Monday, 2017-03-20

suloo/08:53
*** openstackgerrit has quit IRC09:02
tojuvoneHi09:34
*** openstackgerrit has joined #craton09:35
openstackgerritSulochan Acharya proposed openstack/craton master: Include project_id only where needed  https://review.openstack.org/44693809:35
suloour review process needs a review ... we gotta stop doing this +2 +1 by one person11:37
sulothomasem: jimbaker: git-harry: sigmavirus: ^^11:38
suloactually let me add that to todays meeting agenda11:38
sigmavirussulo: sounds like a good topic to me12:15
sigmavirussulo: re: https://review.openstack.org/#/c/446938/3 how does a user transfer something, e.g.,a host, from one project to another12:15
sulosigmavirus: heh good question, dunno ... maybe root user can do it ?12:20
sulothis is assuming we are not doing that12:20
sulobut that sounds like something that might come up12:20
sigmavirusyeah12:20
sigmavirusI don't think it's something our Rackspace users would need, to be fair12:21
sigmavirusBut I wonder if you had a DC that had different projects, so there's a DC-Operations project that assembles everything12:21
suloyeah12:21
sigmavirusThey own the info until they pass it along to Software Opeations12:21
sigmavirusetc12:21
suloyeah i can see the use case tbh, kinda like hw provisioning is handled in core for example12:22
sulosigmavirus: lets -1 that change, ill add that to todays agenda too12:23
sigmavirusRight12:23
sigmavirussulo: no I think that change is a good start12:23
sigmavirusIt just made me think about that use-case12:23
sulosigmavirus: rgr12:23
sigmavirusactually12:23
sigmavirusI do want tests for that change, so I will -1 =P12:23
sigmavirusBut not because I want to chat about the use case of changing project-id =P12:23
sulook12:24
sigmavirus(I do want to chat about that, but I wouldn't -1 this change because of it =P)12:25
sigmavirusthomasem: btw, jimbaker doesn't seem to have replied to everyone on the MySQL 5.7 thread12:35
*** VW has joined #craton13:03
thomasemo/13:04
sigmavirus\o13:07
thomasemsulo: Totally agreed on the reviews. That +2/+1 by a single person was in practice when bandwidth was extremely low and we needed things to happen quick... but even that required at least some +1s. It was more because core folks weren't present, and even that made me a bit uneasy.13:07
thomasemsigmavirus: I don't think jimbaker replied to the thread. The sentiment I was referring to was expressed by him in this channel.13:08
sigmavirusah okay13:08
* sigmavirus misunderstood13:08
thomasemNo problemo. :P13:08
thomasemYeah, it was his first comment when you brought it up originally.13:08
sigmavirusAh, I think I missed that13:09
sigmavirusWhen I'm concentrating on stuff, I lose track of this channel13:09
thomasemsigmavirus: Hmmm, I thought he had said it in channel, but reading through the logs, I don't see it. Well, I could have remembered it incorrectly. Either way, I'm sure he'll provide his thoughts soon enough.13:16
sigmavirusNo worries13:16
thomasemHow was the weekend??13:31
sigmavirusproductive13:31
thomasemExcellent!13:31
sigmavirusGot a good start on some documentation for a project that's had no documentation for almost 3 years13:32
sigmavirusheh13:32
sigmavirusHow was your weekend?13:32
thomasemOh man13:32
thomasemProductive as well. Started working on my dogs social skills downtown a bit; going to make that more of a routine for her, and got a bunch of brush cleaning, trimming, mowing, etc. done so the yard isn't a jungle anymore.13:35
thomasemSo Texas, but there was this big pile left over from a bunch of tree trimming and removal when we first moved in where I was consistently finding snakes (not the friendly kind).13:36
thomasemsigmavirus: which project?13:37
sigmavirusthomasem: but snakes are always friendly! =P13:38
sigmavirushttps://github.com/sigmavirus24/rfc398613:38
thomasemOh awesome13:39
