20:00:16 <markwash> #startmeeting glance
20:00:17 <openstack> Meeting started Thu Feb  6 20:00:16 2014 UTC and is due to finish in 60 minutes.  The chair is markwash. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:18 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:00:20 <openstack> The meeting name has been set to 'glance'
20:00:24 <markwash> roll call!
20:00:24 <arnaud__> hi
20:00:28 <markwash> o/
20:00:35 <flwang> o/
20:01:08 <brianr> o/
20:01:21 <markwash> pehw
20:01:29 <ativelkov> o/
20:01:32 <markwash> s/eh/he/
20:01:39 <jokke_> \o
20:01:52 <markwash> looks like we have some artifact folks today?
20:01:55 <markwash> huzzah
20:02:02 <markwash> agenda is here
20:02:16 <ativelkov> We are here, yeah. At least I am
20:02:19 <markwash> #link https://etherpad.openstack.org/p/glance-team-meeting-agenda
20:02:44 <markwash> I made a pass through all the i3 stuff we care most about and added notes
20:02:51 <markwash> so we'll see how far we can get through that
20:03:02 <markwash> but first I think we had some carryover discussion about glance.store
20:03:05 <markwash> #topic glance.store
20:04:29 <markwash> TBH I'm not sure what discussion exactly is needed at this point
20:04:39 <markwash> and flaper87 is . . not(?) here
20:04:46 <arnaud__> I saw a change
20:04:51 <arnaud__> where is is moving the scrubber
20:05:03 <markwash> indeed, i think that has merged at this point
20:05:03 <arnaud__> not sure if it has been merged yet
20:05:08 <arnaud__> ok
20:05:32 <arnaud__> the point that I would like to discuss is the timeline
20:05:36 <arnaud__> for the change
20:05:48 <markwash> my questions about glance.stores are mostly process-oriented. . how do we set up the repo, and what coordination is needed with other folks / infra etc
20:06:02 <markwash> arnaud__: are you seeing it as more of an early juno thing?
20:06:20 <arnaud__> yes
20:06:36 <markwash> I could see that
20:06:41 <arnaud__> but flaper87 had another point of view
20:06:42 <arnaud__> :)
20:06:45 <markwash> ah
20:06:55 <arnaud__> we can discuss later since he is not here
20:07:07 <flwang> markwash: I don't think it can complete in Icehouse
20:07:24 <markwash> yes, I think perhaps we should do that, arnaud__ can you try to pin flaper87 down to a time to discuss in #openstack-glance ?
20:07:33 <markwash> sometime soon I hope
20:07:33 <arnaud__> ok sure
20:08:08 <markwash> all right, let's hear his point of view a bit later then, and move on for now
20:08:20 <markwash> #topic artifacts api
20:08:49 <markwash> This week I had a f2f meeting with some mirantis folks about how we will move forward with team integration / coordination with the artifacts api
20:09:06 <markwash> I've also been speaking with jbernard about the instance template, which is related
20:09:33 <markwash> We had a few takeaway items from the f2f meeting that I wanted to share here
20:09:58 <markwash> 1) we're going to try to just integrate the artifacts work into the normal glance team rather than having separate teams
20:10:12 <flwang> markwash: cool
20:10:31 <markwash> which will probably mean bringing on more devs to the meeting and core members etc
20:10:50 <markwash> and for a good while there will probably be a bit of an expertise gap as the two silos are joined
20:10:52 <flwang> markwash: nice
20:11:14 <markwash> but I think we can evolve past that, especially as we figure out how to integrate images into the artifacts work
20:11:27 <jbernard> from what ive read so far, the artifact design should accomodate instance-templates nicely
20:11:40 <markwash> 2) we'll discuss the progress of the artifacts stuff from folks who are working on it in the weekly meeting
20:11:59 <ativelkov> Yes, we just don't need to hurry with this: better have slow progress then to break something
20:12:07 <markwash> as a status checkin and as an opportunity to ask questions
20:12:24 <markwash> I put some links to the design work and initial blueprint in the agenda
20:12:32 <markwash> #link https://blueprints.launchpad.net/glance/+spec/metadata-artifact-repository
20:12:41 <markwash> #link https://etherpad.openstack.org/p/MetadataRepository-ArtifactRepositoryAPI
20:12:49 <markwash> #link https://etherpad.openstack.org/p/MetadataRepository-API
20:13:34 <markwash> I think the theory was that tims, a heat guy, would be proposing some code around the api soonish that we could at least look at / keep WIP for a while
20:13:55 <markwash> but I've not seen anything yet, is anyone in a position to contact tims to check on his status?
