic new blueprints20:03
*** openstack changes topic to "new blueprints (Meeting topic: Glance)"20:03
markwashThe summit was great, and we had a lot of fruitful discussions20:03
markwashthanks, session leaders!20:03
markwashI've co-opted the the glance meeting agenda etherpad to try to streamline things today20:04
markwash#link https://etherpad.openstack.org/glance-meeting-agenda20:04
markwashjust curious, do we have ameade nikhil jbresnah present?20:05
jbresnahnod20:05
jbresnahwave20:05
ameadehere20:05
ameadephew20:05
markwashiccha is traveling, correct?20:05
nikhilack20:05
jbresnahyeah, i believe she is20:06
nikhilyes markwash20:06
nikhilactually on vacay20:06
markwash:-)20:06
rosmaitano one is on vacation from openstack!20:06
*** esheffield has joined #openstack-meeting-alt20:06
rosmaitait's like 2 am where she is though20:06
markwash:-)20:06
markwashopenstack will find you wherever you are20:06
markwashSo looking at the etherpad, anyone have thoughts on moving forward on quotas?20:07
nikhiltrue20:07
*** jmaron has quit IRC20:07
*** sarob_ has joined #openstack-meeting-alt20:07
rosmaita(not me)20:07
jbresnahummm20:07
jbresnahis the idea to manage it via keystone?20:08
markwashnot as far as I am concerned20:08
nikhilmarkwash: is that your top priority?20:08
jbresnahis that effort mostly about enforcement then?20:08
markwashnot really, the list is ordered more by summit session than anything else20:08
nikhili'm a little curious about folks assigned to the existing blueprints as such20:08
ameadewhat was happening with the whole centralized quota thing? how would that even work?20:09
markwashdid anybody other than me go to the centralize quota session?20:09
jbresnahi was there20:09
nikhili was there20:09
jbresnahhaving keystone manage the information seemed broken to me20:09
markwashI'm not sure I got a clear impression of it, but it sounds like 1) there are issues with the proposed implementation20:09
jbresnahlike... working around the procedure to create a new service20:09
nikhiland then we sync-ed up later in the common lobby/hall20:09
markwashand 2) even if there weren't, it would take a while to get it rolling20:10
flaper87I don't think that's a good idea, FWIW20:10
ameadenot everyone uses keystone20:10
ameadelets just do it in glance?20:10
flaper87ameade: +!20:10
flaper87ameade: +120:10
markwashI'm at +1 for that20:10
rosmaitahow badly do we need quotas?20:10
nikhilmarkwash: ameade do we need to create an action item to sync with folks who are working on it in keystone?20:10
jbresnahi dunno, i like the idea of having it shared across all projects20:11
rosmaitai really think a multi-component solution would be better20:11
nikhil(agree to non-centralized)20:11
rosmaita(what jbresnah said)20:11
*** sarob has quit IRC20:11
jbresnahso many things will be needed every where20:11
rosmaitano i mean multi-component in the sense of something that multiple components can use, not just glance20:11
flaper87I think the quotas thing should start moving but perhaps we should try to sync with other projects as well. I mean, there's some code duplicated between nova and cinder20:11
flaper87we should contribute back to oslo and use that20:11
ameadeflaper87: +120:12
ameadei would like a general solution, but not a centralized one20:12
*** sarob_ has quit IRC20:12
flaper87I was going to take care of that in the next few weeks since I'm trying to reduce duplicated code throughout openstack20:12
markwashrosmaita: jbresnah: I"m all for centralizing the UX of managing quotas, but I never want to have to make a call across a region to determine a users quota limit20:12
nikhili feel there is a discussion going on somewhere else where folks are pushing towards the other idea20:13
rosmaitawell, a centralized solution is needed if you are worried about totla bandwith usage in the cloud for example20:13
rosmaitais anyone worried about that?20:13
markwashnikhil: I feel that you are right20:13
jbresnahameade: I see20:13
jbresnahyou are looking for a shared code base but not a shared service?20:13
ameaderosmaita: good point, drats, this is a can of worms!20:13
rosmaitaso I agree with the action item to find out more info before working on this in glance20:14
