| *** rodrigods_ has joined #openstack-qa | 00:07 | |
| *** malini1 has joined #openstack-qa | 00:09 | |
| *** kmartin has quit IRC | 00:11 | |
| *** asselin has quit IRC | 00:11 | |
| *** kmartin has joined #openstack-qa | 00:11 | |
| *** asselin has joined #openstack-qa | 00:11 | |
| *** asselin has quit IRC | 00:14 | |
| *** asselin has joined #openstack-qa | 00:15 | |
| *** rodrigods__ has joined #openstack-qa | 00:15 | |
| *** rodrigods_ has quit IRC | 00:15 | |
| *** dmorita has joined #openstack-qa | 00:26 | |
| openstackrecheck | openstack/swift change: https://review.openstack.org/97615 failed gate-grenade-dsvm in the gate queue with an unrecognized error | 00:31 |
|---|---|---|
| *** aimon has joined #openstack-qa | 00:31 | |
| *** asselin has quit IRC | 00:33 | |
| *** xianghui has joined #openstack-qa | 00:34 | |
| *** gmann_ has joined #openstack-qa | 00:37 | |
| *** gmann has quit IRC | 00:38 | |
| *** liusheng-leave is now known as liusheng | 01:00 | |
| openstackrecheck | Console logs not available after 13:25s for gate-swift-docs 88736,32,7a624d7 | 01:04 |
| openstackrecheck | openstack/nova change: https://review.openstack.org/101883 failed gate-nova-python26 in the gate queue with an unrecognized error | 01:04 |
| *** nosnos has joined #openstack-qa | 01:13 | |
| *** nosnos has quit IRC | 01:20 | |
| *** Sukhdev has quit IRC | 01:22 | |
| *** weshay has quit IRC | 01:23 | |
| *** atiwari has quit IRC | 01:28 | |
| *** Obulpathi has joined #openstack-qa | 01:31 | |
| *** Obulpathi has quit IRC | 01:31 | |
| *** marun is now known as marun_afk | 01:39 | |
| openstackrecheck | openstack/nova change: https://review.openstack.org/91727 failed because of: gate-grenade-dsvm: unrecognized error, gate-grenade-dsvm-partial-ncpu: https://bugs.launchpad.net/bugs/1298472 | 01:48 |
| *** rodrigods__ has quit IRC | 01:55 | |
| *** rodrigods_ has joined #openstack-qa | 01:57 | |
| *** Matt1 has quit IRC | 02:09 | |
| *** ekhugen has quit IRC | 02:15 | |
| *** rodrigods_ has quit IRC | 02:24 | |
| openstackrecheck | openstack/nova change: https://review.openstack.org/104009 failed gate-grenade-dsvm in the gate queue with an unrecognized error | 02:25 |
| *** bkopilov has quit IRC | 02:35 | |
| openstackrecheck | openstack/tempest change: https://review.openstack.org/99231 failed because of: gate-tempest-dsvm-postgres-full: https://bugs.launchpad.net/bugs/1286818, gate-grenade-dsvm: unrecognized error | 02:58 |
| openstackrecheck | openstack/tempest change: https://review.openstack.org/98971 failed because of: gate-grenade-dsvm-partial-ncpu: https://bugs.launchpad.net/bugs/1298472 | 03:07 |
| *** SridharG has joined #openstack-qa | 03:10 | |
| *** jyuso has joined #openstack-qa | 03:21 | |
| *** salv-orlando has quit IRC | 03:21 | |
| openstackgerrit | Matt Riedemann proposed a change to openstack-infra/elastic-recheck: Add query for pypi mirror infra bug 1334550 https://review.openstack.org/104432 | 03:28 |
| openstackgerrit | Zhi Kun Liu proposed a change to openstack/tempest: Enable some volumes v2 tests by sharing codes https://review.openstack.org/100193 | 03:32 |
| *** dims has quit IRC | 03:32 | |
| *** mriedem has quit IRC | 03:34 | |
| *** Poornima has joined #openstack-qa | 03:35 | |
| *** sweston has quit IRC | 03:42 | |
| *** malini1 has quit IRC | 03:43 | |
| *** zhiyan_ is now known as zhiyan | 03:47 | |
| openstackgerrit | ajay proposed a change to openstack/tempest: Add three new CLI tests of sahara https://review.openstack.org/100784 | 03:49 |
| openstackgerrit | ajay proposed a change to openstack/tempest: Improve cinder CLI existing tests https://review.openstack.org/101694 | 03:51 |
| openstackgerrit | ajay proposed a change to openstack/tempest: Improve nova CLI existing tests https://review.openstack.org/101767 | 03:51 |
| *** nosnos has joined #openstack-qa | 03:52 | |
| *** mlavalle has joined #openstack-qa | 03:59 | |
| *** bkopilov has joined #openstack-qa | 04:01 | |
| *** Madan has joined #openstack-qa | 04:02 | |
| *** jhesketh has quit IRC | 04:14 | |
| *** ajc_ has joined #openstack-qa | 04:17 | |
| openstackgerrit | Daryl Walleck proposed a change to openstack/tempest: Replaces mutable default parameters with a safe implementation https://review.openstack.org/104443 | 04:21 |
| openstackgerrit | Daryl Walleck proposed a change to openstack/tempest: Replaces mutable default parameters with a safe implementation https://review.openstack.org/104443 | 04:24 |
| *** yfried has quit IRC | 04:27 | |
| *** k4n0 has joined #openstack-qa | 04:39 | |
| *** pcrews has quit IRC | 04:42 | |
| *** jhesketh has joined #openstack-qa | 04:52 | |
| *** psharma has joined #openstack-qa | 04:54 | |
| *** Longgeek has joined #openstack-qa | 04:55 | |
| *** amitpp has joined #openstack-qa | 04:55 | |
| *** harikav has joined #openstack-qa | 04:59 | |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/devstack-gate: Handle proposed/ branches in test-matrix.py https://review.openstack.org/104448 | 05:00 |
| *** nati_ueno has joined #openstack-qa | 05:02 | |
| *** Madan has quit IRC | 05:03 | |
| openstackrecheck | Console logs not available after 13:23s for gate-swift-python26 104441,1,99a7a2c | 05:03 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/devstack-gate: Handle proposed/ branches in test-matrix.py https://review.openstack.org/104448 | 05:10 |
| *** pcrews has joined #openstack-qa | 05:11 | |
| openstackrecheck | Console logs not available after 13:24s for check-tempest-dsvm-neutron-heat-slow 104440,1,34d0248 | 05:17 |
| *** Madan has joined #openstack-qa | 05:19 | |
| *** Madan has joined #openstack-qa | 05:20 | |
| openstackgerrit | A change was merged to openstack/tempest: Remove the skip for test_get_instance_action() https://review.openstack.org/99231 | 05:20 |
| openstackrecheck | openstack/nova change: https://review.openstack.org/104021 failed because of: gate-grenade-dsvm: unrecognized error, gate-tempest-dsvm-full: https://bugs.launchpad.net/bugs/1336755 | 05:20 |
| *** rdekel has joined #openstack-qa | 05:21 | |
| *** Madan1 has joined #openstack-qa | 05:25 | |
| *** yfried has joined #openstack-qa | 05:26 | |
| *** Madan has quit IRC | 05:27 | |
| *** sabeen has joined #openstack-qa | 05:27 | |
| *** pcrews has quit IRC | 05:28 | |
| *** marun_afk has quit IRC | 05:29 | |
| openstackgerrit | A change was merged to openstack/tempest: Correcting 'id' type in compute schema https://review.openstack.org/98971 | 05:29 |
| *** Madan1 has quit IRC | 05:38 | |
| kashyap | afazekas, Morning, have you seen these kind of time outs: | 05:40 |
| kashyap | ./safe-devstack-vm-gate-wrap.sh: line 284: 14756 Killed timeout -s 9 ${REMAINING_TIME}m $BASE/new/devstack-gate/devstack-vm-gate.sh | 05:40 |
| kashyap | http://logs.openstack.org/90/102490/4/check/check-tempest-dsvm-neutron/81659ec/console.html | 05:40 |
| *** Madan has joined #openstack-qa | 05:41 | |
| *** Longgeek has quit IRC | 05:45 | |
| *** Longgeek has joined #openstack-qa | 05:45 | |
| *** sweston has joined #openstack-qa | 05:47 | |
| *** harlowja is now known as harlowja_away | 05:47 | |
| *** Longgeek has quit IRC | 05:50 | |
| *** ianw has quit IRC | 06:05 | |
| *** ianw has joined #openstack-qa | 06:05 | |
| openstackgerrit | Ghanshyam Mann proposed a change to openstack/tempest: Handle 'HTTPNotFound' in scenario tests cleanup https://review.openstack.org/104458 | 06:08 |
| *** coolsvap|afk is now known as coolsvap | 06:09 | |
| *** jhesketh has quit IRC | 06:11 | |
| *** jlibosva has joined #openstack-qa | 06:18 | |
| openstackgerrit | Zhi Kun Liu proposed a change to openstack/tempest: Enable some volumes v2 tests by sharing codes https://review.openstack.org/100193 | 06:20 |
| *** nati_ueno has quit IRC | 06:20 | |
| openstackgerrit | Ghanshyam Mann proposed a change to openstack/tempest: unskip list server actions tests https://review.openstack.org/90278 | 06:23 |
| *** chandan_kumar has joined #openstack-qa | 06:25 | |
| *** Longgeek has joined #openstack-qa | 06:27 | |
| *** mlavalle has quit IRC | 06:30 | |
| openstackgerrit | A change was merged to openstack-dev/devstack: Add ability to enable OSprofiler https://review.openstack.org/103835 | 06:34 |
| *** mlavalle has joined #openstack-qa | 06:36 | |
| *** ildikov has joined #openstack-qa | 06:37 | |
| openstackgerrit | Yair Fried proposed a change to openstack/tempest: Convert scenario test tearDown to addCleanup https://review.openstack.org/62101 | 06:44 |
| morganfainberg | does tempest know what version (e.g. Icehouse) it is running against? I have a change that is pending that (yes I know... changes return, ML thread is available on it) that would change a 204 to a 200 return, so icehouse would be 204, juno and beyond 200. | 06:47 |
| morganfainberg | do i just need to accept both? or can i gate based upon minversion? | 06:47 |
| *** Anju has joined #openstack-qa | 06:49 | |
| openstackgerrit | Ghanshyam Mann proposed a change to openstack/qa-specs: Add qa spec for bp/cli-tests-coverage https://review.openstack.org/102433 | 06:52 |
| openstackgerrit | Ghanshyam Mann proposed a change to openstack/qa-specs: Add qa spec for bp/cli-tests-coverage https://review.openstack.org/102433 | 06:53 |
| openstackgerrit | ajay proposed a change to openstack/tempest: Improve cinder CLI existing tests https://review.openstack.org/101694 | 07:01 |
| *** jhesketh_ has joined #openstack-qa | 07:04 | |
| openstackgerrit | ajay proposed a change to openstack/tempest: Add three new CLI tests of sahara https://review.openstack.org/100784 | 07:05 |
| *** Anju has quit IRC | 07:05 | |
| *** Anju has joined #openstack-qa | 07:06 | |
| openstackgerrit | ajay proposed a change to openstack/tempest: Improve nova CLI existing tests https://review.openstack.org/101767 | 07:06 |
| *** afazekas_ has joined #openstack-qa | 07:15 | |
| *** nati_ueno has joined #openstack-qa | 07:20 | |
| *** gszasz has joined #openstack-qa | 07:27 | |
| *** gmann_ has quit IRC | 07:29 | |
| *** nati_ueno has quit IRC | 07:33 | |
| *** mmaglana has quit IRC | 07:35 | |
| *** mmaglana has joined #openstack-qa | 07:35 | |
| *** mlavalle has quit IRC | 07:36 | |
| *** leseb has joined #openstack-qa | 07:37 | |
| *** safchain has joined #openstack-qa | 07:37 | |
| openstackgerrit | afazekas proposed a change to openstack-infra/elastic-recheck: Add query for pypi index fetching issues https://review.openstack.org/104298 | 07:40 |
| *** mmaglana has quit IRC | 07:40 | |
| *** luqas has joined #openstack-qa | 07:41 | |
| *** mlavalle has joined #openstack-qa | 07:42 | |
| *** luqas has quit IRC | 07:42 | |
| *** ildikov_ has joined #openstack-qa | 07:42 | |
| *** luqas has joined #openstack-qa | 07:45 | |
| *** ildikov has quit IRC | 07:45 | |
| *** jlanoux has joined #openstack-qa | 07:54 | |
| openstackrecheck | Console logs not available after 13:25s for check-grenade-dsvm 89298,16,ff44b07 | 07:55 |
| *** gfidente has joined #openstack-qa | 07:55 | |
| *** samuelbercovici has joined #openstack-qa | 07:55 | |
| *** lanoux_ has joined #openstack-qa | 07:57 | |
| *** jlanoux has quit IRC | 07:57 | |
| *** jpich has joined #openstack-qa | 07:59 | |
| *** vrovachev has joined #openstack-qa | 07:59 | |
| *** nosnos has quit IRC | 08:05 | |
| *** salv-orlando has joined #openstack-qa | 08:05 | |
| *** nosnos has joined #openstack-qa | 08:05 | |
| *** afazekas_ is now known as afzekas | 08:05 | |
| *** afzekas is now known as afazekas_ | 08:07 | |
| *** tkammer has joined #openstack-qa | 08:07 | |
| openstackgerrit | Vadim Rovachev proposed a change to openstack/tempest: Add glance notifications tests for ceilometer https://review.openstack.org/71259 | 08:08 |
| *** samuelbercovici1 has joined #openstack-qa | 08:11 | |
| openstackrecheck | Console logs not available after 13:25s for check-tempest-dsvm-full-icehouse 100193,7,3067802 | 08:11 |
| *** samuelbercovici has quit IRC | 08:14 | |
| *** samuelbercovici1 is now known as samuelbercovici | 08:14 | |
| *** samuelbercovici1 has joined #openstack-qa | 08:18 | |
| openstackgerrit | FeihuJiang proposed a change to openstack/tempest: Adds an orchestration API test for resource types https://review.openstack.org/103775 | 08:18 |
| openstackgerrit | Rabi Mishra proposed a change to openstack/tempest: Templest test change for heat-cli changes https://review.openstack.org/104473 | 08:19 |
| *** samuelbercovici has quit IRC | 08:21 | |
| *** samuelbercovici1 is now known as samuelbercovici | 08:21 | |
| openstackrecheck | openstack/horizon change: https://review.openstack.org/104456 failed gate-grenade-dsvm in the gate queue with an unrecognized error | 08:24 |
| *** zz_johnthetubagu is now known as johnthetubaguy | 08:27 | |
| openstackgerrit | Rabi Mishra proposed a change to openstack/tempest: Templest test change for heat-cli changes https://review.openstack.org/104473 | 08:27 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack-dev/devstack: Ironic: Setup/Configure iPXE https://review.openstack.org/99677 | 08:35 |
| *** luqas has quit IRC | 08:37 | |
| *** luqas has joined #openstack-qa | 08:37 | |
| openstackgerrit | A change was merged to openstack/tempest: Improve neutron CLI existing tests https://review.openstack.org/101707 | 08:37 |
| openstackgerrit | ajay proposed a change to openstack/tempest: Improve nova CLI existing tests https://review.openstack.org/101767 | 08:37 |
| openstackrecheck | Console logs not available after 13:24s for check-tempest-dsvm-full 100473,6,69963c7 | 08:37 |
| openstackgerrit | Zhi Kun Liu proposed a change to openstack/tempest: Enable some volumes v2 tests by sharing codes https://review.openstack.org/100193 | 08:46 |
| openstackgerrit | ajay proposed a change to openstack/tempest: Improve nova CLI existing tests https://review.openstack.org/101767 | 08:46 |
| openstackgerrit | Rabi Mishra proposed a change to openstack/tempest: Templest test changes for heat-cli changes https://review.openstack.org/104473 | 08:49 |
| openstackgerrit | Chmouel Boudjnah proposed a change to openstack-dev/devstack: Hide output when configuring apt retry https://review.openstack.org/103539 | 08:50 |
