16:00:38 <TheJulia> #startmeeting ironic_bfv
16:00:39 <openstack> Meeting started Thu Apr  6 16:00:38 2017 UTC and is due to finish in 60 minutes.  The chair is TheJulia. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:41 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:43 <dtantsur> o/
16:00:43 <openstack> The meeting name has been set to 'ironic_bfv'
16:00:44 <joanna> o/
16:00:59 <hshiina> o/
16:01:12 <TheJulia> Our agenda today, as always can be found on the wiki: https://wiki.openstack.org/wiki/Meetings/Ironic-BFV
16:01:15 <TheJulia> #link https://wiki.openstack.org/wiki/Meetings/Ironic-BFV
16:01:29 <TheJulia> #topic Announcements/Reminders
16:03:05 <TheJulia> I have no announcements today.  Does anyone have any announcements or reminders today?
16:03:32 <hshiina> i have nothing
16:03:37 <TheJulia> Well, actually, I do have one announcement, the common/cinder.py code landed \o/
16:03:53 <joanna> ah I thought it's the current status part :D
16:03:58 <joanna> anyway \o/
16:04:12 <TheJulia> True, but it feels like there should be a party now :)
16:04:27 <TheJulia> Anyway, moving on if nobody has anything else for this part of the meeting.
16:04:27 <joanna> that's so true :)
16:04:56 <TheJulia> #topic Current Status
16:05:46 <joanna> there's a new patch: https://review.openstack.org/#/c/453839 that is a follow up to newly landed common cinder interface
16:05:49 <TheJulia> I updated the etherpad and added the follow-up to the common code, and a WIP that I committed to begin hammering out last week regarding skipping the deployment if BFV makes sense
16:05:58 <TheJulia> #link https://etherpad.openstack.org/p/Ironic-BFV
16:06:06 <TheJulia> #link https://review.openstack.org/#/c/453839
16:06:17 <TheJulia> Folks, please review ^^^
16:07:11 <TheJulia> Also, please take a quick look at the WIP change I posted, https://review.openstack.org/#/c/454243/ and leave any feedback if pertinent.  Naturally I still need to write tests, which I suspect will be quite a bit of code.
16:07:13 <TheJulia> #link https://review.openstack.org/#/c/454243/
16:08:13 <hshiina> TheJulia, sure. thank you for your work
16:08:39 <TheJulia> hshiina: I skimmed over the the client revisions and marked them -1 since we should have a depends-on tag in the commit message to prevent premature merging
16:09:19 <hshiina> TheJulia, i will add the tag
16:09:28 <TheJulia> Thank you!
16:09:38 <TheJulia> Aside from that, does anyone have anything else status wise to report?
16:10:53 <joanna> I'll update the driver, wire, ipxe so they have a proper base, today or early next week
16:11:15 <TheJulia> Awesome, I was going to ask about that
16:11:24 <joanna> :)
16:11:46 <TheJulia> #info joanna is going to update the driver, storage call wire-in, and ipxe template updates to have the proper base in the next few days.
16:12:05 <TheJulia> So I guess onward we go, if there are no objections.
16:12:53 <TheJulia> #topic Planning/Priorities
16:13:58 <TheJulia> From a priorities standpoint, I'd like to see reviews on  https://review.openstack.org/#/c/366197/ and https://review.openstack.org/#/c/406290 this next week.
16:14:08 <joanna> I think since the common patch was merged, we can choose a new one for next week priority list
16:14:32 <TheJulia> I don't think it will take much to get joanna's follow-up rev landed.
16:15:48 <TheJulia> Agreed, I'd like to get a little more velocity personally, but I suspect if we get one patch a week for the main ironic code, we're doing really good.
16:17:11 <TheJulia> hshiina: do you feel that there needs to be any specific attention paid to the nova or client revisions at this time, or do you think we should just try to get some of the changes in the conductor landed for now?
16:18:15 <hshiina> i agree to review conductor patches following dependency
16:18:39 <joanna> hshiina: If you need any reviews, please let me know, I'll be happy to do it :)
16:19:04 <hshiina> joanna, thank you
16:19:10 <TheJulia> Okay, seems like we're good to move on to discussion?  Any objections?
16:19:21 <joanna> +
16:19:25 <hshiina> nothing
16:20:07 <TheJulia> #topic Discussion
16:20:23 <TheJulia> I see we have one item today that hshiina has raised regarding API path.
16:20:35 <TheJulia> hshiina you have the floor
16:21:01 <joanna> looks like my fault... :)
16:21:10 <hshiina> joanna gave a comment to the API patch regarding API path
16:21:39 <dtantsur> commented on the patch already, but I'd prefer structured approach, i.e. /volume/{connectors,targets}. easier for API discovery, and otherwise nicer.
16:22:07 <TheJulia> dtantsur: I saw, thank you!
16:22:14 <joanna> I like the approach where api paths are resource based
16:22:51 <joanna> logical grouping is not resource-based, if there are no objects representing the group
16:23:17 <dtantsur> neither is root URL
16:23:30 <joanna> but that's a specific case
16:23:38 <dtantsur> "object" can be a collection of links, like /v1 links to /v1/nodes, /v1/ports, etc
16:23:47 <joanna> while here we'd introduce a mixup
16:24:10 <joanna> yes, but there's not GET for /v1/../volume
16:24:11 <dtantsur> if somebody wants to traverse Ironic API to know its capabilities, they'll see a group /v1/volume there, jump into it, and see /v1/volume/...
