| *** maurosr has quit IRC | 00:00 | |
| *** maurosr has joined #openstack-qa | 00:01 | |
| *** shwetaap has joined #openstack-qa | 00:02 | |
| *** ociuhandu has quit IRC | 00:05 | |
| *** markmcclain has quit IRC | 00:05 | |
| *** mkollaro has quit IRC | 00:08 | |
| *** markmcclain has joined #openstack-qa | 00:09 | |
| openstackgerrit | Sean Dague proposed a change to openstack/tempest: add cli.has_manage option https://review.openstack.org/75245 | 00:09 |
|---|---|---|
| *** rockyg has quit IRC | 00:11 | |
| ravikumar_hp | sdague: mtreinish: Gate runs full tempest . Right? so where do you we use tests identified with attr=gate | 00:12 |
| sdague | we don't, it was a half implemented feature | 00:13 |
| sdague | so attr=gate should be removed | 00:14 |
| ravikumar_hp | so same is smoke . right? | 00:14 |
| ravikumar_hp | i will remove attr=gate .. | 00:15 |
| *** ociuhandu has joined #openstack-qa | 00:16 | |
| openstackgerrit | Eric Windisch proposed a change to openstack/tempest: Allow LiveBlockMigrationV3Test to be skipped if v3 is disabled https://review.openstack.org/75249 | 00:17 |
| *** jroovers has quit IRC | 00:17 | |
| openstackgerrit | Armando Migliaccio proposed a change to openstack/tempest: Allow IPv6 tests to be disabled https://review.openstack.org/74933 | 00:21 |
| *** ociuhandu has quit IRC | 00:24 | |
| sdague | smoke stays | 00:25 |
| *** StevenK_ is now known as StevenK | 00:25 | |
| *** StevenK is now known as Guest99695 | 00:26 | |
| *** Guest99695 is now known as StevenK | 00:26 | |
| openstackgerrit | Eric Windisch proposed a change to openstack/tempest: Only execute tearDown if setUpClass created cls.volume https://review.openstack.org/75254 | 00:29 |
| *** ociuhandu has joined #openstack-qa | 00:30 | |
| *** matsuhashi has joined #openstack-qa | 00:31 | |
| *** cadenzajon has quit IRC | 00:32 | |
| openstackgerrit | John Griffith proposed a change to openstack/tempest: Ensure detach before teardown in scenario manager https://review.openstack.org/74272 | 00:46 |
| *** salv-orlando_ has joined #openstack-qa | 00:47 | |
| *** matsuhas_ has joined #openstack-qa | 00:48 | |
| *** mlavalle_ has joined #openstack-qa | 00:48 | |
| *** nati_uen_ has joined #openstack-qa | 00:48 | |
| *** dmorita has joined #openstack-qa | 00:50 | |
| *** notmyname_ has joined #openstack-qa | 00:50 | |
| openstackgerrit | Eric Windisch proposed a change to openstack/tempest: Only execute tearDown if setUpClass created cls.volume https://review.openstack.org/75254 | 00:52 |
| openstackgerrit | Anita Kuno proposed a change to openstack/tempest: Removed RestClientXML class https://review.openstack.org/74954 | 00:53 |
| *** nati_ue__ has joined #openstack-qa | 00:53 | |
| anteaya | sdague: I had got some logs from the failures the last time | 00:54 |
| anteaya | the console logs | 00:54 |
| anteaya | do you want them? | 00:54 |
| *** ruhe- has joined #openstack-qa | 00:54 | |
| *** nati_uen_ has quit IRC | 00:54 | |
| *** SergeyLukjanov2 has joined #openstack-qa | 00:55 | |
| *** matsuhashi has quit IRC | 00:55 | |
| *** nati_ueno has quit IRC | 00:55 | |
| *** jaypipes has quit IRC | 00:55 | |
| *** mlavalle has quit IRC | 00:55 | |
| *** jgrimm has quit IRC | 00:55 | |
| *** salv-orlando has quit IRC | 00:55 | |
| *** lifeless has quit IRC | 00:55 | |
| *** SergeyLukjanov has quit IRC | 00:55 | |
| *** ruhe has quit IRC | 00:55 | |
| *** notmyname has quit IRC | 00:55 | |
| *** ewindisch has quit IRC | 00:55 | |
| *** obondarev has quit IRC | 00:55 | |
| *** dtroyer has quit IRC | 00:55 | |
| *** smurashov has quit IRC | 00:55 | |
| *** SergeyLukjanov2 is now known as SergeyLukjanov | 00:55 | |
| *** salv-orlando_ is now known as salv-orlando | 00:55 | |
| *** mlavalle_ is now known as mlavalle | 00:55 | |
| *** notmyname_ is now known as notmyname | 00:55 | |
| *** ruhe- is now known as ruhe | 00:55 | |
| openstackgerrit | Eric Windisch proposed a change to openstack/tempest: Skip EC2VolumesTest if Cinder is not available https://review.openstack.org/75267 | 00:57 |
| *** obondarev has joined #openstack-qa | 00:57 | |
| *** dtroyer has joined #openstack-qa | 00:57 | |
| *** lifeless has joined #openstack-qa | 00:57 | |
| anteaya | sdague: here are the links: http://paste.openstack.org/show/67813/ | 01:00 |
| *** cnesa has quit IRC | 01:00 | |
| *** atiwari has quit IRC | 01:01 | |
| *** jgrimm has joined #openstack-qa | 01:01 | |
| *** ewindisch has joined #openstack-qa | 01:01 | |
| *** obondarev has quit IRC | 01:02 | |
| *** dtroyer has quit IRC | 01:02 | |
| *** lifeless has quit IRC | 01:02 | |
| *** obondarev has joined #openstack-qa | 01:02 | |
| *** dtroyer has joined #openstack-qa | 01:02 | |
| *** lifeless has joined #openstack-qa | 01:02 | |
| openstackgerrit | Ian Wienand proposed a change to openstack-dev/devstack: Add GIT_TIMEOUT variable to watch git operations https://review.openstack.org/74910 | 01:02 |
| *** jaypipes has joined #openstack-qa | 01:03 | |
| *** ChanServ sets mode: +o jaypipes | 01:03 | |
| *** smurashov has joined #openstack-qa | 01:03 | |
| *** bill_az_ has quit IRC | 01:08 | |
| openstackgerrit | A change was merged to openstack-infra/elastic-recheck: Add fingerprint for bug 1282795 https://review.openstack.org/75220 | 01:14 |
| *** weshay has joined #openstack-qa | 01:15 | |
| openstackgerrit | Eric Windisch proposed a change to openstack/tempest: Skip deleting flavors in tearDown if setUpClass did not complete https://review.openstack.org/75274 | 01:16 |
| *** openstackgerrit has quit IRC | 01:18 | |
| *** openstackgerrit has joined #openstack-qa | 01:18 | |
| *** openstackrecheck has quit IRC | 01:22 | |
| *** openstackrecheck has joined #openstack-qa | 01:22 | |
| *** rfolco has joined #openstack-qa | 01:22 | |
| openstackgerrit | Eric Windisch proposed a change to openstack/tempest: Skip deleting flavors in tearDown if setUpClass did not complete https://review.openstack.org/75274 | 01:23 |
| *** ociuhandu has quit IRC | 01:23 | |
| *** nosnos has joined #openstack-qa | 01:34 | |
| openstackgerrit | A change was merged to openstack/tempest: Minor changes to scenario manager https://review.openstack.org/73973 | 01:36 |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/elastic-recheck: Improved timestamp parsing https://review.openstack.org/73741 | 01:39 |
| *** rfolco has quit IRC | 01:46 | |
| *** zhikunliu has joined #openstack-qa | 01:52 | |
| *** Sukhdev has quit IRC | 01:54 | |
| *** mlavalle has quit IRC | 01:57 | |
| zhikunliu | cyeoh: ping | 01:59 |
| cyeoh | zhikunliu, hi | 02:00 |
| zhikunliu | I have updated https://review.openstack.org/#/c/73503/. could you have a look when you have time? | 02:00 |
| cyeoh | zhikunliu: sure - I'm just heading out to lunch now but will have a look when I get back | 02:04 |
| zhikunliu | cyeoh: got it. thank you! :) | 02:05 |
| *** weshay has quit IRC | 02:28 | |
| *** matsuhas_ has quit IRC | 02:34 | |
| *** matsuhas_ has joined #openstack-qa | 02:36 | |
| *** marun has quit IRC | 02:37 | |
| cyeoh | zhikunliu: hi | 02:43 |
| *** markmcclain has quit IRC | 02:46 | |
| openstackgerrit | Ken'ichi Ohmichi proposed a change to openstack/tempest: WIP: Add "delete the volume-attached server" tests https://review.openstack.org/75288 | 02:47 |
| zhikunliu | cyeoh: hi | 02:47 |
| *** bhuvan has joined #openstack-qa | 02:47 | |
| cyeoh | zhikunliu: so I ended up just writing a message on your patch, but I was wondering - did you end up autogenerating the sample file again or edit it manually? | 02:48 |
| zhikunliu | cyeoh: auto generated | 02:48 |
| cyeoh | ok - what I find interesting is how the sample file was generated incorrectly in the first place. | 02:49 |
| cyeoh | I'm thinking there might be a bug in it perhaps | 02:49 |
| cyeoh | because the groupings in the config file didn't change | 02:50 |
| cyeoh | mtreinish: when you get back you might want to look at the conversation above and at https://review.openstack.org/#/c/73503 | 02:50 |
| zhikunliu | cyeoh: what I found is - generator compares opt with the opts in groups to find the group in which the opt is. these 3 opts has same 'name', 'default value', and 'help message'. The first match is '[network-feature-enabled]', so they go to the same group. | 02:58 |
| openstackgerrit | Steve Baker proposed a change to openstack/tempest: NeutronResourcesTestJSON test reasons for stack failure https://review.openstack.org/75289 | 02:58 |
| openstackgerrit | Steve Baker proposed a change to openstack/tempest: Improve ServerCfnInitTestJSON asserts and debugging https://review.openstack.org/69241 | 02:58 |
| cyeoh | zhikunliu: ok, that's definitely a bug then :-) Would you mind submitting a bug report for it? | 02:59 |
| openstackgerrit | Ken'ichi Ohmichi proposed a change to openstack/tempest: Check server terminations on "delete server" tests https://review.openstack.org/75290 | 02:59 |
| zhikunliu | cyeoh: https://bugs.launchpad.net/tempest/+bug/1280085. could we still use this bug to track the problem? | 03:00 |
| cyeoh | zhikunliu: I would just submit another one, being specific about the bug being in the config file generator and referencing 1280085 | 03:01 |
| zhikunliu | cyeoh: ok, thanks :) | 03:02 |
| cyeoh | zhikunliu: thank you! | 03:02 |
| *** matsuhas_ has quit IRC | 03:03 | |
| *** matsuhas_ has joined #openstack-qa | 03:06 | |
| mtreinish | cyeoh: I know that we're kind of out of date on the config generator from oslo | 03:25 |
| mtreinish | it might be worth trying to sync it to the latest to see if that bug has been fixed already | 03:26 |
| openstackgerrit | Ian Wienand proposed a change to openstack-dev/devstack: Add GIT_TIMEOUT variable to watch git operations https://review.openstack.org/74910 | 03:26 |
| openstackrecheck | Console logs not available after 13:32s for check-tempest-dsvm-neutron-full 49166,6,d157b42 | 03:27 |
| *** matsuhas_ has quit IRC | 03:29 | |
| cyeoh | mtreinish: ah cool - thanks | 03:29 |
| openstackgerrit | Ian Wienand proposed a change to openstack-dev/devstack: Add more files to run_tests.sh bash8 check https://review.openstack.org/75295 | 03:35 |
| openstackrecheck | Console logs not available after 13:38s for check-tempest-dsvm-full 74954,4,157de31 | 03:41 |
| *** masayukig has quit IRC | 03:44 | |
| *** harlowja is now known as harlowja_away | 03:45 | |
| *** dkranz has joined #openstack-qa | 03:54 | |
| openstackrecheck | Console logs not available after 13:41s for check-tempest-dsvm-full 69328,2,42151f1 | 03:58 |
| *** liudong has joined #openstack-qa | 04:00 | |
| *** bhuvan has quit IRC | 04:13 | |
| openstackgerrit | Ken'ichi Ohmichi proposed a change to openstack/tempest: WIP: Add "delete the volume-attached server" tests https://review.openstack.org/75288 | 04:14 |
| *** mrodden has quit IRC | 04:24 | |
| *** locke1051 has quit IRC | 04:24 | |
| openstackrecheck | Console logs not available after 13:25s for gate-tempest-dsvm-full 74413,1,fecb122 | 04:25 |
| *** matsuhashi has joined #openstack-qa | 04:25 | |
| *** nati_ue__ has quit IRC | 04:25 | |
| *** om has joined #openstack-qa | 04:26 | |
| openstackgerrit | Ken'ichi Ohmichi proposed a change to openstack/tempest: Do not merge: Reproduce bug/1277027 https://review.openstack.org/74944 | 04:29 |
| *** Madan has joined #openstack-qa | 04:36 | |
| openstackrecheck | Console logs not available after 13:36s for check-tempest-dsvm-neutron-full 75210,1,445be7b | 04:39 |