| openstackrecheck | Console logs not available after 13:22s for check-tempest-dsvm-postgres-full 104189,1,cad0941 | 08:53 |
| openstackgerrit | Vishal kumar mahajan proposed a change to openstack/tempest: Add three new CLI tests of keystone https://review.openstack.org/102109 | 08:56 |
| *** luqas has quit IRC | 08:58 | |
| *** Longgeek has quit IRC | 09:04 | |
| *** dmorita has quit IRC | 09:04 | |
| openstackrecheck | Console logs not available after 13:23s for gate-tempest-dsvm-full 99000,10,67159b3 | 09:10 |
| *** Anju has quit IRC | 09:13 | |
| *** mlavalle_ has joined #openstack-qa | 09:15 | |
| *** mlavalle has quit IRC | 09:15 | |
| *** mlavalle_ is now known as mlavalle | 09:15 | |
| *** mitz has quit IRC | 09:22 | |
| openstackrecheck | Console logs not available after 13:22s for check-grenade-dsvm 96489,3,8596e24 | 09:23 |
| *** samuelbercovici1 has joined #openstack-qa | 09:25 | |
| *** mitz has joined #openstack-qa | 09:26 | |
| *** samuelbercovici has quit IRC | 09:29 | |
| *** samuelbercovici1 is now known as samuelbercovici | 09:29 | |
| *** luqas has joined #openstack-qa | 09:33 | |
| *** samuelbercovici1 has joined #openstack-qa | 09:33 | |
| openstackgerrit | Yair Fried proposed a change to openstack/tempest: Convert scenario test tearDown to addCleanup https://review.openstack.org/62101 | 09:35 |
| *** raies has joined #openstack-qa | 09:36 | |
| *** mmaglana has joined #openstack-qa | 09:36 | |
| raies | afazekas_: hi | 09:36 |
| *** samuelbercovici has quit IRC | 09:36 | |
| *** samuelbercovici1 is now known as samuelbercovici | 09:36 | |
| openstackrecheck | Console logs not available after 13:25s for check-tempest-dsvm-full 103895,3,9c743c8 | 09:37 |
| openstackgerrit | Ishant Tyagi proposed a change to openstack/tempest: Add orchestration scenario test for stack-abandon and stack-adopt https://review.openstack.org/103198 | 09:37 |
| *** mmaglana has quit IRC | 09:40 | |
| *** qba73 has joined #openstack-qa | 09:43 | |
| openstackgerrit | Vadim Rovachev proposed a change to openstack/tempest: Add glance notifications tests for ceilometer https://review.openstack.org/71259 | 09:47 |
| *** samuelbercovici1 has joined #openstack-qa | 09:49 | |
| openstackrecheck | Console logs not available after 13:25s for check-grenade-dsvm 102721,1,c454f85 | 09:51 |
| openstackgerrit | A change was merged to openstack-dev/devstack: Add libyaml-devel package dependancy https://review.openstack.org/102490 | 09:51 |
| *** samuelbercovici has quit IRC | 09:53 | |
| *** samuelbercovici1 is now known as samuelbercovici | 09:53 | |
| *** mlavalle has quit IRC | 09:54 | |
| *** samuelbercovici1 has joined #openstack-qa | 09:57 | |
| *** samuelbercovici has quit IRC | 10:00 | |
| *** samuelbercovici1 is now known as samuelbercovici | 10:00 | |
| *** samuelbercovici1 has joined #openstack-qa | 10:02 | |
| openstackrecheck | Console logs not available after 13:29s for check-tempest-dsvm-postgres-full 104272,2,2e9de64 | 10:04 |
| *** samuelbercovici has quit IRC | 10:06 | |
| *** samuelbercovici1 is now known as samuelbercovici | 10:06 | |
| *** samuelbercovici1 has joined #openstack-qa | 10:08 | |
| *** samuelbercovici has quit IRC | 10:12 | |
| *** samuelbercovici1 is now known as samuelbercovici | 10:12 | |
| *** samuelbercovici1 has joined #openstack-qa | 10:14 | |
| *** samuelbercovici has quit IRC | 10:17 | |
| *** samuelbercovici1 is now known as samuelbercovici | 10:17 | |
| openstackrecheck | Console logs not available after 13:28s for check-tempest-dsvm-postgres-full-icehouse 104123,2,7cf9fdd | 10:18 |
| *** leseb has quit IRC | 10:18 | |
| *** leseb has joined #openstack-qa | 10:19 | |
| *** samuelbercovici1 has joined #openstack-qa | 10:22 | |
| openstackgerrit | Yair Fried proposed a change to openstack/tempest: Fix manager.status_timeout exception https://review.openstack.org/103547 | 10:22 |
| *** leseb has quit IRC | 10:24 | |
| *** gmann has joined #openstack-qa | 10:24 | |
| *** samuelbercovici has quit IRC | 10:25 | |
| *** samuelbercovici1 is now known as samuelbercovici | 10:25 | |
| *** samuelbercovici1 has joined #openstack-qa | 10:26 | |
| *** marun has joined #openstack-qa | 10:28 | |
| sdague | kashyap / DinaBelova I noticed that ceilometer test yesterday is now our top gate bug. Any idea if there is a skip moving through yet? | 10:29 |
| sdague | or if anyone made progress figuring out what it is | 10:29 |
| kashyap | sdague, Morning, | 10:29 |
| sdague | kashyap: morning | 10:29 |
| kashyap | What do you mean"a skip moving through" -- sorry, couldn't parse it | 10:29 |
| sdague | if someone proposed a skip to that test in tempest | 10:29 |
| *** samuelbercovici has quit IRC | 10:30 | |
| *** samuelbercovici1 is now known as samuelbercovici | 10:30 | |
| kashyap | sdague, Didn't check in tempest, let me do a local git pull to see. Was looking at other Neutron failures I saw | 10:30 |
| openstackgerrit | A change was merged to openstack/tempest: Code clean-up of CLI test "cinder list-extensions" https://review.openstack.org/100498 | 10:30 |
| openstackrecheck | Console logs not available after 13:27s for check-tempest-dsvm-full 98164,1,742c4f6 | 10:32 |
| openstackrecheck | openstack/horizon change: https://review.openstack.org/100656 failed gate-horizon-python26, gate-horizon-python27, gate-horizon-python27-django14 in the gate queue with an unrecognized error | 10:33 |
| kashyap | sdague, I don't notice a skip in current Tempest master | 10:33 |
| openstackrecheck | openstack/horizon change: https://review.openstack.org/104456 failed gate-grenade-dsvm in the gate queue with an unrecognized error | 10:33 |
| * kashyap goes to #ceilometer | 10:33 | |
| sdague | kashyap: ok, cool | 10:34 |
| sdague | just curious, as I was waking up :) | 10:34 |
| kashyap | :-) No worries | 10:35 |
| * kashyap didnt' have internet for the last 3 hours, just got it, so catching up on stuff | 10:35 | |
| *** abhi- has joined #openstack-qa | 10:37 | |
| openstackgerrit | Yair Fried proposed a change to openstack/tempest: Convert scenario test tearDown to addCleanup https://review.openstack.org/62101 | 10:38 |
| kashyap | sdague, I don't have a response yet. If it's just that to add a skip, I can submit a trivial patch -- @tempest.test.skip_because(bug="1336755") | 10:43 |
| kashyap | But I'd like a Ceilometer person's input | 10:43 |
| sdague | yeh, I think that's the right call | 10:45 |
| *** samuelbercovici1 has joined #openstack-qa | 10:45 | |
| kashyap | sdague, Which one? :-) Go ahead w/ the trivial patch or waiting for the Ceilo person's input? | 10:45 |
| openstackrecheck | Console logs not available after 13:22s for check-tempest-dsvm-full 102778,4,fb10caf | 10:47 |
| kashyap | Anyway, I proposed it on #openstack-ceilometer too, will wait a bit, probably I guess they're in the mid-cycle meet-up | 10:47 |
| sdague | oh, the patch :) | 10:47 |
| sdague | we'll need test results back on it anyway | 10:47 |
| kashyap | Okay, will submit it, what's the worse - someone will point a resolution, or accept the skip for now | 10:48 |
| *** samuelbercovici has quit IRC | 10:48 | |
| *** samuelbercovici1 is now known as samuelbercovici | 10:48 | |
| *** leseb has joined #openstack-qa | 10:49 | |
| *** samuelbercovici1 has joined #openstack-qa | 10:51 | |
| *** Longgeek has joined #openstack-qa | 10:52 | |
| *** leseb has quit IRC | 10:54 | |
| *** samuelbercovici has quit IRC | 10:55 | |
| *** Rajalakshmi has joined #openstack-qa | 10:55 | |
| *** samuelbercovici1 is now known as samuelbercovici | 10:55 | |
| *** abhi- has quit IRC | 10:57 | |
| *** abhi- has joined #openstack-qa | 10:59 | |
| openstackgerrit | Kashyap Chamarthy proposed a change to openstack/tempest: Skip telemetry test - test_check_nova_notification() https://review.openstack.org/104516 | 11:00 |
| openstackrecheck | Console logs not available after 13:22s for check-grenade-dsvm-partial-ncpu 104486,1,b7a63fe | 11:01 |
| *** samuelbercovici1 has joined #openstack-qa | 11:04 | |
| harikav | abhi- : Hi | 11:04 |
| harikav | abhi- : I wanted to know, is this your submission (https://review.openstack.org/#/c/92524/) ? | 11:06 |
| *** samuelbercovici has quit IRC | 11:07 | |
| *** samuelbercovici1 is now known as samuelbercovici | 11:07 | |
| harikav | If so, implementation of Andrea Frittoli's feedback, is dependent on this submission too (https://review.openstack.org/#/c/90370/) | 11:08 |
| *** gmann has quit IRC | 11:08 | |
| harikav | Can you please confirm on your implementation of the code change in test.py file? | 11:10 |
| *** Anu_ has joined #openstack-qa | 11:11 | |
| abhi- | sorry was away.. that was not me | 11:14 |
| openstackrecheck | Console logs not available after 13:22s for check-grenade-dsvm 99000,10,5705571 | 11:15 |
| *** tkammer has quit IRC | 11:16 | |
| harikav | Ok. Sorry from me too, confused with the nick name abhi thought it would be (Abhijeet Jain). Thanks for your reply. | 11:17 |
| abhi- | np :)..might need to use a more specific nick next time | 11:19 |
| kashyap | sdague, Okay - eglynn mentioned on #openstack-ceilometer he agrees with the skip, and +1'ed the review. You too take a look, when you're done with coffee :-) | 11:20 |
| kashyap | ("the review" == https://review.openstack.org/104516) | 11:21 |
| *** Longgeek has quit IRC | 11:22 | |
| *** Longgeek has joined #openstack-qa | 11:23 | |
| *** harikav has quit IRC | 11:24 | |
| *** Poornima has quit IRC | 11:26 | |
| openstackrecheck | Console logs not available after 13:22s for check-grenade-dsvm-partial-ncpu 89530,8,0975ebe | 11:28 |
| kashyap | They'd un-skip the test next week | 11:33 |
| *** leseb has joined #openstack-qa | 11:35 | |
| *** samuelbercovici1 has joined #openstack-qa | 11:35 | |
| *** tkammer has joined #openstack-qa | 11:37 | |
| openstackgerrit | Andrey Kurilin proposed a change to openstack/qa-specs: Automate tempest config generation https://review.openstack.org/94473 | 11:38 |
| openstackgerrit | Daniel Berrange proposed a change to openstack/tempest: Avoid errors in log when neutron tests are skipped https://review.openstack.org/103857 | 11:38 |
| *** samuelbercovici has quit IRC | 11:38 | |
| *** samuelbercovici1 is now known as samuelbercovici | 11:38 | |
| openstackgerrit | Cedric Brandily proposed a change to openstack/tempest: Add provider extension positive test cases https://review.openstack.org/103498 | 11:40 |
| openstackgerrit | Cedric Brandily proposed a change to openstack/tempest: Add multi-provider extension positive test cases https://review.openstack.org/104526 | 11:40 |
| openstackgerrit | A change was merged to openstack/tempest: update iptables rules for more useful debugging https://review.openstack.org/104220 | 11:42 |
| *** samuelbercovici1 has joined #openstack-qa | 11:42 | |
| *** Longgeek_ has joined #openstack-qa | 11:43 | |
| openstackgerrit | Andrey Kurilin proposed a change to openstack/qa-specs: Automate tempest config generation https://review.openstack.org/94473 | 11:44 |
| *** samuelbercovici has quit IRC | 11:46 | |
| *** samuelbercovici1 is now known as samuelbercovici | 11:46 | |
| *** Longgeek has quit IRC | 11:47 | |
| openstackgerrit | Cedric Brandily proposed a change to openstack/tempest: Add multi-provider extension positive test cases https://review.openstack.org/104526 | 11:48 |
| *** samuelbercovici1 has joined #openstack-qa | 11:53 | |
| openstackgerrit | Cedric Brandily proposed a change to openstack/tempest: Add multi-provider extension positive test cases https://review.openstack.org/104526 | 11:54 |
| *** abhi- has quit IRC | 11:54 | |
| *** nosnos has quit IRC | 11:55 | |
| openstackrecheck | Console logs not available after 13:25s for check-tempest-dsvm-neutron 62101,27,11054fa | 11:56 |
| *** samuelbercovici has quit IRC | 11:56 | |
| *** samuelbercovici1 is now known as samuelbercovici | 11:56 | |
| *** nosnos_ has joined #openstack-qa | 11:58 | |
| *** yfried_ has joined #openstack-qa | 11:58 | |
| *** yfried has quit IRC | 11:59 | |
| *** luqas has quit IRC | 11:59 | |
| *** yfried__ has joined #openstack-qa | 12:00 | |
| *** gszasz has quit IRC | 12:00 | |
| *** yfried has joined #openstack-qa | 12:01 | |
| *** yfried_ has quit IRC | 12:03 | |
| *** Longgeek_ has quit IRC | 12:04 | |
| *** yfried__ has quit IRC | 12:04 | |
| *** dims has joined #openstack-qa | 12:04 | |
| *** Longgeek has joined #openstack-qa | 12:05 | |
| *** jhenner has quit IRC | 12:05 | |
| *** jhenner has joined #openstack-qa | 12:06 | |
| *** Longgeek_ has joined #openstack-qa | 12:06 | |
| *** afazekas_ has quit IRC | 12:07 | |
| *** Anu_ has quit IRC | 12:08 | |
| *** Longgeek has quit IRC | 12:09 | |
| openstackrecheck | Console logs not available after 13:27s for check-tempest-dsvm-full 104493,1,d6fb725 | 12:10 |
| *** gszasz has joined #openstack-qa | 12:10 | |
| *** jyuso has quit IRC | 12:11 | |
| sdague | kashyap: https://review.openstack.org/#/c/104516/ has to be respun | 12:11 |
| sdague | it's actually currently a syntax error | 12:12 |
| *** johnthetubaguy is now known as zz_johnthetubagu | 12:12 | |
| *** samuelbercovici1 has joined #openstack-qa | 12:15 | |
| *** dkranz has joined #openstack-qa | 12:16 | |
| *** yamahata has joined #openstack-qa | 12:16 | |
| *** samuelbercovici has quit IRC | 12:18 | |
| *** samuelbercovici1 is now known as samuelbercovici | 12:18 | |
| *** amotoki has joined #openstack-qa | 12:19 | |
| *** samuelbercovici1 has joined #openstack-qa | 12:22 | |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/elastic-recheck: make ER only go back 10 days https://review.openstack.org/104533 | 12:23 |
| openstackrecheck | Console logs not available after 13:24s for gate-nova-python27 103066,11,e673645 | 12:24 |
| *** afazekas_ has joined #openstack-qa | 12:24 | |
| *** samuelbercovici has quit IRC | 12:25 | |
| *** samuelbercovici1 is now known as samuelbercovici | 12:25 | |
| openstackgerrit | Kashyap Chamarthy proposed a change to openstack/tempest: Skip telemetry test - test_check_nova_notification() https://review.openstack.org/104516 | 12:25 |