20:14:06 * flaper87 is back
20:14:11 <flaper87> sorry I'm late guys
20:14:19 <markwash> better late than never
20:14:34 <markwash> okay maybe I'll try to follow up with tims to see what's going on
20:14:54 <markwash> #action markwash track down tims and get status
20:14:59 <ativelkov> I may ask gokrokve to contant tims
20:15:19 <markwash> ativelkov: good idea
20:15:45 <markwash> jbernard: any status for us on the instance templates stuff? what's your initial angle of attack?
20:15:51 <ativelkov> ok, as soon as he appears around, I'll tell him
20:16:22 <jbernard> markwash: the artifact design should be perfect
20:16:25 <markwash> okay cool
20:16:37 <jbernard> markwash: the type would be 'instance' or something like this
20:16:56 <jbernard> markwash: and the associated metadata would contain the block device mapping
20:16:57 <arnaud__> markwash: some parts are still not very clear to me. Espcially, the upload/download and the metadata part.
20:17:48 <markwash> jbernard: great to hear
20:17:58 <arnaud__> let's say, you want to download an artifact, what happens? what do you get?
20:18:00 <markwash> jbernard: so are you sort of blocked/waiting on more progress with artifacts stuff?
20:18:30 <jbernard> markwash: i was going to raise that here actually
20:18:35 <soniarvind> Hi guys, My name is Arvind and I work as OpenStack PM @ VMware. joining the glance IRC for first time. Look forward to learn more about artifacts and other blueprints.
20:18:46 <markwash> soniarvind: hi good to have you
20:19:14 <markwash> arnaud__: (not ignoring you, I will come back to your point in one sec)
20:19:28 <arnaud__> :) cool thks markwash :)
20:19:38 <jbernard> ativelkov: you published the artifact blueprint/etherpads?
20:19:57 <ativelkov> jbernard: not yet, working on that now
20:20:22 <markwash> jbernard: so the issue right now is, without artifact api structure, there's not really a place for you to code /design stuff, right?
20:20:38 <markwash> or are you raising some other issue?
20:20:51 <jbernard> yeah, we need a consensus on direction
20:20:59 <jbernard> at least, that's my current take
20:21:05 <markwash> okay makes sense
20:21:24 <jbernard> the artifact direction is a good one IMO
20:21:25 <markwash> jbernard: do you think it would make sense to come up with the list of attributes you would expect the instance template to have?
20:21:44 <jbernard> yes
20:22:09 <jbernard> ativelkov: maybe we can sync up when you're at a stopping point in your design
20:22:11 <markwash> I suggest that because it seems it *might* help drive /verify some of the artifact discussion
20:22:41 <markwash> for example, it would be neat if we figure out the set of acceptable answers to arnaud's download/upload question for instances
20:22:56 <ativelkov> jbernard: sure. I'll write to the ML as soon as we are ready
20:23:05 <jbernard> yeah i agree, arnaud__ raises some good questions
20:23:58 <jbernard> ok, based on what I see so far, there is enough to agree on a basic design
20:23:59 <markwash> okay, it seems like jbernard you have a little bit of a way to move in terms of defining what an instance template would need to include, so let's take a quick look at arnaud's question in more detail, and then see in what form we need to follow up after the meeting
20:24:04 <rosmaita> arnaud__: that is a good question ... i think we should follow how it's done with images, /v2/artifact/uuid would give you metadata record, /v2/artifact/uuid/file would give you the data (i.e., the actual artifact)
20:24:10 <jbernard> when ativelkov pushes to the list, then we can hammer out the details
20:24:37 <markwash> rosmaita: that's a valid approach but I'd love to let the folks driving the artifact api make their suggestion as part of a general proposal
20:24:51 <arnaud__> markwash: +1
20:25:15 <markwash> mostly wanting to avoid committee design, preferring instead committee review
20:25:20 <arnaud__> ativelkov: I will let you know my concerns with more details, this way you keep that in mind
20:25:38 <arnaud__> while designing :)
20:25:48 <ativelkov> ok, sure
20:25:58 <markwash> arnaud__: so specifically, it seems like you are concerned about what the path is for a consuming service to get an artifact for use
20:26:08 <jbernard> ativelkov: i will send you what an instance template would need to include
20:26:11 <arnaud__> yes
20:26:20 <markwash> so I think the we can take that as a request for specific info in any design info
20:26:26 <markwash> ugh
20:26:31 <markwash> any design info -> any design doc
20:27:17 <markwash> ativelkov: are you doing the initial design? or is it still a question we need to answer after tracking down tims ?