| openstackgerrit | Ian Wienand proposed a change to openstack-dev/devstack: Un-nest generate_swift_config https://review.openstack.org/75308 | 04:40 |
| *** ken1ohmichi has quit IRC | 04:51 | |
| *** bhuvan has joined #openstack-qa | 04:52 | |
| openstackgerrit | A change was merged to openstack/tempest: Migrate negative test to a different file https://review.openstack.org/69839 | 05:02 |
| *** shwetaap has joined #openstack-qa | 05:02 | |
| openstackgerrit | Masayuki Igawa proposed a change to openstack/tempest: Add qcow2 image support https://review.openstack.org/75312 | 05:05 |
| openstackgerrit | Ian Wienand proposed a change to openstack-dev/devstack: Fix heredoc matching for trailing quote https://review.openstack.org/75314 | 05:16 |
| openstackgerrit | Ian Wienand proposed a change to openstack-dev/devstack: Enforce function declaration format in bash8 https://review.openstack.org/75315 | 05:22 |
| *** matsuhashi has quit IRC | 05:23 | |
| *** jroovers has joined #openstack-qa | 05:24 | |
| *** bhuvan has quit IRC | 05:30 | |
| openstackgerrit | Ken'ichi Ohmichi proposed a change to openstack/tempest: Add "delete the volume-attached server" tests https://review.openstack.org/75288 | 05:30 |
| *** matsuhashi has joined #openstack-qa | 05:33 | |
| openstackrecheck | Console logs not available after 13:39s for check-tempest-dsvm-neutron-full 72798,18,5768dc3 | 05:34 |
| *** locke105 has joined #openstack-qa | 05:40 | |
| *** shwetaap has quit IRC | 05:47 | |
| *** ociuhandu has joined #openstack-qa | 05:48 | |
| openstackgerrit | Zhi Kun Liu proposed a change to openstack/tempest: fix duplicate api_extensions https://review.openstack.org/73503 | 05:54 |
| *** ociuhandu has quit IRC | 05:54 | |
| *** liudong has quit IRC | 06:07 | |
| *** liudong has joined #openstack-qa | 06:09 | |
| *** coolsvap has joined #openstack-qa | 06:13 | |
| *** locke105 has quit IRC | 06:16 | |
| *** thomasbiege has joined #openstack-qa | 06:18 | |
| *** thomasbiege has quit IRC | 06:18 | |
| *** protux has joined #openstack-qa | 06:19 | |
| raies | afazekas: hi | 06:24 |
| openstackgerrit | Ken'ichi Ohmichi proposed a change to openstack/tempest: Do not merge: Reproduce bug/1277027 https://review.openstack.org/74944 | 06:25 |
| *** vkozhukalov has joined #openstack-qa | 06:33 | |
| openstackgerrit | Ken'ichi Ohmichi proposed a change to openstack/tempest: Add "delete the volume-attached server" tests https://review.openstack.org/75288 | 06:34 |
| *** zhikunliu has quit IRC | 06:36 | |
| *** morganfainberg is now known as morganfainberg_Z | 06:36 | |
| *** saju_m has joined #openstack-qa | 06:36 | |
| raies | hi | 06:41 |
| raies | devstack installation is getting failed due to http://paste.ubuntu.com/6969426/ | 06:41 |
| raies | any suggestions ?? ^^ | 06:41 |
| *** om has quit IRC | 06:47 | |
| *** om has joined #openstack-qa | 06:53 | |
| *** afazekas has quit IRC | 07:09 | |
| coolsvap | add IMAGE_URLS=https://launchpad.net/cirros/trunk/0.3.0/+download/cirros-0.3.0-x86_64-disk.img to your localrc and rerun stack.sh | 07:11 |
| *** afazekas_ has quit IRC | 07:14 | |
| *** bhuvan has joined #openstack-qa | 07:18 | |
| *** banix has joined #openstack-qa | 07:21 | |
| *** vkozhukalov has quit IRC | 07:31 | |
| *** bhuvan has quit IRC | 07:37 | |
| *** pbrooko has joined #openstack-qa | 07:42 | |
| *** jamielennox is now known as jamielennox|away | 07:44 | |
| *** bhuvan has joined #openstack-qa | 07:56 | |
| *** afazekas_ has joined #openstack-qa | 07:56 | |
| *** om has quit IRC | 07:58 | |
| *** om has joined #openstack-qa | 08:00 | |
| openstackgerrit | Valeriy Ponomaryov proposed a change to openstack/tempest: Implement pluggability for tempest (config) https://review.openstack.org/69985 | 08:04 |
| openstackgerrit | Valeriy Ponomaryov proposed a change to openstack/tempest: Implement pluggability for tempest (clients) https://review.openstack.org/69983 | 08:04 |
| *** giulivo has joined #openstack-qa | 08:09 | |
| *** devvesa has joined #openstack-qa | 08:09 | |
| openstackgerrit | Ann Kamyshnikova proposed a change to openstack/tempest: Verify more information in API tests for LBaaS https://review.openstack.org/61118 | 08:10 |
| *** bhuvan has quit IRC | 08:14 | |
| *** luqas has joined #openstack-qa | 08:17 | |
| *** shardy_afk is now known as shardy | 08:20 | |
| *** saju_m has quit IRC | 08:31 | |
| *** jordanP has joined #openstack-qa | 08:32 | |
| openstackgerrit | Masayuki Igawa proposed a change to openstack/tempest: Clean up scenario functions https://review.openstack.org/75339 | 08:33 |
| openstackgerrit | Yaroslav Lobankov proposed a change to openstack/tempest: Savanna: add tests for job binaries https://review.openstack.org/74668 | 08:34 |
| openstackgerrit | Yuiko Takada proposed a change to openstack/tempest: Add agents tests for Nova v3 API https://review.openstack.org/74611 | 08:36 |
| raies | coolsvap: keeping it in localrc Internal server error is found | 08:36 |
| openstackgerrit | Masayuki Igawa proposed a change to openstack/tempest: Add qcow2 image support https://review.openstack.org/75312 | 08:40 |
| *** thomasbiege has joined #openstack-qa | 08:42 | |
| *** rossella-s has joined #openstack-qa | 08:43 | |
| *** jlibosva has joined #openstack-qa | 08:45 | |
| openstackgerrit | Ken'ichi Ohmichi proposed a change to openstack/tempest: Add "delete the volume-attached server" tests https://review.openstack.org/75288 | 08:45 |
| *** masayukig has joined #openstack-qa | 08:47 | |
| openstackgerrit | Sylvain Baubeau proposed a change to openstack/tempest: Add Cinder tests for quota sets https://review.openstack.org/73606 | 08:50 |
| *** ociuhandu has joined #openstack-qa | 09:00 | |
| coolsvap | raies: are you configuring swift with your devstack? | 09:00 |
| *** saju_m has joined #openstack-qa | 09:01 | |
| *** banix has quit IRC | 09:02 | |
| *** saju_m has quit IRC | 09:02 | |
| *** rossella-s has quit IRC | 09:04 | |
| *** yassine has joined #openstack-qa | 09:04 | |
| *** thomasbiege has quit IRC | 09:05 | |
| *** vkozhukalov has joined #openstack-qa | 09:07 | |
| *** fbo_away is now known as fbo | 09:09 | |
| *** ociuhandu has quit IRC | 09:11 | |
| *** luqas has quit IRC | 09:17 | |
| *** mkollaro has joined #openstack-qa | 09:21 | |
| *** saju_m has joined #openstack-qa | 09:21 | |
| raies | coolsvap: yes it is also configured | 09:21 |
| openstackgerrit | Daisuke Morita proposed a change to openstack/tempest: Add parametric tests of Swift account API https://review.openstack.org/67071 | 09:21 |
| openstackgerrit | Yuiko Takada proposed a change to openstack/tempest: Separate negative tests for test_images https://review.openstack.org/75345 | 09:22 |
| *** dmorita has quit IRC | 09:23 | |
| *** gszasz has joined #openstack-qa | 09:24 | |
| *** BobBall has left #openstack-qa | 09:31 | |
| openstackgerrit | afazekas proposed a change to openstack/tempest: Add custom pretty tox entry https://review.openstack.org/75351 | 09:32 |
| openstackgerrit | Eiichi Aikawa proposed a change to openstack/tempest: Separate negative tests for test_server_addresses https://review.openstack.org/75352 | 09:36 |
| openstackgerrit | Gavin Brebner proposed a change to openstack-dev/devstack: Create glance-swift for swift or s-proxy https://review.openstack.org/74047 | 09:41 |
| *** Madan1 has joined #openstack-qa | 09:42 | |
| *** max_lobur_afk is now known as max_lobur | 09:43 | |
| *** Madan has quit IRC | 09:44 | |
| openstackgerrit | Masayuki Igawa proposed a change to openstack/tempest: Add qcow2 image support https://review.openstack.org/75312 | 09:49 |
| *** masayukig has quit IRC | 09:51 | |
| *** luqas has joined #openstack-qa | 09:53 | |
| *** liudong has quit IRC | 10:08 | |
| *** psedlak has quit IRC | 10:22 | |
| *** thomasbiege has joined #openstack-qa | 10:28 | |
| *** psedlak has joined #openstack-qa | 10:29 | |
| *** masayukig has joined #openstack-qa | 10:36 | |
| *** rfolco has joined #openstack-qa | 10:36 | |
| *** yassine has quit IRC | 10:40 | |
| *** yassine has joined #openstack-qa | 10:40 | |
| *** masayukig has quit IRC | 10:40 | |
| jroovers | Hi everyone, I'm running tempest against our OS deployment and I'm hitting an issue with the floating IP tests. In particular, our deployment does not have a default floating ip pool, i.e. you always need to specify the name of the pool when creating a new floating ip. I was wondering whether a patchset exists that allows the explicit configuration of the floating ip pool that tempest | 10:42 |
| jroovers | uses. | 10:42 |
| *** mkollaro has quit IRC | 10:42 | |
| *** coolsvap has quit IRC | 10:45 | |
| openstackgerrit | Zhang Gong proposed a change to openstack/tempest: Add Keystone role and service test cases https://review.openstack.org/74950 | 10:47 |
| *** tosky has joined #openstack-qa | 10:48 | |
| *** Madan has joined #openstack-qa | 10:49 | |
| *** Madan1 has quit IRC | 10:50 | |
| *** mkollaro has joined #openstack-qa | 10:51 | |
| *** coolsvap has joined #openstack-qa | 10:52 | |
| *** luqas has quit IRC | 10:56 | |
| *** nosnos has quit IRC | 10:58 | |
| *** matsuhashi has quit IRC | 11:02 | |
| *** vkozhukalov has quit IRC | 11:08 | |
| afazekas_ | jroovers: why you do not specify a default floating ip pool ? | 11:18 |
| afazekas_ | in your deployment | 11:19 |
| *** vkozhukalov has joined #openstack-qa | 11:20 | |
| *** coolsvap has quit IRC | 11:28 | |
| afrittoli | sdague: I just saw bug 1265672 again, so I re-opened the ticket | 11:29 |
| *** masayukig has joined #openstack-qa | 11:32 | |
| *** Madan has quit IRC | 11:46 | |
| *** luqas has joined #openstack-qa | 12:07 | |
| *** luqas has quit IRC | 12:08 | |
| *** luqas has joined #openstack-qa | 12:09 | |
| *** akamyshnikova has quit IRC | 12:11 | |
| openstackgerrit | afazekas proposed a change to openstack/tempest: Fix InvalidInstanceID.NotFound handling https://review.openstack.org/75385 | 12:13 |
| malini | masayukig, chmouel: Thanks for the reviews https://review.openstack.org/#/c/75160/2//COMMIT_MSG | 12:21 |
| malini | I was trying to troubleshoot the failing marconi tests in http://logs.openstack.org/49/63449/16/experimental/check-tempest-dsvm-marconi/be62e9b/console.html | 12:22 |
| masayukig | malini: you're welcome :) | 12:22 |
| malini | This is the issue I had with the tests - devstack was starting marconi in 0.0.0.0:8888 ok, but my tests were failing with a connection error | 12:23 |
| malini | & I am still searching in the dark how to fix tht - since everything looks ok in my devstack server | 12:23 |
| malini | I saw chmouel's comment tht SERVICE_HOST is the standard in all other devstack services | 12:24 |
| malini | Do you think tht could be the reason? | 12:24 |
| chmouel | malini: hello, why is tempest trying to contact 127.0.0.1 ? | 12:25 |
| malini | chmouel:hmm... marconi is running there..should it be something else ? | 12:26 |
| openstackgerrit | afazekas proposed a change to openstack/tempest: EC2 client token is not reuseable for new instance creation https://review.openstack.org/75388 | 12:26 |
| chmouel | malini: where is that patch referenced from? | 12:27 |
| chmouel | malini: so I can check the review | 12:27 |
| malini | chmouel: https://review.openstack.org/#/c/63449/ | 12:27 |
| malini | chmouel: I have an experimental job running the marconi-tempest test | 12:29 |