nikhilthat's a really good idea (not sure how fast it might move in oslo though)20:14
jbresnahwell the other thing here, is who should do the enforcement?20:14
nikhilguessing the best route would be to develop a good solution and then make it generic20:14
jbresnahi am sure there are good OS level things to do this20:14
jbresnahone person in one of the sessions was strongly against trying to do it in python20:15
markwashinteresting20:15
nikhilthink there are a lot of them20:15
*** ErikB has quit IRC20:15
jbresnahif it is at the host level having limits across sevices is easier (i think)20:15
ameadejbresnah: josh harlow was against ratelimiting in python20:15
nikhilmy guess is it's kind of a tie20:15
markwash#action research approaches to centralize quota management, limit storage, enforcement, and usage monitoring20:15
jbresnahameade: was it for streaming limits only?20:15
jbresnahor limiting rate of request/connection also?20:16
markwashokay, let's table this for a bit, looks like we don't have a stong consensus20:16
jbresnahnod20:16
nikhilmarkwash: nice!20:16
ameade+120:16
flaper87+120:16
rosmaita+120:16
jbresnahi can look into that action item also20:16
esheffield+!20:16
jbresnah+120:16
markwashOkay, protected properties (again from the etherpad)20:16
*** SergeyLukjanov has quit IRC20:17
markwashI think this is pretty uncontroversial?20:17
nikhil+120:17
markwashiccha and smclaren and mikal were possible assignees20:17
rosmaita+120:17
nikhilyes20:17
*** devcamcar has quit IRC20:17
markwash#action markwash talk to iccha and smclaren and mikal about who will lead api-v2-property-protection20:17
markwashmoving on20:18
nikhilmarkwash: one sec20:18
nikhilcan we get the etherpad in the bp20:18
nikhillooks like folks are out of sync (at least from what i sense here)20:18
markwashwho in the what now?20:18
ameadeyeah lets update teh blueprints to have all the info20:18
flaper87+120:19
nikhilthe link to the etherpad which has details about this protected properties implementation proposal20:19
markwashah20:19
markwashdo you have that link? I don't recall it?20:19
nikhilbe in the bp *20:19
nikhilone min20:19
markwash#action nikhil to update api-v2-property-protection blueprint with link to etherpad discussion/proposal20:19
markwashanything else before we move on?20:19
*** sarob has joined #openstack-meeting-alt20:20
nikhilhttps://etherpad.openstack.org/public-glance-protected-props20:20
nikhilok20:20
markwashthanks!20:20
markwashUpload/Download Workflow: https://blueprints.launchpad.net/glance/+spec/upload-download-workflow20:20
markwashthis was discussed during the summit, I'm proposing we split it up into two blueprints20:20
markwashbut I'm not sure that I captured all the details we need to cover20:20
markwashrosmaita: any thoughts here?20:21
flaper87mhh, my understanding is that users will be able to download the image directly from swift20:21
flaper87am i right ?20:21
*** IlyaE has joined #openstack-meeting-alt20:21
markwashflaper87: in some deployments yes, in others perhaps not20:22
rosmaita(what mark said)20:22
markwashI think the idea is that direct download and copy-to would have potentially different policies20:22
flaper87and I guess the direct download won't happen using glance's client, right?20:22
jbresnahso the idea is that you put an image some place and then register it with glance?20:22
markwashjbresnah: that's a little different20:22
jbresnah'some place' being something outside of the scope of glance?20:22
markwashthe idea is that you put an image someplace, and then glance downloads it from there20:23
markwashasync20:23
ameadeoh this is copy from?20:23
ameadeoh i see now20:23
markwashameade: that's what I'm proposing, that "import" is really just "copy-from"20:23
markwashand "export" is just "copy-to"20:23
markwashthough I'm not married to any particular verbal scheme here20:23
jbresnahok20:24
rosmaitaso are copy-from and copy-to already implemented?20:24
jbresnahso 'conversion/verification' do not apply here?20:24
rosmaitaor are you talking conceptually?20:24
markwashcopy-from exists in v120:24
ameadewhat is this bp proposing?20:24
ameadeit seems to be a lot20:24
nikhilgeneric workflow20:24
