Monday, 2016-04-04

« Sunday, 2016-04-03 Index Tuesday, 2016-04-05 »
*** EricGonczer_ has joined #openstack-dns00:49
*** EricGonczer_ has quit IRC01:37
*** ducttape_ has joined #openstack-dns01:54
*** ducttape_ has quit IRC01:55
*** ducttape_ has joined #openstack-dns01:55
*** ducttape_ has quit IRC02:33
*** ducttape_ has joined #openstack-dns02:34
*** ducttape_ has quit IRC02:34
*** ducttape_ has joined #openstack-dns02:34
*** ducttape_ has quit IRC02:57
*** ducttape_ has joined #openstack-dns02:58
*** ducttape_ has quit IRC02:58
*** puck has quit IRC04:07
openstackgerritTin Lam proposed openstack/designate-dashboard: Update requirement for Horizon in stable/mitaka  https://review.openstack.org/30093004:11
*** puck has joined #openstack-dns04:15
*** Alex_Stef has joined #openstack-dns04:26
*** jasonsb has joined #openstack-dns04:36
*** Alex_Stef has quit IRC04:53
*** Alex_Stef has joined #openstack-dns04:53
*** jasonsb has quit IRC05:09
*** jasonsb has joined #openstack-dns05:13
*** jasonsb has quit IRC05:24
*** jasonsb has joined #openstack-dns05:27
*** Alex_Stef has quit IRC06:33
*** nyechiel_ has joined #openstack-dns07:02
*** rudrajit has quit IRC07:04
-openstackstatus- NOTICE: Gerrit is going to be restarted due to bad performance07:13
*** jschwarz has joined #openstack-dns07:13
*** adreznec has quit IRC07:39
*** adreznec has joined #openstack-dns07:41
*** fawadkhaliq has joined #openstack-dns07:54
*** fawadkhaliq has quit IRC08:10
*** _shaps_ has joined #openstack-dns08:14
*** Alex_Stef has joined #openstack-dns08:17
*** mdavidson has quit IRC08:39
*** rudrajit has joined #openstack-dns08:45
*** rudrajit has quit IRC08:49
*** jordanP has joined #openstack-dns09:20
*** kei_yama has quit IRC09:25
*** mdavidson has joined #openstack-dns09:35
*** GonZo2K has joined #openstack-dns10:00
*** GonZo2K has quit IRC10:30
*** openstackgerrit has quit IRC10:48
*** openstackgerrit has joined #openstack-dns10:48
*** ChanServ sets mode: +v openstackgerrit10:48
*** krotscheck_vaca is now known as krotscheck11:13
*** jordanP has quit IRC12:00
*** EricGonczer_ has joined #openstack-dns12:07
openstackgerritKiall Mac Innes proposed openstack/designate: Initial layout of designate tempest plugin  https://review.openstack.org/28351112:22
*** EricGonczer_ has quit IRC12:26
openstackgerritFederico Ceratto proposed openstack/designate: Fix rrset serialization, improve mdns tests  https://review.openstack.org/28655212:29
openstackgerritKiall Mac Innes proposed openstack/designate: Initial layout of designate tempest plugin  https://review.openstack.org/28351113:13
*** GonZo2K has joined #openstack-dns13:14
*** GonZo2K has quit IRC13:16
*** GonZo2K has joined #openstack-dns13:16
*** GonZo2K has quit IRC13:16
*** GonZo2K has joined #openstack-dns13:16
openstackgerritKiall Mac Innes proposed openstack/designate: Initial layout of designate tempest plugin  https://review.openstack.org/28351113:20
*** rsyed_away is now known as rsyed13:37
openstackgerritGraham Hayes proposed openstack/designate: Add Enhanced Logging to APIs  https://review.openstack.org/16735113:44
*** GonZo2K has quit IRC13:46
openstackgerritKiall Mac Innes proposed openstack/designate: Initial layout of designate tempest plugin  https://review.openstack.org/28351113:48
Kiallpglass / timsim / elarson / mugsie: I *think* ^ is good to go as a base framework.. reviews appreciated before I start converting existing tests! :)13:52
*** mlavalle has joined #openstack-dns13:58
*** GonZo2K has joined #openstack-dns14:01
*** GonZo2K has joined #openstack-dns14:01
*** venkat has joined #openstack-dns14:08
*** GonZo2K has quit IRC14:14
openstackgerritGraham Hayes proposed openstack/designate: Add Enhanced Logging to APIs  https://review.openstack.org/16735114:16
*** pglass has joined #openstack-dns14:23
*** jasonsb has quit IRC14:25
*** jasonsb has joined #openstack-dns14:26
elarsonmugsie: are we creating a new branch for the workshop repo?14:37
mugsieyeah14:38
