| *** jovon has quit IRC | 00:12 | |
| *** Syed__ has quit IRC | 02:55 | |
| *** valw has joined #craton | 03:23 | |
| *** valw_ has joined #craton | 03:25 | |
| *** valw has quit IRC | 03:28 | |
| *** valw_ has quit IRC | 03:33 | |
| *** tojuvone has joined #craton | 05:52 | |
| sigmavirus | morning all | 13:23 |
|---|---|---|
| *** mdorman has quit IRC | 13:46 | |
| *** mdorman has joined #craton | 13:47 | |
| *** VW has joined #craton | 13:48 | |
| *** VW has quit IRC | 13:49 | |
| *** VW has joined #craton | 13:51 | |
| *** VW has quit IRC | 13:53 | |
| *** VW has joined #craton | 13:53 | |
| thomasem | hihi! | 13:57 |
| thomasem | Thanks for the help on that project vars patch, sigmavirus jimbaker! | 14:02 |
| sigmavirus | Ididn't do very much =P | 14:02 |
| thomasem | Moral support is a thing. | 14:02 |
| thomasem | :P | 14:02 |
| sigmavirus | siblings in merge conflicts? | 14:26 |
| *** VW has quit IRC | 14:28 | |
| *** openstackgerrit has joined #craton | 14:29 | |
| openstackgerrit | Thomas Maddox proposed openstack/craton master: Removes left-over debugging logs https://review.openstack.org/434936 | 14:29 |
| openstackgerrit | Thomas Maddox proposed openstack/craton master: Removes left-over debugging logs https://review.openstack.org/434936 | 14:30 |
| thomasem | no brainer https://review.openstack.org/#/c/434936 | 14:30 |
| thomasem | Oh, right. IRC integration | 14:30 |
| thomasem | NEat. | 14:30 |
| thomasem | sigmavirus: that's right! | 14:32 |
| sigmavirus | yw | 14:33 |
| thomasem | Thanks!! | 14:36 |
| *** VW has joined #craton | 14:36 | |
| *** VW has quit IRC | 14:53 | |
| *** VW has joined #craton | 14:57 | |
| *** VW has quit IRC | 14:58 | |
| *** VW has joined #craton | 14:59 | |
| git-harry | I see there's a functional check now | 15:22 |
| git-harry | but it doesn't work | 15:22 |
| git-harry | https://review.openstack.org/#/c/434936/ | 15:22 |
| git-harry | Are we able to make it non-voting for the time being? | 15:23 |
| jimbaker | i was going to ask the same thing... something is wrong with the setup here it seems | 15:24 |
| jimbaker | sigmavirus, ^^^ | 15:24 |
| sigmavirus | I can flip that switch, sure | 15:25 |
| jimbaker | cool, thanks | 15:30 |
| openstackgerrit | Thomas Maddox proposed openstack/craton master: RBAC-Lite for generic variables implementation https://review.openstack.org/434976 | 15:42 |
| thomasem | Doing some manual testing and writing more tests for ^^, but would appreciate thoughts on the implementation! | 15:44 |
| thomasem | Again, this is temporary since we don't have time for full RBAC support right away. | 15:44 |
| sigmavirus | So I have the change in to make it non-voting | 15:48 |
| thomasem | +1 | 15:48 |
| sigmavirus | But I suspect it's because the database isnt' being configured | 15:48 |
| thomasem | :( aww man | 15:48 |
| jimbaker | re functional testing, maybe some sort of race on container_data['NetworkSettings']['IPAddress']? some checking on availability... i guess we could submit a review to explicitly get more debug info | 15:48 |
| *** Syed__ has joined #craton | 15:49 | |
| sigmavirus | Let me revise my schema change to depend on the infra change | 15:50 |
| *** jovon has joined #craton | 15:52 | |
| sigmavirus | So, my change flicks a switch I think will auto-configure databases for us while also making it non-voting | 15:55 |
| sigmavirus | Which means it may start to succeed after it has become non-voting P | 15:55 |
| tojuvone | Hi | 15:56 |
| sigmavirus | Hi tojuvone ! | 15:56 |
| openstackgerrit | Ian Cordasco proposed openstack/craton master: Add sort_dir and sort_keys to endpoint schemata https://review.openstack.org/434506 | 15:59 |
| tojuvone | Struggling devstack, Craton, py2.7 and py3.5 all in one system. | 15:59 |
| openstackgerrit | Thomas Maddox proposed openstack/craton master: RBAC-Lite for generic resources implementation https://review.openstack.org/434976 | 16:00 |
| jimbaker | thomasem, re https://review.openstack.org/434976 - i would assume is_root=false, is_admin=true would mean that the corresponding user could work with project vars | 16:01 |
| thomasem | jimbaker: yep! | 16:02 |
| thomasem | I thought that made more sense than the original restrictions on projects (where you couldn't even get_by_id without being root) | 16:02 |
| thomasem | But, wanted thoughts on that approach | 16:02 |
| jimbaker | thomasem, actually i'm all mixed up on my truth tables; i'm going to do this systematically before i report back | 16:02 |
| jimbaker | as what actually happens | 16:03 |
| jimbaker | but in general: | 16:03 |
| thomasem | As I understand it, top-level /projects is limited to is_root=True, anything under that pertaining to a specific project requires is_admin=True (meaning they have admin permissions on that project) | 16:03 |
| thomasem | Is that correct? | 16:03 |
| thomasem | I' | 16:03 |
| jimbaker | thomasem, yes | 16:03 |
| jimbaker | and a user associated with a project can always read the vars | 16:04 |
| jimbaker | just not set if they are not admin | 16:04 |
| thomasem | Okay, so that's going to be a snag... the errors we were getting were from oslo_db trying to do exactly that. | 16:04 |