| malini | See check-tempest-dsvm-marconi | 12:29 |
| *** luqas has quit IRC | 12:31 | |
| *** luqas has joined #openstack-qa | 12:31 | |
| chmouel | malini: indeed, i think there is something wrong with your patch and it should be fixed there but I don't know very well tempest to be able to pinpoint it | 12:32 |
| *** om has quit IRC | 12:33 | |
| *** luqas has quit IRC | 12:33 | |
| chmouel | malini: checking out your patch to test it may take a bit | 12:34 |
| malini | chmouel: Does it look like a devstack fix to you ? | 12:35 |
| chmouel | malini: i think the fix to bind marconi to localhost is not the way to do it | 12:36 |
| chmouel | give me a bit as i am restacking and willl see if i can help | 12:36 |
| malini | chmouel: thanks!! | 12:36 |
| *** vkozhukalov has quit IRC | 12:39 | |
| chmouel | malini: may takes a while since i am restarting a fresh vm | 12:42 |
| *** masayukig has quit IRC | 12:42 | |
| *** masayukig has joined #openstack-qa | 12:43 | |
| malini | chmouel: np..Thnks for chking this !! I might be a lil bit on & off - since I am still at home & have a 1 year old hovering around. | 12:43 |
| chmouel | malini: cool no worries, i got to take care of a 8 year old one today for me :) | 12:44 |
| *** masayukig has quit IRC | 12:47 | |
| *** weshay has joined #openstack-qa | 12:51 | |
| *** masayukig has joined #openstack-qa | 12:51 | |
| *** vkozhukalov has joined #openstack-qa | 12:55 | |
| openstackrecheck | Console logs not available after 13:27s for gate-glance-python27 75292,2,f64f670 | 12:56 |
| openstackgerrit | Chmouel Boudjnah proposed a change to openstack/tempest: Add tests for Swift in orchestration https://review.openstack.org/65358 | 13:03 |
| *** zul has joined #openstack-qa | 13:09 | |
| *** mwagner_lap has joined #openstack-qa | 13:11 | |
| openstackgerrit | Masayuki Igawa proposed a change to openstack/tempest: Clarify admin_delete_server_of_others test case https://review.openstack.org/71709 | 13:12 |
| openstackgerrit | Masayuki Igawa proposed a change to openstack/tempest: Clarify admin_delete_server_of_others test case https://review.openstack.org/71709 | 13:13 |
| *** cnesa has joined #openstack-qa | 13:16 | |
| *** ociuhandu has joined #openstack-qa | 13:16 | |
| openstackgerrit | Dirk Mueller proposed a change to openstack/tempest: Determine schema location properly even with symlinks https://review.openstack.org/75408 | 13:20 |
| openstackgerrit | Sean Dague proposed a change to openstack/tempest: add cli.has_manage option https://review.openstack.org/75245 | 13:20 |
| *** malini is now known as malini_afk | 13:30 | |
| openstackgerrit | afazekas proposed a change to openstack/tempest: Fix InvalidInstanceID.NotFound handling https://review.openstack.org/75385 | 13:34 |
| *** luqas has joined #openstack-qa | 13:35 | |
| *** mwagner_lap has quit IRC | 13:38 | |
| *** max_lobur is now known as max_lobur_afk | 13:41 | |
| *** russellb is now known as rustlebee | 13:44 | |
| *** gszasz has quit IRC | 13:46 | |
| *** mriedem has joined #openstack-qa | 13:48 | |
| openstackgerrit | afazekas proposed a change to openstack/tempest: Configure the ec2 zone explicitly https://review.openstack.org/75411 | 13:54 |
| *** sthaha has joined #openstack-qa | 13:54 | |
| *** sthaha has joined #openstack-qa | 13:54 | |
| openstackgerrit | afazekas proposed a change to openstack/tempest: Configure the ec2 zone explicitly https://review.openstack.org/75411 | 13:55 |
| sdague | dkranz: I see you are awake and reviewing :) | 13:56 |
| *** amotoki has quit IRC | 13:57 | |
| dkranz | sdague: Yes, first thing in the morning it all that is working for me lately | 13:57 |
| dkranz | sdague: I think next week may be better :) | 13:57 |
| sdague | by the time I got back to certain reviews you'd already approved them, it was funny :) | 13:58 |
| dkranz | sdague: Same here :) | 13:59 |
| sdague | hey, so I hadn't quite realized how much devstack defaults were all over the tempest.conf until this morning, the boto tests are especially terrible. | 13:59 |
| dkranz | sdague: Yes. I've got a question | 13:59 |
| sdague | I think we should not have defaults for any of those in tempest.conf, and be explicit on setting them in devstack | 13:59 |
| dkranz | sdague: I tripped across that stuff while writing a script to configure tempest.conf for havana in a more automated way | 14:00 |
| dkranz | sdague: I want to submit it to master but there are too many differences between havana and master for that to really work | 14:00 |
| dkranz | sdague: So I think I will submit it to master as WIP for comments | 14:00 |
| dkranz | sdague: And then try to merge it to master after it is all working on havana | 14:01 |
| dkranz | sdague: Our stable/havana policy makes it touch to do anything else | 14:01 |
| sdague | so step #1 is we need to submit the explicit piece in devstack | 14:01 |
| dkranz | sdague: Right | 14:01 |
| sdague | oh, because icehouse refactored so much? | 14:01 |
| dkranz | sdague: Yes, including auto-gen of the sample conf | 14:02 |
| dkranz | sdague: havana tempest.conf actgually has values in the sample that are different from the config.py | 14:02 |
| dkranz | sdague: So the default is wrong | 14:02 |
| openstackgerrit | afazekas proposed a change to openstack/tempest: Configure the ec2 zone explicitly https://review.openstack.org/75411 | 14:02 |
| sdague | yeh, well there is a reason to do that, because we really do need to fix master, otherwise a havana just gets lost | 14:03 |
| dkranz | sdague: ?? | 14:03 |
| dkranz | sdague: Oops. I am supposed to be at a meeting now. | 14:04 |
| sdague | if we fix havana but not master, then we're still broken for the base case :) | 14:04 |
| *** gszasz has joined #openstack-qa | 14:04 | |
| sdague | it only helps downstream folks, and only for one release | 14:04 |
| dkranz | sdague: We need to fix master. | 14:08 |
| sdague | agreed | 14:08 |
| sdague | well, push your patch, I'll go and see about a devstack master fix | 14:08 |
| sdague | I can use it as reference | 14:08 |
| dkranz | sdague: Sorry, my patch was just a script to auto-configure tempest.conf | 14:09 |
| dkranz | sdague: I just don't want to merge it to master yet because it needs more work. | 14:09 |
| dkranz | sdague: But I wanted to make sure we are all comfortable with the basic approach | 14:10 |
| sdague | how are you getting things like the images and endpoints? | 14:10 |
| *** saju_m has quit IRC | 14:10 | |
| dkranz | sdague: You provide the endpoint and it does everything else | 14:10 |
| dkranz | sdague: I'll push it as WIP | 14:10 |
| sdague | and you think you can fill out the entirety of tempest.conf from that? | 14:11 |
| dkranz | sdague: That is my goal | 14:11 |
| *** sthaha has quit IRC | 14:11 | |
| sdague | how do you handle local images? | 14:11 |
| dkranz | sdague: I have to pay attention to my call now. Back at 10 | 14:12 |
| openstackgerrit | afazekas proposed a change to openstack/tempest: pep8 tox entry should relay on site packages https://review.openstack.org/75414 | 14:12 |
| sdague | ok | 14:12 |
| sdague | I'm not convinced you could make this work for the gate because there are internal questions like is resize enabled | 14:13 |
| sdague | that are not auto discoverable | 14:13 |
| *** otherwiseguy has quit IRC | 14:14 | |
| *** banix has joined #openstack-qa | 14:16 | |
| openstackgerrit | Masayuki Igawa proposed a change to openstack/tempest: Cleanup _interface class variables in compute/v3 https://review.openstack.org/75415 | 14:17 |
| *** om has joined #openstack-qa | 14:19 | |
| *** zul has quit IRC | 14:20 | |
| openstackgerrit | Elena Ezhova proposed a change to openstack/tempest: Add tests for binding extended attributes for ports https://review.openstack.org/64271 | 14:23 |
| openstackgerrit | Sean Dague proposed a change to openstack/tempest: add cli.has_manage option https://review.openstack.org/75245 | 14:27 |
| *** rossella-s has joined #openstack-qa | 14:29 | |
| openstackgerrit | afazekas proposed a change to openstack/tempest: pep8 tox entry should not depend on site packages https://review.openstack.org/75414 | 14:30 |
| salv-orlando | sdague, mtreinish: I see you've enable the full job for neutron as non-voting superseding the patch for the parallel job | 14:32 |
| salv-orlando | which is good, but can I be a bit pedant and ask for the parallel job as well? I mean the smoke tests ran in parallel fashion | 14:32 |
| sdague | salv-orlando: it's actually parallel | 14:33 |
| salv-orlando | I'd like to keep it running for 2-3 weeks and then enable parallelism on voting jobs after icehouse is released (this way there won't be last minute drama) | 14:33 |
| sdague | that's the full parallel job | 14:33 |
| salv-orlando | I know but it will fail always before of the tests not implemented... | 14:33 |
| salv-orlando | sorry the tests not supported | 14:33 |
| sdague | it's passing | 14:33 |
| salv-orlando | uhm nope | 14:33 |
| salv-orlando | fails 100% | 14:34 |
| *** masayukig has quit IRC | 14:34 | |
| sdague | link? | 14:34 |
| sdague | http://logs.openstack.org/45/75245/2/check/check-tempest-dsvm-neutron-full/1473fb1/console.html is very clearly running in parallel | 14:35 |
| sdague | and passed | 14:35 |
| salv-orlando | oh you're right. What the hell was I looking at then? | 14:36 |
| salv-orlando | sorry for wasting your time then... | 14:36 |
| sdague | hey, we were as surprised as anyone yesterday on this | 14:36 |
| sdague | afazekas_: pointed out that it was passing on some experimental jobs | 14:36 |
| salv-orlando | so miracles happen | 14:36 |
| sdague | it looks like, from overnight, the regular neutron check job fails about %12 of the time | 14:36 |
| sdague | and the parallel one fails 20% | 14:37 |
| sdague | so it's not yet to the point where I'd want to make it vote, but if there was a hard push to clear up some of the fails, we could get there by i3 | 14:37 |
| salv-orlando | yeah I'm looking at those failures. I found an instance which might be of infra interest since the job fails at startup | 14:37 |
| salv-orlando | like before even starting devstack | 14:37 |
| sdague | yep, there have been some nodepool issues | 14:37 |
| salv-orlando | sdague: right but the one I'm talking about is in the zuul queue now : https://jenkins07.openstack.org/job/check-tempest-dsvm-neutron-isolated/339/console | 14:38 |
| *** jgrimm has quit IRC | 14:41 | |
| *** otherwiseguy has joined #openstack-qa | 14:42 | |
| sdague | ok, yeh, there is failures in another log there, it was probably an infra fail | 14:42 |
| *** Nikolay_St has joined #openstack-qa | 14:43 | |
| openstackrecheck | openstack-infra/devstack-gate change: https://review.openstack.org/72365 failed gate-tempest-dsvm-postgres-full in the gate queue with an unrecognized error | 14:46 |
| *** jordanP has quit IRC | 14:47 | |
| salv-orlando | sdague: thanks for looking at it. I would like to classify the failures of the full job, as well as the standard jobs. Pretty much I would just like to see all the failure for these jobs in the past 24 hours, see what are unclassified, triage them, file bugs, and whatever | 14:48 |
| openstackgerrit | Dirk Mueller proposed a change to openstack/tempest: Use Python 2.6.x compatible syntax for dict comprehension https://review.openstack.org/75422 | 14:49 |
