Tuesday, 2017-02-14

*** VW has quit IRC00:32
*** VW has joined #craton00:44
*** VW has quit IRC00:46
*** `_^gk`_^1wm`_^ has joined #craton00:52
*** `_^gk`_^1wm`_^ has quit IRC00:54
*** VW has joined #craton02:43
*** VW has quit IRC02:48
*** david-lyle_ has joined #craton03:56
*** david-lyle has quit IRC03:58
*** VW has joined #craton05:10
*** VW has quit IRC05:11
*** VW has joined #craton05:11
*** VW has quit IRC05:22
*** VW has joined #craton07:38
*** VW has quit IRC08:04
*** VW has joined #craton15:06
*** Syed__ has joined #craton15:28
*** jovon has joined #craton15:31
thomasemDoh! I remember what the problem with the functional tests is now.15:46
jimbaker?15:46
thomasemNot going through the API for initial project setup didn't set up the proper variable association15:46
thomasemI guess somewhere along the way my fix for that got lost in some refactoring :(15:47
thomasemAlright, fixing again15:47
thomasemThere were these weird failures that seemed to only happen in the functional tests, so it had to be related to how they're set up.15:47
thomasemLocally it worked as expected, but when the functional test came along, it bombed.15:47
thomasemAnd that made me think of resolution order, since the only thing that changed in my patch was adding the project variables stuff, but the failures were on hosts tests.15:48
jimbakeryep, that would make sense15:48
* thomasem puts on spelunking gear15:48
jimbakerbecause the functional tests are the only ones that really exercise some of those code paths. i think we can do a better job with the db tests fwiw15:48
jimbakerthere's no reason in principle they cannot15:49
thomasemYeah. The out-of-band setup that reaches into the DB is a bit of a no-no, too.15:49
jimbakeryes. don't do that15:49
jimbakerat best reserve it for "bootstrap" or whatever you ended calling it. or will call it15:49
thomasemThat's just how it's doing things in master15:49
thomasemNeed to refactor that whole piece15:50
jimbakerso the change you posted is still WIP?15:50
thomasemfor a moment15:50
jimbakerhttps://review.openstack.org/#/c/427777/15:50
jimbakerack. will hold off then15:50
thomasemFeel free to code review, it's mostly done.15:50
thomasemJust know functional tests are going to bomb until I fix this15:50
thomasemjimbaker: ^^15:50
thomasemAny issues I can get in the next revision I push up will save some time15:51
thomasemissues I can get resolved*15:51
sigmavirusthomasem: jimbaker did y'all come up with priorities?15:53
thomasemsigmavirus: it's on the etherpad https://etherpad.openstack.org/p/cmdb_prototype_meeting_2017_02_0915:54
thomasemjimbaker: did you ever get a bug going for that CLI work?15:54
thomasemWanted to link it in the etherpad15:55
thomasemI see this one https://bugs.launchpad.net/python-cratonclient/+bug/165911015:56
openstackLaunchpad bug 1659110 in Craton's Python Client "Variables should be available from cli" [High,New] - Assigned to Sulochan Acharya (sulochan-acharya)15:56
thomasemjimbaker: Okay, it's gtg now!16:17
thomasemAll fixed up16:17
*** david-lyle has joined #craton16:26
*** david-lyle_ has quit IRC16:29
Syed__thomasem: just checked16:31
Syed__still failing some tests :/16:31
thomasemSyed__: hmmm?16:32
jimbakerthomasem, now assigned to me. i will add some additional notes to it16:32
thomasemjimbaker: awesome, thanks!16:32
thomasemSyed__: what has failing tests?16:32
Syed__let me put it in pastebin16:34
Syed__a min16:34
thomasemgit-harry: FYI, the reason I was fixing that inconsistent ID type for project was because it was causing all sorts of problems in tests and it got more painful working around the inconsistency than just fixing the problem.16:35
Syed__thomasem: https://gist.github.com/ahsan518/6f1dd188a4fa60e2ddf50d8b7b9a9d5e16:35
thomasemSyed__: that looks like a Docker problem on your end.16:36
Syed__docker16:36
thomasemCheck your daemon and be sure you can actually run the tests on, say, master16:36
Syed__let me check again16:36
thomasemAttributeError: 'APIV1RegionTest' object has no attribute 'service_ip'16:36
thomasemTells me that something got messed up when trying to set up your Docker daemon for the functional tests. That's where `service_ip` gets set.16:37
thomasemBy the way, this project vars patch is blocking Syed16:39
thomasemSo, that's why you're probably going to see me get annoying about reviews for it.16:39
