Thursday, 2016-11-17

« Wednesday, 2016-11-16 Index Friday, 2016-11-18 »
*** diogogmt has joined #openstack-fwaas00:27
*** hoangcx has joined #openstack-fwaas00:37
*** hoangcx has quit IRC02:23
*** mickeys has quit IRC02:26
*** lnicolas has joined #openstack-fwaas02:56
*** mickeys has joined #openstack-fwaas03:11
*** hoangcx has joined #openstack-fwaas04:24
*** mickeys has quit IRC05:05
*** mickeys has joined #openstack-fwaas05:06
*** amotoki has joined #openstack-fwaas05:51
*** codertux has joined #openstack-fwaas07:03
*** korzen has joined #openstack-fwaas07:28
*** korzen has quit IRC07:56
*** codertux has quit IRC09:02
*** mickeys has quit IRC09:32
*** codertux has joined #openstack-fwaas09:44
*** mickeys has joined #openstack-fwaas09:50
*** codertux has quit IRC09:53
*** mickeys has quit IRC09:56
*** hoangcx has quit IRC10:17
*** codertux has joined #openstack-fwaas10:23
*** korzen has joined #openstack-fwaas10:46
*** codertux has quit IRC11:14
*** korzen has quit IRC11:26
*** SarathMekala has joined #openstack-fwaas12:41
*** andrein has joined #openstack-fwaas12:53
*** chandanc_ has joined #openstack-fwaas12:55
*** chandanc_ has quit IRC13:20
*** andrein has quit IRC13:22
*** andrein has joined #openstack-fwaas13:25
*** diogogmt has quit IRC13:58
*** andrein has quit IRC14:31
*** andrein has joined #openstack-fwaas14:37
*** yamamoto has joined #openstack-fwaas15:03
*** SarathMekala has quit IRC15:33
*** diogogmt has joined #openstack-fwaas15:46
*** andrein has quit IRC15:47
SridarK_yamamoto: ping15:51
*** SarathMekala has joined #openstack-fwaas15:57
SridarK_SarathMekala: hi15:59
*** andrein has joined #openstack-fwaas16:03
*** amotoki has quit IRC16:07
*** andrein has quit IRC16:20
*** andrein has joined #openstack-fwaas16:25
*** mickeys has joined #openstack-fwaas16:26
*** mickeys has quit IRC16:32
*** yamamoto has quit IRC16:34
*** yamamoto has joined #openstack-fwaas16:43
*** SarathMekala has quit IRC16:48
*** yamamoto has quit IRC16:48
*** mickeys has joined #openstack-fwaas17:23
mfranc213SridarK_: ping17:30
SridarK_mfranc213: hi17:30
SridarK_mfranc213: in a mtg, slow response but pls go ahead17:30
mfranc213SridarK_: nope, my question is not critical.  will talk to you later :)17:31
SridarK_mfranc213: ok will ping when i am done in abt an hour17:31
mfranc213SridarK_: sounds good.  have a good meeting!17:31
SridarK_:-)17:31
*** SumitNaiksatam has joined #openstack-fwaas17:48
*** yamamoto has joined #openstack-fwaas18:13
*** yamamoto has quit IRC18:19
*** SumitNaiksatam has left #openstack-fwaas18:54
*** diogogmt has quit IRC18:55
SridarK_mfranc213: hi18:55
mfranc213Sridar_K: hello18:55
mfranc213SridarK_: john schwarz was asking me early about the https://review.openstack.org/#/c/391320/18:56
mfranc213SridarK_: http://eavesdrop.openstack.org/irclogs/%23openstack-neutron/%23openstack-neutron.2016-11-17.log.html#t2016-11-17T15:15:5618:56
*** SumitNaiksatam has joined #openstack-fwaas18:57
*** diogogmt has joined #openstack-fwaas18:57
mfranc213SridarK_: by any chance would you have an update i could pass along to jonathan?18:57
*** SumitNaiksatam has quit IRC19:00
SridarK_mfranc213: sure just went thru the logs19:00
mfranc213SridrK_: it seems your local env is good, but jenkins has the wrong config value?19:01
SridarK_mfranc213: On https://review.openstack.org/#/c/391320/, the tests are all passing but on Jenkins it fails because all the api extensions are enabled19:01
mfranc213SridarK_: do we (the royal we) know how to fix?19:02
SridarK_so when the tests run for the v1 job, v1 tests pass as expected and v2 tests are supposed to be skipped19:02
mfranc213yeah, makes sense19:02
SridarK_but they are not19:02
*** diogogmt has quit IRC19:02
SridarK_and vice versa for the v2 job19:02
mfranc213right.19:03
SridarK_i was abt to ping njohnston for help on this as this is possibly something on the Jenkins job config19:03
SridarK_the skips happen locally19:03
mfranc213SridarK_: something tells me njohnston will appear momentarily :)19:04