| salv-orlando | only thing is that I have to admit my ignorance in not knowing a way for simply retrieve every failure for a given job | 14:49 |
| *** max_lobur_afk is now known as max_lobur | 14:49 | |
| openstackgerrit | A change was merged to openstack/tempest: Separate the deleting server tests https://review.openstack.org/74936 | 14:49 |
| sdague | salv-orlando: so, I can help with that, let me check in on a few other things first | 14:50 |
| sdague | that's basically how we build this - http://status.openstack.org/elastic-recheck/data/uncategorized.html | 14:50 |
| salv-orlando | no worries I'm not in a rush. Thanks | 14:50 |
| salv-orlando | yes indeed that where I usually look | 14:51 |
| afazekas_ | sdague: I need this change https://review.openstack.org/#/c/75388/ before I can add non DEBUG https://review.openstack.org/#/c/71474/ | 14:51 |
| openstackgerrit | Steven Hardy proposed a change to openstack-dev/devstack: Set stack_user_domain config correctly https://review.openstack.org/75424 | 14:51 |
| openstackgerrit | David Kranz proposed a change to openstack/tempest: Script to generate tempest.conf for a deployment https://review.openstack.org/75425 | 14:52 |
| openstackgerrit | A change was merged to openstack/tempest: Add "delete a shelved server" tests https://review.openstack.org/74939 | 14:58 |
| rahmu | hey, can anyone confirm they can reproduce this bug I just filed? https://bugs.launchpad.net/tempest/+bug/1283083 | 14:58 |
| openstackgerrit | A change was merged to openstack/tempest: Makes some Swift API tests optional https://review.openstack.org/75046 | 14:58 |
| *** malini_afk is now known as malini | 15:00 | |
| openstackgerrit | Mate Lakat proposed a change to openstack/tempest: Experimental patch to exclude some tempest tests https://review.openstack.org/75120 | 15:00 |
| sdague | afazekas_: while you are fixing boto things, the tempest config for boto really shouldn't have devstack defaults in it | 15:02 |
| sdague | those should be set by devstack | 15:02 |
| *** jordanP has joined #openstack-qa | 15:02 | |
| chmouel | malini: it was just a transient error (i could not replicate your issue as well on my newly spinned devstack) | 15:03 |
| afazekas_ | sdague: I wanted to do on both side | 15:04 |
| *** masayukig has joined #openstack-qa | 15:05 | |
| malini | chmouel: I'll try re-runnng the experimental job in jnekins.I saw it a couple of times in jenkins..But hope its transient | 15:05 |
| afazekas_ | I am also planning to add on option for explicitly specify the network id | 15:05 |
| malini | chmouel: it would have been a lot easier, if we cud repro this locally in devstack :( | 15:05 |
| dkranz | sdague: So I submitted https://review.openstack.org/#/c/75425/ which I hope will clarify what I am trying to do. | 15:05 |
| chmouel | malini: jenkins sets his service_host to 127.0.0.1 | 15:05 |
| chmouel | malini: so i guess marconi didn't run properly or crashed in some way so tempest could not contact it properly | 15:06 |
| sdague | afazekas_: so basically, we should never default filenames or service urls in tempest.conf | 15:06 |
| sdague | those need to be blank | 15:06 |
| afazekas_ | ok | 15:07 |
| *** afazekas_ has quit IRC | 15:07 | |
| *** luqas has quit IRC | 15:07 | |
| malini | chmouel: the logs say it started finehttp://logs.openstack.org/49/63449/16/experimental/check-tempest-dsvm-marconi/be62e9b/logs/screen-marconi-server.txt.gz .It cud have crashed..But I have never seen the same thing happen in my devstack servers ..& Itried quite a few times | 15:07 |
| *** shwetaap has joined #openstack-qa | 15:08 | |
| chmouel | malini: yep saw that but I don't see why binding to 127.0.0.1 from 0.0.0.0 would fix it | 15:08 |
| *** Nikolay_1t has joined #openstack-qa | 15:09 | |
| *** masayukig has quit IRC | 15:09 | |
| malini | chmouel: It was just a last resort try ..I am not confident tht wud change anything either | 15:09 |
| *** Nikolay_St has quit IRC | 15:09 | |
| sdague | dkranz: yeh, though I think this changes one set of embedded assumptions for another. A few less assumptions here, but still plenty of them. | 15:09 |
| dkranz | sdague: Well, it is a draft :) | 15:10 |
| *** luqas has joined #openstack-qa | 15:10 | |
| sdague | yeh | 15:10 |
| malini | chmouel: meanwhile I can update it to use service_host to .So we are consistent with other services | 15:10 |
| dkranz | sdague: serioulsy, it would help if you could enumerate the issues you know about | 15:10 |
| *** coolsvap has joined #openstack-qa | 15:10 | |
| dkranz | sdague: The fact that you cannot autodiscover resize is a but in the api | 15:10 |
| sdague | like the fact that you create flavors in there | 15:10 |
| dkranz | sdague: bug | 15:11 |
| dkranz | sdague: The creation is optional | 15:11 |
| chmouel | malini: cool, sounds a good idea | 15:11 |
| sdague | dkranz: well, plenty of compute features aren't discoverable in advance | 15:11 |
| sdague | like resize working or migration | 15:11 |
| *** rossella-s has quit IRC | 15:11 | |
| sdague | live migration being testable | 15:11 |
| dkranz | sdague: Right, those are bugs that need to be worked around | 15:11 |
| dkranz | sdague: My target was not really the gate though that would be nice to have | 15:12 |
| dkranz | sdague: and is why there is a patch file argument so you can customize the starting point | 15:12 |
| sdague | so that's my objection | 15:12 |
| sdague | if we can't come up with a tool that both works with the gate, and works with the vendor setups that exist | 15:13 |
| sdague | how do we think we have a tool that can work with any vendor setup | 15:13 |
| sdague | at which point I'd just much rather say - look, he's the config file, have at | 15:13 |
| dkranz | sdague: Well, if this approach is rejected, so be it | 15:14 |
| *** bhuvan has joined #openstack-qa | 15:14 | |
| dkranz | sdague: But I disagree that it is useful to tell folks doing deployments to have at the config file | 15:14 |
| dkranz | sdague: This tool would have to be customized by each vendor, which is why there is the --patch argument | 15:15 |
| dkranz | sdague: It is not nirvana but, IMO, much better than the current state | 15:15 |
| *** yfried has quit IRC | 15:15 | |
| afrittoli | drankz, sdague: about this I was thinking we could integrate with the tool which is actually doing the deployment - like in the gate, where devstack is deploying everything | 15:15 |
| dkranz | afrittoli: In that sense the gate is actually easier because there is only one deployment | 15:16 |
| afrittoli | dkranz, sdague: official openstack deployment tools is tripleo / heat, so it would be good to have something that works with that | 15:16 |
| sdague | afrittoli: which don't test in the gate :) | 15:16 |
| dkranz | sdague: Yet, but supposedly we will be | 15:16 |
| sdague | so here's the issue, it's about testability | 15:16 |
| sdague | if this tool can't work for the gate, then it's not testable | 15:17 |
| dkranz | sdague: It can work for the gate. | 15:17 |
| dkranz | sdague: At worst, no discovery is needed there because we control the deploy as well | 15:17 |
| sdague | ok, so start with proposing an experimental job which would use a tool like this | 15:17 |
| dkranz | sdague: But this sort of thing is impossible to really test in the gate because it dies on the multi-config issue. | 15:18 |
| sdague | and demonstrate we could do that, then we could talk about switching over | 15:18 |
| dkranz | sdague: It is a worst-case example of that | 15:18 |
| sdague | dkranz: what do you mean about multi-config? | 15:18 |
| *** relaxdiego has joined #openstack-qa | 15:18 | |
| dkranz | sdague: The fact that this script is supposed to deal with configuration details but we only use a few in the gate. | 15:19 |
| sdague | well it at least ensures that those don't break | 15:19 |
| dkranz | sdague: So it will be impossible to test it in the gate except for a narrow subset, just like we do for the rest of OpenStack | 15:19 |
| dkranz | sdague: Right | 15:19 |
| dkranz | sdague: I am not disagreeing with you except for one thing | 15:20 |
| dkranz | sdague: My priority right now is to get this working for real deployments | 15:20 |
| sdague | I'm just concerned in moving from a "we know this works" to "use this tool, but we don't test it, so meh" | 15:20 |
| dkranz | sdague: I would not do that at all | 15:20 |
| sdague | also, apparently enovance has a tool like that | 15:20 |
| dkranz | sdague: do you have a pointer? | 15:21 |
| sdague | which I only found out about by their posting | 15:21 |
| openstackgerrit | Rabi Mishra proposed a change to openstack-dev/devstack: Implements fix to run lbaas service on fedora with devstack https://review.openstack.org/75433 | 15:21 |
| sdague | https://www.openstack.org/vote-atlanta/Presentation/whatthestack-a-tool-based-on-tempest-to-verify-your-deployments | 15:22 |
| sdague | chmouel: you know anything about that? | 15:22 |
| sdague | and where one might find the source | 15:22 |
| dkranz | Is it this https://github.com/enovance/tempest-report | 15:22 |
| chmouel | sdague: hold on on the call will talk to you about it | 15:22 |
| dkranz | sdague: All I was saying about testing is that if we used something like this with devstack in the gate, then by definition it would keep working with devstack. | 15:23 |
| sdague | chmouel: cool, thanks | 15:24 |
| dkranz | sdague: But hard to make sure it kept working with every variant | 15:24 |
| sdague | dkranz: ok, sure. So that's the direction I'd like to see this evolve. Start with a new job that triggers this to config and overwrite the tempest conf | 15:24 |
| sdague | then evolve this and figure out if we can get parity to the existing approach | 15:24 |
| sdague | and if the answer is yes, then I'm for it | 15:25 |
| chmouel | yep so that's something that christian scheweed and nick barcet has been working on | 15:25 |
| chmouel | I don't think we have announced it properly yet excet via this presentation | 15:25 |
| dkranz | chmouel: Can you look at the patch I just pushed and tell me if it is "the same" or different? | 15:25 |
| chmouel | dkranz: about? | 15:26 |
| sdague | chmouel: so... realistically, if you guys don't get your code out there now, we're just going to implement the same thing | 15:26 |
| sdague | and make your stuff irrelevant | 15:26 |
| *** rossella-s has joined #openstack-qa | 15:26 | |
| chmouel | sdague: good point, I think that github repo is the latest one | 15:26 |
| sdague | tempest-report? | 15:26 |
| chmouel | do you want us to move it to stackfoge? | 15:26 |
| dkranz | chmouel: https://review.openstack.org/#/c/75425/1 | 15:26 |
| sdague | well, I'll let dkranz make a call on all that | 15:27 |
| chmouel | sdague: tempest-report is the repo i think whatthestack is the website | 15:27 |
| dkranz | chmouel: This is the first I heard of what you are doing. I would appreciate if you could look at what I just pushed and tell me if we should be collaborating or that what you are doing is different | 15:27 |
| chmouel | which is a nice UI | 15:27 |
| chmouel | dkranz: let me check with the engineer who's been working on that | 15:28 |
| dkranz | chmouel: ok | 15:28 |
| dkranz | chmouel: THe problem I was solving has no use for a ui | 15:28 |
| dkranz | chmouel: But there is a part around service-discovery where I could see use for that in a more general sense | 15:28 |