thomasemThat and it's one of our priority items for ze week :P16:39
thomasemGetting really good at these keyboard shortcuts for Gerrit, though.16:40
Syed__thomasem: works fine now, looks good. Thanks16:42
thomasemSyed__: You bet! And thank you for taking a look at it. :)16:43
jovonwill notes be included in networks or contained to regions cells and devices?16:52
Syed__jovon: No, since it is not present in the schema16:55
Syed__jovon: https://github.com/openstack/craton/blob/master/craton/api/v1/schemas.py16:56
git-harrythomasem: okay, might have been nice to have done that separately16:57
thomasemHonestly? The way code reviews have been working on this team made me feel that would probably push this out by days.16:57
thomasembut, I do agree wholeheartedly with the goal there and want to do it that way.16:58
thomasemfarid: this meeting is actually on Vidyo in Craton-recordable17:00
faridah thanks thomasem17:01
*** VW has quit IRC17:11
*** VW has joined #craton17:11
jovonhttp://docs.openstack.org/contributor-guide/api-guides.html17:11
sigmavirusthomasem: so I have it narrowed down to one last functional test17:24
sigmaviruscombination of pagination + vars filtering17:24
sigmavirusgetting a 200 with 0 hosts17:24
sigmavirusalso, should we time-box these discussions?17:25
sigmavirusthomasem: inspired that by timeboxing Jim's standup17:27
sigmavirus:)17:27
sigmavirusI LEARNED IT FROM WATCHING thomasem17:28
sigmavirusjimbaker: so, we're not getting 500 out of paginated filtering with vars, but I haven't yet determined why we get 0 hosts back17:56
sigmavirus:)17:56
jimbakerlove the timeboxing, and thanks!17:56
jimbakersigmavirus, re pagination - it certainly simplifies the work if the resulting collection always has zero items17:57
sigmaviruslol17:57
sigmavirusjimbaker.enable_helper_mode()17:57
jimbakerbut have you considered further optimizing the query by just avoiding any sqlalchemy to begin with? ;)17:58
thomasemlol sigmavirus17:59
sigmavirushm the query just works fine18:00
* sigmavirus just stepped through the query generation and filtering18:00
* sigmavirus glares18:00
sigmavirussure, when I debugger through that monstrosity it works fine18:01
* sigmavirus begins to suspect something is wrong with the functional test18:05
jimbakerdoubting the test is a good place to start18:09
sigmavirusyeah, when I just use requests, I can get filtering by vars + pagination working fine18:20
thomasemWhat's the error you're getting? (I usually comment out the container cleanup in __init__.py and docker logs...)18:28
jimbakeri'm warming up to /v1/devices18:29
jimbakeranyway, RPC meeting. back in 90 min or so18:30
thomasemSame18:30
sigmavirusthomasem: I'm not getting any errors. Everything comes back ok except that there are 0 hosts19:05
thomasemOH sorry, I see it now in the scrollback19:05
sigmavirusalso, my testing box is on RAX pub cloud and ssh'ing to it on the vpn is a headache19:05
sigmavirusand I keep getting dropped19:05
thomasemSerrriously19:06
thomasemMine's been dropping all morning19:06
sigmavirusI should just work with two laptops19:06
thomasemRight.19:17
jimbakerback19:54
jimbakeri looked a bit into devices, just so i can better understand what's going on for supporting /v1/devices. should be very straightforward, with minimal api work, and a bit of change in the client to instantiate the right resource19:58
jimbakerhttps://bugs.launchpad.net/craton/+bug/166469619:58
openstackLaunchpad bug 1664696 in craton "Support GET on a collection of devices" [Undecided,New]19:58
jimbakergit-harry, ^^^20:01
jimbakersigmavirus, thomasem - one thing i noticed in doing this is that queries on /v1/hosts really need to be more lazy20:01
jimbakeri assume this is a common thing with out queries - we too aggressively load the object graph, even when details are not being returned20:02
jimbakeranyway, we should add a tuning bug20:03
jimbakerto at least get at some of the common one. or at least, let's track as work to be done :)20:04
jimbakerhttps://bugs.launchpad.net/craton/+bug/166470720:19
openstackLaunchpad bug 1664707 in craton "SQLAlchemy queries are not lazy enough" [Undecided,New]20:19
sigmavirusFound the bug in the functional tests20:28
sigmavirusjimbaker: let's optimize that when we find it to be a real problem20:28
jimbakersigmavirus, i did mark it as high20:29