20:27:48 <ativelkov> I would prefer to sync up with tims, yes
20:28:12 <markwash> okay great
20:28:20 <markwash> so, to summarize
20:28:58 <markwash> #action ativelkov markwash track down tims somehow and get status, answer the quesiton who is designing this thing
20:29:18 <markwash> #action jbernard look into instance template requirements and make any thoughts available to artifact api workers
20:29:21 <markwash> okay
20:29:31 <jbernard> sounds good
20:29:38 <markwash> anything else on this topic before we jump into blueprint hell ? :-)
20:29:56 <ativelkov> not from my side :)
20:29:56 * markwash opens nether portal
20:30:09 <markwash> #topic I-3 blueprints
20:30:21 <markwash> First one: image-recover (flwang)
20:30:33 <markwash> aren't you glad your baby woke you up for this? :-)
20:30:47 <markwash> #link https://blueprints.launchpad.net/glance/+spec/image-recover
20:30:51 <flwang> yep, since I don't want to miss this one
20:30:55 <markwash> there is a changeset
20:31:03 <markwash> #link https://review.openstack.org/61680
20:31:05 <markwash> but its abandoned
20:31:05 <flwang> I will submit a patch soon to restore
20:31:15 <markwash> I have a few questions
20:31:17 <flwang> it will depends on zhiyan's location status patch
20:31:19 <flwang> sure
20:31:29 <markwash> 1) how does a user find out an image/ image location is pending delete ?
20:31:36 <markwash> because I thought we didn't want to expose that to users at all
20:31:52 <zhiyan> flwang: may i know what's kind of dependencies there?
20:32:28 <markwash> 2) should this be admin-only? how do we make it admin-only? will it be a burden if it is admin-only but still shows up in the api docs as a valid path for users?
20:32:29 <flwang> markwash: I'd like to let the admin can list the images which are in 'pending-delete' status
20:32:57 <flwang> zhiyan: recover the location from 'pending-delete' to 'active' as well
20:33:06 <flwang> markwash: 2) yes
20:33:39 <markwash> flwang: can we do it in a way where it does not affect the user-facing api then?
20:33:53 <zhiyan> flwang: IMO even we have no location status, we can also do pending-delete => active, no?
20:33:56 <flwang> 2) if the user is not admin, he will run into NotFound error IIRC
20:33:58 <markwash> I'm probably just being a bit stodgy, but I'm very hesitant to change the api for this without more operations folks breaking down my door for this feature
20:34:34 <flwang> zhiyan: yes, but it will be ugly and more work
20:35:20 <markwash> flwang: I seem to recall an ML discussion around this, am I correct?
20:35:26 <arnaud__> flwang: are you also working on the glance-client part?
20:35:32 <markwash> *this feature
20:35:34 <flwang> markwash: I'm going to add a new api like images/xxx/recover
20:36:18 <flwang> markwash: nope? but I'd like to send it if it's necessary :)
20:36:34 <markwash> the api design for this feature doesn't feel right to me
20:36:53 <markwash> I can't really make a clear argument as to why, however
20:36:56 <zhiyan> flwang: i'm a little confused tbh, seems you are trying to do location-recover but image-recover, am i right? i think they are different
20:37:15 <arnaud__> zhiyan: +1
20:37:15 <flwang> zhiyan: nope
20:37:16 <markwash> I guess it seems like image-recover should be separate from the user api completely
20:37:42 <zhiyan> flwang: since if you doing latter one, maybe* we need a api to list "pending-delete" location instead of image ..
20:37:53 <markwash> flwang: yeah, an ML discussion, especially if you could include openstack-operators would be nice
20:37:55 <flwang> zhiyan: when the image is in 'pending-delete', what's the status of location?