| thomasem | Since it expects project_id to be present to limit the scope that way | 16:05 |
| jimbaker | also 401 errors should always report back with the corresponding error messge | 16:05 |
| thomasem | Are they not in this patch? | 16:05 |
| jimbaker | right now it's silent on the return | 16:05 |
| thomasem | Hmm, weird... wasn't for me? | 16:05 |
| thomasem | Mind sharing the call you made? | 16:05 |
| thomasem | And associated user configuration | 16:06 |
| jimbaker | $ curl -w "%{http_code}" http://127.0.0.1:8080/v1/projects/b9f10eca-66ac-4c27-9c13-9d01e65f96b4 -H "Content-Type: application/json" -H "X-Auth-Token: bootstrap" -H "X-Auth-User: bootstrap" -H "X-Auth-Project: b9f10eca66ac4c279c139d01e65f96b4" | 16:07 |
| thomasem | But, the more concerning issue at hand - if we're wanting a regular non-admin project user to be able to view the project vars like that, I think we're going to have some issues with oslo_db. | 16:07 |
| jimbaker | maybe that's just a bug then, and we fix with rbac | 16:07 |
| thomasem | Yeah, I think we'll have to do it that way. | 16:08 |
| thomasem | What this patch I'm doing aims to do is restore what was previously done. | 16:08 |
| thomasem | and prevent the 500s | 16:08 |
| jimbaker | for cmdb timeline purposes, i believe this should suffice | 16:08 |
| thomasem | Excellent | 16:08 |
| jimbaker | also make sure we configure the right users, with the right properties, in our generate_fake_data.py script | 16:09 |
| thomasem | $ curl -s http://127.0.0.1:8080/v1/projects/b9f10eca-66ac-4c27-9c13-9d01e65f96b4 -H "Content-Type: application/json" -H "X-Auth-Token: demo" -H "X-Auth-User: demo" -H "X-Auth-Project: b9f10eca66ac4c279c139d01e65f96b4" | jq | 16:09 |
| thomasem | { | 16:09 |
| thomasem | "message": "This action requires the 'admin' role", | 16:09 |
| thomasem | "status": 401 | 16:09 |
| thomasem | } | 16:09 |
| thomasem | was what I was getting, btw | 16:09 |
| jimbaker | similar proviso applies to making sure it puts in parents - i have a small patch for that, which i used to test git-harry's recent work | 16:09 |
| thomasem | jimbaker: do we want demo to have is_admin=True? | 16:10 |
| thomasem | Since it's been False all this time | 16:10 |
| thomasem | OR I could create an `admin` user. | 16:10 |
| thomasem | :D | 16:10 |
| thomasem | then we'd have all three | 16:10 |
| jimbaker | thomasem, interesting - i just get back a blank, but i was mucking around with is_admin, is_root settings (that's why i said truth table) | 16:10 |
| thomasem | Gotcha | 16:10 |
| jimbaker | demo_root, demo_admin, demo | 16:10 |
| thomasem | I can make those easy enough | 16:11 |
| thomasem | Got a lot of practice in that last patch. :P | 16:11 |
| jimbaker | something like that? we still have root implies admin, right? which means the two columns is useless... but whatever | 16:11 |
| jimbaker | tokens = user names, for consistency | 16:11 |
| thomasem | You mean how is_root overrides is_admin, making it pointless? | 16:11 |
| jimbaker | yes | 16:11 |
| thomasem | tokens = user name? You lost me. | 16:12 |
| jimbaker | the token for demo_root will be 'demo_root', etc | 16:12 |
| thomasem | Oh yeah | 16:12 |
| thomasem | definitely, make life easy | 16:12 |
| thomasem | Well | 16:12 |
| thomasem | easier, anyway | 16:12 |
| jimbaker | yes, and hopefully make it obvious it's not even close to be sure. thanks, scott/tiger | 16:13 |
| jimbaker | close to being secure ;) | 16:13 |
| thomasem | LOL | 16:13 |
| jimbaker | tojuvone, re your config, that sounds great | 16:18 |
| tojuvone | jimbaker, Just have keystone conflicting. Get mixed python versions it seems | 16:20 |
| jimbaker | tojuvone, yes, that could cause potential conflict. should be easy enough to resolve i would think | 16:21 |
| tojuvone | jimbaker, Yeah, not one of my special skills, but yes :) | 16:22 |
| openstackgerrit | Thomas Maddox proposed openstack/craton master: RBAC-Lite for generic resources implementation https://review.openstack.org/434976 | 16:22 |
| sigmavirus | jimbaker: thomasem does http://logs.openstack.org/06/434506/3/check/gate-craton-tox-functional-ubuntu-xenial/2b1ccf3/console.html#_2017-02-16_16_02_44_870561 seem suspicious to you at all? | 16:22 |
| thomasem | sigmavirus: docker container setup failed some how | 16:23 |
| thomasem | That's where service_ip comes from | 16:23 |
| jimbaker | sigmavirus, it does. but i find it interesting that the code gets past this | 16:23 |
| sigmavirus | Right | 16:23 |
| jimbaker | specifically in terms of its container checks | 16:23 |
| sigmavirus | I don't see docker in the logs | 16:23 |
| sigmavirus | I wonder | 16:23 |
| jimbaker | maybe multiple docker daemons running? and causing problems? | 16:24 |
| jimbaker | sorry, or one daemon running | 16:24 |
| sigmavirus | to http://codesearch.openstack.org/ to see how others use docker | 16:24 |
| thomasem | Oh wow. I've never seen this site before. | 16:24 |
| jimbaker | does look useful. brb | 16:25 |
| * thomasem begins singing "A Whole New World | 16:25 | |