mugsieI am trying to get neutron installed atm14:38
elarsonso you haven't pushed the branch just yet then?14:38
mugsienot yet14:39
*** ducttape_ has joined #openstack-dns14:41
openstackgerritMerged openstack/designate: Update config samples to show Identity v3 params  https://review.openstack.org/29729914:45
openstackgerritMerged openstack/designate: Updated from global requirements  https://review.openstack.org/30076014:47
openstackgerritMerged openstack/designate-dashboard: Update requirement for Horizon in stable/mitaka  https://review.openstack.org/30093014:47
*** james_li has joined #openstack-dns14:49
elarsonmugsie: :coolio:14:50
*** rsyed is now known as rsyed_away14:57
*** mlavalle has quit IRC15:00
*** Alex_Stef has quit IRC15:00
*** mlavalle has joined #openstack-dns15:03
*** EricGonczer_ has joined #openstack-dns15:04
mugsieelarson: timsim pglass - can you read http://lists.openstack.org/pipermail/openstack-dev/2016-April/091295.html and let me know what you think?15:05
mugsie+ Kiall15:05
pglassmugsie: do any other projects keep tests in a separate repo?15:09
mugsienot yet15:09
mugsiebut, I dont like the idea of keeping it in repo because no one else did15:10
mugsietempest explisitly supports out of repo tests15:11
elarsonmugsie: I think it is a resonable way forward, especially if it means you can point the tests at any install without having to download / configure a designate env.15:12
elarsongiven, that isn't a huge deal, but it does simplify things15:12
mugsiecool.15:12
elarsonassuming I'm understanding correclty15:12
mugsieelarson: want to reply to list ? :P15:12
elarsoncorrectly*15:12
elarsonsure!15:12
pglassthe only thing there is that the gate depends on the tests. so then you need to manually coordinate changes to tests and code.15:12
mugsieelarson: you are15:12
KiallSo, out of tree has some nicities for us @ HP when it comes to easy of integrating into tooling.. BUT - It also means we can start doing "branchless tempest". Zones API v2 should have the same stuff in L as it does in M15:12
mugsiepglass: which is a good thing, to make sure we do not break APIs :)15:13
KiallBut that in turn forces us to expose via the API that $new shiny feature is supported or not15:13
mugsiedoes it 404? then it is not supported15:14
mugsie(I joke)15:14
elarsonKiall: you could run the tests with filters to exclude tests you expect to fail15:14
Kialllol15:14
mugsiekinda15:14
KiallYou can, but that wouldn't be in the spirit of it ;)15:14
Kiallhttp://docs.openstack.org/developer/tempest/HACKING.html#branchless-tempest-considerations explains it better than I15:14
* elarson reads up15:14
* elarson has mainly used some pglass magic with tempest to test envs, so might be skewed as to the use cases15:15
Kialllol15:15
pglasswell i'm just thinking about like... oslo something pbr breaks the world - then you need changes to both repos. or you merge tests before a bug is fixed and then everything is broken.15:15
elarsonpglass: that is a good point15:16
Kiallthe test repo would be gated on the tests passing, so you can't land a broken testy15:16
pglassthat's true15:16
Kiallre requirements like pbr etc, those would break us either way.. in-tree or separate repo, same explosions15:16
elarsonthis is also true, pbr breaks everything™15:17
KiallThe core services like nova/neutron/cinder etc have been keeping their tests in the tempest repo itself for years, so it's certainly managable.15:17
pglassi mean, i'm fine with it15:18
KiallWhile all the new services keeping stuff in tree are in for 1 hell of a suprise IMO15:18
elarsonit seems like you'd want to publish a "here is the baseline tempest tests" that you run from a separate repo while still having some tempest tests that are release specific for new features, etc.15:18
Kialli.e. when tempest (master) requires PBR > 2.0 and there stable/mitaka requires PBR < 2.015:18
elarsonrather than breaking things up by technology (ie tempest) you break them up via function15:18