thomasemsigmavirus: Lol, so very.13:40
sigmavirusThere wasn't much to document previously13:41
sigmavirusBut I've revamped it a bit (in a backwards-compatible way in the 1.0 PR) to make it a much better overall experience for pip/warehouse and so I'm bothering to document it now13:41
sigmavirusAlso, figure that if OpenStack ever upgrades/reads the release notes, they might change the nonsense they use it for too13:41
thomasemHah, gotcha. Could put a big section about how not to use this thing. :P13:47
fsaadmorning all13:47
* fsaad sads for sigmavirus13:47
thomasemhey fsaad13:47
fsaadhi thomasem13:48
fsaadhey I have today's craton meeting cancelled but no replacement invite for it13:48
fsaadthe 10am ct13:48
thomasemfsaad: lemme forward13:49
thomasemfsaad: ahh... I didn't get jimbaker's invite for that one.13:49
fsaadthat's two of us! :)13:49
thomasemI was trying to consolidate by canceling mine since I thought jimbaker had sent something to replace it.13:49
fsaadplan on intro'ing mr anonymike since he joins team today!13:49
anonymikehey guys!13:50
thomasemDidn't want you to do this: https://i.imgur.com/9Kr6AAH.gif when you opened your calendar.13:50
fsaadlol13:50
thomasemhey anonymike13:50
thomasemWelcome to the party13:50
anonymikeQuick intro for some of you. My name is Michael Porras, I'm a developer at Rackspace who worked on our public cloud inventory system and I'm starting on the Craton team today :D13:51
* fsaad hands sombreros to each member13:51
anonymikethanks thomasem, excited to be here13:51
* anonymike maracas and empanadas 13:51
fsaadlol13:51
openstackgerritSulochan Acharya proposed openstack/craton master: Create user under project_id passed in request  https://review.openstack.org/44697213:54
thomasemMan, just ate breakfast and now I'm hungry again.13:59
fsaadtwo breakfast mondays!13:59
thomasemLol13:59
thomasemWe've had one breakfast, yes. What about second breakfast?14:00
suloanonymike: good stuff, welcome !!14:00
anonymikeThanks sulo!14:03
openstackgerritSulochan Acharya proposed openstack/craton master: Create user under project_id passed in request  https://review.openstack.org/44697214:05
sigmavirusthomasem: are you perhaps part hobbit?14:13
*** valw has joined #craton14:22
openstackgerritIan Cordasco proposed openstack/python-cratonclient master: Add Betamax for testing  https://review.openstack.org/44216514:25
*** valw has quit IRC14:26
sulolooks like rax mirror is down causing failures in gate14:28
thomasemsigmavirus: lol, it would explain some things, wouldn't it?14:32
thomasemShould we specify which version of Ubuntu and CentOS/Fedora we're talking about in INSTALL.rst? (as I'm updating docs for MySQL 5.7)14:33
thomasemI'll at least add a note for the version used when creating the documentation.14:35
thomasemI didn't know about betamax until these patches, and I like it.14:38
sigmavirusWe've been using 16.04 Ubuntu14:38
sigmavirusDunno if we've been using a version of fedora although I'd advocate 2514:38
thomasemOkay, I'll use that in my testing before updating those docs.14:39
thomasemCentOS 7, I imagine?14:39
anonymikeI had a dev environment setup in inova but I'm weary that I'll lose access soon. Y'all have any tips on how and where to setup my dev environment? Previously, I followed the "read the docs" article is that still the authority?14:43
anonymikeI have a completely comped account on our public side. I was just going to rebuild a fairly large server there and run through the setup again14:45
suloanonymike: yeah i flip between public cloud vm and my laptop14:48
suloboth work well14:48
thomasemsulo: Linux laptop?14:48
sulothomasem: no, mac14:49
thomasemReally? How do you run functional tests?14:49
thomasemOr do you do that on your VM?14:49
anonymikeohh, interested in local dev as well14:49
suloi dont14:49
suloits on vm14:50