| sigmavirus | could be the jenkins user isn't set-up with the right docker daemon permissions? | 16:28 |
| openstackgerrit | git-harry proposed openstack/craton master: Add devices endpoint https://review.openstack.org/435005 | 16:39 |
| sigmavirus | https://review.openstack.org/434979 should disable voting | 16:40 |
| sigmavirus | until then, I'm looking for other projects that use docker | 16:40 |
| openstackgerrit | Jovon McCloud proposed openstack/craton master: Adds Networks to docs https://review.openstack.org/429110 | 16:50 |
| thomasem | Another meeting going over. Will be in shortly. | 17:01 |
| jimbaker | #startmeeting craton | 17:02 |
| openstack | Meeting started Thu Feb 16 17:02:19 2017 UTC and is due to finish in 60 minutes. The chair is jimbaker. Information about MeetBot at http://wiki.debian.org/MeetBot. | 17:02 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 17:02 |
| openstack | The meeting name has been set to 'craton' | 17:02 |
| jimbaker | #chair sigmavirus sulo jimbaker thomasem | 17:02 |
| openstack | Current chairs: jimbaker sigmavirus sulo thomasem | 17:02 |
| jimbaker | #link https://etherpad.openstack.org/p/craton-meetings | 17:02 |
| jimbaker | sigmavirus, around? | 17:03 |
| jimbaker | if not we can have thomasem chair again... | 17:04 |
| thomasem | I don't mind. It'll be my first time, if you don't mind offering up some halps with the commands. | 17:04 |
| jimbaker | thomasem, sounds good | 17:04 |
| thomasem | #topic Roll Call | 17:04 |
| jimbaker | and looks like you have it down already | 17:05 |
| jimbaker | :) | 17:05 |
| jimbaker | o/ | 17:05 |
| thomasem | o/ | 17:05 |
| jovon | hello | 17:05 |
| * jimbaker should not chair, is the general consensus of the group, i believe :) | 17:05 | |
| thomasem | Welcome to the party, jimbaker, jovon | 17:05 |
| thomasem | #topic Agenda | 17:06 |
| thomasem | #undo | 17:06 |
| openstack | Removing item from minutes: #topic Agenda | 17:06 |
| thomasem | #topic Action Items | 17:06 |
| thomasem | #link http://eavesdrop.openstack.org/meetings/craton/2017/craton.2017-01-30-15.00.html | 17:06 |
| thomasem | #undo | 17:07 |
| openstack | Removing item from minutes: #link http://eavesdrop.openstack.org/meetings/craton/2017/craton.2017-01-30-15.00.html | 17:07 |
| thomasem | I think that one was old | 17:07 |
| jimbaker | http://eavesdrop.openstack.org/meetings/craton/2017/craton.2017-02-13-15.00.html | 17:07 |
| thomasem | Here we are | 17:07 |
| jimbaker | #link http://eavesdrop.openstack.org/meetings/craton/2017/craton.2017-02-13-15.00.html | 17:07 |
| sigmavirus | I'm chairing the OSSP meeting | 17:07 |
| thomasem | Thanks, jimbaker | 17:07 |
| sigmavirus | sorry | 17:07 |
| thomasem | We'll miss you, sigmavirus | 17:07 |
| sigmavirus | yeah, unlikely | 17:08 |
| sigmavirus | =P | 17:08 |
| sigmavirus | Working on the docker fix right now for infra too | 17:08 |
| thomasem | Excellent! | 17:08 |
| jimbaker | sigmavirus, very nice. i think i can repeat here, +1000 | 17:08 |
| thomasem | jimbaker: turn Dusty's document into etherpad, I guess our poorly name etherpad is sufficient? | 17:08 |
| thomasem | https://etherpad.openstack.org/p/cmdb_prototype_meeting_2017_02_09 | 17:09 |
| thomasem | That guy | 17:09 |
| jimbaker | #link https://etherpad.openstack.org/p/cmdb_prototype_meeting_2017_02_09 | 17:09 |
| jimbaker | yes, that one | 17:09 |
| jimbaker | let's just paste dusty's doc at the end | 17:10 |
| jimbaker | and we will work on making it better. sounds good? | 17:10 |
| thomasem | Yeah, sounds good to me. | 17:10 |
| thomasem | jimbaker: Add reviewing said etherpad as a standard agenda item to our meeting template | 17:11 |
| jimbaker | dusty's doc added to cmdb tracking doc | 17:11 |
| thomasem | Awesome. Let's also add that as a standing item for our meeting template | 17:12 |
| thomasem | To review progress-wise, I imagine | 17:12 |
| jimbaker | right, the doc as added is not terribly useful for the review process, but we will put it in a form to do so | 17:13 |
| thomasem | #action thomasem to write BP regarding deployment as a starting point for iterating on a suggested deployment model | 17:14 |
| jimbaker | this is something that toan effectively asked me to do - to match requirements to work | 17:14 |
| thomasem | Right | 17:14 |
| thomasem | So, do we want to turn that into an action item? | 17:14 |
| jimbaker | thomasem, sure, let's do that | 17:14 |
| thomasem | #action jimbaker to map Dusty's requirements to work or existing features of Craton, especially with respect to short-term deliverable (~2 weeks remaining) | 17:15 |
| jimbaker | thomasem, +1 | 17:15 |
| thomasem | sigmavirus's pagination stuff got finished and merged | 17:16 |
| jimbaker | and to be clear - this will be selective - only will consider short term stuff (first 3 reqs iirc) | 17:16 |
| jimbaker | for this mapping | 17:16 |
| thomasem | jimbaker: I would confirm that... I still think some folks are expecting more than what we are. | 17:17 |
| thomasem | Do you mean UC1-3 are all that's expected? | 17:17 |
| jimbaker | thomasem, https://etherpad.openstack.org/p/cmdb_prototype_meeting_2017_02_09, starting at line 89 | 17:18 |