| dkranz | chmouel: The commit message clearly describes my goal. | 15:29 |
| chmouel | dkranz: I am getting christian to jump over here | 15:29 |
| *** bhuvan has quit IRC | 15:30 | |
| chmouel | as I wasn't much involve with this | 15:30 |
| *** cschwede_ has joined #openstack-qa | 15:30 | |
| openstackgerrit | Elena Ezhova proposed a change to openstack/tempest: Add tests for binding extended attributes for ports https://review.openstack.org/64271 | 15:30 |
| chmouel | dkranz, sdague: but from the first look the two have different goals | 15:31 |
| chmouel | (between the review and whatthestack) | 15:31 |
| cschwede_ | dkranz: i was just looking at your recent patch | 15:32 |
| *** thomasbiege1 has joined #openstack-qa | 15:32 | |
| cschwede_ | dkranz: whatthestack uses part of https://github.com/enovance/tempest-report | 15:34 |
| dkranz | cschwede_: Can you state your goal? | 15:35 |
| dkranz | cschwede_: Is it to "visualize" what the stack is providing? | 15:35 |
| *** thomasbiege1 has quit IRC | 15:36 | |
| *** thomasbiege has quit IRC | 15:36 | |
| dkranz | cschwede_: There was a reference to a "presentation". Is that available? | 15:36 |
| cschwede_ | dkranz: yes, to easily run tempest tests and get a quick glance on what is working and what not | 15:36 |
| cschwede_ | dkranz: presentation? | 15:36 |
| dkranz | cschwede_: chmouel mentioned it above | 15:36 |
| chmouel | dkranz: the whatthestack website? | 15:37 |
| dkranz | cschwede_: My patch is about how you setup tempest to run | 15:37 |
| *** jgrimm has joined #openstack-qa | 15:37 | |
| cschwede_ | dkranz: no, there is no public presentation yet. presentation will be hopefully during the summit :) | 15:37 |
| dkranz | cschwede_: Perhaps they are complements and not trying to do the same thing | 15:37 |
| chmouel | dkranz: oh yeah i mean the summit talk sorry for my frenglish | 15:38 |
| dkranz | chmouel: not a frenglish issue at all. I just didn't kknow you were referring to summit | 15:38 |
| dkranz | chmouel, cschwede_ : So I stated a clear goal and pushed a draft | 15:39 |
| cschwede_ | dkranz: yes, saw your patch, currently reading | 15:39 |
| dkranz | Can you do at least the first now? | 15:39 |
| cschwede_ | stating our goal? | 15:40 |
| dkranz | cschwede_: Yes | 15:40 |
| dkranz | cschwede_: You saw my goal in the commit message | 15:40 |
| dkranz | cschwede_: Most of the code is just details | 15:41 |
| *** mkoderer has quit IRC | 15:41 | |
| dkranz | cschwede_: Except for the argparse part | 15:41 |
| cschwede_ | dkranz: quite similar to yours: just enter your keystone credentials and let the tool discover available services and settings and finally write a tempest.conf | 15:41 |
| cschwede_ | dkranz: except that whattestack is a webbased-service | 15:42 |
| dkranz | cschwede_: Is this service actually running tempest too? | 15:42 |
| cschwede_ | dkranz: yes | 15:43 |
| dkranz | cschwede_: So what that really sounds like is rally | 15:43 |
| dkranz | cschwede_: Part of my motivation for this script was to enhance tempest/rally integration | 15:43 |
| *** max_lobur is now known as max_lobur_afk | 15:45 | |
| dkranz | cschwede_: what kind of service is this? paid, free, open-source, proprietary, etc? | 15:45 |
| *** jroovers|afk has joined #openstack-qa | 15:47 | |
| *** Nikolay_1t has quit IRC | 15:50 | |
| *** jroovers has quit IRC | 15:50 | |
| *** om has quit IRC | 15:52 | |
| *** jroovers|afk has quit IRC | 15:53 | |
| *** jordanP has quit IRC | 15:56 | |
| relaxdiego | hi, can anybody point me to the devstack local.conf used for the gating process? | 15:56 |
| cschwede_ | dkranz: tempest-report itself is opensource. afaik about rally use cases are different? so use case for rally is more benchmarking? | 15:59 |
| dkranz | cschwede_: At the moment. But I think the methodology of rally is similar. Such is the way of open source :) | 16:00 |
| *** markmcclain has joined #openstack-qa | 16:01 | |
| openstackgerrit | andrea-frittoli proposed a change to openstack/tempest: Refactor Managers to a common base class https://review.openstack.org/73704 | 16:02 |
| openstackgerrit | andrea-frittoli proposed a change to openstack/tempest: Defines a Credentials class https://review.openstack.org/74387 | 16:02 |
| dkranz | sdague, mtreinish: I am about to send out the message about contributing to negative test replacement but it would be good to get a +A on https://review.openstack.org/#/c/73982/ if possible. | 16:03 |
| cschwede_ | dkranz: yes, very true :) | 16:04 |
| *** masayukig has joined #openstack-qa | 16:05 | |
| *** coolsvap has quit IRC | 16:08 | |
| *** coolsvap1 has joined #openstack-qa | 16:08 | |
| *** coolsvap1 has quit IRC | 16:08 | |
| *** shwetaap has quit IRC | 16:09 | |
| openstackgerrit | A change was merged to openstack/tempest: Remove unsed client variables https://review.openstack.org/74893 | 16:10 |
| *** masayukig has quit IRC | 16:10 | |
| *** luqas has quit IRC | 16:13 | |
| *** coolsvap1 has joined #openstack-qa | 16:14 | |
| *** atiwari has joined #openstack-qa | 16:15 | |
| *** jordanP has joined #openstack-qa | 16:21 | |
| *** jlibosva has quit IRC | 16:22 | |
| *** otherwiseguy has quit IRC | 16:28 | |
| openstackgerrit | A change was merged to openstack/tempest: Refactor cross_tenant to security_groups_basic_ops https://review.openstack.org/69567 | 16:37 |
| *** mlavalle has joined #openstack-qa | 16:37 | |
| openstackgerrit | andrea-frittoli proposed a change to openstack/tempest: Refactor Managers to a common base class https://review.openstack.org/73704 | 16:39 |
| openstackgerrit | A change was merged to openstack/tempest: Splits network_basic_ops to fully isolated test cases https://review.openstack.org/66879 | 16:39 |
| cschwede_ | dkranz: coming back to your patch for tempest.conf (https://review.openstack.org/75425): i will try to integrate our code from tempest-report also, thus we have a common code base | 16:42 |
| dkranz | cschwede_: How will you do that? | 16:42 |
| *** shwetaap has joined #openstack-qa | 16:44 | |
| cschwede_ | dkranz: reviewing your code and then submit possible additional code as a follow up patch | 16:44 |
| dkranz | cschwede_: Ah, ok great! | 16:44 |
| dkranz | cschwede_: I think what you are doing overall has other goals but it would be good to separate the "configure tempest" part into a module that can be called or invoked as script. | 16:45 |
| openstackgerrit | Malini Kamalambal proposed a change to openstack-dev/devstack: Bind Marconi to SERVICE_HOST & add health check https://review.openstack.org/75160 | 16:45 |
| dkranz | cschwede_: You can then use that in the rest of your code. Does that make sense? | 16:45 |
| cschwede_ | dkranz: yes, that was my idea also. and then I will submit the CLI tool that also runs the tests and summarizes the results | 16:46 |
| dkranz | cschwede_: I am going to create a blueprint in tempest for this. | 16:46 |
| cschwede_ | dkranz: great! | 16:46 |
| *** otherwiseguy has joined #openstack-qa | 16:47 | |
| openstackgerrit | Julien Leloup proposed a change to openstack/tempest: Stress Test for server availability through SSH https://review.openstack.org/74067 | 16:48 |
| malini | relaxdiego: I believe you can get the localrc under the logs directory | 16:48 |
| cschwede_ | dkranz: and finally whatthestack will use that code also, so if someone doesn't want to use the CLI you can use the web app (which itself is a free service) to run tests against an openstack deployment | 16:48 |
| dkranz | cschwede_: sure, sounds good | 16:48 |
| *** cnesa has quit IRC | 16:49 | |
| relaxdiego | thanks @malini! i was actually look at that just now. | 16:49 |
| *** max_lobur_afk is now known as max_lobur | 16:54 | |
| salv-orlando | hi quality people! I remember some tempest cores a while ago pointed out log_ip_ns was a temporary thing and had eventually to go. With the better logging neutron now has, it seems really redundant. Personally I have not used those information in weeks. Is it ok for you if I push a quick patch for removing it? It's dumping so much stuff in the console log, that it's more of a hindrance than a help. | 16:55 |
| openstackgerrit | David Kranz proposed a change to openstack/tempest: Script to generate tempest.conf for a deployment https://review.openstack.org/75425 | 16:57 |
| malini | relaxdiego: yw | 16:57 |
| *** cnesa has joined #openstack-qa | 16:57 | |
| *** cschwede_ has quit IRC | 17:00 | |
| openstackgerrit | Dirk Mueller proposed a change to openstack/tempest: Fix s3 url default https://review.openstack.org/75463 | 17:00 |
| openstackgerrit | David Kranz proposed a change to openstack/tempest: Script to generate tempest.conf for a deployment https://review.openstack.org/75425 | 17:01 |
| *** rossella-s has quit IRC | 17:04 | |
| *** cnesa has quit IRC | 17:04 | |
| *** rossella-s has joined #openstack-qa | 17:06 | |
| *** masayukig has joined #openstack-qa | 17:06 | |
| *** rossella-s has quit IRC | 17:07 | |
| *** GGiora has quit IRC | 17:07 | |
| *** otherwiseguy has quit IRC | 17:08 | |
| *** masayukig has quit IRC | 17:10 | |
| *** cschwede_ has joined #openstack-qa | 17:15 | |
| *** relaxdiego has quit IRC | 17:16 | |
| *** mkollaro has quit IRC | 17:24 | |
| *** cschwede_ has quit IRC | 17:26 | |
| *** pbrooko has quit IRC | 17:27 | |
| *** jog0 is now known as flashgordon | 17:30 | |
| *** mkollaro has joined #openstack-qa | 17:32 | |
| *** relaxdiego has joined #openstack-qa | 17:37 | |
| *** devvesa has quit IRC | 17:37 | |
| openstackgerrit | Elena Ezhova proposed a change to openstack/tempest: Assign floating ip to a server in load balancer scenario https://review.openstack.org/71898 | 17:37 |
| *** dtalton has joined #openstack-qa | 17:39 | |
| *** cadenzajon has joined #openstack-qa | 17:39 | |
| *** vrovachev has left #openstack-qa | 17:40 | |
| *** dtalton has left #openstack-qa | 17:40 | |
| *** atiwari has quit IRC | 17:40 | |
| *** vkozhukalov has quit IRC | 17:40 | |
| *** dansmith is now known as damnsmith | 17:52 | |
| *** jordanP has quit IRC | 17:53 | |
| *** max_lobur is now known as max_lobur_afk | 17:55 | |
| *** harlowja_away is now known as harlowja | 17:56 | |
| *** wendar has quit IRC | 17:57 | |
| *** YorikSar has quit IRC | 18:00 | |
| *** otherwiseguy has joined #openstack-qa | 18:03 | |
| *** shwetaap has quit IRC | 18:04 | |
| *** masayukig has joined #openstack-qa | 18:07 | |
| *** hemnafk is now known as hemna | 18:09 | |
| *** masayukig has quit IRC | 18:11 | |
| *** marun has joined #openstack-qa | 18:12 | |
| openstackrecheck | openstack/tempest change: https://review.openstack.org/73503 failed check-tempest-dsvm-postgres-full, check-tempest-dsvm-cells-full because of: https://bugs.launchpad.net/bugs/1262153 | 18:14 |
| *** damnsmith is now known as busymany | 18:18 | |
| *** busymany is now known as damnsmith | 18:18 | |
| *** cadenzajon has quit IRC | 18:21 | |
| *** vponomaryov has left #openstack-qa | 18:24 | |
| *** morganfainberg_Z is now known as morganfainberg | 18:24 | |
| *** cadenzajon has joined #openstack-qa | 18:28 | |
| *** atiwari has joined #openstack-qa | 18:32 | |
| *** YorikSar has joined #openstack-qa | 18:38 | |
| *** markmcclain has quit IRC | 18:52 | |
| *** shwetaap has joined #openstack-qa | 19:07 | |