SridarK_when njohnston is avail - could we may be take a look at this ?19:04
SridarK_mfranc213: :-)19:04
SridarK_njohnston: ping19:04
mfranc213SridarK_: i have found that njohnston is everywhere at all times, wherever and whenever njohnston is needed19:05
SridarK_mfranc213: absolutely could not agree more19:05
mfranc213maybe njohnston is santa clause?19:05
SridarK_:-)19:05
SridarK_i pushed the patch late last night after verifying that tests were running locally but was a bit unsure on jenkins19:06
SridarK_just had to get on calls in the am and was going to look for njohnston anyways19:06
SridarK_it seems there should be some knob that sets this in the tempest.conf19:07
SridarK_locally it does pick it up from devstack local.conf - or kind of derives it based on what plugins are set19:07
SridarK_and i would have thought that it should be the same on Jenkins19:08
mfranc213SridarK_: i suspect nate will know the answer to this mystery.19:08
SridarK_yes it is all pixie dust, dark magic for me :-)19:09
SridarK_when njohnston is around - we can take a quick look to triage this19:10
njohnstonI'm back19:10
SridarK_njohnston: hi19:10
mfranc213well, whaddaya know!  hi nate :)19:11
SridarK_will let u catch up on the text above - u may know the answer already19:11
njohnstonI am reading through http://logs.openstack.org/20/391320/8/check/gate-neutron-fwaas-v1-dsvm-tempest/8979b90/console.html ... I see fwaas v1 configured ... and I see the fwaas v2 tests like neutron_fwaas.tests.tempest_plugin.tests.api.test_fwaasv2_extensions.FWaaSv2ExtensionTestJSON.test_create_show_delete_firewall_group running and failing...19:13
SridarK_yes they should be skipped19:13
SridarK_they are not skipped because api_extensions = all in tempest.conf19:15
SridarK_so the skip logic finds that the fwaas_v2 ext is probab set and runs that19:15
njohnstonI see.  That sounds to me like we should change the devstack plugin to manage the api_extensions line19:16
SridarK_hmm, devstack will set the fwaas plugin - which in turn will load the correct ext19:17
SridarK_as happens in my local env19:17
njohnstonso the tests pass in your local devstack?19:18
SridarK_yes19:18
SridarK_the skips happen correctly19:18
SridarK_maybe a quick call so i can share my screen19:18
njohnstonno, I don't think that would help, yet19:19
SridarK_ok19:20
SridarK_let me paste the contents of tempest.conf as in my local env19:20
SridarK_http://paste.openstack.org/show/589636/19:22
SridarK_this is the section in my local tempest.conf corresponding to api_ext19:23
SridarK_compare that with:19:24
SridarK_http://logs.openstack.org/20/391320/8/check/gate-neutron-fwaas-v2-dsvm-tempest-nv/e95eb51/logs/tempest_conf.txt.gz19:24
SridarK_u will see api_extensions = all19:24
njohnstonright19:26
njohnstonso we need that to be set to something we control19:26
SridarK_it seems so, i would have thought that all this should be controlled by the devstack tag for v1 or v2 that is set in the fwaas jobs19:27
SridarK_clearly that does take effect as the v2 tests pass in the v2 job and v1 tests pass in the v1 job19:27
njohnstonI think we need to escalate this to infra, or else figure out a different way to skip19:28
njohnstonbut look at this: api_extensions19:28
njohnstonoops19:28
njohnstonthis: https://github.com/openstack-infra/devstack-gate/blob/master/devstack-vm-gate.sh#L802-L81019:28
njohnstonit seems to me that the stanza I just linked to would exit if api_extensions was set to 'all' when on master19:30
SridarK_but it does seem to run thru19:30
njohnstonright, which confounds me19:31
SridarK_sigh yes :-(19:31
njohnstonall right, let me approach infra with this19:32
SridarK_njohnston: many thanks19:32
njohnstonperhaps sc68cal can help, it seems he has deep devstack magic19:32
SridarK_yes that is for sure, but i think this may also be an infra thing as well19:33
SridarK_because devstack clearly does the right thing locally19:33
sc68calhmm?19:33
njohnstonhello sc68cal19:33
SridarK_sc68cal: hi, ur fwaas buddies hound u :-)19:34
njohnstonwe are confounded by the way devstack-vm-gate.sh works19:34