20:38:05 <zhiyan> deleted, currently
20:38:24 <markwash> I'm tempted to suggest we move this one to J just becuase I"m not yet comfortable with exposing /v2/images/xxx/recover in the main images api
20:38:44 <flwang> markwash: I'm ok
20:38:45 <zhiyan> +0.5 :)
20:38:55 <markwash> flwang: you're okay with moving it to J?
20:39:00 <rosmaita> +0.5 == +1
20:39:10 <flwang> markwash: I'm ok
20:39:13 <markwash> heh
20:40:02 <markwash> adding notes in etherpad
20:40:04 <zhiyan> rosmaita: 0.5 : J is make more sense. but i personally think it will be nice if we can think about my above question..
20:40:11 <markwash> "Move to J to address concerns about where this is exposed and how it affects pending-deleted locations"
20:40:26 <markwash> zhiyan: yes, I agree with the idea behind that question
20:40:43 <markwash> it would be neat if we had an admin only api that said "here are the pending-delete locations" and put /recover on that
20:40:54 <markwash> but let's defer that conversation, a lot more bps to look at
20:41:07 <markwash> Next up! glance.stores flaper87
20:41:08 <flwang> markwash: yes
20:41:18 <flaper87> o/
20:41:18 <zhiyan> markwash: sure
20:41:22 <flwang> pending-delete location maybe another topic
20:41:44 <markwash> flaper87: there was  a motion to push off the glance.stores integration to J, I'm not actually sure what steps we need to take and when they could happen
20:41:57 <markwash> flaper87: do you wanna discuss outside this meeting or do you have some quick notes for us?
20:42:19 <markwash> (if we don't get through all the blueprints I can track down folks offline)
20:42:31 <flaper87> markwash: I can do both, whatever is best for the meeting. Perhaps, I could start sharing the steps I'm trying to follow
20:42:37 <flaper87> and we can discuss this based on that
20:42:44 <markwash> flaper87: that would be fantastic
20:42:48 <flaper87> ok
20:42:58 <flaper87> very quickly so I don't get other folks time
20:43:00 <markwash> #action flaper87 to share steps / timeline for glance.stores for future followup
20:43:33 <flaper87> markwash: should I do it now? or later in the channel ?
20:43:37 <flaper87> I can write an ehterpad
20:43:40 <flaper87> that would be even better
20:43:49 <markwash> let's go for later
20:43:57 <flaper87> +1
20:44:04 <markwash> flaper87: maybe you can try to coordinate a time to discuss with us as well?
20:44:07 <zhiyan> flaper87: maybe we can reuse https://etherpad.openstack.org/p/right_place_for_glance.store_modules
20:44:10 <markwash> I should be generally available ish :-)
20:44:38 <markwash> okay I'm going to skip around to try to hit the easier ones
20:44:58 <markwash> the following in my view just need more review attention
20:45:14 <markwash> cross-service-request-id tracking
20:45:17 <markwash> #link https://review.openstack.org/#/c/68524/
20:45:31 <markwash> expose-owner-in-v2
20:45:32 <markwash> #link https://review.openstack.org/#/c/69887/
20:45:53 <markwash> image-locaiotn-selection-strategy
20:46:05 <markwash> #link https://review.openstack.org/#/c/58482/
20:46:19 <markwash> and related to that is getting glance tests on testrepository which is working okay for me now
20:46:28 <markwash> #link https://review.openstack.org/#/c/59699/
20:46:42 <flwang> markwash: I saw there is a +2 from zhiyan, so it would be nice if there is anyone can approve it soon
20:47:18 <flwang> https://review.openstack.org/#/c/69887/
20:47:21 <markwash> okay
20:47:29 <markwash> I'll look again, maybe that's all we need
20:47:35 <zhiyan> flwang:  for https://review.openstack.org/#/c/69887/ , yes, i think it wil be better if we add a "docImpect" flag
20:47:39 <flwang> markwash: cool
20:47:53 <markwash> next: i18n-messages
20:48:03 <markwash> flwang: any plans for this one in the next few weeks? or defer it to J ?
20:48:17 <markwash> #link https://blueprints.launchpad.net/glance/+spec/i18n-messages
20:48:29 <flwang> markwash: pls don't defer it :D
20:48:37 <markwash> okay cool
20:48:44 <markwash> but we'll look again next week if there is anything to review ;-)
20:48:52 <flwang> markwash: we can complete it if the reviewers can follow up :)
20:49:11 <flwang> markwash: I think it's the only patch
20:49:15 <markwash> next: community-level-v2-sharing (rosmaita )
20:49:25 <markwash> #link https://blueprints.launchpad.net/glance/+spec/community-level-v2-image-sharing
20:49:34 <markwash> rosmaita: let's go ahead and make iccha the assignee, sound okay?