16:24:20 <dtantsur> there is, just like there is get for /v1
16:25:00 <joanna> but /v1 is our root for available APIs
16:25:26 <TheJulia> For some reason I also think /v1 is optional, but I might be confusing that part of the api code with something else
16:25:45 <joanna> in other cases path are resource based, and grouping is resource-wise
16:25:49 <dtantsur> joanna, right, and /v1/volume is a root for volume API
16:26:15 * dtantsur wishes devananda was here, he's a big fan of traverseable API
16:26:17 <joanna> but there's not "volume" object returned by the API, right?
16:27:06 <TheJulia> There is not
16:27:18 <dtantsur> yes, and I don't see it as a problem
16:27:46 <TheJulia> a get could return what is available, I think we do something similar for provision/state
16:28:08 <joanna> there is no other case where we use a resource that does not represent actual resource just to group stuff
16:28:23 <joanna> or is there? (not sure)
16:28:52 <dtantsur> joanna, / and /v1 as I already noticed
16:28:58 <dtantsur> at least, dunno about something else
16:29:45 <joanna> but /v1 and / are something that may be compared to new index.html :) I don't want to introduce something non-standard that will be misleading
16:30:29 <dtantsur> I don't agree with "non-standard"
16:30:50 <joanna> and having all the resources represent something, having one just for grouping would make me surprised (and probably I'd think that I am doing something wrong as I cannot retrieve volume object from volume resource)
16:30:53 <dtantsur> the / and /v1 are not special, we are just used to them
16:31:15 <dtantsur> joanna, heh, you did not work with Redfish protocol, it's built completely on this principle :)
16:31:29 <dtantsur> anyway, I'm not -2 on volume-connectors, kinda -0.5 rather
16:32:10 <TheJulia> Personally, I'd prefer what was agreed to in the spec.
16:32:32 <joanna> dtantsur: I didn't :(
16:32:33 <TheJulia> since otherwise, we need to go rev the spec, and all of the patches to change it.
16:33:32 <dtantsur> joanna, there is a whole idea of making HTTP API traverseable in this fashion, so that you start at / and walk you way up to the resource you need. but this is kinda off-topic now
16:34:06 <joanna> TheJulia, dtantsur: if that will create additional work, I'm fine with /volume/volume*
16:34:45 <joanna> dtantsur: is there a RFE/bug/spec for that? It looks interesting :)
16:36:11 <dtantsur> joanna, kinda like https://en.wikipedia.org/wiki/HATEOAS
16:36:26 <TheJulia> I don't remember a bug off-hand.  There was hope of making a v2 api with consistency in that regard, but that hasn't really gone anywhere.
16:36:43 <TheJulia> Anyway, I think we are getting off-topic
16:36:50 <joanna> dtantsur: so based on: https://tools.ietf.org/html/rfc5988 ?
16:37:05 <joanna> TheJulia: sorry :)
16:37:24 <TheJulia> joanna: no, it is good to discuss!  Just, we shouldn't spend the entire meeting on it :)
16:38:19 <jlvillal> Drive by link: https://lwn.net/SubscriberLink/718801/1b8957dbeaa0d0ca/
16:38:30 <jlvillal> LWN article about booting from remote storage
16:39:52 <joanna> suming up - we keep /volume/ for grouping, right? :)
16:40:05 <TheJulia> I think so yes
16:40:08 <joanna> jlvillal: thank you :)
16:41:06 <TheJulia> jlvillal: oh the fun of booting systems from remote storage! \o/ :0
16:41:21 * TheJulia hears crickets and thinks that is agreement
16:41:34 <hshiina> should 'get /v1/volume' return links to connectors and targets?
16:41:53 <dtantsur> hshiina, yes, it must have a meaningful response
16:42:22 <hshiina> dtantsur, sure, thanks. i will fix the patch
16:42:40 <TheJulia> #agreed Keep /volume/ for grouping in the api.
16:42:56 <TheJulia> So if there is nothing else, time for Open Discussion?
16:43:28 <TheJulia> #topic Open Discussion
16:44:32 <TheJulia> So I forgot one thing in announcements.  I will at the community leadership training next week, and on vacation the following week. joanna has volunteered to run the meeting next week.
16:44:41 <TheJulia> #info joanna to run the next meeting.
16:45:14 <TheJulia> #info TheJulia will be moderately unavailable next week, and on vacation the following week. Returning the 24th.
16:45:32 <TheJulia> Anything random for discussion today?
16:47:02 <TheJulia> I guess we're done for the day then...
16:47:49 <joanna> thank you TheJulia, dtantsur, hshiina, jlvillal :)
16:48:12 <TheJulia> Thank you everyone!  Have a wonderful day and talk to you, at least some, next week!
16:48:32 <hshiina> thank you everyone!
16:48:48 <dtantsur> thanks!
16:48:49 <TheJulia> #endmeeting