| *** weshay has joined #openstack-qa | 12:26 | |
| yfried | dkranz: mtreinish: sdague: https://review.openstack.org/#/c/62101/ | 12:28 |
| *** mkollaro has joined #openstack-qa | 12:28 | |
| *** samuelbercovici1 has joined #openstack-qa | 12:29 | |
| yfried | ^ it was approved yesterday, but I had to rebase it. could you please give it a special treatment? | 12:30 |
| *** ajc_ has quit IRC | 12:30 | |
| *** coolsvap is now known as coolsvap|afk | 12:31 | |
| *** zz_johnthetubagu is now known as johnthetubaguy | 12:32 | |
| *** samuelbercovici has quit IRC | 12:33 | |
| *** samuelbercovici1 is now known as samuelbercovici | 12:33 | |
| *** afazekas_ has quit IRC | 12:34 | |
| *** yfried_ has joined #openstack-qa | 12:35 | |
| *** yfried__ has joined #openstack-qa | 12:35 | |
| *** yfried__ has quit IRC | 12:35 | |
| rdekel | I'm trying to run stress tests. Do anyone know how (if it's impossible) to run them with pycharm? | 12:36 |
| *** jhenner has quit IRC | 12:36 | |
| rdekel | I meant if it's possible.... | 12:36 |
| *** zigo has quit IRC | 12:37 | |
| *** ekhugen has joined #openstack-qa | 12:37 | |
| openstackrecheck | Console logs not available after 13:24s for check-tempest-dsvm-postgres-full 102838,9,2e5ee0b | 12:37 |
| *** zigo has joined #openstack-qa | 12:37 | |
| *** samuelbercovici1 has joined #openstack-qa | 12:39 | |
| openstackgerrit | Yair Fried proposed a change to openstack/tempest: Fix manager.status_timeout exception https://review.openstack.org/103547 | 12:39 |
| *** yfried has quit IRC | 12:39 | |
| *** bkopilov has quit IRC | 12:41 | |
| *** zigo has quit IRC | 12:41 | |
| *** zigo has joined #openstack-qa | 12:42 | |
| *** samuelbercovici has quit IRC | 12:42 | |
| *** samuelbercovici1 is now known as samuelbercovici | 12:42 | |
| *** markmcclain has joined #openstack-qa | 12:42 | |
| *** adalbas has joined #openstack-qa | 12:43 | |
| *** k4n0 has quit IRC | 12:44 | |
| *** zigo has quit IRC | 12:45 | |
| *** mtreinish has quit IRC | 12:45 | |
| *** afazekas_ has joined #openstack-qa | 12:47 | |
| *** jamielennox is now known as jamielennox|away | 12:50 | |
| *** jordanP has joined #openstack-qa | 12:50 | |
| *** jhenner has joined #openstack-qa | 12:50 | |
| *** Longgeek_ has quit IRC | 12:51 | |
| openstackrecheck | Console logs not available after 13:23s for check-tempest-dsvm-full-icehouse 103547,3,d5bb108 | 12:51 |
| *** Longgeek has joined #openstack-qa | 12:52 | |
| *** luqas has joined #openstack-qa | 12:53 | |
| *** jordanP has quit IRC | 12:58 | |
| *** dims_ has joined #openstack-qa | 12:58 | |
| *** jordanP has joined #openstack-qa | 12:59 | |
| *** yfried__ has joined #openstack-qa | 13:00 | |
| *** dims has quit IRC | 13:02 | |
| *** yfried has joined #openstack-qa | 13:02 | |
| *** yfried_ has quit IRC | 13:02 | |
| openstackgerrit | Joseph Lanoux proposed a change to openstack/tempest: Add python-ceilometerclient in requirements https://review.openstack.org/104542 | 13:03 |
| openstackrecheck | Console logs not available after 13:24s for check-tempest-dsvm-full 104504,1,3b681b5 | 13:05 |
| *** yfried__ has quit IRC | 13:05 | |
| openstackgerrit | Joseph Lanoux proposed a change to openstack/tempest: Add python-ceilometerclient in requirements https://review.openstack.org/104542 | 13:05 |
| *** Sondi has joined #openstack-qa | 13:08 | |
| *** Longgeek_ has joined #openstack-qa | 13:12 | |
| openstackgerrit | Emilien Macchi proposed a change to openstack-dev/grenade: Add initial support for Heat https://review.openstack.org/86978 | 13:12 |
| *** oomichi has quit IRC | 13:14 | |
| Sondi | can someone explain me the difference in openstack PUBLIC_INTERFACE and FLAT_INTERFACE? | 13:14 |
| *** Longgeek has quit IRC | 13:15 | |
| Sondi | i didn't mean openstack, i mean devstack | 13:17 |
| *** amitpp has quit IRC | 13:17 | |
| *** pcrews has joined #openstack-qa | 13:18 | |
| *** chandan_kumar is now known as chandankumar | 13:18 | |
| *** mriedem has joined #openstack-qa | 13:20 | |
| *** Longgeek has joined #openstack-qa | 13:21 | |
| *** Longgeek_ has quit IRC | 13:24 | |
| *** gmann has joined #openstack-qa | 13:25 | |
| *** psharma has quit IRC | 13:26 | |
| *** malini1 has joined #openstack-qa | 13:26 | |
| openstackgerrit | Ishant Tyagi proposed a change to openstack/tempest: Move orchestration scenario test templates in seperate directory https://review.openstack.org/104546 | 13:26 |
| *** Madan has quit IRC | 13:27 | |
| Sondi | can someone explain me the difference in devstack between the PUBLIC_INTERFACE and FLAT_INTERFACE? | 13:29 |
| *** gmann has quit IRC | 13:33 | |
| *** jaypipes has joined #openstack-qa | 13:35 | |
| *** ChanServ sets mode: +o jaypipes | 13:35 | |
| *** nosnos_ has quit IRC | 13:37 | |
| openstackrecheck | Console logs not available after 13:23s for check-tempest-dsvm-full 95680,4,a8f2329 | 13:44 |
| *** rdekel has quit IRC | 13:47 | |
| openstackgerrit | David Kranz proposed a change to openstack/tempest: First part of moving success response checking to identity client https://review.openstack.org/104290 | 13:47 |
| *** Obulpathi has joined #openstack-qa | 13:48 | |
| *** nosnos has joined #openstack-qa | 13:49 | |
| openstackgerrit | David Kranz proposed a change to openstack/tempest: First part of moving success response checking to identity client https://review.openstack.org/104290 | 13:52 |
| *** mriedem has quit IRC | 13:54 | |
| *** mriedem has joined #openstack-qa | 13:56 | |
| *** arnaud__ has joined #openstack-qa | 13:57 | |
| *** zigo has joined #openstack-qa | 13:58 | |
| *** nosnos has quit IRC | 13:59 | |
| *** Obulpathi has quit IRC | 14:00 | |
| *** dkranz has quit IRC | 14:00 | |
| *** Obulpathi has joined #openstack-qa | 14:00 | |
| maurosr | dkranz: hey, morning, are you cordinatin the check 'response code checking in the clients' somewhere or doing it all by your own? | 14:02 |
| openstackrecheck | Console logs not available after 13:31s for check-tempest-dsvm-full 100288,14,004200b | 14:04 |
| *** rbak has joined #openstack-qa | 14:04 | |
| *** luqas has left #openstack-qa | 14:07 | |
| *** luqas has joined #openstack-qa | 14:08 | |
| openstackgerrit | Ishant Tyagi proposed a change to openstack/tempest: Remove duplicate code from orchestration scenario tests https://review.openstack.org/104555 | 14:08 |
| *** ylobankov has quit IRC | 14:09 | |
| *** yfried_ has joined #openstack-qa | 14:10 | |
| *** BobBall has quit IRC | 14:12 | |
| *** yfried has quit IRC | 14:13 | |
| *** dims_ has quit IRC | 14:16 | |
| *** mtreinish has joined #openstack-qa | 14:16 | |
| *** bkopilov has joined #openstack-qa | 14:18 | |
| *** andreaf has joined #openstack-qa | 14:22 | |
| openstackgerrit | Emilien Macchi proposed a change to openstack-dev/grenade: Add initial support for Heat https://review.openstack.org/86978 | 14:23 |
| *** yfried_ is now known as yfried | 14:24 | |
| *** jgrimm has joined #openstack-qa | 14:27 | |
| *** dkranz has joined #openstack-qa | 14:27 | |
| openstackgerrit | Ishant Tyagi proposed a change to openstack/tempest: Add orchestration scenario test for stack-abandon and stack-adopt https://review.openstack.org/103198 | 14:29 |
| openstackrecheck | Console logs not available after 13:24s for check-tempest-dsvm-neutron-heat-slow 95680,5,8ccbb5b | 14:29 |
| *** mtreinish has quit IRC | 14:30 | |
| *** mtreinish has joined #openstack-qa | 14:30 | |
| *** mtreinish has quit IRC | 14:31 | |
| *** mtreinish has joined #openstack-qa | 14:33 | |
| afazekas_ | raies: hi | 14:36 |
| *** chandankumar has quit IRC | 14:38 | |
| *** leseb has quit IRC | 14:38 | |
| *** leseb has joined #openstack-qa | 14:39 | |
| openstackgerrit | Ishant Tyagi proposed a change to openstack/tempest: Move orchestration scenario test templates in separate directory https://review.openstack.org/104546 | 14:42 |
| *** leseb has quit IRC | 14:43 | |
| *** leseb has joined #openstack-qa | 14:43 | |
| openstackrecheck | Console logs not available after 13:22s for check-tempest-dsvm-postgres-full 104473,3,3ddc2d9 | 14:45 |
| Sondi | can someone explain me the difference in devstack between the PUBLIC_INTERFACE and FLAT_INTERFACE? | 14:48 |
| *** dkranz has quit IRC | 14:49 | |
| *** coolsvap|afk is now known as coolsvap | 14:55 | |
| openstackgerrit | Kashyap Chamarthy proposed a change to openstack/tempest: Skip telemetry test - test_check_nova_notification() https://review.openstack.org/104516 | 14:58 |
| openstackrecheck | Console logs not available after 13:23s for check-grenade-dsvm-partial-ncpu 104530,1,60cef7e | 14:58 |
| openstackgerrit | Matthew Treinish proposed a change to openstack/qa-specs: Move branchless tempest to implemented specs https://review.openstack.org/104575 | 14:59 |
| openstackgerrit | Matthew Treinish proposed a change to openstack/qa-specs: Remove nova-api-test-inheritance spec https://review.openstack.org/104576 | 14:59 |
| *** samuelbercovici has quit IRC | 15:00 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack-dev/devstack: Ironic: Setup/Configure iPXE https://review.openstack.org/99677 | 15:01 |
| *** tkammer has quit IRC | 15:04 | |
| *** dims_ has joined #openstack-qa | 15:04 | |
| morganfainberg | mtreinish, sdague, does tempest know what version (e.g. Icehouse) it is running against? I have a change that is pending that (yes I know... changes return, ML thread is available on it) that would change a 204 to a 200 return, so icehouse would be 204, juno and beyond 200. | 15:07 |
| morganfainberg | mtreinish, sdague, this is re: http://lists.openstack.org/pipermail/openstack-dev/2014-July/039132.html | 15:07 |
| *** ylobankov has joined #openstack-qa | 15:08 | |
| *** ylobankov has quit IRC | 15:08 | |
| *** ylobankov has joined #openstack-qa | 15:08 | |
| *** dims_ has quit IRC | 15:08 | |
| mtreinish | morganfainberg: http://docs.openstack.org/developer/tempest/overview.html#bug-fix-on-core-project-needing-tempest-changes | 15:09 |
| nikhil__ | mtreinish: mriedem : o/ | 15:09 |
| morganfainberg | mtreinish, right | 15:09 |
| morganfainberg | mtreinish, the issue is icehouse and juno have different return codes | 15:09 |
| morganfainberg | mtreinish, i don't see an icehouse branch? | 15:09 |
| mtreinish | morganfainberg: the short answer is no tempest doesn't know and shouldn't know whether it's running against juno vs icehouse | 15:09 |
| morganfainberg | mtreinish, the whole branchless thing. | 15:09 |
| mtreinish | to get the change landed you'll have to make it consistent between the releases | 15:09 |
| openstackgerrit | A change was merged to openstack/qa-specs: Add spec for bp:post-run-cleanup https://review.openstack.org/91725 | 15:10 |
| morganfainberg | mtreinish, not doable. | 15:10 |
| *** markmcclain has quit IRC | 15:10 | |
| nikhil__ | just had a quick convo in the #openstack-glance channel and it seems that we've some sort of acceptance to move part of glance tests into tempest (especially ones which need to start a api service) | 15:10 |
| morganfainberg | mtreinish, this is REST api htttp status change. | 15:10 |
| morganfainberg | mtreinish, well unless we really want to back port a nasty fix like this. | 15:10 |
| morganfainberg | mtreinish, i'm concerned about backporting. | 15:10 |
| nikhil__ | so, will be creating a glance-spec review and link it to you guys soon | 15:10 |
| nikhil__ | (fyi) | 15:11 |
| mtreinish | morganfainberg: ok, then the thing is if the api is going from async to sync is it really the same api? | 15:11 |
| mtreinish | nikhil__: ok cool | 15:11 |
| mtreinish | nikhil__: thanks for driving that | 15:11 |
| mtreinish | morganfainberg: because that seems like a pretty big change in the behavior | 15:11 |
| morganfainberg | mtreinish, i don't know if i can convince stable maintenance to accept this. i do know we should be fixing this. | 15:12 |
| morganfainberg | mtreinish, and yes. it's a "we can't deploy keystone under httpd and be consistent" issue. | 15:12 |
| morganfainberg | mtreinish, bascially keystone under mod_wsgi results in very different behavior on a couple API calls. | 15:13 |
| morganfainberg | mtreinish, specifically returned HTTP status. | 15:13 |
| morganfainberg | mtreinish, i can try and get stable maintenance to accept, but my concern is we'll get juno accepted, and stable will balk. | 15:13 |
| morganfainberg | mtreinish, and we'll lose some coverage in tempest then. | 15:14 |
| nikhil__ | np | 15:14 |
| mtreinish | morganfainberg: hmm, I'm wondering if we should cheat then and call deployable under httpd a feature... | 15:14 |
| morganfainberg | mtreinish, let me poke at a few more people to look at that ML thread before i go down this path. | 15:14 |
| mtreinish | and say icehouse doesn't support it | 15:14 |
| mtreinish | morganfainberg: ok well I'm on stable maint and I'd +2 | 15:15 |
| morganfainberg | mtreinish, ++ ok that is good :) | 15:15 |
| mtreinish | although, I really don't care about stable maint | 15:15 |
| mtreinish | I was added to the team kinda against my will :) | 15:15 |
| morganfainberg | mtreinish, let me ask dolph his opinion on backporting that fix to icehouse | 15:16 |
| morganfainberg | mtreinish, PTL, right? | 15:16 |
| morganfainberg | mtreinish ;) | 15:16 |
| mtreinish | no actually before that :) | 15:16 |
| morganfainberg | ah | 15:16 |
| mtreinish | I fixed abunch of stable bugs with tempest | 15:16 |
| openstackgerrit | Sébastien Han proposed a change to openstack-dev/devstack: Implement Ceph backend for Glance / Cinder / Nova https://review.openstack.org/65113 | 15:17 |
| mtreinish | and they said you're on the team now because they needed someone who works on tempest | 15:17 |
| EmilienM | sdague: hey Sean, quick question. In Grenade / heat patch (https://review.openstack.org/#/c/86978/) - it seems python-qpid is missing after the upgrade from Havana to Icehouse. I did a patchset to manually install it, and Jenkins is now happy. Could you let me know where I should check the pip requirements? | 15:17 |