| afrittoli | mtreinish, dkranz, giulivo, sdague: I rebased 73704 (and thus lost mtreinish +2). Check is OK. May I ask for re-review and merge if the change is ok, so I can progress on the multiauth bp ? | 19:08 |
| *** wendar has joined #openstack-qa | 19:11 | |
| *** locke105 has joined #openstack-qa | 19:12 | |
| dkranz | afrittoli: I gave +2 again | 19:12 |
| *** shwetaap has quit IRC | 19:14 | |
| *** shwetaap has joined #openstack-qa | 19:18 | |
| *** tosky has quit IRC | 19:20 | |
| openstackgerrit | Aaron Rosen proposed a change to openstack-dev/devstack: Ensure die is called if git fetch/checkout fails https://review.openstack.org/75486 | 19:20 |
| *** vkozhukalov has joined #openstack-qa | 19:25 | |
| *** cnesa has joined #openstack-qa | 19:27 | |
| *** cnesa has quit IRC | 19:31 | |
| *** masayukig has joined #openstack-qa | 19:35 | |
| *** ociuhandu has quit IRC | 19:37 | |
| *** masayukig has quit IRC | 19:39 | |
| *** mrodden has joined #openstack-qa | 19:40 | |
| *** bill_az_ has joined #openstack-qa | 19:43 | |
| afrittoli | dkranz: thanks! | 19:52 |
| *** protux_ has joined #openstack-qa | 19:57 | |
| *** protux has quit IRC | 19:57 | |
| *** protux_ has joined #openstack-qa | 19:57 | |
| *** harlowja is now known as harlowja_away | 19:59 | |
| *** markmcclain has joined #openstack-qa | 20:02 | |
| *** harlowja_away is now known as harlowja | 20:08 | |
| *** malini is now known as malini_afk | 20:13 | |
| *** cnesa has joined #openstack-qa | 20:16 | |
| *** afrittoli has quit IRC | 20:18 | |
| *** YorikSar has quit IRC | 20:27 | |
| *** shwetaap has quit IRC | 20:28 | |
| *** otherwiseguy has quit IRC | 20:29 | |
| *** cadenzajon has quit IRC | 20:29 | |
| *** shwetaap has joined #openstack-qa | 20:29 | |
| *** cnesa has quit IRC | 20:34 | |
| *** masayukig has joined #openstack-qa | 20:36 | |
| *** andreaf has quit IRC | 20:37 | |
| *** andreaf has joined #openstack-qa | 20:37 | |
| *** cnesa has joined #openstack-qa | 20:38 | |
| *** ociuhandu has joined #openstack-qa | 20:40 | |
| *** masayukig has quit IRC | 20:41 | |
| *** YorikSar has joined #openstack-qa | 20:48 | |
| *** ekarlso has quit IRC | 20:50 | |
| *** ekarlso has joined #openstack-qa | 20:50 | |
| *** atiwari has quit IRC | 20:50 | |
| *** shwetaap has quit IRC | 20:54 | |
| *** bradsturtevant has joined #openstack-qa | 21:00 | |
| *** ociuhandu has quit IRC | 21:01 | |
| *** atiwari has joined #openstack-qa | 21:03 | |
| *** rfolco has quit IRC | 21:07 | |
| *** cadenzajon has joined #openstack-qa | 21:13 | |
| *** cnesa has quit IRC | 21:13 | |
| *** cnesa has joined #openstack-qa | 21:18 | |
| *** dtalton has joined #openstack-qa | 21:21 | |
| *** dtalton2 has joined #openstack-qa | 21:22 | |
| openstackgerrit | A change was merged to openstack/tempest: Separate non-admin AZ tests from admin test files https://review.openstack.org/74898 | 21:25 |
| *** dtalton has quit IRC | 21:26 | |
| *** markmcclain has quit IRC | 21:31 | |
| *** shwetaap has joined #openstack-qa | 21:33 | |
| *** masayukig has joined #openstack-qa | 21:36 | |
| mriedem | are there any plans for moving tempest to use python-openstackclient? | 21:40 |
| *** masayukig has quit IRC | 21:41 | |
| openstackgerrit | David Kranz proposed a change to openstack/tempest: Script to generate tempest.conf for a deployment https://review.openstack.org/75425 | 21:43 |
| *** jeblair has quit IRC | 21:45 | |
| *** markmcclain has joined #openstack-qa | 21:46 | |
| *** dtalton2 has quit IRC | 21:46 | |
| *** jeblair has joined #openstack-qa | 21:50 | |
| *** dtalton has joined #openstack-qa | 21:52 | |
| *** dtalton2 has joined #openstack-qa | 21:53 | |
| *** yassine has quit IRC | 21:54 | |
| *** dtalton3 has joined #openstack-qa | 21:55 | |
| *** dtalton3 has left #openstack-qa | 21:56 | |
| *** dtalton has quit IRC | 21:58 | |
| *** dtalton2 has quit IRC | 21:59 | |
| *** gszasz has quit IRC | 22:05 | |
| *** lyxus has joined #openstack-qa | 22:05 | |
| lyxus | Hello Folks, I was wondering what was the difference. If I am running something like testr run tempest.api.network.test_networks.NetworksIpV6TestJSON.test_list_networks, the test will be run without problems. However if the [gate,smoke] is appended at the end. The test won't be run. Any ideas/pointers? | 22:05 |
| clarkb | lyxus: oh actually now that I think about it more, the string you pass testr run is a regex | 22:11 |
| clarkb | you might need to escape the []s | 22:11 |
| boris-42_ | dkranz hi | 22:14 |
| dkranz | boris-42_: hello | 22:15 |
| boris-42_ | dkranz I just saw your patch about tempest config | 22:15 |
| boris-42_ | dkranz it will absolutely unusable in case of Rally | 22:16 |
| boris-42_ | will be* | 22:16 |
| dkranz | boris-42_: What do you think? It is still missing things | 22:16 |
| boris-42_ | yep | 22:16 |
| dkranz | boris-42_: Why | 22:16 |
| dkranz | ? | 22:16 |
| boris-42_ | let me tyr to describe case | 22:16 |
| boris-42_ | how I would like to be able to use it | 22:16 |
| boris-42_ | with tempest.config.InitCloudForTempest(): | 22:17 |
| *** jgrimm has quit IRC | 22:17 | |
| boris-42_ | or even another | 22:17 |
| boris-42_ | with tempest.config.GenerateConifg as conf: | 22:17 |
| boris-42_ | call tempest with this config | 22:18 |
| dkranz | boris-42_: So you want to separate the module from the script? | 22:18 |
| dkranz | boris-42_: That is pretty trivial | 22:18 |
| boris-42_ | in __enter__ you will do all work around crating resources | 22:18 |
| boris-42_ | in __exit__ cleanup | 22:18 |
| dkranz | boris-42_: ok | 22:18 |
| boris-42_ | and I don't see cleanup stuff here ..' | 22:18 |
| boris-42_ | another thing probably you should split | 22:19 |
| boris-42_ | https://review.openstack.org/#/c/75425/4/tools/config_tempest.py | 22:19 |
| boris-42_ | to something like "lib" | 22:19 |
| dkranz | boris-42_: No, because this script is not about running tempest, but about configuring it | 22:19 |
| dkranz | boris-42_: Yes, I agree. | 22:19 |
| boris-42_ | that has API to do different steps | 22:19 |
| dkranz | boris-42_: THe module should be inside tempest. | 22:19 |
| boris-42_ | and call of it | 22:20 |
| boris-42_ | dkranz not sure that I get latest sentence.. | 22:20 |
| dkranz | boris-42_: You should put comments in the review... | 22:20 |
| dkranz | boris-42_: You are raising two issues | 22:20 |
| dkranz | 1. The bulk of the code should be in a module inside tempest, and the tools/.. script should call it | 22:21 |
| dkranz | Agreed | 22:21 |
| *** masayukig has joined #openstack-qa | 22:21 | |
| dkranz | 2. This code should be a class that supports more methods than create-conf-file: it should run, cleanup, etc. | 22:21 |
| dkranz | boris-42_: Is that right? | 22:21 |
| boris-42_ | dkranz yep both right | 22:22 |
| boris-42_ | dkranz when I say "lib" | 22:22 |
| boris-42_ | I thought separated files for running stuff script | 22:22 |
| dkranz | boris-42_: So the first is a simple code reorg which I will do | 22:22 |
| boris-42_ | and bunch of methods that do all works | 22:22 |
| boris-42_ | everything except women is simple in this life=) | 22:22 |
| dkranz | boris-42_: I suggest you add a review comment with the methods you would like to see supported. | 22:23 |
| *** fbo is now known as fbo_away | 22:24 | |
| dkranz | boris-42_: I am not so sure these should all be in one class as opposed to having a function that configures, another one that cleans up, etc. | 22:24 |
| boris-42_ | dkranz about ContexClass | 22:24 |
| dkranz | boris-42_: what is that? | 22:25 |
| boris-42_ | with SomeStuffThatMakesMeHappy: | 22:27 |
| boris-42_ | dkranz like we have in Rally https://github.com/stackforge/rally/blob/master/rally/benchmark/runner.py#L66 | 22:28 |
| boris-42_ | so you are creating users inside context object | 22:29 |
| boris-42_ | and if something went wrong you are always deleting them | 22:30 |
| boris-42_ | without spaghetti of try/except | 22:30 |
| dkranz | boris-42_: So I was trying to limit the scope of this but if you can put in the blueprint exactly how you think this should work we can discuss that. | 22:34 |
| boris-42_ | dkranz sure (not sure that I am able today ) | 22:34 |
| dkranz | boris-42_: np | 22:34 |
| boris-42_ | but on weekend I will add how to make it useful for Rally | 22:34 |
| boris-42_ | one thing that I would like is to help to adopt tempest | 22:35 |
| dkranz | boris-42_: The only reason I combined creation of resources with creation of tempest.conf at all is because they have to match | 22:35 |
| boris-42_ | to better integration | 22:35 |
| dkranz | boris-42_: right | 22:35 |
| dkranz | boris-42_: But there are other use cases than rally for this functionality | 22:35 |
| boris-42_ | sure | 22:35 |
| boris-42_ | Rally is a just 1 of users | 22:36 |
| dkranz | boris-42_: So we have to first understand them all | 22:36 |
| boris-42_ | let we make at least useful for 1 project) | 22:36 |
| openstackgerrit | A change was merged to openstack/tempest: EC2 client token is not reuseable for new instance creation https://review.openstack.org/75388 | 22:36 |
| dkranz | boris-42_: ok, we can check back on Monday | 22:36 |
| boris-42_ | definitely | 22:37 |
| boris-42_ | dkranz what i would like as well is to have results of tempest in JSON | 22:37 |
| boris-42_ | dkranz what do you think where is the best place to make this? tempest or Rally? | 22:38 |
| dkranz | boris-42_: So that is a totally different issue | 22:38 |
| boris-42_ | dkranz sure | 22:38 |
| boris-42_ | dkranz but anyway ?) | 22:38 |
| dkranz | boris-42_: I think tempest already generates xunit | 22:38 |
| dkranz | boris-42_: BUt it is really the test runner | 22:38 |
| lifeless | boris-42_: tempest output is in subunit | 22:39 |
| dkranz | boris-42_: tempest itself has no output except for log messages | 22:39 |
| lifeless | boris-42_: you can convert subunit to xunit (but thats lossy, so why would you?) | 22:39 |
| dkranz | lifeless: He says he wants json | 22:39 |
| boris-42_ | ya | 22:39 |
| boris-42_ | "JSON" | 22:39 |
| lifeless | oh | 22:39 |
| dkranz | lifeless: I just suggested he convert that from xunit | 22:39 |
| boris-42_ | cause "JSON" I can analyze | 22:39 |
| dkranz | lifeless: or subunit | 22:39 |
| dkranz | boris-42_: The point is that it is not tempest, it is the runner | 22:40 |
| lifeless | boris-42_: analyzing in python? there's good python libraries for subunit, or we can do a subunit2json (I've been meaning to do that anyway) | 22:40 |
| boris-42_ | lifeless i would like as well to store this stuff in DB | 22:40 |
| boris-42_ | lifeless and i think that at the end DB will be mongo | 22:40 |
| boris-42_ | lifeless so storing in JSON makes sense | 22:40 |
| lifeless | sure, if you like | 22:41 |
| boris-42_ | lifeless dkranz so I think this is part of Rally? | 22:41 |
| boris-42_ | we will run and process results inside it | 22:41 |
| lifeless | boris-42_: please tell me rally won't depend on mongo | 22:41 |
| * clarkb is with lifeless | 22:41 | |
| boris-42_ | lifeless no it won't' depend on mongo | 22:42 |
| lifeless | whew :) | 22:42 |
| boris-42_ | lifeless but I would like to make support for mongo as well | 22:42 |
| boris-42_ | implementing one more db backend | 22:42 |
| boris-42_ | lifeless so no worries=) | 22:42 |
| boris-42_ | lifeless the key reason for mongo is next | 22:43 |
| *** shwetaap has quit IRC | 22:43 | |
| lifeless | next what? | 22:43 |