suloyep14:50
thomasemGotcha14:50
thomasemanonymike: yeah, the functional tests can't run locally on OS X because Docker for Mac doesn't set up networking in the way the tests expect.14:51
jimbakerthomasem, you got the invite for the tues sprint planning/retrospective meetings, right? (outlook says you have accepted)14:51
jimbakerfsaad, i will resend those to you14:52
thomasemjimbaker: yes, but not Monday & Thursday.14:52
jimbakermonday?14:52
* sigmavirus might be late to the meeting today14:52
thomasemBack up.14:52
thomasemjimbaker: I thought you had a series for each of the Craton meetings - Monday, Tuesday, and Thursday.14:52
thomasemAnd that you were going to add us to those series.14:52
thomasemIs that not the case?14:52
jimbakerok, i will send that to you as well. i do have it in my calendar, but we have been tracking through openstack's meeting cal for the monday one14:53
jimbakeri don't have it as a meeting invite for monday, i guess because of that reason14:53
thomasemAhhh, gotcha... Okay, so I think we misunderstood each other. I thought we had decided to generate .ics attachments for all Craton meetings so we're all using the same format.14:54
fsaadjimbaker: thanks!14:54
jimbakerthomasem, yes, that's a good next step14:55
jimbakerfor publication14:55
thomasemCool. Thanks for handling that. Sorry for the confusion.14:55
jimbakerbut first get it done in exchange14:55
jimbakercool14:55
thomasemRight :)14:55
anonymikemind sending me those invites jimbaker?14:55
* thomasem has lots of reviews to do today.14:55
jimbakeranonymike, sure14:55
anonymikethanks!14:56
fsaadhm, not sure what you sent but can't seem to add to calendar , at least via webmail14:56
jimbakerfsaad, hmmm14:57
jimbakeranonymike, forwarded the series to you14:57
thomasemOh the woes of trying to straddle internal exchange and external calendars...14:57
anonymikegot em14:57
jimbakerhopefully you will not have a problem14:57
jimbakerfsaad, curiously i only use the web interface for exchange14:57
jimbakerit seems to work well enough14:57
fsaadyeah me too, will try in a sec, says email has one attachment but doesn't properly show it14:58
thomasemWeird14:58
jimbakerbrb - i'm going to be a few minutes late for the meeting14:58
fsaadyeah don't wait for me either plz, I'll review transcript14:58
* thomasem is on track to meet with himself for the first 5 minutes...14:59
thomasem:P14:59
fsaadlol14:59
anonymikeyeah, "weekly craton irc meeting" not adding. any issues with manually setting that up on my cal?14:59
thomasemanonymike: FYI, it's in #openstack-meeting-4 right now.15:00
anonymikewew thanks15:01
thomasemSure thing!15:10
* thomasem goes to make tea15:45
thomasemback15:51
jimbakeranonymike, no issues with monday's meeting, since we track here - http://eavesdrop.openstack.org/#Craton_Team_Meeting15:53
jimbakerwith convenient ICS file :)15:53
anonymikenice :)15:53
tojuvoneHi, try to catch tomorrow meeting and say couple of words about Milan15:54
antonymanonymike: i had to double click it to import, it didnt auto import in outlook15:55
anonymikemine opens to a dead event link15:56
anonymikeantonym: in owa?15:56
anonymike(I just made my own recurring reminder)15:56
antonymoutlook client, it imported once i opened it15:56
sigmavirusanonymike: yeah, outlook on the desktop works well with this kind of workflow15:57
sigmavirusI understand people liek OWA better though15:57
sigmavirusOutlook App has better shortcuts though15:57
antonymoutlook used to be awful on mac but i think it's gotten a ton better15:58
anonymikeI haven't used the native app in over a year. Maybe I'll give it another shot15:59
sigmavirusGotta get down on monday16:00
*** Syed__ has joined #craton16:02
*** VW_ has joined #craton16:12
thomasemanonymike: https://gist.github.com/jimbaker/e74a7b98bc60519033fd455a22163ad2#gistcomment-2023979 and https://gist.github.com/jimbaker/9089e381af38d9db6bc4af59baed6d4716:15