| jimbaker | not another set of stories we have been recently looking at | 17:18 |
| jimbaker | which have the UCn numbering | 17:18 |
| thomasem | Oh, okay | 17:19 |
| jimbaker | clear? i'm hoping to make it so :) | 17:19 |
| jimbaker | for my sanity's sake at least | 17:19 |
| thomasem | Haha, indeed. Not entirely clear to me yet. I'll review it in more detail. | 17:19 |
| thomasem | I've read through all of the things, but they continue to have additional scope that I don't think we can meet with Craton, at least. | 17:19 |
| jimbaker | going from reqs to something actionable is always a challenge | 17:19 |
| thomasem | Definitely | 17:20 |
| jimbaker | thomasem, feel free to add your comments accordingly | 17:20 |
| jimbaker | so we can converge | 17:20 |
| thomasem | And scoping that by responsibility so Craton doesn't become the Australian from Futurama. | 17:20 |
| thomasem | #action thomasem to review Dusty and Bjorn's stories/use-cases and add notes on concerns or questions | 17:21 |
| jimbaker | and now i know that meme... | 17:21 |
| thomasem | I hope it's the one I was thinking of, and not something bad. | 17:21 |
| thomasem | Well... tasteless. | 17:21 |
| jimbaker | i believe we have a joint understanding. but hey, brains | 17:21 |
| thomasem | Anyway, does anyone know the status of the CLI testing that sigmavirus had an action item for? | 17:22 |
| thomasem | Lol, indeed | 17:22 |
| thomasem | I'll carry that one forward until we can get a status on it. | 17:22 |
| thomasem | #action sigmavirus to finish up testing on cli | 17:23 |
| jimbaker | that will be a good hamster wheel for sigmavirus | 17:23 |
| sigmavirus | o_O | 17:23 |
| thomasem | Do we then do Stand Up in this meeting? | 17:23 |
| Syed__ | o/ | 17:23 |
| thomasem | #topic Stand Up | 17:23 |
| jimbaker | yes. but sigmavirus is here, so let's cover that action item? | 17:24 |
| jimbaker | #undo | 17:24 |
| openstack | Removing item from minutes: #topic Stand Up | 17:24 |
| thomasem | #undo | 17:24 |
| openstack | Removing item from minutes: #action sigmavirus to finish up testing on cli | 17:24 |
| thomasem | Pardon the onslaught of pings | 17:24 |
| jimbaker | some handy commands here... | 17:24 |
| thomasem | I dunno, jimbaker, I think he's busy. | 17:25 |
| jimbaker | sigmavirus just pops in with emoticons as necessary | 17:25 |
| thomasem | Yep | 17:25 |
| thomasem | Alright, let's move on. | 17:26 |
| jimbaker | +1 | 17:26 |
| thomasem | #action sigmavirus to finish up testing on cli | 17:26 |
| thomasem | #topic Stand Up | 17:26 |
| thomasem | #info each team member briefly describes what they are working on this week, and describes blockers (if there are any) | 17:26 |
| thomasem | #topic Stand Up :: jimbaker | 17:26 |
| jimbaker | finish up WIP on vars in client/CLI; map reqs to craton tasks, focused on short term for the cmdb milestone; review stuff | 17:27 |
| jimbaker | done | 17:27 |
| thomasem | #topic Stand Up :: thomasem | 17:28 |
| thomasem | project vars merged (yay!); working on some bugs found during that work, then moving back to adding clouds; reviewing queue; review user stories and use-cases for concerns/questions I have so we can improve communication and expectations there. | 17:29 |
| thomasem | done | 17:29 |
| thomasem | #topic Stand Up :: jovon | 17:29 |
| jovon | cleaning some current doc patches as well as investigating current doc tools available doc generating more autoimatic | 17:30 |
| jovon | automatic* | 17:31 |
| jimbaker | +1 | 17:31 |
| thomasem | awesome | 17:31 |
| thomasem | Anything else, jovon, or 'done'? :) | 17:32 |
| openstackgerrit | Ian Cordasco proposed openstack/craton master: Fix up functional testing https://review.openstack.org/435038 | 17:32 |
| jovon | done | 17:32 |
| sigmavirus | ^ should be fun to watch | 17:32 |
| jovon | sorry | 17:32 |
| thomasem | No problem at all! | 17:32 |
| thomasem | #topic Stand Up :: Syed__ | 17:32 |
| Syed__ | Working on patch for update project and users | 17:32 |
| Syed__ | tests are broken now and its a mess :/ | 17:33 |
| Syed__ | but well, hoping to get it going today | 17:33 |
| jimbaker | sigmavirus, nice about that possible bug fix | 17:33 |
| Syed__ | apart from that working over CLI tests | 17:33 |
| sigmavirus | be back in ~10 min | 17:33 |
| thomasem | Syed__: mind linking the review? I'll pull it down when I get a moment and see if anything jumps out at me that might help? | 17:34 |
| Syed__ | thomasem: https://review.openstack.org/#/c/425463/ | 17:34 |
| Syed__ | thomasem: thanks | 17:34 |
| jimbaker | this is an important thing to get fixed | 17:34 |
| thomasem | #link https://review.openstack.org/#/c/425463/ | 17:34 |
| thomasem | You bet! | 17:34 |
| Syed__ | once i get this merged, have couple of minor reviews in the queue i would like y'all to check them out | 17:34 |
| Syed__ | then will focus towards CLI users and projects | 17:35 |
| Syed__ | done | 17:35 |
| thomasem | #topic Open Discussion | 17:35 |
| jimbaker | basically all the refactoring that has hit projects... it's good, but definitely a lot of stepping on each other, given centrality | 17:35 |