| boris-42_ | lifeless when we finish base steps I will write results share system | 22:43 |
| boris-42_ | lifeless like pastebing | 22:43 |
| boris-42_ | pastebin | 22:43 |
| boris-42_ | lifeless http://pavlovic.me/rally/new_snapshot.html | 22:43 |
| lifeless | sure; we want to do machine learning for tempest in the gate too | 22:43 |
| boris-42_ | ^ but you will get link to something like that ^ | 22:43 |
| lifeless | its on our copious todo lists | 22:43 |
| lifeless | I'd really like something native for that though, so testr can plug straight into it | 22:44 |
| clarkb | also put the json data in postgres >_> | 22:44 |
| boris-42_ | clarkb we are putting now=) | 22:44 |
| *** shwetaap has joined #openstack-qa | 22:44 | |
| boris-42_ | lifeless btw as you here | 22:44 |
| boris-42_ | lifeless so what about installing Rally with cloud in tripleO ? | 22:45 |
| boris-42_ | lifeless as well we can make tripleO based engine inside Rally (as now we have devstack, and going to have Fuel) | 22:47 |
| lifeless | sounds great | 22:47 |
| boris-42_ | lifeless I will ask Roman to help me a bit with installing Rally in TripleO cause I am totally newbie.. | 22:48 |
| lifeless | :) | 22:48 |
| boris-42_ | lifeless btw would you like to see something aghh | 22:48 |
| lifeless | on saturday morning... not really :) | 22:49 |
| *** malini_afk is now known as malini | 22:50 | |
| boris-42_ | lifeless http://pavlovic.me/rally/create_and_delete_user.html | 22:52 |
| *** masayukig has quit IRC | 22:52 | |
| boris-42_ | lifeless why keystone spends so much time to create user? | 22:53 |
| *** masayukig has joined #openstack-qa | 22:53 | |
| lifeless | that is arggh :) | 22:53 |
| lifeless | dunno, it doesn't take that long for tripleo | 22:54 |
| boris-42_ | it is 30 concurrent requests | 22:54 |
| boris-42_ | sometimes it works 1sec | 22:54 |
| boris-42_ | sometimes 14 | 22:54 |
| lifeless | oh, because keystone and concurrency are not friends | 22:54 |
| boris-42_ | lol | 22:54 |
| boris-42_ | and keystone & production? | 22:54 |
| lifeless | like, its fixable, someone just needs to roll their sleeves up and focus on it | 22:54 |
| morganfainberg | keystone eventlet is single threaded | 22:54 |
| lifeless | rather than new APIs or whatever, make it better. | 22:55 |
| lifeless | morganfainberg: shouldn't matter | 22:55 |
| lifeless | morganfainberg: like in principle - work with me - how much /actual/ logic is needed to make a user? 10ms worth maybe? + lets allow 5ms to flush a commit to disk ? | 22:55 |
| lifeless | morganfainberg: so 30 requests - 450ms. | 22:55 |
| morganfainberg | lifeless, eh. | 22:55 |
| lifeless | morganfainberg: as a target goal. | 22:55 |
| lifeless | morganfainberg: but we take longer -> there's room to optimise. | 22:56 |
| morganfainberg | lifeless, i think there are other more glaring performance issues than user creation | 22:56 |
| lifeless | morganfainberg: sure are ;) | 22:56 |
| morganfainberg | lifeless, that should be addressed first, and that is largely mitigated by running under apache (Yes, i know... still needs to be done for devstack/etc) | 22:56 |
| boris-42_ | seems like I should stop working around graphs | 22:57 |
| boris-42_ | and continue working around OpenStack profiler | 22:57 |
| boris-42_ | so we will get where there problem is | 22:57 |
| lifeless | morganfainberg: what is the reference deployment? | 22:57 |
| morganfainberg | aannnyway | 22:57 |
| lifeless | morganfainberg: we run keystone standalone today like all the other openstack services, | 22:57 |
| morganfainberg | lifeless, the best practice deployment is keystone under mod_wsgi | 22:57 |
| lifeless | morganfainberg: with an ssl frontend | 22:57 |
| *** masayukig has quit IRC | 22:57 | |
| morganfainberg | lifeless, but devstack (obstensibly reference) is single eventlet | 22:57 |
| clarkb | "best practice" <- note this isn't tested 15k times a day | 22:57 |
| morganfainberg | clarkb, i know. | 22:58 |
| lifeless | morganfainberg: so let me rephrase. | 22:58 |
| morganfainberg | clarkb, we need to fix that | 22:58 |
| lifeless | morganfainberg: tripleo. upstream reference production deployment. | 22:58 |
| morganfainberg | clarkb, long long list :( higher on my list but still blocking on other thigns | 22:58 |
| lifeless | morganfainberg: we're installing whatever setup.py install does. | 22:58 |
| lifeless | morganfainberg: which is standalone at the moment. | 22:58 |
| morganfainberg | lifeless, setup.py can't do apache | 22:58 |
| morganfainberg | sorry, that would be a bad idea imo | 22:58 |
| morganfainberg | very bad :P | 22:59 |
| morganfainberg | but i would say, if at all possible, running under mod_wsgi would be my target for "Reference" deployment | 22:59 |
| lifeless | morganfainberg: http://docs.openstack.org/developer/keystone/installing.html | 22:59 |
| lifeless | documents 'keystone-all' | 22:59 |
| lifeless | for 'using keystone' | 22:59 |
| morganfainberg | but... i don't want that until we test it more (e.g. gate) | 22:59 |
| clarkb | morganfainberg: I know. I am just going to continue to give you guys a hard time about it until reality reflects what you say :) | 22:59 |
| morganfainberg | clarkb, ^ | 22:59 |
| lifeless | morganfainberg: tripleo has check jobs running this on every tripleo change now | 23:00 |
| morganfainberg | clarkb, i wont get on much more than the talking point until we fix devstack for gate | 23:00 |
| lifeless | morganfainberg: and 'check experimental' in keystone will run the same tripleo jobs. | 23:00 |
| clarkb | keystone can say wsgi + apache is the reference recommended best practice but until it is actually tested it really isnt | 23:00 |
| morganfainberg | clarkb, and i wont propose a change to triple-o stuff until we gate on it | 23:00 |
| lifeless | morganfainberg: I don't think devstack has to do apache (in fact I'd argue it shouldn't) | 23:00 |
| clarkb | lifeless: well the port issue has forced the issue | 23:00 |
| morganfainberg | lifeless, i think devstack should support that out of the box well, and i think we need to gate on it | 23:01 |
| lifeless | morganfainberg: you can get CI on it in the tripleo jobs - they aren't up to gating yet, but we're working on it | 23:01 |
| morganfainberg | lifeless, in fact, i think we should be gating on both tbh | 23:01 |
| morganfainberg | clarkb, ^ | 23:01 |
| morganfainberg | unless we're dropping eventlet | 23:01 |
| clarkb | morganfainberg: thats fine if we can fix the port issue | 23:01 |
| lifeless | morganfainberg: yes, just saying that where something is in conflict with developer friendliness, I think trpleo is a viable gate strategy for it | 23:01 |
| lifeless | clarkb: whats the port issue? | 23:01 |
| clarkb | lifeless: 35357 is a linux ephemeral port | 23:01 |
| morganfainberg | lifeless, IANA and linux ephemeral ports don't match | 23:01 |
| morganfainberg | lifeless, we are assigned 35357 | 23:01 |
| morganfainberg | linux says that's ephemeral even though IANA says otherwise | 23:02 |
| lifeless | you can fix the linux ephemeral range when you install | 23:02 |
| lifeless | no? | 23:02 |
| morganfainberg | clarkb, what would you think of a devstack-g for keystone that is eventlet, but everything else gets httpd? | 23:02 |
| morganfainberg | lifeless, it needs to be baked into the image to be effective | 23:02 |
| lifeless | morganfainberg: because you need to reboot? | 23:03 |
| morganfainberg | lifeless, nope, just anything running could consume the port | 23:03 |
| morganfainberg | lifeless, so installing the sysctl override at stack.sh time doens't solve the issue | 23:03 |
| lifeless | morganfainberg: so put it in prepare_node | 23:03 |
| lifeless | in devstack_gate | 23:03 |
| morganfainberg | lifeless, still doesn't solve it unless that is what roles the image we print nodes from | 23:03 |
| morganfainberg | ? | 23:03 |
| lifeless | it is | 23:03 |
| boris-42_ | lifeless clarkb morganfainberg dkranz btw is somebody interesting to hear about Rally I making about 1 hangout call per wiki for developers | 23:04 |
| morganfainberg | annnd. i've been told that is suboptimal | 23:04 |
| boris-42_ | deep dive into code + live demo | 23:04 |
| boris-42_ | just to understand idea of project=0 | 23:04 |
| lifeless | boris-42_: maybe after feature freeze | 23:04 |
| lifeless | right now human bandwidth is streeeetched | 23:04 |
| morganfainberg | lifeless, i was told we shouldn't "fix" the ephemeral ports, we should fix keystone | 23:04 |
| boris-42_ | as usually =) | 23:04 |
| clarkb | lifeless: so that can fix it for the gate but not for devstack | 23:04 |
| morganfainberg | boris-42_, ++ post FF would be better for me. | 23:04 |
| clarkb | we want to actually test devstack | 23:04 |
| lifeless | morganfainberg: by using a different port than the IANA gave you ? | 23:05 |
| clarkb | any changes to the node images to unbreak bugs in devstack is a bug | 23:05 |
| boris-42_ | Okay guys I will ping you=) after FF | 23:05 |
| morganfainberg | lifeless, at that point we should use apache | 23:05 |
| boris-42_ | but if you will be ready before just ping me=) | 23:05 |
| morganfainberg | clarkb, i disagree since this is a linux issue...but | 23:05 |
| clarkb | lifeless: I proposed the change to use a different port but people really like the IANA assignment | 23:05 |
| lifeless | clarkb: mmm, linux using ports IANA say they shouldn't is not a devstack bug. | 23:05 |
| clarkb | and my change was shot down quit quikcly | 23:05 |
| lifeless | morganfainberg: how does using apache fix the port issue ? | 23:05 |
| clarkb | lifeless: right its an IANA bug | 23:05 |
| morganfainberg | clarkb, no, if we use a different port we should just abandon the port completely | 23:05 |
| morganfainberg | lifeless, port 80/443 | 23:05 |
| morganfainberg | clarkb, no linux bug imo | 23:06 |
| morganfainberg | clarkb, linux is the one refusing to adhere to the authority on port assignments | 23:06 |
| clarkb | morganfainberg: its both and neither side will budge I am sure | 23:06 |
| lifeless | IANA and Linux distros can beat each other up over it | 23:06 |
| morganfainberg | lifeless, yep | 23:06 |
| clarkb | morganfainberg: right bu linux is making a decision based on performance and stuff | 23:06 |
| openstackgerrit | Chris Buccella proposed a change to openstack/tempest: Skip test_get_instance_action test for nova v3 https://review.openstack.org/73044 | 23:06 |
| clarkb | like you need more ports than IANA says you can have | 23:06 |
| lifeless | clarkb: unlike the IANA ? :P | 23:06 |
| lifeless | clarkb: next you'll be blaming the IETF | 23:07 |
| clarkb | lifeless: everyone :P | 23:07 |
| morganfainberg | lifeless, since we can use port 80/443, we can be sure nothing else is there but httpd. | 23:07 |
| clarkb | also why isn't ipv6 loopback a /64 or something | 23:07 |
| openstackgerrit | Joe Gordon proposed a change to openstack-infra/elastic-recheck: Add fingerprint for bug 1283283 https://review.openstack.org/75533 | 23:07 |
| morganfainberg | unless someone is doing something that is against VERY well established | 23:07 |
| clarkb | this /128 is really annoying :P | 23:07 |
| lifeless | clarkb: I know right | 23:07 |
| morganfainberg | clarkb, ugh. ipv6 | 23:07 |
| * morganfainberg shoves ipv6 soapbox under the desk. | 23:07 | |