*** VW__ has joined #craton16:16
*** VW has quit IRC16:16
*** VW__ has quit IRC16:17
*** VW_ has quit IRC16:17
*** VW has joined #craton16:17
openstackgerritgit-harry proposed openstack/craton master: Ensure JSON responses result from failure  https://review.openstack.org/44758016:22
anonymikethanks thomasem16:26
thomasemYou bet!16:26
openstackgerritgit-harry proposed openstack/craton master: Ensure JSON responses result from failure  https://review.openstack.org/44758016:30
thomasemlunch16:31
fsaad+116:31
*** VW has quit IRC17:00
*** VW has joined #craton17:01
openstackgerritgit-harry proposed openstack/craton master: Ensure JSON responses result from failure  https://review.openstack.org/44758017:24
thomasemsigmavirus: so, for recording betamax stuff, what's your process? Deploy the master craton with `craton-dbsync bootstrap`, set the env vars, and then run tox -e py35?17:31
thomasemThe env vars being CRATON_DEMO_USER, CRATON_DEMO_TOKEN, CRATON_DEMO_PROJECT17:32
sigmavirusthomasem: that's what I would do, yes17:37
sigmavirusThat's not exactly what I was doing, but it works just as well that way17:37
thomasemWhat were you doing? (I'm getting 401 and about to dig in)17:37
sigmavirusWell it was simpler before setting those vars17:38
sigmavirus:)17:38
sigmavirusAlso simpler with an older version of craton that I had with the demo user bootstrapped =P17:38
thomasemHaha, yeah. The bootstrap part adds another layer of complexity here.17:38
sigmavirusI'd make sure the CRATON_... env vars are being passed through17:38
thomasemYeah, that's what I'm wondering.17:38
sigmavirusI'm not unconvinced I didn't bungle that syntax in the tox.ini17:38
thomasemWondering if the defaults are overriding.17:39
thomasemYep17:39
thomasemIt's all good. I'll see what I come up with here and report back.17:39
jimbakerthomasem, for setting up those env vars, the wrapper code is a good start17:40
jimbakermight be a nice addition to https://gist.github.com/jimbaker/9089e381af38d9db6bc4af59baed6d4717:41
jimbakerwhich i'm going to start thinking about where it might actually live, since it seems to be broadly applicable17:41
jimbakercould just be an addition to our tools/ dir17:42
thomasemSounds good. Yeah, handy dev tooling17:42
jimbakeri need to get that sed/tempdir working, and it should be good for universal usage17:43
thomasemAhhhh, so this isn't going to work. We need demo credentials to be added specifically because it's recorded.17:55
thomasemSo, we actually DO need a demo user with known creds and project for this recording to work properly.17:55
thomasemOtherwise we run the risk of checking in recordings and then when it tries to run against that recording with different credentials (generated by the boostrap command), you'll get 401s.17:56
thomasemsigmavirus: ^^17:56
sigmavirusthomasem: so let me look into someting17:57
thomasemmmkay17:57
sigmavirusI have a trick of doing this with Betamax that doesn't have keystoneauth in between it and me though17:57
sigmavirusbasically https://github.com/sigmavirus24/github3.py/blob/develop/tests/conftest.py#L20 and https://github.com/sigmavirus24/github3.py/blob/develop/tests/conftest.py#L2717:57
sigmavirusthat will negate what you're seeing, but I'm surprised you're seeing that honestly17:58
thomasemYou recorded with the old demo user that used to be there, now that we're using dbsync bootstrap it sets up a new user with new creds.17:59
thomasemthe demo user no longer is created for you17:59
thomasemMaybe I'm misunderstanding18:00
thomasemBut, it makes sense to me that if the recording is all according to one set of credentials and a different project, and then you try to make calls against that cassette, it makes sense to me that a 401 would happen.18:00