| thomasem | Yes. It's caused a fair amount of pain for all involved. | 17:35 |
| thomasem | But, we're trucking through it. I am curious how everyone's feeling? We've been going at a bit of a pace here. :) | 17:36 |
| thomasem | Working early/late/weekends | 17:36 |
| jimbaker | thomasem, well, if i tried to sustain yesterday's pace for much longer, i will fall over | 17:36 |
| git-harry | I could do with some reviews on https://review.openstack.org/#/c/435005/ | 17:36 |
| git-harry | before I start on the code | 17:37 |
| jimbaker | so no, not sustainable. but i think it was important to pitch in here to break logjams | 17:37 |
| thomasem | #link https://review.openstack.org/#/c/435005/ | 17:37 |
| openstackgerrit | Ian Cordasco proposed openstack/craton master: Fix up functional testing https://review.openstack.org/435038 | 17:37 |
| jimbaker | git-harry, thanks, will take a look | 17:37 |
| jimbaker | to summarize, git-harry captures in that spec our discussion about a heterogeneous device collection, to accommodate for now net devices and hosts, but other devices in the future | 17:38 |
| jimbaker | so important stuff to be able to fully query with respect to a given switch for example | 17:39 |
| thomasem | Right, since /hosts doesn't support that? | 17:39 |
| thomasem | And we don't want /[switches,firewalls,hosts,etc.], rather just /devices | 17:40 |
| jimbaker | thomasem, correct - /hosts not surprisingly only returns host objects | 17:40 |
| thomasem | LOL yes, I was surprised at least.. | 17:40 |
| * thomasem kids | 17:40 | |
| jimbaker | :) | 17:40 |
| thomasem | Excellent | 17:40 |
| jimbaker | so switches, firewalls - presumably network devices | 17:40 |
| thomasem | Right | 17:40 |
| thomasem | Cool. I will take a look also | 17:40 |
| jimbaker | the intent is that /devices will also in the future cover such usage as AWS resources, which are not hosts | 17:41 |
| jimbaker | but are they devices... | 17:41 |
| jimbaker | ;) | 17:41 |
| jimbaker | we may have to accept that generalization may mean we are not going to refactor our names. i don't know | 17:42 |
| thomasem | refactor our names? | 17:42 |
| jimbaker | s/device/resource/ in our code base, or something like that | 17:42 |
| thomasem | Ahhhhhhhhh, gotcha | 17:42 |
| thomasem | Alright, cool. Any other topics of discussion? | 17:43 |
| jimbaker | anyway, for rackspace private cloud (potential customer #1 for craton), this is moot | 17:43 |
| thomasem | Yep | 17:43 |
| jovon | i like the idea | 17:43 |
| jimbaker | jovon, i like it too, other than the pain it will cause | 17:43 |
| thomasem | c'est la vie! | 17:44 |
| git-harry | I think we should make the functional testing non-voting to unblock the gate. | 17:44 |
| jimbaker | but i was on a project at canonical where we decided to change the name from ensemble to juju. this was a very pervasive change, and impacted tooling, bug trackers, launchpad projects, etc | 17:45 |
| git-harry | I know sigmavirus is trying to fix the problem but we risk not being able to merge anything for a while if the issue ends up being problematic to solve. | 17:45 |
| jimbaker | git-harry, +1 | 17:45 |
| jovon | +1 | 17:45 |
| thomasem | Yeah, let's be sure it's working before gating on it. | 17:45 |
| jimbaker | i believe sigmavirus stated he had submitted something to that effect | 17:46 |
| jimbaker | https://review.openstack.org/#/c/435038/ can be done in parallel | 17:46 |
| Syed__ | i feel like functional testing is an important aspect and once its out there it will be greate | 17:46 |
| git-harry | jimbaker: https://review.openstack.org/#/c/434979/ | 17:46 |
| Syed__ | great ** | 17:46 |
| git-harry | It's been marked as -W | 17:46 |
| jimbaker | by sigmavirus, yeah we need to reverse that | 17:46 |
| thomasem | Added notes to that effect on the review | 17:47 |
| jimbaker | in retrospect, a bit of settling in with a nonvoting gate would have been nice. but funct testing had been stable quite recently | 17:47 |
| thomasem | Yeah. This is always a concern when executing in a different environment, though. | 17:48 |
| thomasem | I think in the future it'd be nice to see it pass there at least once before making it voting. | 17:48 |
| jimbaker | so i would say non-voting for a week or so might be the right way to do this | 17:48 |
| jimbaker | once the gate starts *working* of course | 17:49 |
| thomasem | yep | 17:49 |
| thomasem | Alright, cool. Well, let's follow up with sigmavirus about that. | 17:49 |
| thomasem | Any other topics? Or do we want 10 minutes back? | 17:49 |
| jimbaker | thomasem, maybe cli testing? | 17:50 |
| thomasem | Integration testing? | 17:50 |
| jimbaker | of course the vagueness of the action item for sigmavirus suggests that it is his forever & ever | 17:50 |
| jimbaker | eg hamster wheel | 17:50 |
| thomasem | Pretty much. | 17:50 |
| jimbaker | but yes, integration testing | 17:50 |
| thomasem | Wait, so you mean to tell me he doesn't want to do that forever? | 17:50 |
| thomasem | So, how has integration testing typically been done in OpenStack, especially between client and API? | 17:51 |