| *** sabeen has joined #openstack-qa | 15:17 | |
| *** dkranz has joined #openstack-qa | 15:21 | |
| *** asselin has joined #openstack-qa | 15:24 | |
| *** mtreinish has quit IRC | 15:25 | |
| *** niedbalski has joined #openstack-qa | 15:31 | |
| *** yfried has quit IRC | 15:32 | |
| *** stratuspaulg has joined #openstack-qa | 15:33 | |
| *** arnaud__ has quit IRC | 15:45 | |
| *** jaypipes_ has joined #openstack-qa | 15:48 | |
| *** jaypipes_ has quit IRC | 15:49 | |
| *** ylobankov has quit IRC | 15:49 | |
| openstackgerrit | Dean Troyer proposed a change to openstack-dev/devstack: Support multiple Cinder backend types https://review.openstack.org/100271 | 15:50 |
| *** jpich has quit IRC | 15:50 | |
| openstackrecheck | Console logs not available after 13:22s for check-tempest-dsvm-full 103547,4,3cdd6a0 | 15:54 |
| openstackrecheck | openstack/heat change: https://review.openstack.org/97974 failed because of: gate-tempest-dsvm-postgres-full: https://bugs.launchpad.net/bugs/1336755 | 15:54 |
| openstackgerrit | gordon chung proposed a change to openstack-dev/devstack: Revert "Add missing config option os_auth_url" https://review.openstack.org/104597 | 15:56 |
| *** aimon has quit IRC | 15:56 | |
| *** nati_ueno has joined #openstack-qa | 15:56 | |
| openstackgerrit | Sébastien Han proposed a change to openstack-dev/devstack: Implement Ceph backend for Glance / Cinder / Nova https://review.openstack.org/65113 | 15:57 |
| *** ildikov has joined #openstack-qa | 15:59 | |
| *** mtreinish has joined #openstack-qa | 16:00 | |
| *** ildikov_ has quit IRC | 16:02 | |
| asvechnikov | sdauge: Hi, I have a question, is it possible to overwrite config of some service for single test in tempest? | 16:02 |
| *** mmaglana has joined #openstack-qa | 16:02 | |
| *** ildikov has quit IRC | 16:03 | |
| *** mtreinish has quit IRC | 16:04 | |
| *** dhellmann is now known as dhellmann_ | 16:04 | |
| *** mtreinish has joined #openstack-qa | 16:04 | |
| *** atiwari has joined #openstack-qa | 16:04 | |
| *** dims_ has joined #openstack-qa | 16:05 | |
| *** lanoux_ has quit IRC | 16:05 | |
| openstackrecheck | Console logs not available after 13:26s for check-tempest-dsvm-postgres-full 103627,3,29778f9 | 16:08 |
| *** dims_ has quit IRC | 16:09 | |
| *** safchain has quit IRC | 16:10 | |
| openstackgerrit | David Shrewsbury proposed a change to openstack/tempest: Fix disk_available_least in JSON schemas https://review.openstack.org/104603 | 16:17 |
| *** nati_ueno has quit IRC | 16:17 | |
| *** nati_uen_ has joined #openstack-qa | 16:17 | |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/devstack-gate: Handle proposed/ branches in test-matrix.py https://review.openstack.org/104448 | 16:19 |
| *** leseb has quit IRC | 16:19 | |
| *** leseb has joined #openstack-qa | 16:20 | |
| *** dims_ has joined #openstack-qa | 16:21 | |
| openstackgerrit | David Shrewsbury proposed a change to openstack/tempest: Fix disk_available_least in JSON schemas https://review.openstack.org/104603 | 16:24 |
| *** leseb has quit IRC | 16:24 | |
| openstackgerrit | Morgan Fainberg proposed a change to openstack/tempest: Add a skip for bug #1334368 https://review.openstack.org/104610 | 16:26 |
| *** zhiyan is now known as zhiyan_ | 16:29 | |
| *** aimon has joined #openstack-qa | 16:30 | |
| *** luqas has quit IRC | 16:32 | |
| *** qba73 has quit IRC | 16:32 | |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/devstack-gate: Handle proposed/ branches in test-matrix.py https://review.openstack.org/104448 | 16:34 |
| openstackrecheck | Console logs not available after 13:24s for check-tempest-dsvm-postgres-full 102886,3,a16c3ee | 16:34 |
| openstackgerrit | Adam Gandelman proposed a change to openstack/tempest: Fixes ephemeral disk flavor test for non-virt https://review.openstack.org/104069 | 16:35 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/devstack-gate: Handle proposed/ branches in test-matrix.py https://review.openstack.org/104448 | 16:38 |
| *** amitpp has joined #openstack-qa | 16:39 | |
| *** Longgeek has quit IRC | 16:41 | |
| *** vrovachev has quit IRC | 16:42 | |
| *** ylobankov has joined #openstack-qa | 16:43 | |
| *** dhellmann_ is now known as dhellmann | 16:44 | |
| *** afazekas_ has quit IRC | 16:47 | |
| openstackrecheck | Console logs not available after 13:24s for check-tempest-dsvm-full 104222,2,ff2e505 | 16:48 |
| *** leseb has joined #openstack-qa | 16:50 | |
| openstackgerrit | Anant Patil proposed a change to openstack-dev/devstack: Fixed tunnel ranges variable inconsistency. https://review.openstack.org/104618 | 16:50 |
| *** jlanoux has joined #openstack-qa | 16:51 | |
| *** k4n0 has joined #openstack-qa | 16:51 | |
| *** leseb has quit IRC | 16:52 | |
| *** arnaud has joined #openstack-qa | 16:52 | |
| *** leseb has joined #openstack-qa | 16:52 | |
| *** adalbas has quit IRC | 16:56 | |
| *** leseb has quit IRC | 16:57 | |
| *** prkumar has joined #openstack-qa | 16:57 | |
| *** gszasz has quit IRC | 17:02 | |
| openstackrecheck | Console logs not available after 13:22s for check-tempest-dsvm-postgres-full 91722,33,3044ee3 | 17:02 |
| *** leseb has joined #openstack-qa | 17:02 | |
| openstackrecheck | openstack/nova change: https://review.openstack.org/104010 failed gate-grenade-dsvm in the gate queue with an unrecognized error | 17:02 |
| *** leseb has quit IRC | 17:02 | |
| *** leseb has joined #openstack-qa | 17:03 | |
| *** ylobankov has quit IRC | 17:06 | |
| *** leseb has quit IRC | 17:07 | |
| *** jlibosva has quit IRC | 17:08 | |
| openstackgerrit | A change was merged to openstack/tempest: Extending quota support for neutron LBaaS entities https://review.openstack.org/60008 | 17:09 |
| *** aimon has quit IRC | 17:10 | |
| *** adalbas has joined #openstack-qa | 17:10 | |
| *** jordanP has quit IRC | 17:11 | |
| *** aimon has joined #openstack-qa | 17:11 | |
| *** bnemec is now known as bnemec_appt | 17:15 | |
| openstackrecheck | Console logs not available after 13:22s for check-tempest-dsvm-full 104543,1,fb10876 | 17:16 |
| *** k4n0 has quit IRC | 17:23 | |
| *** k4n0 has joined #openstack-qa | 17:30 | |
| *** nati_uen_ has quit IRC | 17:31 | |
| *** nati_ueno has joined #openstack-qa | 17:31 | |
| *** yfried has joined #openstack-qa | 17:36 | |
| openstackrecheck | Console logs not available after 13:22s for check-tempest-dsvm-postgres-full 104304,1,c19de26 | 17:42 |
| sdague | EmilienM: we don't use qpid in the gate, why does heat require it? | 17:44 |
| *** amitpp has quit IRC | 17:45 | |
| *** yfried has quit IRC | 17:47 | |
| *** arnaud has quit IRC | 17:52 | |
| adam_g | mtreinish, devananda care to chat ironic after the qa meeting wraps up? | 17:52 |
| *** Sondi has quit IRC | 17:52 | |
| devananda | adam_g: ++ | 17:52 |
| mtreinish | adam_g: works for me | 17:54 |
| openstackgerrit | Morgan Fainberg proposed a change to openstack/tempest: Add a skip for bug #1334368 https://review.openstack.org/104610 | 17:55 |
| openstackrecheck | Console logs not available after 13:24s for check-grenade-dsvm 104490,2,49de37e | 17:56 |
| openstackrecheck | openstack/nova change: https://review.openstack.org/103164 failed because of: gate-grenade-dsvm-partial-ncpu: https://bugs.launchpad.net/bugs/1298472 | 17:56 |
| *** marun is now known as marun_afk | 17:57 | |
| k4n0 | Hi , i am trying to get some new tests added to neutron subnets api and network api (provider networks) | 18:01 |
| *** malini1 has quit IRC | 18:01 | |
| k4n0 | should i propose a bug or a spec or what? | 18:02 |
| mtreinish | k4n0: it depends on the scope of the effort | 18:02 |
| openstackgerrit | Cedric Brandily proposed a change to openstack/tempest: Add multi-provider extension positive test cases https://review.openstack.org/104526 | 18:02 |
| mtreinish | it's either a spec or nothing | 18:02 |
| openstackgerrit | Cedric Brandily proposed a change to openstack/tempest: Add provider extension positive test cases https://review.openstack.org/103498 | 18:02 |
| mtreinish | k4n0: https://github.com/openstack/qa-specs/blob/master/README.rst#specs-for-new-tests | 18:02 |
| *** arnaud has joined #openstack-qa | 18:02 | |
| *** arnaud has quit IRC | 18:02 | |
| k4n0 | mtreinish, got it, ill try to get a estimate of work and rough spec | 18:03 |
| k4n0 | thanks | 18:03 |
| mtreinish | k4n0: just a heads up there is a spec proposal for some ipv6 stuff including provider networks | 18:04 |
| mtreinish | I can't remember who submitted it | 18:04 |
| k4n0 | mtreinish, will check that | 18:04 |
| *** johnthetubaguy is now known as zz_johnthetubagu | 18:07 | |
| *** k4n0 has quit IRC | 18:07 | |
| *** jlanoux has quit IRC | 18:07 | |
| *** coolsvap is now known as coolsvap|afk | 18:08 | |
| openstackgerrit | Steven Hardy proposed a change to openstack/tempest: orchestration add volume backup/restore scenario test https://review.openstack.org/90143 | 18:09 |
| *** liusheng has quit IRC | 18:10 | |
| *** liusheng has joined #openstack-qa | 18:11 | |
| adam_g | mtreinish, okay so.. ironic, got a few? | 18:16 |
| adam_g | devananda, Shrews ^ | 18:16 |
| mtreinish | adam_g: sure | 18:16 |
| Shrews | adam_g: here | 18:17 |
| adam_g | mtreinish, so, i just wanted to touch base and run by what we've been working on, and whatever concerns you had for the rest of the cycle, so we can get lined up well in QA for (we hope) ironic graduation later | 18:17 |
| adam_g | https://etherpad.openstack.org/p/IronicCI | 18:17 |
| adam_g | first and foremost, we're working on getting rid of the multiple confusing ironic jobs that run, and having a single base tempest job that tests ironic in the same way it'd test nova. | 18:18 |
| mtreinish | well a big +2 on number 1 on that list :) | 18:18 |
| adam_g | shrews and i have been chipping away at tempest issues we've been finding that tie tests too closely to libvirt. | 18:19 |
| openstackrecheck | Console logs not available after 13:27s for check-grenade-dsvm-partial-ncpu 104546,2,b303a30 | 18:19 |
| adam_g | this will probably require adding a bunch of new compute feature flags, is there any objection to that? | 18:19 |
| devananda | adam_g: o/ | 18:19 |
| mtreinish | yeah that's good work not just ironic but people working on other drivers | 18:19 |
| mtreinish | adam_g: no I expected that to happen at some point | 18:19 |
| adam_g | we seem to have only a limited subset of whats covered @ https://wiki.openstack.org/wiki/HypervisorSupportMatrix#Hypervisor_feature_support_matrix | 18:19 |
| adam_g | okay coolio | 18:19 |
| adam_g | this may or may not require some infra work to get a devstack slave to run tests that require booting multiple servers in parallel | 18:20 |
| mtreinish | adam_g: yeah BobBall and some of the other guys running 3rd party ci were complaining about the same basic thing. | 18:20 |
| mtreinish | but no one stepped up to map config options where needed | 18:21 |
| adam_g | sweet! | 18:21 |
| devananda | if we need to boot >1 node with ironic, it will very likely require devstack slaves with >8GB RAM | 18:21 |
| adam_g | devananda, ive been testing locally with an 8GB VM and it seems okay, haven't tested on either cloud provider yet tho | 18:21 |
| adam_g | mtreinish, are there any plans to do away with non-isolated credentials, or running tests serially? we kinda heavily rely on that | 18:22 |
| devananda | adam_g: interesting. how many nodes booting at once? each is 1GB now, I believe | 18:22 |
| adam_g | devananda, 3 is the maximum ive found in the compute API tests | 18:22 |
| adam_g | devananda, tho ive seen issues where ironic->libvirt ssh connections start to get flaky | 18:22 |
| devananda | adam_g: right. there's paramiko/eventlet fun to be had there | 18:22 |
| Shrews | devananda: they should be half that (512) | 18:22 |
| mtreinish | adam_g: so the only plan related to the creds for running tests is basically this: https://review.openstack.org/#/c/98400/ | 18:23 |
| adam_g | Shrews, we need 1GB to run ironic's deploy ramdisk :| | 18:23 |
| devananda | so if we add concurrent boot tests, we will be introducing significant races to the tests | 18:23 |
| Shrews | ugh | 18:23 |
| adam_g | devananda, they are not concurrent. tests usually spawn instance, wait for active, spawn nexzt | 18:23 |
| devananda | mtreinish: that's a terrible commit message | 18:23 |
| mtreinish | adam_g: so basically allow a list of creds instead of tenant isolation | 18:23 |
| mtreinish | devananda: heh, it's not my commit :) | 18:23 |
| adam_g | mtreinish, cool. ill read that spec in a bit | 18:23 |
| mtreinish | adam_g: but the goal there is to force everyone to run in parallel | 18:24 |
| mtreinish | and to stop concentrating on serial execution | 18:24 |
| adam_g | mtreinish, is the fact that we will be running in serial indefinitely going to be an issue? nova has a very finite amount of compute resources available, just enough to run tests serially | 18:26 |
| devananda | adam_g: any reason you see that our tests could not be run in parallel, as long as it is limited to num clients <= num nodes? | 18:27 |
| devananda | granted, if num nodes == 1, then it has to be serial | 18:27 |
| adam_g | devananda, what do you mean by limiting num clients? | 18:27 |
| devananda | parallel tempest execution threads | 18:27 |
| *** arnaud__ has joined #openstack-qa | 18:27 | |
| Shrews | some tests require multiple nodes, iirc | 18:27 |
| devananda | oh | 18:27 |
| adam_g | ^ yea | 18:27 |
| devananda | nvm | 18:28 |
| devananda | serial it is | 18:28 |
| devananda | unless we had infinite resources :) | 18:28 |
| * Shrews submits requests for infinite resources to -infra | 18:28 | |
| clarkb | Shrews: please submit them to your employer | 18:28 |
| adam_g | this also brings up the issue of test runtime. if we enable compute API tests against ironic, we're looking at ~2 min per nova boot. | 18:28 |
| Shrews | clarkb: heh | 18:29 |
| devananda | adam_g: good point, thanks for bringing that up. | 18:29 |
| mtreinish | adam_g: so the thing with serial only is there are a set of issues that only realistically get exposed if you run in paralle | 18:29 |