thomasemthen you try to make calls against that cassette with different credentials and project(s)*18:01
sigmaviruswill need to re-record everything though18:01
sigmavirusso, I'd understand that you would be seeing issues if we were matching against headers as well18:02
thomasemBut that's my concern. dbsync bootstrap creates a new user every time18:02
thomasemthe only thing that stays the same is the username18:02
thomasemthe token and project ID are different18:02
sigmavirusSure, so, the thing I want to do will make that not a problem18:02
thomasemGotcha18:02
sigmavirusBut without matching against headers, the existing tests should continue to work18:02
thomasemBut, what if we just use the user API and the root creds that get passed in to make the demo user for the tests?18:02
sigmavirusoh wait18:02
sigmavirusdoes keystoneauth talk to the server out of band or something?18:02
sigmavirusthomasem: eh, we could do that. This way isn't much work and will prevent people from recording personal credentials, for example18:03
thomasemGotcha18:03
sigmavirusIt's how I record tests against GitHub's actual API without leaking my creds18:03
thomasemOkay, yeah. Then if we can essentially set some specific credentials (demo) for recordings18:04
thomasemAnd always use those when talking to the cassettes, we should be good18:04
thomasemis that what you're doing?18:04
thomasemWhat I was doing was deleting one of the cassettes because I wanted to see it record one, and that's how I ran into this.18:05
sigmavirusyep18:05
sigmavirusThat's exactly what I'm doing18:05
thomasemCool18:05
sigmavirusWill have to verify it'll work for us as the keystoneauth bit in there might complicate things slightly18:05
thomasemAnd it can just inject those at recording time, without needing that actual user/project on the API?18:05
thomasemGotcha. Okay, I'll let you run with that for a bit and try again when I see the changes go in?18:06
thomasemsigmavirus: btw, the credentials were not getting passed by tox. I had to sub, for instance, 'demo' for '{env:CRATON_DEMO_USERNAME:demo}' to get it to pass existing env vars through.18:08
thomasem   CRATON_DEMO_USERNAME={env:CRATON_DEMO_USERNAME:demo}18:08
thomasemin tox.ini18:08
thomasemRather, I had to sub '{env:CRATON_DEMO_USERNAME:demo}' for 'demo'18:09
thomasem:)18:09
sigmavirusahhh right18:17
thomasemjimbaker: so what's left on https://review.openstack.org/#/c/441644/?18:17
* sigmavirus will incorporate that too18:17
jimbakerthomasem, just some testing18:58
thomasemjimbaker: okay, cool18:58
jimbakerit's mostly tested as is by existing testing, but there are a couple of things where i made the constraints stronger to reflect what the model states. so this is about deletes18:59
jimbakersimilar to what we saw with host deletes18:59
anonymikeGot everything set up and working. (I think lol) I found a few changes to make for the docs. Currently brushing up on Gerrit19:19
thomasemanonymike: lol, good! Yeah, looking forward to reviewing the changes. :) Feel free to create a bug for them in LP: https://bugs.launchpad.net/craton19:43
thomasemThen you can use https://docs.openstack.org/infra/manual/developers.html#working-on-bugs to allow commits to push the bug through its lifecycle in LP19:44
thomasemanonymike: ^^19:44
anonymikeah cool19:44
thomasemPointing specifically at Closes-Bug/Partial-Bug19:44
thomasemWhat version of CentOS has our INSTALL.rst been tested with?19:48
thomasemIf any... it actually looks like it was tested on Fedora, but not CentOS.19:48
thomasemConsidering package names and such. Might need to do a separate section for CentOS, if that's something we want to document.19:48
thomasemI'19:52
thomasemI've tried 6 and 7 and once I get to the yum install python3-devel step, the packages are different for default repositories.19:54
antonymi can give 7 a try and see about updating docs19:58