jimbakernot critical as a consequence20:29
jimbakerbut retrieving every variable setting for every host that is returned, even with pagination... sounds potentially expensive even if we throw it all away on the output to the rest api20:30
jimbakerespecially since this pruning should be a simple change :)20:30
sigmavirusjimbaker: that's fair20:30
sigmavirusjimbaker: do you have 5 minutes for a video call? I suspect thomasem is out20:31
sigmavirusfarid: you might want to join us if you have a few minutes?20:31
faridsure thing20:32
jimbakerok20:32
jimbakercraton room?20:32
faridsounds good, gimme a min to fill up my cup20:33
sigmavirusyep20:33
sigmavirusDo we want a voting functional gate for Craton20:50
Syed__I guess that would be useful and handy ?20:52
Syed__to check functional testing at gate to verify if we break stuff or not,20:52
sigmavirusSyed__: yeah20:52
sigmavirusWould have caught some breakage with my patch sooner for me because I forgot about `tox -e functional` :020:52
Syed__sigmavirus: just playing around with your patch20:53
Syed__sigmavirus: works fine till now :)20:53
sigmavirusSyed__: ssssh don't jinx it20:53
Syed__yeah, my vote is yes20:53
sigmavirusI won't pull a jimbaker and +2+A it myself ;)20:53
jimbaker;)20:54
git-harryjimbaker: thomasem adding parent links to hosts/network-devices is going to conflict with the pagination patch from sigmavirus. If it's important to get his patch in, it could do with some reviews today otherwise basing my patch off his will likely cause delays.21:32
git-harryDelays in the critical items that is.21:33
jimbakergit-harry, let's get sigmavirus' patch in21:42
jimbakerhe's suffered enough21:43
jimbakergit-harry, if we haven't done so, we do need a bug re PUT for device_id update21:44
git-harryjimbaker: parent_id?21:45
jimbakersorry, parent_id21:46
jimbaker:)21:46
git-harryOkay, I can sort that out tomorrow21:48
jimbakergit-harry, yeah, i think this is the best prioritization21:50
thomasemWhat'd I miss?22:32
thomasemjimbaker, farid, sigmavirus: did I miss anything in your vidyo call?22:34
faridhey thomasem22:34
faridwe were just discussing and creating a milestone in the craton project for the next 3 weeks22:34
*** jovon_ has joined #craton22:35
thomasemOh okay22:35
faridyou can see it here https://launchpad.net/craton/+milestone/v0.1.022:35
*** izaakk_ has joined #craton22:35
Syed__That would be really helpful22:35
thomasemExcellent22:35
thomasemYes, thank you!22:35
faridessentially of what we had discussed we should focus on, just to have a central place to track.22:35
faridjimbaker / sigmavirus anything to add ?22:35
thomasemdo we also need one on the client LP project?22:36
faridgood point, item #2 is for the client huh22:36
thomasemYep, and #3, partially22:36
thomasemSpecifically, 3.322:37
faridcool, sounds like a good idea22:38
faridjumping in impromptu meeting, brb and can check on that22:38
thomasemjimbaker: should we just consider that part of https://bugs.launchpad.net/python-cratonclient/+bug/1659110?22:39
openstackLaunchpad bug 1659110 in Craton's Python Client "Variables should be available from cli" [High,New] - Assigned to Jim Baker (jimbaker)22:39
thomaseminstead of https://bugs.launchpad.net/craton/+bug/164862622:39
openstackLaunchpad bug 1648626 in craton "Projects should support variables, which other entities use in resolution" [Critical,In progress] - Assigned to Thomas Maddox (thomas-maddox)22:39
thomasemsince that's for Craton API22:39
thomasemI don't mind doing the work, but it seems silly to make another bug since it's the same thing, really.22:39
*** sballe__ has joined #craton22:40
*** jovon has quit IRC22:42
*** wirehead_ has quit IRC22:42
*** izaakk has quit IRC22:42
*** sballe_ has quit IRC22:42
*** mhayden has quit IRC22:42
*** jovon_ is now known as jovon22:42
*** izaakk_ is now known as izaakk22:43
*** sballe__ is now known as sballe_22:45
*** klindgren has quit IRC22:46
*** jimbaker has quit IRC22:46
*** mhayden has joined #craton22:50
*** klindgren has joined #craton22:52
*** jimbaker has joined #craton22:52
*** adams.freenode.net sets mode: +o jimbaker22:52
*** VW_ has joined #craton23:22
*** VW has quit IRC23:25
*** VW_ has quit IRC23:26
*** jovon has quit IRC23:42
*** wirehead_ has joined #craton23:57

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