| lifeless | morganfainberg: so, I'm fine with whatever port you choose :) but I still don't see how thats related to running in mod_wsgi specifically | 23:07 |
| morganfainberg | lifeless, mod_wsgi also gives us multiple keystone processes | 23:08 |
| morganfainberg | lifeless, it's a 2-fold win, sane ports, and better concurrency | 23:08 |
| lifeless | morganfainberg: I am not a huge fan of mod_wsgi (or really any such tightly coupled servers) | 23:08 |
| lifeless | morganfainberg: so, I probably need to sit down and understand why people think more processes is a good first thing to fix the performance issues we see | 23:08 |
| morganfainberg | lifeless, we could use unicorn, or any WSGI thing, but mod_wsgi also plays into the fact we want to use mod_shib (?) for SAML assertion processing for federation | 23:08 |
| lifeless | morganfainberg: because it makes no sense to me | 23:09 |
| morganfainberg | lifeless, we have some issues with serialization as well due to crypto work and relying on it to sign tokens | 23:09 |
| lifeless | morganfainberg: openssl? its terrrible ain't it :( | 23:09 |
| morganfainberg | lifeless, yep | 23:09 |
| lifeless | morganfainberg: but - a multiprocessing worker pool will handle that beautifully | 23:09 |
| lifeless | morganfainberg: get all the openssl code out of your process | 23:10 |
| morganfainberg | lifeless, yeah, but we do use apache for other things, and eventlet afaict does not play nice with that | 23:10 |
| morganfainberg | lifeless, we offer apache as a way to provide external auth (SAML processing, mod_basic_auth, etc) | 23:10 |
| morganfainberg | lifeless, it's a many many pronged reason | 23:10 |
| *** banix has quit IRC | 23:10 | |
| morganfainberg | lifeless, there is a lot of work to move towards sane pools that can do that. | 23:11 |
| lifeless | morganfainberg: ok, so sure - but you mentioned performance, and IME multiple processes for performance just makes liveness issues worse *unless* the specific issue that was being encountered was GIL threading fairness | 23:11 |
| lifeless | if its bad(O) code in the server, more processes just causes process thrashing (bad) unless you tune timeslices and then you get an unresponsive server (bad) | 23:12 |
| morganfainberg | lifeless, sure, except GIL reasons are only one factor when you have multiple backends to deal with. some data in ldap, some in sql, etc | 23:12 |
| morganfainberg | lifeless, we have plenty of bad O code | 23:12 |
| morganfainberg | lifeless, but we also aggregate data across a ton of systems. | 23:13 |
| morganfainberg | lifeless, s/ton/variety | 23:13 |
| morganfainberg | lifeless, if it was always SQL and always the same set of SQL, it is easier | 23:13 |
| morganfainberg | lifeless, be we do have deployments where some stuff is LDAP, some SQL, some in-mem, some in memcache, and we need to get the data from all of them. there is still a limit to the benefit of coroutine-esque performance gains in those scenarior | 23:14 |
| morganfainberg | os* | 23:14 |
| morganfainberg | grr. | 23:14 |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/devstack-gate: Add timestamps to devstack-gate output https://review.openstack.org/75534 | 23:14 |
| lifeless | morganfainberg: so, what I'd do is have a service per system you're connecting do and do RPC calls from the API to each backend | 23:14 |
| openstackgerrit | A change was merged to openstack-infra/elastic-recheck: Add fingerprint for bug 1283283 https://review.openstack.org/75533 | 23:14 |
| morganfainberg | lifeless, that is an architecture change i'm not opposed to | 23:14 |
| lifeless | morganfainberg: make each one small, lean, focused and easy to analyse (and report on) | 23:14 |
| openstackgerrit | A change was merged to openstack/tempest: Clean up scenario functions https://review.openstack.org/75339 | 23:14 |
| morganfainberg | lifeless, but we're a bit from being able ot do that | 23:14 |
| openstackgerrit | A change was merged to openstack/tempest: Add qcow2 image support https://review.openstack.org/75312 | 23:15 |
| morganfainberg | lifeless, but yes. it is on my "what i'd like to see" kind of setup (ish) | 23:15 |
| lifeless | cool | 23:15 |
| morganfainberg | lifeless, so, for immediacy, i see three things: 1: multiple backends, 2: OpenSSL *sigh*, 3: bad(O) code | 23:16 |
| lifeless | morganfainberg: did you see | 23:16 |
| morganfainberg | lifeless, 1 and 2 do absolutely benefit from multiple processes as we work on better™ approaches | 23:16 |
| morganfainberg | 3 should be addressed in either case | 23:16 |
| morganfainberg | damn it :P lifeless , i swore to my self i'd leave my soapbox at home today :P | 23:17 |
| lifeless | morganfainberg: https://bugs.launchpad.net/tripleo/+bug/1282842 | 23:17 |
| morganfainberg | lifeless, nope hadn't seen that one yet | 23:17 |
| lifeless | so 2 - I don't understand why you can't use a multiprocessing pool from eventlet | 23:17 |
| lifeless | morganfainberg: there's some reason to think its keystone giving up the ghost | 23:17 |
| lifeless | morganfainberg: and keystones default logging is - excuse me I don't mean to be mean - awful. | 23:18 |
| morganfainberg | lifeless, we could. in fact we should support that too. but if you want external auth/federation w/ SAML we need mod_shib | 23:18 |
| morganfainberg | lifeless, i think we have some patches in flight to help w/ logging | 23:18 |
| *** mriedem has quit IRC | 23:18 | |
| morganfainberg | lifeless, in fact.. i am fairly certain i saw one or two | 23:18 |
| lifeless | morganfainberg: well I'm saying that the openssl signing stuff is still an issue if its in your process at all because its just - well, you know ;). So I'd run a process pool even with mod_wsgi | 23:19 |
| morganfainberg | lifeless, yep. | 23:19 |
| lifeless | morganfainberg: and I can't imagine that its *not* appropriately setup to just split out | 23:20 |
| morganfainberg | lifeless, and we need multiple ways to handle it nicely | 23:20 |
| lifeless | or even run as a RPC service locally on the same machine | 23:20 |
| morganfainberg | lifeless, it isn't setup well for that atm. | 23:20 |
| morganfainberg | lifeless, i hope we are "fixing" that in J. | 23:20 |
| morganfainberg | lifeless, one of my main targets for J, be 100% performance focused | 23:20 |
| morganfainberg | let other people work on features | 23:20 |
| morganfainberg | but i want to streamline keystone a lot | 23:21 |
| lifeless | \o/ | 23:21 |
| morganfainberg | lifeless, i'll know how much i'll be focused on that for J after i have a chat w/ the guys who pay my bills next week :) | 23:22 |
| morganfainberg | lifeless, ok wtf is with those logs | 23:22 |
| morganfainberg | thats... bad | 23:22 |
| morganfainberg | give me a moment to finish some quick copy/pasta for a review and i'll take a closer look | 23:23 |
| morganfainberg | lifeless, ^ | 23:23 |
| lifeless | morganfainberg: this is our config https://git.openstack.org/cgit/openstack/tripleo-image-elements/tree/elements/keystone/os-apply-config/etc/keystone/keystone.conf | 23:23 |
| lifeless | morganfainberg: we take the sample, then for things folk want to change we templateise them, everything else we leave as is | 23:24 |
| morganfainberg | 5 mins to finish adding helpstrings for our sample keystone config and post it :P | 23:24 |
| lifeless | more-or-less | 23:24 |
| lifeless | we try anyhow | 23:24 |
| *** openstackrecheck has quit IRC | 23:24 | |
| openstackgerrit | A change was merged to openstack-infra/elastic-recheck: Suppress bug 1269940 from the graphs page https://review.openstack.org/72731 | 23:24 |
| *** openstackrecheck has joined #openstack-qa | 23:24 | |
| *** malini is now known as malini_afk | 23:26 | |
| morganfainberg | lifeless, ok looking at the sample now | 23:27 |
| morganfainberg | erm | 23:27 |
| morganfainberg | your config | 23:27 |
| morganfainberg | ah, | 23:27 |
| morganfainberg | ok i think what has happened here is we moved to the standard oslo logging constructs | 23:28 |
| morganfainberg | and got rid of our bad icky stuff | 23:28 |
| lifeless | so we default to no settings at all | 23:28 |
| morganfainberg | basically, we're getting the bare-minimum logging | 23:28 |
| morganfainberg | warning+ only i think | 23:28 |
| morganfainberg | which seems about right for what you're seeing | 23:28 |
| morganfainberg | let me see what the settings should be now | 23:28 |
| lifeless | For something handling HTTP requests, I'd expect an httpd-style log by default, no ? | 23:29 |
| morganfainberg | lifeless, well, it's going to look more like nova's now? | 23:29 |
| lifeless | you still have # Print more verbose output (set logging level to INFO instead | 23:29 |
| lifeless | # of default WARNING level). (boolean value) | 23:29 |
| lifeless | in keystone.conf.sample | 23:29 |
| lifeless | #verbose=false | 23:29 |
| lifeless | *anyhow*, point is - the default, just whines about API versions | 23:30 |
| morganfainberg | lifeless, yeah | 23:30 |
| lifeless | on every request or something | 23:30 |
| lifeless | we can't see how much traffic (no http request logs) | 23:30 |
| lifeless | so there's no way to get even the slightest hint about behaviour without changing the defaults... this seems wrong to me | 23:30 |
| lifeless | If I may be so bold | 23:31 |
| lifeless | the deprecation warning should log once per process lifetime, or perhaps at startup if the v2 API is enabled. | 23:31 |
| morganfainberg | lifeless, that is a fair assessment | 23:31 |
| lifeless | it could add an HTTP Warning header to the response | 23:31 |
| lifeless | that way folk using the old API could tell, if they wanted to. Sysadmins don't give a rats. | 23:32 |
| lifeless | [and have no direct control either] | 23:32 |
| lifeless | (also - we're seeing that implies that neutron / heat/nova - something - is using the v2, so we're internally using a deprecated API. ouch). | 23:32 |
| lifeless | and secondly, I'd really like to see a one-line HTTPd style log item per HTTP request | 23:33 |
| lifeless | so that I can run any vanilla http log aggregator against it. | 23:33 |
| morganfainberg | lifeless, hm.. | 23:33 |
| morganfainberg | if we use httpd, you'll absolutely get that | 23:33 |
| morganfainberg | lifeless, i'm not surprised people are using v2 still | 23:34 |
| *** openstackrecheck has quit IRC | 23:34 | |
| *** openstackrecheck has joined #openstack-qa | 23:34 | |
| morganfainberg | s/people/services | 23:35 |
| lifeless | bbiaw | 23:35 |
| *** rustlebee is now known as russellb | 23:36 | |
| morganfainberg | lifeless, i'll see if i can figure out what settings will hlep for immidiate time for your logs in keystone | 23:36 |
| morganfainberg | lifeless, talk w/ ya later (monday at the latest) | 23:37 |
| *** shwetaap has quit IRC | 23:45 | |
| *** shwetaap has joined #openstack-qa | 23:46 | |
| *** damnsmith is now known as dansmith | 23:46 | |
| *** giulivo has quit IRC | 23:48 | |
| openstackrecheck | Console logs not available after 13:24s for gate-keystone-python33 75526,1,d971da3 | 23:51 |
| *** masayukig has joined #openstack-qa | 23:53 | |
| openstackgerrit | Joe Gordon proposed a change to openstack-infra/elastic-recheck: Add fingerprint for bug 1282876 https://review.openstack.org/75542 | 23:55 |
| *** jhenner has quit IRC | 23:55 | |
| *** masayukig has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!