antonymwanted to get this playbook running on centos 7 too19:59
thomasemantonym: awesome. Then I'll focus on Fedora/Ubuntu for MySQL 5.7 install instructions.20:02
thomasem:)20:02
thomasemFedora 25 / Ubuntu 16.0420:02
antonymguess everything is python34-devel now20:08
thomasemGotcha20:09
antonymthomasem: also no mysql-server in default repo, have to add community it looks like20:13
thomasemsaddyface20:13
thomasemOkay20:13
antonymthanks oracle :/20:15
*** anonymike_ has joined #craton20:16
*** pwnall138 has joined #craton20:16
*** anonymike has quit IRC20:18
*** anonymike_ is now known as anonymike20:18
*** palendae_ has joined #craton20:22
thomasemyeah :\20:22
*** palendae_ is now known as palebndae20:23
antonymgoing to give centos it's own section20:28
*** zz_pwnall1337 has quit IRC20:29
*** palendae has quit IRC20:29
*** palebndae is now known as palendae_20:29
antonymi'm guessing you'd still use the MySQL-python for the interface?  don't really see any python3 specific packages for that20:29
sigmavirusthomasem: uh, how do you get user credentials out of master now?20:31
antonymor is the mysql-connector-python ?20:31
sigmavirusantonym: I think it is MySQL-ypthon20:31
thomasemsigmavirus: it gets dumped in the logs :\20:32
sigmavirusI think it's Py2/Py3 compatibile20:32
thomasemWhen the service starts.20:32
antonymah, ok, wasn't sure if there was a specific py3 version20:32
sigmavirusah20:32
sigmavirusseems i needed to build it a second time20:32
antonymwould having the ability to pass the user/pass/project during bootstrap be possible?20:33
antonymso that it's not so random?20:33
thomasemI see no reason we couldn't add that.20:33
thomasemWould make this easier.20:34
sigmaviruswhat fresh hell is this? I can't connect to craton running it in docker for mac?20:34
sigmavirusThis was working last week20:34
sigmavirusto the VMs20:34
antonymthomasem: yeah, keep finding my creds change lol20:34
thomasemantonym: Yeah, lol20:34
thomasemIt's a matter of updating the migration utils and dbsync to accept those parameters.20:34
thomasemAnd, otherwise, generating for you.20:35
sigmavirusantonym: that's for your security ;)20:35
sigmaviruskeeps you on your toes20:35
thomasemLol20:35
antonymhah, it's working too well :)20:37
sigmavirusantonym: I think we could better secure those credentials by double rot13'ing them before printing them out, thoughts? ;)20:38
antonymhah20:39
jimbakerxor?21:00
jimbakeragreed, being able to specify the project_id and api_key make a lot of sense. most testing can ignore the randomization by extracting, but changing the betamax plays each time doesn't make sense21:02
openstackgerritAntony Messerli proposed openstack/craton master: [WIP] Updates documentation for setting up a dev env  https://review.openstack.org/44769721:24
openstackgerritAntony Messerli proposed openstack/craton master: [WIP] Updates documentation for setting up a dev env  https://review.openstack.org/44769721:24
openstackgerritAntony Messerli proposed openstack/craton master: [WIP] Updates documentation for setting up a dev env  https://review.openstack.org/44769722:00
antonymthomasem: went ahead and tossed some additions for ubuntu and fedora, and reformatted a bit22:01
anonymikecan I +1? those were some of the changes I was going to add lol  re the pip3 shenanigans22:09
anonymikeor does WIP mean to hold off :/ hue22:13
jimbakeranonymike, so comments are welcome during WIP, but probably best to hold any actual +1/-1 (given WIP means explicitly "not ready")22:16
jimbakerwe might give a -2 if the direction is completely pointless. i don't know, it hasn't happened22:17
anonymikegotcha, I zeroed out22:18
antonymyeah, probably still needs some work but just wanted to get it up there22:23
anonymikeCool, then I'll hold off on mine :)22:44
*** VW has quit IRC22:47

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