| jimbaker | thomasem, i don't know. let's ask him? sigmavirus, it's ok if you work on that task forever? remember silence means yes | 17:51 |
| thomasem | Cruel | 17:51 |
| jimbaker | ;) | 17:52 |
| jimbaker | ok, so yeah, i mean actually testing the client/CLI in some reasonably robust way against the api sever | 17:52 |
| jimbaker | tox -e integration | 17:52 |
| thomasem | Is it a usual pattern for that to live in the client project, or API server? | 17:53 |
| jimbaker | we could start with the generate fake data as a way of loading up fixture data, suitably modified to support the changes we make | 17:53 |
| jimbaker | so we decided not to make it a dependency on the api server. because circularity | 17:53 |
| jimbaker | craton itself should be tested via rest tests, as we now do in tox -e functional | 17:54 |
| jimbaker | that's the only true contract it provides | 17:54 |
| thomasem | So, then you'd tox -e integration in python-cratonclient project? | 17:54 |
| thomasem | And that would... set up a craton-api and go to town exercising the code paths, or would it be a mock craton API? | 17:55 |
| jimbaker | yes, and it can potentially take advantage of projects for suitable isolation | 17:55 |
| thomasem | Ah | 17:55 |
| thomasem | Yeah, generate a couple of projects specific to the integration test and then mutate those projects | 17:55 |
| jimbaker | we already have some form of mock craton api testing going on the client. that's good, but we need stronger | 17:55 |
| thomasem | and all of their descendants. | 17:55 |
| jimbaker | i'm sure we can get some fun stuff going, but i just want to verify stuff works. such as the recent pagination stuff comes to mind | 17:56 |
| jimbaker | passing sort_keys should be tested, and verified it goes end-to-end | 17:57 |
| jimbaker | hence tox -e integration | 17:57 |
| thomasem | I don't think anyone is going to disagree with that. | 17:57 |
| jimbaker | cool. and again, silence means a chorus of resounding yeses ;) | 17:57 |
| thomasem | My only concern is how the python-cratonclient project now has to know how to deploy a craton-api for testing. | 17:58 |
| jimbaker | thomasem, i'm ok if we put in a separate project as well | 17:58 |
| thomasem | I was imagining that's what, like, tempest tests did. | 17:58 |
| jimbaker | but that's just extra stuff in all likelihood | 17:58 |
| jimbaker | so for tempest, that makes a lot of sense, because there are multiple projects | 17:59 |
| thomasem | As there are here | 17:59 |
| jimbaker | right now, we just have one client. so until we grow more, easier to track in place | 17:59 |
| jimbaker | in one place | 17:59 |
| jimbaker | can always refactor by pulling out | 17:59 |
| thomasem | Sure. I do anticipate it moving out, though. With a separate project, you can more easily manage versions and such that are being tested. | 17:59 |
| jimbaker | good points indeed | 18:00 |
| thomasem | But, I'm not going to plant my feet over it | 18:00 |
| thomasem | And that may take up valuable time | 18:00 |
| jimbaker | yeah, i just want one test script that can grow over time | 18:00 |
| thomasem | So, I appreciate where you're coming from. | 18:00 |
| jimbaker | and start with generate_fake_data.py, and start updating it | 18:00 |
| thomasem | Honestly, a lot of the stuff from our functional tests can apply here. | 18:01 |
| thomasem | Just would need to be leveraged by python-cratonclient | 18:01 |
| thomasem | Since that sets up a craton-api | 18:01 |
| jimbaker | i'm pretty sure we can just use in some fashion. it should be dockerized | 18:01 |
| thomasem | And has all of the logic for create/teardown | 18:01 |
| thomasem | and such | 18:01 |
| jimbaker | and it's available as an import, etc | 18:01 |
| thomasem | Yeah, I guess it is. | 18:01 |
| thomasem | didn't think about that | 18:02 |
| thomasem | handy | 18:02 |
| jimbaker | there's no reason the test harness in the integration testing cannot just use | 18:02 |
| thomasem | Yep | 18:02 |
| jimbaker | yeah, those pieces can continue to live in the craton project | 18:02 |
| thomasem | So, as long as we maintain that contract, the integration tests will be fine. | 18:02 |
| thomasem | Save for legitimate breakage. :P | 18:03 |
| thomasem | I'm going to end the meeting (over time) | 18:03 |
| thomasem | but we can continue this chat | 18:03 |
| thomasem | #endmeeting | 18:03 |
| openstack | Meeting ended Thu Feb 16 18:03:29 2017 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 18:03 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/craton/2017/craton.2017-02-16-17.02.html | 18:03 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/craton/2017/craton.2017-02-16-17.02.txt | 18:03 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/craton/2017/craton.2017-02-16-17.02.log.html | 18:03 |
| thomasem | Well, well, well. Chaired my first OpenStack meeting today. | 18:03 |
| thomasem | How'd I do? | 18:03 |
| thomasem | And I need to go take care of some errands... Oh boy. | 18:05 |
| thomasem | brb | 18:05 |
| jimbaker | thomasem, you did great | 18:06 |