rosmaitamarkwash: what is your proposal for splitting?20:24
markwashjbresnah: I think copy-from may have some room for conversion / verification support20:24
nikhilit a conceptual ameade20:25
ameadegot it, so it would have a bunch of dep bps20:25
markwashrosmaita: two blueprints, one for adding copy-from like behavior to v2, another for adding copy-to like behavior to v220:25
rosmaitathat makes sense to me!20:25
nikhilprolly could be an action item here ameade (get all the stuff implemented noted down and add bp for things to be created) good point20:26
jbresnahcan we spell out use cases for this?20:26
jbresnahi am having some trouble getting my head around what it is trying to do20:26
flaper87jbresnah: you beat me20:26
markwashsure20:26
ameadeyeah, i dunno why we need copy to if we have copy from20:26
jbresnahand i have some interest in surrounding areas20:26
nikhilupload/download20:26
markwashwell, in the usual case, you can just directly upload the image bits on creation20:27
markwashbut in the "import" case, the implication is that the bits you upload are not necessarily the bits that get stored20:27
markwashb/c of possible conversion, etc.20:27
ameade(almost halfway through meeting time btw)20:27
ameadewanna add the usecases to the bp?20:28
ameadethat could be an action item20:28
markwashperhaps a few of us should take this offline and repackage it to better explain to others?20:28
nikhil+120:28
flaper87+120:28
jbresnah+120:28
rosmaita+120:28
esheffield+120:28
markwashrosmaita, how about you and me?20:28
nikhillet's have a team speak channel for glance?20:28
nikhilmay be i went a bit too optimistic on that20:28
*** ErikB has joined #openstack-meeting-alt20:28
markwash#action rosmaita & markwash to clarify upload/download workflow bp20:29
rosmaitasure20:29
markwashHow about rate limits? Summit discussion was mostly "nope" as I understood it20:29
*** dmitryme has quit IRC20:29
rosmaitaagreed, nope20:30
jbresnahi think that one will fall out of the quota reseach20:30
markwashjbresnah: I agree they're related20:30
flaper87+!20:30
flaper87+120:30
markwash:-)20:30
markwashso, can I mark it superseded by quotas, but say consensus is leading towards not implementing it in glance?20:31
ameadeyeah the lazy consensus at the summit was we dont have to worry about it in glance20:31
flaper87I'd say superseded by quotas, period :P20:31
markwashit == https://blueprints.launchpad.net/glance/+spec/transfer-rate-limiting20:31
rosmaitaright, it seems better done  by something else so glance never sees the requests in the first place20:31
markwashkk, now on to image workers https://blueprints.launchpad.net/glance/+spec/async-glance-workers20:32
markwashwhich is sort of a new one, sort of an old one20:32
nikhillike!20:33
rosmaita+120:33
ameadegotta be done somehow20:33
ameadelol20:33
markwashthis would dovetail into the upload/download workflow to give us a chance at image conversion20:33
rosmaita+120:33
flaper87I like the idea20:34
nikhilmarkwash: is there a strong focus on queue based async?20:34
jbresnahit seems to be the same line of thinking as upload/download...20:34
markwashand also possibly https://blueprints.launchpad.net/glance/+spec/iso-image-metadata20:34
jbresnahi have not processed this one yet20:34
markwashnikhil: I'm not sure. . I think its a good idea in some possible implementations20:34
markwashnikhil: but I could also see a devstack-focused one that doesn't bother with a queue20:34
jbresnahi know i may sound like a broken record on this but i feel glance needs to get its story strat on replica management before it should do conversion20:35
*** jcru has quit IRC20:35
ameade+120:35
jbresnahfor example: is a qcow2 image that was converted from a raw image a replica?20:35
nikhilmarkwash: cool, was just gunna ask about metadata sync and exceptions on the queue based proposal20:35
markwashif the answer is "no" does that make things simpler?20:35
jbresnahdoes a user look to glance to: find me all ubuntu11.11 base images?20:36
jbresnahummm20:36
jbresnahit might, but then it makes me ask 'why is glance doing conversion'20:36
flaper87I think the async-workers bp can be related to conversion but it isn't. It's useful for other operations as well.20:36