20:49:43 <rosmaita> +1
20:49:59 <markwash> and I think we'll want at least an initial patch in the next 12 days
20:50:31 <rosmaita> she's travelling right now, but if she were here, she would say "no problem"
20:50:49 <markwash> nice
20:50:54 <ashwini> she probably wont be able to get started on this until mid next week though
20:50:56 <markwash> I'll trust you
20:51:18 <markwash> ashwini: :-)
20:51:24 <markwash> let's give it a shot
20:51:30 <markwash> next: workers  / tasks
20:51:40 <markwash> I see there's a lot of discusison about this on in the agenda etherpad
20:51:43 <nikhil> added some comments in the etherpad
20:51:47 <nikhil> yep
20:52:07 <markwash> it sounds like we expect patches soonish?
20:52:13 <nikhil> ya
20:52:28 <nikhil> markwash: when do they need to be merged in?
20:52:38 <markwash> do we still need async-processing bp? or can we kick that one out and just have it under Import Workflow ?
20:52:46 <markwash> or Upload workflow, I think it is called
20:52:48 <nikhil> we need a baseline patch by 16th I suppose?
20:52:59 <markwash> nikhil: by the 18th
20:53:04 <nikhil> okay, thanks
20:53:48 <markwash> okay let's follow up on these bps on monday
20:54:00 <markwash> next, image-location-status (zhiyan)
20:54:10 <flaper87> ops, my laptop died
20:54:11 <flaper87> :(
20:54:19 <flaper87> I'll work on the etherpad
20:54:19 <markwash> I think this is mostly ready to review, but maybe there is a bit of concern about the first patch
20:54:29 <markwash> #link https://review.openstack.org/#/c/67079/
20:54:34 <markwash> flaper87: okay great thanks
20:54:36 <zhiyan> yes, migration
20:55:28 <zhiyan> what about others? markwash, do you think their shape is good?
20:55:50 <markwash> zhiyan: I looked, and while I like the idea of a little more clarity about what each is supposed to accomplish they looked mostly good
20:56:29 <markwash> yeah I think I was generally pleased, it may be the only issue is the migration
20:57:04 <markwash> zhiyan: I wouldn't mind just taking a crack at the migration, trying to get it to avoid being nullable, would I be stepping on your toes though?
20:58:11 <zhiyan> markwash: actually i think nullable is ok, but maybe i missunderstanding your concerns..
20:58:24 <markwash> flaper87: let us know if you need help or just reviews on bug https://review.openstack.org/#/c/59150/
20:58:38 <markwash> okay
20:58:41 <arnaud__> "but might complicate future operations" could you share what you have in mind markwash ?
20:58:51 <flaper87> markwash: I need help in the other one
20:58:54 <flaper87> the virtual_size
20:59:01 <flaper87> I don't get why it keeps raising that error
20:59:03 <markwash> flaper87: ah okay yes I missed that one
20:59:05 <flaper87> I can't replicate the issue
20:59:13 <flaper87> flwang: tried as well and he wasn't able
20:59:20 <flaper87> -infra guys can't give me access to that box
20:59:22 <markwash> #link https://review.openstack.org/#/c/65499/
20:59:27 <flaper87> I tried creating a venv with ubuntu precise
20:59:36 <markwash> ^^ needs help figuring out testing issues
20:59:36 <flaper87> but nein, it won't replicate the issue
20:59:47 <flaper87> and the funny thing is that it's a really trivial migration
20:59:58 <flaper87> it adds an BigInt field and removes it
21:00:04 <flaper87> no idea why it fails in the gate
21:00:17 <markwash> zhiyan: I'll try to follow up with you on the migration, ping me if I forget
21:00:26 <markwash> okay I htink we need to make way for the next meeting
21:00:34 <flaper87> markwash: thanks
21:00:43 <markwash> thanks everybody for helping with the icehouse-3 triage push!
21:00:44 <zhiyan> markwash: ok, thanks.
21:00:51 <jokke_> thanks markwash
21:00:54 <markwash> #endmeeting