KiallWell, tempest's way of doing that is to decorate tests with:15:19
Kiall@test.requires_ext(extension='ext-gw-mode', service='network')15:20
Kialland your service is expected to advertise the "things" it supports, or tempest.conf can be set to say which extensions are enabled.15:20
*** venkat has quit IRC15:20
pglassyeah, tempest has test attributes you can filter on... but it does make "which tests are supposed to pass for mitaka?" a bit more fuzzy15:21
*** rsyed_away is now known as rsyed15:21
Kiallpglass: yea - it does, but which tests should pass for a release should == every test executed ;)15:22
KiallAnyway - I honestly can't see how in-tree works long term as a tempest plugin, thanks to requirements.txt.15:23
pglassright. when you release, all tests should pass. but a year later, when there are new tests for new features it's less clear.15:23
pglassbut maybe nobody cares about that.15:24
pglassi don't.15:24
pglassi'm +1 on a separate repo15:24
mugsietimsim: last call to -2 the idea?15:26
pglasstimsim is pto, baseball15:26
KiallIdeally, those new tests for new features should ideally be keyed off something we expose in our API - e.g. say we added "status" to the output of a zone list API call, we'd want to expose a method for saying "status-on-zones" is supported.15:26
Kialli.e. being really explicit about API changes15:26
Kiall(which, by itself is a Good Thing to be explicit about)15:27
*** nkinder has joined #openstack-dns15:27
pglassnormally, a version would work for that.15:28
pglassbut we just have "v2" that never changes15:29
mugsiepglass: we could do microversions :P15:33
Kiall-1015:33
mugsie(for the record i am -1000 on that)15:33
pglassit would be fine if the version wasn't in the url15:34
Kiallbut.. we could have a v2/extensions that lists out the "extra adde stuff"15:34
Kialladded*15:34
mugsiethats the only place you should have a microversion URL15:34
Kiall(as nothing should ever be removed)15:34
mugsieKiall: we could add it at the root of /v2/15:34
pglassi mean, we're effectively releasing new microversions of v2 anyway. it wouldn't be horrible if we started including a "v2.0.9" in the versions list on a GET /15:36
pglassbut that's probably not openstack's way15:36
Kiallpglass: openstack in general started down this microversions idea, and its... awkward and klunky at best15:36
pglassit's what like all other software does15:37
mugsiepglass: what you *should* do in microversions is v2/ always points to the latest. You then have /v2.0.9 versions availible, for people who need to pin to a partiuclar release15:43
pglassyes15:43
pglassi would be fine with that15:44
mugsieopenstack has decided to go with a header and (for nova for example) /v2.115:44
*** ccneill has joined #openstack-dns15:45
mugsieso you GET /v2.1 with a header of X-OpenStack-Nova-API-Version: 2.11415:46
*** bpokorny has joined #openstack-dns15:46
*** jschwarz has quit IRC16:01
*** EricGonczer_ has quit IRC16:14
*** ccneill_ has joined #openstack-dns16:27
*** EricGonczer_ has joined #openstack-dns16:28
*** rudrajit has joined #openstack-dns16:28
*** ccneill has quit IRC16:29
*** barra204 has quit IRC16:32
*** rudrajit has quit IRC16:41
*** EricGonczer_ has quit IRC16:42
*** EricGonczer_ has joined #openstack-dns16:43
openstackgerritKiall Mac Innes proposed openstack/designate: Initial layout of designate tempest plugin  https://review.openstack.org/28351116:44
*** _shaps_ has quit IRC16:50
*** rsyed is now known as rsyed_away16:53
*** jasonsb has quit IRC17:04
*** EricGonczer_ has quit IRC17:07
*** gord0 has joined #openstack-dns17:15
mlavallemugsie: ping17:17