jbresnahat some level glance is a discovery service for images20:36
markwashjbresnah: sorry, didn't mean for that to come off as jerk-ish :-)20:36
jbresnahit didnt20:36
jbresnah(didn't even entry my mind that way)20:37
markwashphew20:37
jbresnahif glance is just tracking image blobs20:37
jbresnahthen the formats would not be replicas at all20:37
jbresnahbut if it is just doing blobs, why is it in the business of conversion20:37
flaper87TBH, I'd also like glance to be more "image aware" and be capable of introspecting images (extracting metadata and useful info from them)20:37
jbresnahi have no objection either way, i just think it should be well defined20:38
flaper87but that maybe could be discussed in another meeting20:38
markwashI was definitely in the "just blobs" mindset20:38
jbresnahwe need to have a strong typing on the service or feature creap happens20:38
nikhil12 mins to go20:38
jbresnahjust blobs works, but then why does glance convert?20:38
markwashbut then there would still be a justification for conversion, in that a given deployer knows that he can't actually run qcow2 internally20:38
markwashfor example20:39
jbresnahso, is the idea is to have a image conversion service under the glance umbrella?20:39
markwashmaybe there is more discussion here that should follow on or be a part of the upload/download workflow?20:39
flaper87I'm not sure about having conversions in Glance, TBH20:39
markwashjbresnah: not quite20:39
ameadedoes glance deal with bits or not?20:40
ameadewe just need to define that20:40
ameadethen we can go headstrong into whatever20:40
jbresnahyeah, i think glance's role could be more clearly defined20:40
markwashso, action item is to refine the pitch for conversion?20:40
flaper87as for now, I like to think it doesn't20:40
jbresnahand if it was, i think i would be siginificantly less of a PITA :-(20:40
nikhilis glance == metadata + image data?20:40
jbresnahperhaps a specific meeting to discuss glance's hats?20:41
esheffieldor glance == metadata + image directory20:41
rosmaitai think that glance should allow users to upload images in wome enumeration of formats and should make sure that when the image is used, it's in a format that the cloud the glance is in can use20:41
flaper87jbresnah: +120:41
nikhilthere is yes and no there either way esheffield20:41
nikhil?20:41
markwashquick, somebody #action something! :-)20:42
nikhilyou might have done that markwash20:42
nikhil!20:42
nikhilon "something" :)20:43
jbresnahheh20:43
markwash#action markwash schedule meeting to further discuss image conversion and glance's role with image data20:43
nikhilawesome20:43
ameade+120:43
flaper87+120:43
jbresnah+120:43
esheffield+120:43
markwashokay, we aren't going to get through all of the ones on the list today, but that's def okay20:43
markwashI'd like to skip out of the summit session bps down to the "Other new Blueprints" section of https://etherpad.openstack.org/glance-meeting-agenda20:44
*** IlyaE has quit IRC20:44
markwashand, going out of order, flaper87 had some discussion points around functional testing20:44
nikhilwow, that domain model will need a lot of work20:44
nikhilguessing20:45
markwashflaper87: functional test bp notes?20:45
*** jcru has joined #openstack-meeting-alt20:46
flaper87well, not many. THere are a couple of things I think we should improve there: 1) Reduce the number of tests that need an up&running glance service (perhaps converting some to unittests)20:46
markwash+120:46
flaper87jbresnah: has been doing an amazing job on improving fuinctional tests performance20:46
markwash+100020:46
jbresnahreally i jsut did the impl, markwash did the hard part20:47
ameadeyeah shoutout to jbresnah20:47
jbresnahredirects to markwash20:47
markwashwell, and also the fix to port selection has landed20:47
nikhilo/20:47
flaper87oh another one20:47
markwashI'm curious if folks are still seing a lot of ECONNREFUSED jenkins blowouts20:47
flaper872) We should stop using glance-control for functional tests20:48
flaper87T_T20:48
jbresnahyeah i would like to know that too20:48
jbresnahECONN20:48
ameadeone thing i wanted to mention with regard to reviews, which has been better lately but still, is requiring unit tests and not just functional20:48