| mtreinish | l | 18:29 |
| devananda | adam_g: do you have an estimate of total test run time? | 18:29 |
| adam_g | devananda, not atm, i can come up with a rough estimate based on num instances spawned in tempest.api.compute.* | 18:29 |
| devananda | mtreinish: there are certainly known races in ironic right now, if you run in parallel -- and I'd love to test them. the problem is resources. | 18:30 |
| devananda | adam_g: rough is good | 18:30 |
| devananda | clarkb: i think we talked at one point (or maybe it was with mordred ) about getting devstack node with > 8 GB ram | 18:32 |
| mtreinish | devananda: yeah I mean if it's not possible to run in parallel in a gate slave then we don't have a choice | 18:32 |
| *** nati_ueno has quit IRC | 18:32 | |
| clarkb | devananda: I don't think that was me | 18:32 |
| devananda | clarkb: k | 18:32 |
| clarkb | devananda: we intentionally limit it so that we don't place crazy burdens on our devs | 18:32 |
| clarkb | devananda: in fact half of our slaves do have 30GB of ram | 18:32 |
| clarkb | we just boot using only 8gb of it | 18:33 |
| openstackrecheck | Console logs not available after 13:34s for check-tempest-dsvm-postgres-full 101794,4,1563889 | 18:33 |
| devananda | clarkb: could Ironic tests be run in such a way that they could use all of it? | 18:33 |
| clarkb | devananda: technicaly yes | 18:33 |
| clarkb | devananda: but we have explicitly prevented that | 18:33 |
| devananda | clarkb: :) | 18:34 |
| clarkb | so we would need to figure out if your explicit prevention is wrong | 18:34 |
| *** Rajalakshmi has quit IRC | 18:34 | |
| clarkb | er explicit prevention for you is wrong | 18:35 |
| clarkb | english hard | 18:35 |
| Shrews | it ok. you speak goodly enuf | 18:35 |
| devananda | clarkb: tldr; we clearly can't do parallel tempest testing of ironic in 8 GB because we can't test a PXE deployment to a VM with less than ~ 1GB RAM. if we can get more RAM, it may be worth looking into that so that ironic can get parallel tempest testing eventually | 18:37 |
| devananda | clarkb: or we continue solving concurrency problems downstream and accept the limitation | 18:37 |
| clarkb | devananda: 8GB total isn't enough for two 1GB nodes? | 18:37 |
| devananda | clarkb: see discussion above. it seems to be enough for 3. which is enough for a serial test of tempest, not parallel | 18:38 |
| clarkb | oh we are talking about parallel tempest | 18:39 |
| clarkb | I get it now | 18:39 |
| clarkb | so why couldn't you have a testresource (or similar) that manages 3 nodes? | 18:39 |
| clarkb | and do parallel with that limitation? | 18:39 |
| adam_g | devananda, omitting test files i know we'll skip entirely, the compute api tests run in here spawn about 69 servers. http://logs.openstack.org/27/103227/2/check/check-tempest-dsvm-ironic/0f5e4b1/logs/testr_results.html.gz | 18:40 |
| *** arnaud__ has quit IRC | 18:40 | |
| devananda | deva@9470m:/opt/source/openstack/tempest/tempest/api/compute⟫ git grep 'create_test_server(' | wc -l | 18:40 |
| devananda | 131 | 18:40 |
| devananda | so low end, 70 boots. high end, 130 | 18:40 |
| adam_g | devananda, it'll probably be lower than 70 | 18:41 |
| adam_g | anyway, it'll be long regardless | 18:41 |
| adam_g | i dont know if thats going to be an issue | 18:41 |
| sdague | devananda: so our experience is that until you enable parallel in the gate, you miss *really* big bugs | 18:41 |
| adam_g | clarkb, is there a timeout on devstack jobs? | 18:41 |
| clarkb | adam_g: yes, it varies by job | 18:42 |
| adam_g | ah | 18:42 |
| clarkb | adam_g: usually ~65 minutes iirc | 18:42 |
| sdague | clarkb: it's 120 | 18:42 |
| sdague | 125 actually | 18:42 |
| clarkb | sdague: by the way the timeout plugin is up to date everywhere | 18:42 |
| clarkb | sdague: so you can do the implicit timeout variable stuff now | 18:42 |
| sdague | clarkb: oh, neat | 18:42 |
| devananda | sdague: *nod* and I accept that there are really big bugs right now, which are only being found by folks like OnMetal and TripleO | 18:43 |
| devananda | sdague: who have the hardware available to *occasionally* test large parallel deployment on real hardware | 18:43 |
| sdague | devananda: right, but you don't need large parallel to find the bugs | 18:44 |
| sdague | you need small parallel | 18:44 |
| sdague | and lots of runs | 18:44 |
| sdague | and they are right there | 18:44 |
| sdague | this assumption that you need large parallel is massively misguided | 18:44 |
| devananda | sdague: define "small parallel" | 18:44 |
| clarkb | also you can do like the largeops jobs | 18:44 |
| clarkb | and just fake it out if the bugs are in exercising internal bits | 18:44 |
| sdague | devananda: honestly, every openstack service we've made parallel (which used to be 2x) broke horribly | 18:45 |
| devananda | clarkb: so API testing with the "fake" driver should work fine with parallel execution | 18:45 |
| devananda | by fine i mean, no resource constraint. it may hit bugs :) | 18:45 |
| sdague | yeh, for large parallel you can do the largeops on fake driver | 18:45 |
| sdague | but doing 2x on heat, for instance, made it fail 25% of the time | 18:46 |
| adam_g | devananda, the compute API tests do lots of things that require something real booting | 18:46 |
| clarkb | out of curiousity why does PXE need so much ram? | 18:46 |
| devananda | adam_g: ahh. right :( | 18:46 |
| clarkb | should you be able to pxe cirros? and be tiny? | 18:46 |
| adam_g | clarkb, its the deploy ramdisk | 18:46 |
| clarkb | adam_g: why is yuor deploy ramdisk so big? | 18:46 |
| clarkb | adam_g: I would think that thing should be as tiny as possible | 18:46 |
| devananda | deploy ramdisk itself is ~ 30 MB, last I looked | 18:46 |
| sdague | clarkb: because that's how ironic works | 18:46 |
| clarkb | smaller than cirros even | 18:46 |
| adam_g | https://bugs.launchpad.net/diskimage-builder/+bug/1311987 | 18:46 |
| openstackrecheck | Console logs not available after 13:22s for check-grenade-dsvm 99677,9,0e05eb2 | 18:46 |
| *** stratuspaulg has quit IRC | 18:47 | |
| clarkb | oh you use trusty for that | 18:47 |
| clarkb | well theres your problem :) | 18:47 |
| devananda | heh | 18:47 |
| sdague | yeh, what is the min you guys actually need in these? | 18:47 |
| adam_g | currently 1GB min. on the ironic node vms | 18:48 |
| sdague | adam_g: no, I mean what do you minimally need in the ironic compute OS | 18:49 |
| sdague | lets go backwards | 18:49 |
| sdague | because running all of tempest with trusty guests in the gate will not be quick | 18:49 |
| devananda | sdague: minimal for what? | 18:49 |
| sdague | to be an ironic compute | 18:49 |
| adam_g | package wise, we need either trusty or saucy + UCA | 18:50 |
| sdague | adam_g: that's not an answer | 18:50 |
| adam_g | im not sure your question, TBH | 18:50 |
| sdague | what is the minimum set of binaries you need | 18:50 |
| sdague | you need an ssh server that does keys, right? what else do you minimally need. | 18:51 |
| devananda | sdague: minimally for testing in the gate environment? or minimal in general? | 18:51 |
| sdague | just minimal in general | 18:51 |
| devananda | we don't need ssh server in general | 18:51 |
| sdague | can coreos be an ironic compute? | 18:51 |
| devananda | huh? | 18:52 |
| sdague | I want to know what beyond a linux kernel you need for something to be an ironic compute :) | 18:52 |
| adam_g | for pxe_ssh (whats tested currently): xinetd/tftpd, openvswitch (current version), libvirt (current version) | 18:52 |
| devananda | sdague: wait - your statement earlier didn't make sense. let me go back to this: | 18:52 |
| devananda | 18:49:30 < sdague> because running all of tempest with trusty guests in the gate will not be quick | 18:52 |
| devananda | we're not running trusty in nested VMs | 18:52 |
| adam_g | devananda, the ramdisk is trusty-based | 18:53 |
| devananda | sdague: "ironic compute" is not a term we use. what do you mean | 18:53 |
| adam_g | which is what introduced the 1024GB requirement | 18:53 |
| adam_g | er, 1GB | 18:53 |
| sdague | devananda: right, as adam_g said, you are trusty based | 18:53 |
| sdague | and I want to come back to what's the minimum you need | 18:53 |
| sdague | adam_g: why do you need libvirt in that image? | 18:54 |
| devananda | sdague: the deploy ramdisk is trusty based | 18:54 |
| sdague | devananda: yes | 18:54 |
| adam_g | sdague, we dont, i was speaking of the devstack 'host' | 18:54 |
| devananda | it had better not hhave libvirt in it | 18:54 |
| adam_g | theres 3 OSes in play i guess: devstack 'host', ramdisk, booted instance. | 18:54 |
| adam_g | devstack == trusty (i believe everywhere now?) | 18:54 |
| devananda | sdague: https://github.com/openstack/diskimage-builder/blob/master/elements/deploy-ironic/binary-deps.d/deploy-ironic | 18:54 |
| adam_g | ramdisk == trusty | 18:54 |
| adam_g | booted instance == cirros | 18:54 |
| devananda | adam_g: that's the list of binary deps for the deploy ramdisk AIUI | 18:55 |
| devananda | yes/no? | 18:55 |
| adam_g | devananda, looks about right | 18:55 |
| sdague | ok, so I have no idea how busybox + tgtd == 1 GB requirement | 18:55 |
| adam_g | i believe /w DIB we get a full debootstrap / minimal trusty userspace | 18:56 |
| sdague | right, so that's a big part of the problem | 18:56 |
| adam_g | which ends up ~256MB uncompressed, hitting that tmpfs issue | 18:56 |
| devananda | adam_g: so really we dont need 1GB, we just need a hair over 512MB | 18:56 |
| sdague | because looking at https://github.com/openstack/diskimage-builder/blob/master/elements/deploy-ironic/binary-deps.d/deploy-ironic that looks like add tgtd to cirros | 18:56 |
| sdague | and a kernel module or two | 18:57 |
| adam_g | this is splling over to a DIB issue. im sure they have reasons why they tie ramdisks to a supported distro | 18:57 |
| sdague | right, it is | 18:57 |
| devananda | yep | 18:57 |
| adam_g | even if we were to eliminate that, and squeeze a few more baremetal VMs into our devstack gates, parallel testing would still be an issue | 18:57 |
| *** sweston has quit IRC | 18:57 | |
| sdague | adam_g: the size is only part of the issue | 18:58 |
| sdague | a minimal trusty userspace is still an order of magnitude bigger / slower than what we use for our test computes | 18:58 |
| devananda | sdague: "test compute" | 18:59 |
| mtreinish | devananda: the 2nd level images we boot for tests | 18:59 |
| openstackrecheck | Console logs not available after 13:22s for check-grenade-dsvm-partial-ncpu 101883,3,269c0cb | 19:00 |
| adam_g | so | 19:00 |
| clarkb | right it takes about 35 minutes to boot trusty nested iirc | 19:00 |
| devananda | so we use cirros for that, or what ever devstack is using | 19:00 |
| clarkb | or it did when I was doing nested puppet testing | 19:00 |
| adam_g | part of the devstack+ironic test is a test of diskimage-builder, to produce the deploy ramdisks | 19:01 |
| devananda | clarkb: it takes < 2 minutes to boot the deploy ramdisk AND deploy to it | 19:01 |
| clarkb | devananda: in nested on hpcloud? | 19:01 |
| devananda | yes | 19:01 |
| clarkb | devananda: wow | 19:01 |
| clarkb | what magic is this | 19:01 |
| adam_g | clarkb, its just booting a kernel and ramdisk | 19:01 |
| sdague | yeh, the no disk io is huge | 19:01 |
| * devananda picks a random recent test run | 19:02 | |
| devananda | clarkb: http://logs.openstack.org/27/103227/2/check/check-tempest-dsvm-virtual-ironic/a7fcfbf/logs/screen-ir-cond.txt.gz?level=INFO#_2014-06-27_20_37_48_372 | 19:02 |
| clarkb | devananda: sure I believe you | 19:02 |
| devananda | clarkb: look at timestamps between when ironic began downloading the images from glance and when deploy complete -- that includes a pxe boot of the deploy ramdisk and copying the image to it | 19:02 |
| devananda | sdague: so we could possibly make that faster, sure | 19:02 |
| sdague | devananda: so honestly, I think we get back to this place where ironic is doing things different enough from nova, that honestly we keep going around in circles. I think we need pictures of the flows you think this does, especially with times where you know them | 19:03 |
| devananda | sdague: there are other issues I think it's worth discussing besides whether or not we can support parallel tempest | 19:03 |
| devananda | sdague: unless that is a blocker for graduation which I'm not aware of | 19:03 |
| *** ekhugen has quit IRC | 19:03 | |
| sdague | honestly, I don't ever want to graduate another project that doesn't have it's API exercised in parallel. It continues to plague neutron and heat to a lesser degree. | 19:04 |
| devananda | sdague: .... | 19:04 |
| mtreinish | sdague: I thought we switched heat back not too long ago? | 19:04 |
| sdague | mtreinish: we have | 19:04 |
| sdague | it was non voting for almost 12 weeks though | 19:04 |
| mtreinish | yeah, fair enough | 19:05 |
| devananda | sdague: graduation is a mark of process integration, not production-worthiness. but PLEASE let's not bikeshed even more | 19:05 |
| sdague | devananda: being able to do 2 things at once is table stakes | 19:05 |
| clarkb | and from what I am told we can test that | 19:06 |
| clarkb | we can test 3 things at once no problem | 19:06 |
| sdague | yep | 19:06 |
| clarkb | so lets do iut | 19:06 |
| adam_g | the issue with ironic parrallel API testing is that the API tests are not strictly API tests. they require real resources on the backend to do real things, and switching virt types introduces new constraints.. around resources, in ironic's case | 19:06 |
| devananda | tempest/api/compute/* are not strictly API tests | 19:06 |
| devananda | they're booting VMs | 19:07 |
| *** arnaud__ has joined #openstack-qa | 19:07 | |
| dkranz | adam_g: They are real api tests, just ones that run against "real" systems | 19:07 |
| sdague | devananda: because that's what the API does | 19:07 |
| devananda | sure | 19:07 |
| dkranz | THey are not unit tests | 19:07 |