| jimbaker | just got pulled out at the end by my discussion with the insurance adjuster who was looking into my roof | 18:07 |
| jimbaker | which understandably i'm very much interested in being in good shape | 18:07 |
| jimbaker | vs just getting torn up by the occasional 70 to 100 mph winds i get here because we live on the edge of the city - great views of the boulder county open space that this area is noted for, but nothing between us and the mountains to protect us! | 18:08 |
| openstackgerrit | Ian Cordasco proposed openstack/craton master: Fix up functional testing https://review.openstack.org/435038 | 18:11 |
| farid | $ craton region-list --all | 18:12 |
| farid | ERROR: list indices must be integers, not str | 18:12 |
| farid | what gives | 18:12 |
| farid | 70-100mph winds? the humanity | 18:13 |
| sigmavirus | jimbaker: thomasem https://review.openstack.org/435038 should be sufficient to fix up our functional tests | 18:15 |
| sigmavirus | nevermind | 18:16 |
| jimbaker | sigmavirus, ack, hopefully we will see jenkins +1 momentarily | 18:16 |
| sigmavirus | still can't get the service-ip | 18:16 |
| sigmavirus | jimbaker: we won't | 18:16 |
| sigmavirus | i watched it live | 18:16 |
| sigmavirus | it failed with the same issue | 18:16 |
| jimbaker | sigmavirus, ok, np | 18:16 |
| sigmavirus | un workflowed my infra patch to make it non-voting | 18:17 |
| sigmavirus | you should go nag openstack infra to approve it | 18:17 |
| sigmavirus | #openstack-infra | 18:17 |
| sigmavirus | I have an errand to run quickly | 18:17 |
| jimbaker | farid, the chinook windstorms happen every few years here. but this winter, it has happened a few times. interestingly, such winds are always warm | 18:17 |
| jimbaker | so it's "an enough is an enough" frustration | 18:17 |
| jimbaker | sigmavirus, acj | 18:18 |
| jimbaker | ack | 18:18 |
| *** VW has quit IRC | 18:21 | |
| *** VW has joined #craton | 18:23 | |
| jimbaker | brief discussion over on #openstack-infra | 18:37 |
| jimbaker | re fixing up the functional tests | 18:37 |
| jimbaker | https://review.openstack.org/#/c/435038/ | 18:38 |
| jimbaker | specific file: https://github.com/openstack-infra/project-config/blob/master/jenkins/data/bindep-fallback.txt | 18:38 |
| jimbaker | as to why our docker probe would succeed... i don't know. i'm still suspicious of the specific logic, but i haven't had a chance to investigate | 18:39 |
| *** VW has quit IRC | 18:43 | |
| openstackgerrit | Jim Baker proposed openstack/craton master: Fix up functional testing https://review.openstack.org/435038 | 18:46 |
| jimbaker | no luck with the change i made: http://logs.openstack.org/38/435038/4/check/gate-craton-tox-functional-ubuntu-xenial/20456fe/console.html#_2017-02-16_18_51_05_997722 | 18:56 |
| sigmavirus | Hey I have an idea, what if we don't use docker for our functional tests? | 19:06 |
| jimbaker | and just do it direct? | 19:06 |
| sigmavirus | Yeah, let me see if I can figure out what we're doing with docker-py though | 19:10 |
| jimbaker | sigmavirus, i believe i followed AndreaJ's instructions | 19:12 |
| sigmavirus | jimbaker: I see that :) | 19:12 |
| jimbaker | oops, AJaegar | 19:12 |
| jimbaker | so my doubt here is with respect to the docker client | 19:13 |
| jimbaker | i haven't had a chance to look at this, maybe every read from container_data['NetworkSettings']['IPAddress'] checks the current settings | 19:14 |
| jimbaker | but certainly some suspicion around why that whole container probe is supposedly succeeding | 19:15 |
| jimbaker | and then immediately failing when it tries to run tests | 19:15 |
| sigmavirus | jimbaker: I skimmed docker-py's code | 19:15 |
| sigmavirus | I see where it's generating the log messages | 19:15 |
| sigmavirus | I want to see how we're using it | 19:15 |
| sigmavirus | It seems like CI boxes are different | 19:16 |
| jimbaker | there is quite a mix in bindep | 19:17 |
| thomasem | docker-py has historically been pretty sensitive to docker versions, fyi. | 19:21 |
| thomasem | I remember getting back different responses from different versions of docker with it | 19:21 |
| thomasem | Which, naturally, broke things. | 19:22 |
| sigmavirus | https://github.com/docker/docker-py/blob/997e583ea9a7b33113edd91d5bee73d25d720448/docker/client.py#L30..L40 | 19:26 |
| jimbaker | thomasem, that would make sense, i think docker-py is just a thin wrapper around the docker REST API (or whatever variant we call it, given it's streaming) | 19:26 |
| * sigmavirus needs something with some power to it | 19:26 | |
| sigmavirus | so my xenial box is using 1.11.2 | 19:26 |
| sigmavirus | and docker-py 1.10.6 | 19:26 |
| thomasem | Mine's running 1.13.0 | 19:26 |
| sigmavirus | I'm addign logging to our functional tests | 19:26 |
| sigmavirus | and upating our functional test script | 19:27 |
| sigmavirus | Going to nix Andreas' suggestion about our bindep file thoug | 19:27 |
| sigmavirus | gate should work as soon as our change merges | 19:27 |
| sigmavirus | to run the functional tests you'll need to do `check experimental` | 19:28 |
| sigmavirus | I'm going to hack on these till they work | 19:28 |