*** ccneill_ has quit IRC17:22
*** pglass has quit IRC17:31
*** pglass has joined #openstack-dns17:32
*** rudrajit has joined #openstack-dns17:34
*** rudrajit_ has joined #openstack-dns17:43
*** ccneill_ has joined #openstack-dns17:45
*** rudrajit has quit IRC17:46
*** bpokorny has quit IRC17:59
*** james_li has quit IRC18:00
*** james_li has joined #openstack-dns18:11
*** rsyed_away is now known as rsyed18:14
*** pglass has quit IRC18:16
*** pglass has joined #openstack-dns18:17
*** ccneill__ has joined #openstack-dns18:18
*** ccneill__ is now known as ccneill18:19
*** ccneill_ has quit IRC18:21
*** EricGonczer_ has joined #openstack-dns18:27
*** bpokorny has joined #openstack-dns18:30
*** nyechiel_ has quit IRC18:36
*** rudrajit has joined #openstack-dns18:39
*** rudrajit_ has quit IRC18:43
*** darkxploit has joined #openstack-dns18:46
*** EricGonczer_ has quit IRC18:48
*** EricGonczer_ has joined #openstack-dns18:48
*** james_li has quit IRC18:56
*** EricGonczer_ has quit IRC19:18
openstackgerritKiall Mac Innes proposed openstack/designate: Initial layout of designate tempest plugin  https://review.openstack.org/28351119:29
Kiallpglass: I've made a bunch of changes to the tempest plugin, including some onion unpeeling, but the unpeeling resulted in manual .addCleanups vs magic tracking :( But overall, I'm personally liking where it's at, care to have another look?19:29
openstackgerritKiall Mac Innes proposed openstack/designate: Initial layout of designate tempest plugin  https://review.openstack.org/28351119:30
Kiallderp - forgot to fix mugsie's review comment ;)19:30
Kiallmlavalle: he's elbow deep in popcorn watching Batman v Superman19:31
pglass`from oslo_serialization import jsonutils as json`19:32
Kiallwhat about it? :)19:34
pglassi shouldn't be surprised19:34
pglassat what oslo will wrap or rewrite19:34
Kialllol, yea.19:34
mlavalleKiall: LOL, thanks19:34
Kiallit's mostly for handling things native json.loads/dumps barfs on.. like dattimes.19:35
Kialldatetimes*19:35
pglassand it looks like it just forwards along to json (or anyjson, if it finds it)19:35
Kialljsonutils.loads passes through to https://github.com/openstack/oslo.utils/blob/master/oslo_utils/encodeutils.py#L2719:36
KiallI'm not entirely sure what value it provides ;) habbit to use the oslo versions because there's always something that barfs if you don't these days :)19:37
KiallAnyway, If I don't get any objections to what's there now, I'm going to start looking at padding out pieces I've glossed over (testing updates needs some base methods etc defined..) then port over all the existing zone tests and start down the path of moving all the things over.19:40
pglassKiall: those `credentials = ['primary', 'admin']` bits are compatible with all this? http://docs.openstack.org/developer/tempest/configuration.html#credential-provider-mechanisms19:42
KiallYep, they are.19:43
*** EricGonczer_ has joined #openstack-dns19:48
*** shakamunyi has joined #openstack-dns19:51
*** ccneill has left #openstack-dns19:52
*** shakamunyi has quit IRC19:57
*** rsyed is now known as rsyed_away20:07
Kiallpglass: Q - What exactly do you guys return from your service catalog for Designate?20:19
Kiall(refreshing my memory on https://github.com/openstack/designate/commit/9bdcb63904c3fb2f7964def9dcdd74eb20fd470b )20:19
pglassKiall: https://global.dns.api.rackspacecloud.com/v2/98317420:21
pglassthe number at the end is our tenant id20:21
Kiallwhere the 983.. is a tenant id?20:21
KiallAnd for nova etc? Same pattern?20:21
pglassyes. that's every service20:21
pglassso i imagine someone has already figured out that bit20:21
KiallThat's what I'm thinking. But don't want to write it off and have you redo again.20:22
KiallI'm also thinking the request hooks you added should be easy to port, if you use it and if tempest doesn't offer something built in that provides the same functionality.20:23
pglasswell, that was really just a quick hack to support some kind of rate limiting against our envs20:25
pglassthe right solution for that was to rate limit the client by respecting the retry-after header.20:26
KiallThe hooks? As in, not used/needed anymore?20:26
pglasswell, we still use them20:26
pglassbut there's a better solution which was to get tempest-lib to retry on 413s, based on the retry-after header20:27
pglasswhich tempest-lib does, but it only handles one kind of time format20:27