| sdague | right, exactly, they are not unit tests | 19:07 |
| devananda | they're black box functional tests | 19:07 |
| sdague | yep | 19:07 |
| sdague | and integration tests | 19:08 |
| adam_g | dkranz, right, and the constraints change when the "real" system on the backend is configured differently. | 19:08 |
| sdague | so if DIB contained less, we'd be able to run 6 things? | 19:08 |
| sdague | because that should be enough to run 2way | 19:08 |
| adam_g | sdague, whats a 'thing' | 19:09 |
| sdague | compute | 19:09 |
| *** dims_ has quit IRC | 19:09 | |
| adam_g | maybe? | 19:09 |
| sdague | our high water mark is 3 computes per test | 19:09 |
| sdague | from what I can remember | 19:09 |
| adam_g | sdague, but how many concurrent tests | 19:09 |
| devananda | sdague: so perhaps the discussion here really needs to be about "how is ironic different from nova" | 19:09 |
| adam_g | yeah 3 sounds right | 19:09 |
| devananda | since,well, it is | 19:09 |
| mtreinish | that sounds about right | 19:09 |
| sdague | adam_g: 2 | 19:09 |
| sdague | we're talking about 2 tempest workers, that would be parallel | 19:10 |
| adam_g | sdague, is that the current testr concurrency for parallel tempest? or one that would satisfy the requirement? | 19:10 |
| sdague | the current is cpu / 2 | 19:10 |
| mtreinish | yeah, just 2 workers would find a surprising number of problems | 19:10 |
| sdague | which is usually 4 | 19:10 |
| adam_g | ah | 19:10 |
| sdague | but 2 would satisfy me | 19:10 |
| adam_g | that seems like something we can try to work for | 19:11 |
| clarkb | but you can always do syncing in the jobs | 19:11 |
| clarkb | so that when those particular tests run you are effectively 2x | 19:11 |
| clarkb | and testing keystone with the other workers | 19:11 |
| adam_g | on the topic of the DIB ramdisk | 19:11 |
| sdague | because, like I said before, small parallel + volume of runs finds the same issues as large parallel | 19:11 |
| adam_g | devstack+ironic currently depends on DIB to produce its ramdisks at run time | 19:11 |
| sdague | for the devstack case, can we hand tune that? | 19:12 |
| adam_g | as in, not use DIB? | 19:12 |
| sdague | yeh | 19:12 |
| adam_g | yeah thats what i was getting at | 19:12 |
| devananda | well | 19:12 |
| devananda | we can change DIB | 19:12 |
| adam_g | making ramdisks leaner is something that may or may not work in DIB | 19:12 |
| adam_g | we can talk about perhaps adding cirros support as a DIB base | 19:12 |
| sdague | devananda: or we can hand tune it | 19:12 |
| adam_g | failing that, we can create our own ramdisk strictly for upstream testing? | 19:13 |
| devananda | if we need to test ironic with !DIB, we shouldn't be recommending that deployers use DIB | 19:13 |
| sdague | we have very specific images that get pulled into devstack today | 19:13 |
| clarkb | adam_g: I would think that everyone would want a leaner ramdisk | 19:13 |
| devananda | clarkb: right | 19:13 |
| clarkb | adam_g: it is really odd to me that anyone would want trusty for that | 19:13 |
| adam_g | clarkb, why is that odd? saucy EOLs soon, trusty is an LTS | 19:13 |
| clarkb | because ubuntu is not light in any way | 19:13 |
| openstackrecheck | Console logs not available after 13:23s for check-grenade-dsvm 99992,5,d4e635a | 19:13 |
| clarkb | you end up with a lot of stuff you just don't need | 19:14 |
| clarkb | that slwos down boot time and copying of files and potentially makes your builds more error prone | 19:14 |
| sdague | yeh, smashing a full linux into an embedded env for something like this is kind of silly in my opinion. | 19:14 |
| clarkb | if you boil down to a minimal set of needs its a lot easier to test and deploy | 19:14 |
| clarkb | and debug | 19:14 |
| adam_g | well, right. i imagine it was something about keeping things similar across all levels of the stack. it all predates my involvement | 19:14 |
| clarkb | infra runs into this all the time | 19:14 |
| clarkb | random thing over there updates and everything else is broken | 19:14 |
| clarkb | it isn't fun | 19:15 |
| sdague | yeh, realistically, I think dib in the gate adds too much variability | 19:15 |
| sdague | we change those cirros images maybe once every 9 months | 19:15 |
| sdague | having a package update in an upstream impact things in weird ways, isn't really something I'm interested in signing up to debug | 19:16 |
| adam_g | SpamapS, lifeless ping ^ opinions about adding a cirros option for DIB DISTRO_NAME | 19:17 |
| devananda | sdague: fyi https://review.openstack.org/#/c/92256/ | 19:17 |
| devananda | proposal to add dib to devstack | 19:17 |
| sdague | devananda: yep, i know | 19:17 |
| devananda | k | 19:17 |
| sdague | and I've been back and forth that in my head a lot in the last month | 19:17 |
| sdague | https://review.openstack.org/#/c/94827/ | 19:17 |
| adam_g | sdague, actually i believe we can pass pre-built images to devstack to avoid the DIB run. | 19:18 |
| sdague | adam_g: yep | 19:18 |
| sdague | that's what I'm saying | 19:18 |
| devananda | adam_g: yep ++ | 19:18 |
| adam_g | okay, adding notes to pad | 19:18 |
| devananda | i suggested that a while ago | 19:18 |
| *** arnaud__ has quit IRC | 19:18 | |
| sdague | and that image we can also hand tune to pull stuff out of it | 19:18 |
| devananda | but it depends on DIB (or $something) putting out those images | 19:18 |
| sdague | or build it up differently, whatever | 19:18 |
| sdague | devananda: we're only going to change them every 6 months | 19:19 |
| adam_g | devananda, if we can get DIB to support cirros, we can generate one and cache it on the slaves? | 19:19 |
| sdague | so hand tuning should be fine | 19:19 |
| sdague | or tarballs.openstack.org, like the trove image | 19:19 |
| devananda | adam_g: probably | 19:19 |
| devananda | have DIB emit that artefact on a successful run | 19:19 |
| *** SridharG has quit IRC | 19:19 | |
| devananda | was in the plans when we started discussing all this | 19:19 |
| devananda | *on a successful run of ironic with that deploy ramdisk | 19:20 |
| dkranz | mtreinish: Is there a python reason a cls.skipException in a setUpClass can't be the first line, before a call to super setUpClass? | 19:22 |
| adam_g | is tenant isolation required for parallel? | 19:23 |
| sdague | adam_g: yes | 19:23 |
| mtreinish | dkranz: no, we do that on purpose in someplaces. It depends on when we want to raise the skip compared to running setUpClass for the parent class | 19:23 |
| mtreinish | dkranz: at one point we needed the parent setupclass to eval config | 19:23 |
| mtreinish | so it had to come after the super() | 19:23 |
| mtreinish | but that's not the case anymore | 19:23 |
| dkranz | mtreinish: That's what I was hoping. Thanks. | 19:24 |
| devananda | i'm not familiar withthe implications of tenant isolation - any pointers? | 19:24 |
| adam_g | devananda, adding a note to the wiki about it now | 19:24 |
| adam_g | er, pad | 19:24 |
| mtreinish | devananda: you could watch my summit user talk I explained it there :) | 19:25 |
| devananda | mtreinish: ack | 19:26 |
| devananda | mtreinish: around what time in the talk? | 19:26 |
| mtreinish | devananda: err let me check | 19:27 |
| sdague | devananda: tl;dr - tempest parallel processes need to run as separate tenants so they don't interfere with each other | 19:27 |
| mtreinish | devananda: but basically it's just a separate tenant user pair is created for each test class in tempest | 19:27 |
| adam_g | devananda, we essentially need the isolated tenants created network to be functional WRT netboot in the same way the current demo tenant's | 19:27 |
| devananda | ah | 19:27 |
| openstackrecheck | Console logs not available after 13:23s for check-tempest-dsvm-full 99927,2,5e08a12 | 19:27 |
| devananda | right. the network bridge setup so ironic can PXE boot the VMs | 19:28 |
| adam_g | i'm not exactly sure how to accomplish that, but it *should* be doable--the host-level OVS/network config of ironic+devstack predates me | 19:28 |
| Shrews | adam_g: do we need to quickly touch on the suggested duplicated effort of our ironic scenrio tests vs compute api tests? ex) https://review.openstack.org/94439 | 19:29 |
| mtreinish | Shrews: well, didn't adam_g add a rebuild flag already? | 19:30 |
| Shrews | if api tests aren't going to be using the fake driver, then i can understand the argument | 19:30 |
| Shrews | mtreinish: not that i'm aware of. it's in that review, though | 19:31 |
| Shrews | adam_g: ?? | 19:31 |
| adam_g | mtreinish, snapshot flag is in review, added a rescue flag last week that merged | 19:31 |
| mtreinish | adam_g: yeah that's what I thought I remember giving it a +2 | 19:31 |
| sdague | so it occurs to me that we probably don't actually want to skip on these flags | 19:32 |
| sdague | we actually want to make the call, and say that an Unsupported is kosher | 19:32 |
| adam_g | ? | 19:32 |
| sdague | because you still have to implement the API | 19:32 |
| devananda | um | 19:32 |
| mtreinish | sdague: well we don't have that resp from nova do we? | 19:32 |
| mtreinish | that's why I said go with a feature flag | 19:32 |
| devananda | sdague: we're talking about Nova features here | 19:32 |
| sdague | devananda: yep | 19:33 |
| adam_g | and mtreinish is right, most are async ops that will return a 200/201 from API and fail in the backend | 19:33 |
| devananda | sdague: Nova has accepted this driver knowing full well those are not supported | 19:33 |
| sdague | devananda: right, but that's supposed to actually bubble up | 19:33 |
| devananda | ironic's virt driver raises NotImplemented() | 19:33 |
| sdague | mtreinish: so the theory was always that this would bubble up, maybe that's currently broken in nova | 19:33 |
| devananda | what Nova does with that, we can't exactly control | 19:34 |
| sdague | devananda: yep | 19:34 |
| sdague | this actually isn't ironic specific | 19:34 |
| devananda | sure | 19:34 |
| sdague | it just occured to me recently | 19:34 |
| devananda | the hypervisor support matrix is also a fun topic | 19:34 |
| mtreinish | sdague: sure, I just think it's broken (and I'm not sure it ever worked) | 19:34 |
| adam_g | Shrews, is there an API test for rebuild currently? | 19:34 |
| sdague | mtreinish: that's fine, that's actually a good topic for the nova midcycle | 19:34 |
| mtreinish | it's not like a run things alot without libvirt though.. | 19:34 |
| Shrews | adam_g: i believe so... someone pointed it out to me the other day... lemme check | 19:35 |
| mtreinish | adam_g: yes there are nova api rebuild tests | 19:35 |
| sdague | because as we get compute drivers that are substantially less function, 201 is kind of a terrible API response :) | 19:35 |
| devananda | sdague: fwiw, ironic is trying to solve this internally too, as we get hardware drivers that can't do certain things | 19:36 |
| devananda | sdague: nothing that would be exposed to Nova yet | 19:36 |
| devananda | sdague: just something to earmark for the broader "how do openstack APIs represent feature disparaty in back ends" discussion | 19:36 |
| devananda | *disparity | 19:36 |
| Shrews | adam_g: https://github.com/openstack/tempest/blob/master/tempest/api/compute/servers/test_server_actions.py#L124 | 19:37 |
| devananda | Shrews: looks like that doesn't cover preserve-ephemeral | 19:37 |
| Shrews | adam_g: mtreinish: I don't think they test preserve_ephemeral, though, which is what my review is doing | 19:37 |
| Shrews | devananda: yep | 19:37 |
| sdague | Shrews: preserve_emphemeral is pass how? | 19:38 |
| adam_g | Shrews, yeah. i wonder if that would fit better in the API tests, but is baremetal the only thing that supports preserve_ephemeral? | 19:38 |
| Shrews | adam_g: *shrug* | 19:39 |
| Shrews | sdague: should be an option to rebuild | 19:39 |
| sdague | So the answer of does libvirt do it would be good | 19:40 |
| *** ajo|mac has joined #openstack-qa | 19:40 | |
| sdague | because I could see adding that to rebuild tests | 19:40 |
| adam_g | looks like its a baremetal/ironic-only thing | 19:40 |
| Shrews | panda:nova shrews$ grep nova/virt/libvirt/preserve_ephemeral *.py | 19:41 |
| Shrews | panda:nova shrews$ | 19:41 |
| openstackrecheck | Console logs not available after 13:22s for check-tempest-dsvm-full 97257,12,4eec8e8 | 19:41 |
| devananda | huh | 19:41 |
| Shrews | wow, that's a major fail | 19:41 |
| devananda | libvirt doesn't seem to implement rebuild at all? | 19:41 |
| adam_g | rebuild yes | 19:41 |
| adam_g | preserve_ephemeral no | 19:42 |
| adam_g | IIRC it was added specifically for baremetal, as a way of piggybacking on rebuild to allow image upgrades | 19:42 |
| Shrews | grep preserve_ephemeral nova/virt/libvirt/*.py | 19:42 |
| Shrews | was what i meant :) | 19:42 |
| devananda | grep "rebuild" in nova/virt/libvirt | 19:42 |
| devananda | nothing | 19:42 |
| devananda | and the base class imp is raise NotImplemented | 19:42 |
| *** enikanorov_ has joined #openstack-qa | 19:43 | |
| Shrews | so, there is a *default* rebuild implementation, if the driver doesn't implement it | 19:43 |
| devananda | ahh https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L2436 | 19:44 |
| devananda | right | 19:44 |
| sdague | devananda: yeh, it's implemented above the driver | 19:44 |
| devananda | which explicitly does not support preserve ephemeral | 19:44 |
| sdague | because it can be built out of composite calls | 19:44 |
| devananda | yep | 19:44 |
| devananda | so that's a baremetal only capability in nova now :) | 19:44 |
| sdague | except in your whacky pants use case of making the ephemeral disk non ephemeral :) | 19:45 |
| devananda | i blame lifeless | 19:45 |
| sdague | I do too | 19:45 |
| sdague | so is that really something tripleo needs, or is it just the tripleo use case? | 19:45 |
| devananda | and the nova team for accepting it there | 19:45 |
| devananda | lifeless would say they need it | 19:46 |
| devananda | to preserve local state when upgrading an instance | 19:46 |