markwashah yes, I forgot alla bout glance-control over night20:48
jbresnahameade: what do you mean?20:49
flaper87markwash: I haven't sent the email because it's holiday here and my gf took advantage of that :P20:49
flaper87I'll send it later20:49
markwash:-)20:49
markwashcool20:49
nikhil:)20:49
markwash#action flaper87 send that glance-control email20:49
ameadejbresnah: i just mean that in the past i've seen functional test coverage be good enough for a patch and that shouldnt be20:49
ameadenot a big deal since i havent noticed lately20:50
nikhilmore than 50 mins now20:50
nikhilmarkwash: ^^20:50
flaper87about ECONN, I haven't seen it lately20:50
jbresnahoh i see20:50
jbresnahyeah20:50
markwashameade: that is a really good point20:50
jbresnah+120:50
markwashthere are a few items on open discussion20:50
markwash#topic open discussion20:50
*** openstack changes topic to "open discussion (Meeting topic: Glance)"20:50
nikhili have one question20:50
markwashgo for it20:50
nikhilscott was assign to public glance and wondering if he's working on it?20:51
nikhilassigned*20:51
markwashnikhil: good point, we need to follow up with him20:51
nikhilthanks, action item?20:52
markwash#action markwash follow up with smoser on public glance bp20:52
flaper87I'd like to know if we plan to support legacy commands. They've been deprecated since folsom, AFAIK and some of them have given users some issues20:52
nikhilperfect!20:52
flaper87I think we should drop them based on the fact that new commands have been around for 2 releases20:52
flaper87and we should support those20:52
nikhil+120:53
ameadeflaper87: +120:53
nikhilflaper87: ^^20:53
* jbresnah agrees20:53
markwashits not an issue to drop support for them?20:53
ameademarkwash: well we version glance-client dont we?20:53
flaper87they're all covered by the new commands20:53
flaper87so, a big NOTE in the change log20:53
flaper87and release notes should be enough20:54
flaper87IMHO20:54
markwashthis would be a major version change I think20:54
flaper87plus, they won't see them anymore :P20:54
jbresnahuser legacy scripts could be scripting around the old calls20:54
ameademarkwash: yes definitely20:54
jbresnahbut they have been deprecated for some time20:54
ameadewe dont do "releases" for clients do we?20:54
markwash#action bcwaldon consider dropping legacy shell commands in glance client and let us know why or why not20:54
jbresnahand setting a cultural understanding that deprecation means it is probably a good thing20:54
markwashbcwaldon really still owns this for now, I haven't taken the time to get up to speed20:55
nikhilreleases for glance client too? (good point ameade)20:55
markwashjbresnah: did you want to try to schedule a meeting for talking about the transfer service?20:55
jbresnahi do20:55
jbresnahbut i thnk i will send email about it20:56
flaper87I also wanted to let folks now that I submitted 2 patches for review related to the new registry's api. I'd love to have some feedback from you and make sure we agree with what have been written there20:56
markwashcool20:56
jbresnahi want to have a few more thoughts clear and available for pre-meeting consumption20:56
markwashand in general I'd like to remind folks that20:56
flaper87Registry's V2 https://review.openstack.org/#/c/27359/20:56
markwash#info glance review and bug squashing day tomorrow!20:56
flaper87ww000000000000000000000000000000000ttttttttttttttttt20:56
* jbresnah is looking forward to it!20:56
ameadeheck yeah20:57
flaper87In an hour it will be tomorrow for me20:57
ameadesome hard core coding and reviewing action20:57
jbresnahflaper87: i was looking over that review and i think i need some interactive help understanding some things20:57
rosmaitalet the squashing begin!20:57
markwashalso, its probably obvious that this meeting format is still a little under-construction20:57
jbresnahflaper87: perhaps later today if not too late for you?20:57
rosmaitaflaper87: where are you located physically?20:57
flaper87rosmaita: Italy20:57
flaper87jbresnah: it's never late :P20:57
markwashanybody got quick feedback? also you can email me longer feedback as well20:57