KiallAh, and if that use case is solved in tempest, there's no longer a use of hooks?20:28
pglassthat's right.20:28
Kiall(I'm trying not to copy things for the sake of coping them, if there's a use for them, they come, if not, no need.)20:28
KiallAny idea if there's traction in tempest to fix that? a patch etc?20:28
pglassthat makes sense20:28
pglassi haven't checked. we were in a bit of a deadline rush at the time.20:29
KiallIf there's a patch, there's a good chance we can get it merged by pinging the right people :)20:29
pglasswell, you can leave out the rackspace-specific bits for now, and I'll get around to worrying about them and handling things in better ways at some point20:30
KiallOkay, So.. I'm going to leave both of those patches for the moment. For the URI tenent_id thing, we will need to do something, as we build the v2/zones - where the tenant id needs to land in the middle, but we have a good and easy place to put that code.20:31
KiallFor the hooks retry after use case, it'd be awesome to see a bug +patch to tempest, and I can nag the tempest folks to merge it! Less code for us + a win.20:32
Kiall= a win*20:32
pglassit was this bit here: https://github.com/openstack/tempest/blob/master/tempest/lib/common/rest_client.py#L631-L64020:32
pglass`delay = int(resp['retry-after'])`20:33
openstackgerritKiall Mac Innes proposed openstack/designate: Initial layout of designate tempest plugin  https://review.openstack.org/28351120:33
Kiall^ just changes adm -> admin20:33
*** rsyed_away is now known as rsyed20:33
pglassbut our retry-after headers are in another format. i'll get around to fixing that at some point, assuming everything still uses that tempest.lib client20:34
Kiall       Retry-After: Fri, 31 Dec 1999 23:59:59 GMT20:34
Kiall       Retry-After: 12020:34
Kiallyours is the first format I guess?20:34
pglassyep20:34
Kiallyea, it does use it.20:34
KiallBTW - Have you read AND understood http://docs.openstack.org/developer/tempest/HACKING.html#negative-tests ?20:37
KiallI see there's some sort of specific way to do negative tests, but I haven't found any examples of it used in real life ;)20:38
Kialle.g. "The class decorator SimpleNegativeAutoTest will automatically generate test cases out of the given schema in the attribute _schema."20:38
Kiall^'s that bit I've never seen.20:38
*** nyechiel_ has joined #openstack-dns20:52
*** rsyed is now known as rsyed_away20:59
openstackgerritKiall Mac Innes proposed openstack/designate: Initial layout of designate tempest plugin  https://review.openstack.org/28351121:10
openstackgerritKiall Mac Innes proposed openstack/designate: WIP: Broken attempt at using Tempest NegativeAutoTest  https://review.openstack.org/30135221:10
pglassjust because it exists, doesn't mean we have to use it...21:13
pglasshttps://github.com/openstack/tempest/blob/master/tempest/tests/negative/test_negative_generators.py#L29-L4121:19
pglassthere's a compute example as well: https://github.com/openstack/tempest/blob/8843f0f0768019c76be72b4be2f6a156cdbe3d78/tempest/api_schema/request/compute/v2/flavors.py21:21
pglasshttps://github.com/openstack/tempest/blob/8843f0f0768019c76be72b4be2f6a156cdbe3d78/tempest/api/compute/flavors/test_flavors_negative.py#L2721:21
*** rsyed_away is now known as rsyed21:30
*** bpokorny has quit IRC21:35
*** richm has joined #openstack-dns21:38
*** gord0 has quit IRC21:56
*** nyechiel_ has quit IRC21:57
*** ducttape_ has quit IRC22:12
*** pglass has quit IRC22:32
*** EricGonczer_ has quit IRC22:35
*** km has joined #openstack-dns23:00
*** mlavalle has quit IRC23:00
*** chlong has joined #openstack-dns23:18
*** ducttape_ has joined #openstack-dns23:26
*** ducttape_ has quit IRC23:28
*** kei_yama has joined #openstack-dns23:31
*** kei_yama has quit IRC23:51
*** kei_yama has joined #openstack-dns23:52
*** rsyed is now known as rsyed_away23:55
« Sunday, 2016-04-03 Index Tuesday, 2016-04-05 »

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