| devananda | since the tripleo story for upgrades is "rebuild" not "apt-get distupgrade" | 19:46 |
| sdague | and an instance here means.... | 19:46 |
| devananda | an instance of the undercloud == any piece of the control plane which makes up the overcloud | 19:47 |
| *** enikanorov__ has quit IRC | 19:47 | |
| devananda | which is where this is used by tripleo | 19:47 |
| * devananda avoids bikeshedding on that | 19:48 | |
| Shrews | anyway... since ironic is the only one to support that option, keep it as a scenario test, or move to an api test w/ a new option to enable it? | 19:49 |
| mtreinish | Shrews: what about a compute api test service tagged for ironic | 19:49 |
| mtreinish | a scenario test would only be necessary if the preserved ephemeral disk is needed for something down the line | 19:50 |
| sdague | honestly, scenario is probably simpler | 19:50 |
| sdague | because otherwise we're going to have to flip this compute test off for the non ironic case | 19:50 |
| mtreinish | sdague: a service tag does that automatically | 19:50 |
| Shrews | mtreinish: possible we could build on this test (to avoid more time-consuming rebuilds) | 19:50 |
| sdague | mtreinish: not entirely | 19:51 |
| adam_g | CONF.baremetal.driver_enable could be used as the toggle | 19:51 |
| devananda | so, related question | 19:51 |
| sdague | adam_g: sure | 19:51 |
| devananda | when we can get the api/compute tests to run against ironic, what need is there for a separate scenario test? | 19:52 |
| sdague | Shrews: yeh, I think I lean on making a single baremetal scenario test that does the set of complicated things that only ironic cares about | 19:52 |
| devananda | i'm probably being obtuse here -- i really should get some food soon | 19:52 |
| sdague | devananda: honestly, mostly there is no need | 19:52 |
| devananda | but what am i missing? what does a scenario test give us that .... | 19:52 |
| devananda | right, ok :) | 19:52 |
| sdague | I think for mental organizing given that "NotImplemented" is poorly reflected, it might make it simpler to keep all the ironic specific behavior that no other compute can do in 1 scenario test | 19:53 |
| *** dims_ has joined #openstack-qa | 19:53 | |
| sdague | but that's just a mental sorting preference | 19:53 |
| sdague | and honestly, I could go the other way | 19:54 |
| sdague | with compute/api/servers/baremetal | 19:54 |
| openstackgerrit | Dolph Mathews proposed a change to openstack-infra/elastic-recheck: add query for dox publish failure https://review.openstack.org/104664 | 19:54 |
| mtreinish | sdague: my concern with doing it all in scenario is the official client thing | 19:54 |
| openstackrecheck | Console logs not available after 13:23s for gate-tempest-dsvm-neutron-large-ops 104476,3,175264e | 19:54 |
| sdague | mtreinish: yeh, but you know how I'm leaning on that | 19:54 |
| Shrews | so, i hear votes in both directions :( | 19:55 |
| mtreinish | yeah, but your inclination is still not a project policy change... | 19:55 |
| sdague | Shrews: mtreinish is the PTL, I'll defer to his preference | 19:55 |
| sdague | mtreinish: that's definitely true | 19:56 |
| Shrews | if we move it to an api test, then we might as well remove ironic scenario tests b/c they hold no value then | 19:56 |
| devananda | so one thing to call out | 19:56 |
| devananda | the api/compute tests do not actually touch ironic's api | 19:56 |
| devananda | we will still need api/ironic/admin tests | 19:56 |
| sdague | api/baremetal/admin | 19:57 |
| sdague | yes | 19:57 |
| devananda | er, right | 19:57 |
| mtreinish | Shrews: I'd say an API test for those features because that gives better verification of the api response for those options | 19:57 |
| devananda | with parallel tempest, can we prevent these two classes of tests running together? | 19:57 |
| mtreinish | devananda: just lock it | 19:57 |
| sdague | devananda: with locking | 19:57 |
| Shrews | mtreinish: ok. i'll have to look into how to do the "tagging" thing you suggested, but I'll WIP that review and change it. thx for the feedback :) | 19:57 |
| devananda | ok | 19:57 |
| mtreinish | devananda: or put it in one class | 19:57 |
| devananda | we could run multiple tests within each class in parallel, but not tests from both classes at once | 19:58 |
| openstackgerrit | Dean Troyer proposed a change to openstack-dev/devstack: Support multiple Cinder backend types https://review.openstack.org/100271 | 19:58 |
| mtreinish | devananda: tempest tests are serial within a test class | 19:58 |
| mtreinish | the test classes are executed in parallel | 19:58 |
| devananda | mtreinish: gotcha, thanks | 19:58 |
| sdague | devananda: I think realistically we'll have to figure out what the interaction is | 19:59 |
| *** ajo|mac has quit IRC | 19:59 | |
| devananda | api/baremetal/admin will be adding/deleting resources that would be used by api/compute | 19:59 |
| devananda | hmm | 19:59 |
| sdague | devananda: so it will have to be clever about it, like not delete everything. But I think we can come up with a way | 20:00 |
| devananda | actually we could use a scheduler filter to exclude those resources from the api/compute tests | 20:00 |
| sdague | yeh, that sounds reasonable | 20:00 |
| sdague | so... here is a question | 20:00 |
| devananda | sdague: we prepopulate ironic with the actual VM metadata today (like the MAC addresses for VM NICs) for the PXE_SSH driver | 20:00 |
| sdague | is the content of the api/baremetal/admin so precondition specific that it's not useful to anyone else? | 20:01 |
| sdague | because if the answer is yes, I'd just say, do it in unit tests | 20:01 |
| devananda | should be easy to exclude any transitory nodes created by api/baremetal/admin by setting their properties to unschedulable values (like 1MB RAM) | 20:01 |
| devananda | sdague: i'm not sure what you mean by "precondition specific" | 20:01 |
| sdague | are the tests you are imaging something rax would find value in running against OnMetal regularly | 20:02 |
| *** arnaud has joined #openstack-qa | 20:03 | |
| devananda | i dont know what rax would find useful | 20:03 |
| devananda | they test the interaction between ironic-api, ironic-conductor, the database, the RPC bus, etc | 20:03 |
| mtreinish | sdague: there is still api consistency to think about. Unit testing in tree wouldn't prevent breaking api changes | 20:03 |
| sdague | because what I don't want us to do is build a really complicated setup, and a ton of effort to run these tests in tempest realizing that no one cares, and they don't really help us find isues | 20:03 |
| mtreinish | although I guess if your argument that the api is too specific to the backend it doesn't matter because the api would never be consistent between 2 deployments | 20:04 |
| sdague | mtreinish: it's an admin api, so I'm a little less concerned on it | 20:04 |
| mtreinish | that's something to think about too | 20:04 |
| devananda | and they validate API responses | 20:04 |
| sdague | devananda: yeh, but you can validate API responses in tree | 20:04 |
| devananda | sdague: what's the litmus test for "this test would be useful to rax" mean? | 20:06 |
| devananda | sdague: what kinds of things are/not useful in this context? | 20:06 |
| sdague | devananda: you were talking about creating all kinds of custom schedulers / filters to make this stuff work | 20:06 |
| devananda | no | 20:07 |
| devananda | so lets step backa moment | 20:07 |
| devananda | https://github.com/openstack/tempest/blob/master/tempest/api/compute/keypairs/test_keypairs.py | 20:07 |
| devananda | that is a simple "can i create an artefact and then retrieve it" test | 20:07 |
| devananda | for nova | 20:07 |
| sdague | sure | 20:07 |
| devananda | the api/baremetal/admin tests are the same | 20:07 |
| devananda | can i create this artefact, then manipulate it in predefined ways | 20:07 |
| devananda | and when I do bad-things, do i get the expected error codes | 20:07 |
| devananda | i think that's valuable | 20:07 |
| sdague | test_keypairs is actually a bad example | 20:08 |
| devananda | since the compute/api tests will never exercise those fault conditions | 20:08 |
| openstackrecheck | Console logs not available after 13:24s for check-tempest-dsvm-full 102665,4,f3616ff | 20:08 |
| sdague | because it's resource is only a db artifact | 20:08 |
| devananda | compute/api does not and can not do Creat/Delete testing of resources in Ironic | 20:08 |
| devananda | no, it's a great example | 20:08 |
| devananda | if you think api/baremetal/admin tests shouldn't be in tempest, then delete that one from nova please | 20:09 |
| sdague | devananda: if they are just exercising the db layer, I don't think they are a good fit | 20:09 |
| devananda | then why is that test in tempest? | 20:10 |
| sdague | if create resource causes a thing to materialize | 20:10 |
| devananda | what's "a thing" | 20:10 |
| sdague | devananda: because that's a user est | 20:10 |
| devananda | operators are users too | 20:10 |
| sdague | you would point that at hp public end point with user creds, and it better damn work | 20:11 |
| *** mmaglana has quit IRC | 20:11 | |
| devananda | yes, ironic is unique -- the first service to not have an end-user-facing API | 20:11 |
| sdague | yep | 20:11 |
| devananda | that doesn't mean we shouldn't hold the APi to the same testing standards | 20:11 |
| devananda | i find it odd that i'm arguing for more API testing than you :) | 20:11 |
| sdague | we don't hold admin APIs to the same standards | 20:11 |
| devananda | .... | 20:12 |
| mtreinish | sdague: since when? | 20:12 |
| *** Obulpathi has quit IRC | 20:12 | |
| mtreinish | that's news to me | 20:12 |
| sdague | since have you seen how much those get fiddle in nova | 20:12 |
| sdague | like the events stuff | 20:12 |
| devananda | ok, an admin example | 20:12 |
| devananda | https://github.com/openstack/tempest/blob/master/tempest/api/compute/admin/test_flavors.py | 20:12 |
| sdague | most of it doesn't get tested | 20:12 |
| devananda | flavors are nothing but db artefacts too | 20:12 |
| mtreinish | sdague: yeah but there are coverage gaps everywhere because the api surface is big. It's not an admin api specific thing | 20:13 |
| devananda | and only admins can modify them | 20:13 |
| sdague | mtreinish: if you talk with the nova team, it is about being admin specific. The compat rules are thought of differently | 20:13 |
| sdague | so, honestly, it's late on a virtual friday | 20:14 |
| devananda | sdague: how is api/compute/admin/test_flavors different? | 20:14 |
| adam_g | sdague, aren't you supposed to be in vermont? | 20:14 |
| sdague | adam_g: tonight | 20:15 |
| adam_g | ah | 20:15 |
| sdague | I'm at bard college right now | 20:15 |
| adam_g | so, i think we've come away with enough actionable things for the time being | 20:15 |
| adam_g | broadly | 20:15 |
| sdague | yeh | 20:15 |
| devananda | i feel like we made progress on the discussion about getting compute tests to run with the ironic driver, which is great, but i'm still stuck on this "dont test an admin api" bit | 20:15 |
| adam_g | 1) get rid of virtual-ironic, get the current ironic dsg jobs passing against VIRT_TYPE=ironic | 20:15 |
| adam_g | 2) get tempest running in parallel, with concurrency >= 2 | 20:16 |
| adam_g | 2 will require addressing DIB ramdisk size, isolated credentials | 20:16 |
| sdague | devananda: I'm not actually saying that. I'm saying that we're getting to the point of general bloat in tempest that I think we're going to need to start having considerations about whether admin api testing is actually best done in tempest or not | 20:17 |
| adam_g | devananda, sorry, i wasn't #endmeeting'ing | 20:17 |
| sdague | and as ironic is different in a lot of ways it's probably worth at least thinking about it | 20:17 |
| Shrews | mtreinish: should these changes for ironic be going through a -specs repo? | 20:17 |
| sdague | because at 5 integrated projects, we could clearly say Tempest owns all this | 20:17 |
| sdague | but at 13 | 20:18 |
| *** sweston has joined #openstack-qa | 20:18 | |
| sdague | 15 | 20:18 |
| sdague | 20 | 20:18 |
| *** Sukhdev has joined #openstack-qa | 20:18 | |
| sdague | with very few people working finding issues when this goes wrong | 20:18 |
| mtreinish | Shrews: you should probably have one big ironic spec to rule them all | 20:18 |
| sdague | and no one can merge code | 20:18 |
| openstackgerrit | Morgan Fainberg proposed a change to openstack/tempest: Add a skip for bug #1334368 https://review.openstack.org/104610 | 20:18 |
| openstackgerrit | Morgan Fainberg proposed a change to openstack/tempest: Re-enable 'check_trust_roles' and look for HTTP 200 https://review.openstack.org/104673 | 20:18 |
| mtreinish | at least for adding tests | 20:18 |
| sdague | that becomes something we should think about if it makes sense to validate closer to the project | 20:19 |
| mtreinish | Shrews: if there is improvments or other features aside from adding tests they could be there own spec too | 20:19 |
| devananda | sdague: so there's another conversation about graduating projects (and thus cogating with them) too readily | 20:19 |
| devananda | sdague: which I would love to have with you another time | 20:19 |
| sdague | devananda: sure | 20:19 |
| Shrews | mtreinish: qa-specs repo? | 20:19 |
| mtreinish | Shrews: yeah in the qa-specs repo | 20:19 |
| devananda | sdague: but that doesn't preclude giving thempest the capability to exercise all the APIs in the world | 20:20 |
| mtreinish | there is a new-ish section in the qa-specs readme about specs for new tests too | 20:20 |
| sdague | devananda: maybe, exercising all the APIs in the world is expensive | 20:20 |
| Shrews | mtreinish: thanks | 20:20 |
| devananda | sdague: perhaps just through out-of-tree tests | 20:20 |
| *** zz_johnthetubagu is now known as johnthetubaguy | 20:20 | |
| sdague | and has a huge maintenance cost | 20:20 |
| sdague | and right now, I'm very pessimistic on people paying the maintenance costs that we currently have | 20:21 |