| jimbaker | sigmavirus, awesome | 19:28 |
| *** VW has joined #craton | 19:29 | |
| sigmavirus | so yeah, I'll let you know when to start rechecking other stuff | 19:30 |
| sigmavirus | the change is almost through the check queue | 19:31 |
| sigmavirus | It's in the gate | 19:33 |
| sigmavirus | jimbaker: thomasem go ahead and recheck | 19:42 |
| jimbaker | sigmavirus, so abandoning https://review.openstack.org/#/c/435038/ ? | 19:42 |
| sigmavirus | absolutely not | 19:43 |
| openstackgerrit | Ian Cordasco proposed openstack/craton master: Fix up functional testing https://review.openstack.org/435038 | 19:43 |
| sigmavirus | I'm going to continue using that to fix our functional CI | 19:43 |
| sigmavirus | but until then | 19:43 |
| sigmavirus | just recheck the rest of your stuff | 19:43 |
| jimbaker | ok, got it | 19:43 |
| sigmavirus | so the Gate is using docker 1.12.3 and API 1.24 | 19:47 |
| sigmavirus | ah but now I see: http://logs.openstack.org/38/435038/5/check/gate-craton-tox-functional-ubuntu-xenial/0576d3e/console.html#_2017-02-16_19_45_14_199275 | 19:47 |
| openstackgerrit | Ian Cordasco proposed openstack/craton master: Fix up functional testing https://review.openstack.org/435038 | 19:52 |
| openstackgerrit | Ian Cordasco proposed openstack/craton master: Fix up functional testing https://review.openstack.org/435038 | 20:03 |
| openstackgerrit | Thomas Maddox proposed openstack/craton master: RBAC-Lite for generic resources implementation https://review.openstack.org/434976 | 20:29 |
| thomasem | sigmavirus: I see that whenever there's permissions issues on the Docker socket. | 20:30 |
| sigmavirus | Yeah | 20:31 |
| sigmavirus | Best I can find online is to log out and log back in but lol | 20:31 |
| thomasem | Lol! I've always had to restart the Docker daemon after setting up perms. | 20:31 |
| thomasem | Like, I've always had to add my user to the docker group and then restart the daemon... no idea why? | 20:31 |
| thomasem | Seems like it ought to work as soon as you're in the group that has access to the socket. | 20:32 |
| sigmavirus | thomasem: exactly, doesn't seem to work | 20:53 |
| thomasem | Lame | 20:53 |
| *** VW has quit IRC | 20:57 | |
| openstackgerrit | Ian Cordasco proposed openstack/craton master: Fix up functional testing https://review.openstack.org/435038 | 20:57 |
| jimbaker | thomasem, i think we will implement RBAC before we mixin vars support for users | 21:00 |
| jimbaker | which will avoid the bootstrap problem in say this code, https://review.openstack.org/#/c/434976/4/tools/docker_run.sh | 21:00 |
| jimbaker | (so it's fine for now!) | 21:01 |
| thomasem | jimbaker: well, good thing, because... yeah :) Kinda need that to preserve what we had going on already until we do fix RBAC. :P | 21:01 |
| thomasem | Like proper fix. | 21:01 |
| jimbaker | yes | 21:02 |
| thomasem | Sounds good to me. | 21:02 |
| jimbaker | each iteration of work shows how important this is; fwiw, see https://blueprints.launchpad.net/craton/+spec/craton-rbac-support | 21:02 |
| openstackgerrit | Ian Cordasco proposed openstack/craton master: Fix up functional testing https://review.openstack.org/435038 | 21:03 |
| jimbaker | i made some additional progress on a local branch but foolishly didn't post it WIP. will get back to it | 21:03 |
| thomasem | Yeah, the hoop jumping is going to get out of hand. | 21:08 |
| openstackgerrit | Ian Cordasco proposed openstack/craton master: Fix up functional testing https://review.openstack.org/435038 | 21:11 |
| openstackgerrit | Ian Cordasco proposed openstack/craton master: Fix up functional testing https://review.openstack.org/435038 | 21:16 |
| jimbaker | thomasem, ok, the rbac-lite looks great, and i approved the change | 21:19 |
| *** VW has joined #craton | 21:19 | |
| jimbaker | we will presumably have to recheck as necessary | 21:19 |
| jimbaker | thanks for some very nice cleanup there on that bug | 21:19 |
| jimbaker | and i'm going to get in my run. my really really late lunch run | 21:20 |
| thomasem | jimbaker: Thanks a 'mil! | 21:31 |
| *** acabot has quit IRC | 22:48 | |
| farid | any clue what I'm missing here? https://gist.github.com/faridsaad/c793828f8b5fad2d141fcf3a051e5f35 | 22:50 |
| farid | logs are empty :S | 22:50 |
| farid | allright don't know why it's not logging the error to the file, but here's what I got | 22:55 |
| farid | HTTP/1.0 400 BAD REQUEST\r\nDate: Thu, 16 Feb 2017 22:54:52 GMT\r\nServer: WSGIServer/0.2 CPython/3.5.2\r\nContent-Type: application/json\r\nContent-Length: 108\r\nx-openstack-request-id: req-b2b95017-7727-4c65-a6ff-e32d985ab1da\r\n\r\n{\"errors\": [\"Additional properties are not allowed ('sort_dir' was unexpected)\"], \"message\": \"Bad Request\"}\n | 22:55 |
| farid | so I guess the client is adding this sort_dir stuff in... | 22:56 |
| farid | yeah using curl I got an empty host-list which is what I expect... | 22:59 |
| *** acabot has joined #craton | 23:01 | |
| *** mdorman has quit IRC | 23:42 | |
| *** mdorman has joined #craton | 23:50 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!