flaper87jbresnah: sounds great!20:57
jbresnahmarkwash: i am really happy about these meetings FYI20:58
nikhilmarkwash: like on the meetings!20:58
ameademarkwash: i think they are going well20:58
* flaper87 is happy too20:58
jbresnahmarkwash: it is making it so much easier for me to get involved in things20:58
ameadewe just have so much to discuss atm20:58
rosmaita+120:58
jbresnahnod20:58
markwashokay cool. . feel free to share ideas for improvment though. I'm just improvising at this point :-)20:58
ameadewell good job so far markwash20:58
nikhilfree food20:59
flaper87+120:59
jbresnahnod20:59
markwashlets's close it there, we can discuss the bps we missed next week or offline during the week20:59
rosmaitaflaper87: ciao!20:59
markwashnikhil: lol!20:59
ameadekk20:59
flaper87rosmaita: ciao :D20:59
markwash#endmeeting20:59
markwashbam20:59
*** openstack changes topic to "OpenStack meetings (alternate) || Development in #openstack-dev || Help in #openstack"20:59
openstackMeeting ended Thu Apr 25 20:59:28 2013 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)20:59
openstackMinutes:        http://eavesdrop.openstack.org/meetings/glance/2013/glance.2013-04-25-20.01.html20:59
openstackMinutes (text): http://eavesdrop.openstack.org/meetings/glance/2013/glance.2013-04-25-20.01.txt20:59
openstackLog:            http://eavesdrop.openstack.org/meetings/glance/2013/glance.2013-04-25-20.01.log.html20:59
flaper87awesome! great meeting folks!20:59
ameadeesheffield: https://wiki.openstack.org/wiki/Meetings/Glance21:00
nikhilflaper87: tschao21:00
nikhil:)21:00
flaper87:P21:00
ameadeon over to #openstack-glance21:00
*** cp16net is now known as cp16net|away21:00
*** jbresnah has left #openstack-meeting-alt21:01
*** sacharya has quit IRC21:01
*** sarob has quit IRC21:01
*** sarob has joined #openstack-meeting-alt21:02
*** sarob has quit IRC21:07
*** HenryG has joined #openstack-meeting-alt21:11
*** jcooley_ has joined #openstack-meeting-alt21:14
*** jcooley has quit IRC21:15
*** jcooley_ is now known as jcooley21:15
*** sarob has joined #openstack-meeting-alt21:24
*** bdpayne has joined #openstack-meeting-alt21:29
*** IlyaE has joined #openstack-meeting-alt21:30
*** yidclare has quit IRC21:31
*** yidclare has joined #openstack-meeting-alt21:33
*** mikal has quit IRC21:38
*** sarob has quit IRC21:48
*** sarob has joined #openstack-meeting-alt21:48
*** amyt has quit IRC21:49
*** sarob has quit IRC21:53
*** sarob has joined #openstack-meeting-alt22:04
*** eglynn has quit IRC22:04
*** markwash has quit IRC22:09
*** bdpayne has quit IRC22:11
*** yidclare has quit IRC22:11
*** yidclare has joined #openstack-meeting-alt22:13
*** hartsocks has quit IRC22:15
*** akshayms has quit IRC22:15
*** hartsocks has joined #openstack-meeting-alt22:15
*** akshayms has joined #openstack-meeting-alt22:15
*** dhellmann is now known as dhellmann-away22:15
*** ErikB has quit IRC22:20
*** hartsocks has quit IRC22:27
*** jcru is now known as jcru|away22:32
*** bdpayne has joined #openstack-meeting-alt22:34
*** hartsocks has joined #openstack-meeting-alt22:38
*** IlyaE has quit IRC22:40
*** jcru|away is now known as jcru22:43
*** ErikB has joined #openstack-meeting-alt22:50
*** jcru has quit IRC22:51
*** djohnstone has quit IRC22:55
*** sarob has quit IRC22:57
*** sarob has joined #openstack-meeting-alt22:57
*** ErikB has quit IRC22:59
*** sarob_ has joined #openstack-meeting-alt22:59
*** sarob has quit IRC23:00
*** hartsocks has quit IRC23:01
*** flaper87 has quit IRC23:04
*** hartsocks has joined #openstack-meeting-alt23:04
*** russell_h has quit IRC23:09
*** russell_h has joined #openstack-meeting-alt23:09
*** yidclare has quit IRC23:17
*** kagan has joined #openstack-meeting-alt23:19
*** hartsocks has quit IRC23:20
*** hartsocks has joined #openstack-meeting-alt23:20
*** cloudchimp has quit IRC23:25
*** hartsocks has quit IRC23:26
*** sacharya has joined #openstack-meeting-alt23:29
*** hartsocks has joined #openstack-meeting-alt23:36
*** hartsocks has quit IRC23:37
*** hartsocks has joined #openstack-meeting-alt23:45
*** ErikB has joined #openstack-meeting-alt23:51
*** kagan has quit IRC23:53
*** ErikB has quit IRC23:56

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!