| devananda | sdague: *nod* | 20:21 |
| sdague | as I litterally have done very little except that for the last month | 20:21 |
| openstackrecheck | Console logs not available after 13:22s for check-tempest-dsvm-full 88067,6,cfc4011 | 20:21 |
| morganfainberg | mtreinish, https://review.openstack.org/#/q/status:open+topic:bug/1334368,n,z tempest and keystone changes for the API modifications | 20:22 |
| morganfainberg | mtreinish, thanks for the discussion earlier today | 20:22 |
| devananda | sdague: the graduation system creates this, though, and i do agree that we need to address it more broadly | 20:22 |
| mtreinish | morganfainberg: cool so the backport was fair game? | 20:22 |
| morganfainberg | working on getting the buy-in for stable and master for that change | 20:22 |
| morganfainberg | mtreinish, easy backport, but we'll see what people say | 20:22 |
| morganfainberg | mtreinish, but i'm keeping the changes in lock-step. | 20:22 |
| sdague | devananda: I suppose. But at the same time you argue very strongly for graduating a new program. :) | 20:22 |
| devananda | sdague: /me points to prior statement about "another time" :) | 20:23 |
| sdague | :) | 20:23 |
| devananda | sdague: ah. i wouldn't care as much if not for the pressure to remove the nova.virt.baremetal driver | 20:23 |
| devananda | sdague: if nova would just delete that from the tree, i'd be thrilled | 20:23 |
| devananda | and not worry as much | 20:23 |
| mtreinish | morganfainberg: ok, cool. Well if you need a +2 on the backport just ping me. I'm sure it LGTM :) | 20:25 |
| mtreinish | morganfainberg: the skip is blocked until there is a keystone +2 on the master change | 20:25 |
| morganfainberg | mtreinish, right. i'll ping you once we have other buy in. yep. | 20:25 |
| morganfainberg | mtreinish, actually should be blocked until we have the +2 and the corresponding buy in for stable | 20:25 |
| devananda | anyhow, it's 1:30 and i haven't eaten in 7 hours now. running out of steam... | 20:25 |
| mtreinish | morganfainberg: isn't it fun trying to line all this up? :) | 20:26 |
| devananda | sdague: let's continue ^^that discussion over a drink some time :) | 20:26 |
| mtreinish | morganfainberg: yeah that's probably a fair thing to wait on | 20:26 |
| morganfainberg | mtreinish, i tagged tempest in the bug itself so it's tracked officially there as well in LP | 20:26 |
| sdague | devananda: sounds good | 20:26 |
| morganfainberg | mtreinish, well, i have a bigger set of changes to come in master once we get buy-in for this. notably, anytime a GET is done, make the same request with HEAD and make sure things work the same (minus the body) | 20:27 |
| *** bnemec_appt is now known as bnemec | 20:27 | |
| devananda | thanks, all. have a great (long) weekend! | 20:27 |
| morganfainberg | mtreinish, but i'll tackle that one once this blocker for getting mod_wsgi keystone deployed is fixed | 20:28 |
| morganfainberg | devananda, cheers, have a godo weekend yourself | 20:28 |
| *** jgrimm has quit IRC | 20:28 | |
| *** jaypipes has quit IRC | 20:29 | |
| *** marun_afk is now known as marun | 20:32 | |
| openstackrecheck | Console logs not available after 13:32s for check-tempest-dsvm-neutron 104026,2,8fd165f | 20:48 |
| openstackrecheck | openstack/nova change: https://review.openstack.org/104021 failed because of: gate-tempest-dsvm-full: https://bugs.launchpad.net/bugs/1298472 | 20:48 |
| openstackrecheck | openstack/tempest change: https://review.openstack.org/90278 failed gate-grenade-dsvm in the gate queue with an unrecognized error | 20:48 |
| *** rbak has quit IRC | 20:51 | |
| *** ildikov has joined #openstack-qa | 20:53 | |
| openstackgerrit | Dean Troyer proposed a change to openstack-dev/devstack: Generate recent changes list from git log https://review.openstack.org/104051 | 20:58 |
| openstackrecheck | Console logs not available after 13:22s for check-grenade-dsvm 100271,9,457b0a4 | 21:02 |
| *** johnthetubaguy is now known as zz_johnthetubagu | 21:09 | |
| openstackgerrit | A change was merged to openstack-infra/devstack-gate: Handle proposed/ branches in test-matrix.py https://review.openstack.org/104448 | 21:10 |
| *** aimon has quit IRC | 21:11 | |
| *** yfried has joined #openstack-qa | 21:13 | |
| openstackrecheck | Console logs not available after 13:22s for gate-nova-python26 97529,12,6e6eab9 | 21:15 |
| openstackrecheck | openstack-dev/devstack change: https://review.openstack.org/104125 failed because of: gate-grenade-dsvm-partial-ncpu: unrecognized error, gate-tempest-dsvm-neutron-heat-slow: https://bugs.launchpad.net/bugs/1297560 https://bugs.launchpad.net/bugs/1311066 | 21:26 |
| openstackgerrit | A change was merged to openstack/tempest: Allowed Address Pair API Tests Enhancements https://review.openstack.org/101522 | 21:36 |
| openstackgerrit | A change was merged to openstack/tempest: VPNaaS API Tests Enhancements https://review.openstack.org/83627 | 21:36 |
| *** jhesketh_ is now known as jhesketh | 21:37 | |
| openstackrecheck | Console logs not available after 13:30s for check-tempest-dsvm-neutron-icehouse 104610,1,900ae55 | 21:40 |
| *** dkranz has quit IRC | 21:42 | |
| *** ajo|mac has joined #openstack-qa | 21:43 | |
| *** prkumar has quit IRC | 21:48 | |
| *** ajo|mac has quit IRC | 21:48 | |
| *** ajo|mac has joined #openstack-qa | 21:52 | |
| openstackrecheck | Console logs not available after 13:24s for check-tempest-dsvm-full 103892,2,c54801f | 21:54 |
| *** gszasz has joined #openstack-qa | 21:55 | |
| lifeless | sdague: FWIW adding preserve-ephemeral to kvm is someone in my todo list; there was support for doing that when I proposed it on the list | 21:56 |
| openstackrecheck | Console logs not available after 13:29s for gate-tempest-dsvm-large-ops 102352,2,fc20c80 | 22:09 |
| openstackrecheck | openstack/nova change: https://review.openstack.org/101883 failed because of: gate-grenade-dsvm: https://bugs.launchpad.net/bugs/1334142 | 22:09 |
| *** dims__ has joined #openstack-qa | 22:13 | |
| *** dims_ has quit IRC | 22:15 | |
| *** Sukhdev has quit IRC | 22:18 | |
| ianw | sdague: any blockers on https://review.openstack.org/#/c/104438/ (Revert "demote f20 to experimental") | 22:26 |
| ianw | lifeless: happy to take suggestions in a further patch for https://review.openstack.org/#/c/101110/ (Track last allocations to ensure forward-progress) but i think your comment was incomplete | 22:27 |
| openstackrecheck | Console logs not available after 13:34s for check-tempest-dsvm-neutron 104123,2,0a9821b | 22:28 |
| lifeless | ianw: hmm, gertty must have not synced I guess | 22:31 |
| lifeless | ianw: I'm roaming... | 22:31 |
| devananda | sdague, adam_g: oh, we had had a goal to enable ironic voting on nova as a third-party driver. that seems to have been stalled within infra: https://review.openstack.org/#/c/97411/ | 22:32 |
| adam_g | devananda, is it stalled or just in progress? | 22:33 |
| adam_g | lifeless, thoughts on enabling DIB to produce thinner deploy ramdisks for CI purposes, possibly based on cirros? | 22:34 |
| lifeless | adam_g: testing other than what we use in prod seems like a good way to let bugs through | 22:35 |
| adam_g | lifeless, this is specifically for devstack-gate, not tripleo CI | 22:35 |
| lifeless | adam_g: I don't understand how that invalidates my comment. | 22:35 |
| adam_g | lifeless, we were discussing ways of improving ironic CI in prep for graduation and it becoming part of the integrated gate. generating a deployment ramdisk based on a constantly changing ubuntu cloud image seems like a good way to let thru many bugs that are not relevant to that. qa folks seem keen on having another moving part in there. | 22:37 |
| *** andreaf has quit IRC | 22:37 | |
| adam_g | the idea was to have somethign thinner that can be cached as a test resource similar cirros or the trove image. a known working ramdisk. | 22:38 |
| lifeless | adam_g: so that seems like they are saying they want something stable and non-moving, and to test that. | 22:38 |
| devananda | adam_g: perhaps "slow progress" is the right term | 22:38 |
| lifeless | I have no objection to that. | 22:38 |
| lifeless | I object to it being ! what we tell users to use. | 22:38 |
| lifeless | CI's job is to detect errors that users would otherwise encounter. | 22:39 |
| devananda | i share that objection. making a leaner deploy ramdisk is a good thing in general | 22:39 |
| devananda | testing something in the gate which is not what we deploy with is less good | 22:39 |
| devananda | *what we recommend to | 22:39 |
| clarkb | why wouldn't you deploy with the same ramdisk? | 22:40 |
| clarkb | I would want to use it in production for the same reasons it makes sense for qa | 22:40 |
| devananda | clarkb: right | 22:40 |
| adam_g | i agree, i'm reiterating points brought up by others 2 hours ago | 22:41 |
| devananda | lifeless: so you didn't answer the initial question from adam_g | 22:41 |
| *** yamahata has quit IRC | 22:41 | |
| devananda | we need to "enable DIB to produce thinner deploy ramdisks" | 22:41 |
| lifeless | devananda: the question was about CI | 22:41 |
| devananda | lifeless: the context is CI | 22:42 |
| lifeless | devananda: if you remove CI from the question, it becomes totally different | 22:42 |
| devananda | the question is still valid | 22:42 |
| lifeless | The ramdisks are already mean to be lean, with just the kernel, busybox and drivers. | 22:42 |
| devananda | and yet it won't run in a 1GB VM | 22:42 |
| devananda | not exactly lean | 22:42 |
| lifeless | WAT | 22:42 |
| * adam_g wonders how much leaner they'd beby making the -generic drivers optional | 22:42 | |
| devananda | that's the problem | 22:43 |
| openstackrecheck | Console logs not available after 13:25s for check-grenade-dsvm-partial-ncpu 103854,1,315bc93 | 22:43 |
| lifeless | first I've heard of it not working in 1GB VMs | 22:43 |
| lifeless | thats clearly a bug | 22:43 |
| openstackrecheck | openstack/tempest change: https://review.openstack.org/101694 failed because of: gate-grenade-dsvm: https://bugs.launchpad.net/bugs/1298472 | 22:43 |
| devananda | lifeless: https://bugs.launchpad.net/diskimage-builder/+bug/1311987 | 22:43 |
| lifeless | -rw-r--r-- 1 robertc robertc 83M May 29 08:33 deploy-ramdisk-ironic.initramfs | 22:43 |
| devananda | lifeless: min ram size was bumped from 512 to 1024 because of this | 22:43 |
| devananda | uncompressed size | 22:44 |
| devananda | read bug | 22:44 |
| lifeless | sure, looking | 22:44 |
| lifeless | so adam_g I don't understand comment 7 in that bug | 22:46 |
| lifeless | it seems like we could just set the parameter | 22:46 |
| lifeless | to trigger ramfs rather than tmpfs | 22:46 |
| lifeless | I don't understand why, having gotten that answer, you didn't use it | 22:49 |
| *** rodrigods_ has joined #openstack-qa | 22:54 | |
| *** asselin has quit IRC | 22:55 | |
| lifeless | anyhow for clarity, I'm all in favor of smaller deploy ramdisks; its one of the biiig concerns I have about IPA | 22:56 |
| openstackrecheck | Console logs not available after 13:23s for check-tempest-dsvm-full 103091,4,9bb1822 | 22:56 |
| lifeless | and why I keep harping on about footprint in the ilo deploy spec | 22:57 |
| *** Sukhdev has joined #openstack-qa | 23:00 | |
| devananda | lifeless: correting a previous typo: it wont run in <1GB VMs | 23:01 |
| lifeless | devananda: apparently it will if ironic adds rootfstype=ramfs to the command line | 23:02 |
| devananda | lifeless: w.r.t. making it smaller, it sounds like we're all in agreement that's good. do you have any particular thoughts on how? | 23:02 |
| lifeless | devananda: no, first thing would be to get a broad breakdown on where the size goes today. | 23:02 |
| devananda | lifeless: to the kernel cmdline? I think that's trivially easy to add | 23:03 |
| lifeless | devananda: so do I | 23:03 |
| lifeless | devananda: I was just pondering pushing up a patch to add it | 23:03 |
| devananda | :-/ | 23:03 |
| devananda | it'd take 2 patches to really test it | 23:03 |
| devananda | 1. add to ironic. 2. change VM RAM size in devstack | 23:03 |
| lifeless | yup | 23:03 |
| devananda | ++ | 23:04 |
| lifeless | holy cow, gertty is now at 37MB syncing | 23:04 |
| *** jamielennox|away is now known as jamielennox | 23:04 | |
| lifeless | no wonder I'm hitting my roaming 200M limit | 23:05 |
| adam_g | lifeless, we have no way of setting boot-time and deploy-time parameters separately. appending root=ramfs to a boot time kernel is going to end its boot in the ramdisk | 23:06 |
| adam_g | at least via ironic config settings | 23:06 |
| devananda | adam_g: why not hard add it to the options https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/pxe_config.template#L11 | 23:07 |
| devananda | no need for config option change | 23:07 |
| lifeless | adam_g: that surprises me since the deploy time parameters are not currently sent to the runtime ramdisk | 23:07 |
| adam_g | devananda, i was just looking there. | 23:07 |
| adam_g | the pxe_append_params are | 23:08 |
| adam_g | taking ti #openstack-ironic, actually | 23:09 |
| *** ildikov has quit IRC | 23:10 | |
| *** nithyag__ has quit IRC | 23:11 | |
| *** nithyag__ has joined #openstack-qa | 23:12 | |
| openstackgerrit | Dean Troyer proposed a change to openstack-dev/devstack: Support multiple Cinder backend types https://review.openstack.org/100271 | 23:12 |
| *** gszasz has quit IRC | 23:15 | |
| openstackrecheck | Console logs not available after 13:32s for check-tempest-dsvm-neutron-heat-slow 100271,10,f152f51 | 23:20 |
| *** mtreinish has quit IRC | 23:31 | |
| openstackrecheck | Console logs not available after 13:23s for check-tempest-dsvm-full 104606,2,4e4fe2d | 23:34 |
| *** rodrigods_ has quit IRC | 23:35 | |
| *** mtreinish has joined #openstack-qa | 23:39 | |
| openstackrecheck | Console logs not available after 13:29s for check-tempest-dsvm-full 102064,9,b903a78 | 23:50 |
| *** oomichi has joined #openstack-qa | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!