* sc68cal reads scrollback19:35
SridarK_sc68cal: let me summarize19:39
SridarK_1) we have a v1 job that runs v1 tempest tests and a v2 job that runs v2 tempest19:39
SridarK_2) by virtue of devstack settings for which plugin is loaded, on the v1 job, v2 tests will be skipped and on the v2 job, v1 tests  will be skipped19:39
SridarK_3) The skip is controlled by the ext loaded (which should be either v1 or v2) but it seems tempest.conf on jenkins has api_extensions set to all so the skip does not happen.19:40
SridarK_4) works fine on local devstack env19:40
sc68calright but d-g sets api_extensions to all19:40
SridarK_so something about how api_extensions get set on the gate19:40
sc68calso you guys end up having failures19:40
SridarK_sc68cal: yes exactly19:40
njohnstonsc68cal: what is the meaning of this stanza?  It looks like it will die if api_extensions=all https://github.com/openstack-infra/devstack-gate/blob/master/devstack-vm-gate.sh#L802-L81019:41
sc68calTake a look at NETWORK_API_EXTENSIONS setting in devstack19:42
sc68calit defaults to "all"19:42
njohnstonok so if we set that in the devstack plugin we should be able to work19:42
sc68calyou might be able to define NETWORK_API_EXTENSIONS=router,fwaas[v1,v2] in your job configs19:42
sc68calnot in the plugin19:42
sc68calin project-config19:42
njohnstonah, ok19:43
sc68calI think.19:43
sc68cali'm speed grepping19:43
SridarK_ok that makes sense - there are not many scenarios where we may need an either or on api ext19:43
sc68calI think at the very least NETWORK_API_EXTENSIONS will need to be something like19:44
sc68calNETWORK_API_EXTENSIONS="router,fwaas[v1,v2]"19:44
SridarK_sc68cal: will network, port, subnet get picked up by default or we need to specify them too19:45
sc68calSridarK_: those are core api resources19:45
SridarK_ok so that should be good19:45
sc68calhttps://github.com/openstack-infra/project-config/blob/master/jenkins/jobs/neutron-fwaas.yaml#L78 ?19:46
SridarK_sc68cal: many thanks19:48
njohnstonhttps://review.openstack.org/39920719:48
njohnstonsc68cal: Thanks much!19:49
SridarK_njohnston: many thanks, just reviewed19:50
SridarK_njohnston: how would we test this19:51
njohnstonI think it is untestable, since adding a Depends-On to your patch won't make the project-config take effect just for your change, in my impression.19:51
SridarK_yes exactly19:52
njohnstonwe just have to wait for it to merge19:52
SridarK_ok that works - we can iterate if needed - looking at the tempest.conf will tell us for sure19:53
njohnstonrevved it19:53
njohnstonindeed19:53
SridarK_dang hang on, we may need one more ext for v119:55
njohnstonsomething l3 related?19:58
SridarK_for fwaas v1 we also need the routerinsertion ext20:00
SridarK_i am hoping with this and all the core stuff we are good20:00
njohnstonfirewallrouterinsertion or just routerinsertion20:01
njohnstonSridarK_: ^^20:02
SridarK_hmm checking20:03
SridarK_https://github.com/openstack/neutron-fwaas/blob/master/neutron_fwaas/services/firewall/fwaas_plugin.py#L14620:04
SridarK_i had the right thing posted in the review20:05
SridarK_fwaasrouterinsertion20:05
SridarK_njohnston: http://paste.openstack.org/show/589636/20:06
SridarK_this is from my local devstack20:06
njohnstonfixed20:06
SridarK_i went thru the list and i think most of the other things may not really be needed20:06
njohnstonthat paste shows fwaas_v2; I thought yous aid this was only needed for v1?20:07
SridarK_although i see project-id20:07
SridarK_yes that is from a v2 deployment20:08
SridarK_we dont need fwaasrouterinsertion for that20:08
njohnstonright20:08
SridarK_now i am not sure we need project-id20:08
SridarK_lets go with this perhaps some folks on infra can add some light too20:10
njohnstonok20:10
mfranc213well, i'll just keep +1'ing it  :)20:11
SridarK_:-)20:11
SridarK_since u are the infra liason so u should be able to get this in quickly,20:12
*** andrein has quit IRC20:12
njohnstonpower has it's privileges ;-)20:12
SridarK_:-)20:17
SridarK_sc68cal: , njohnston: thx for the learning20:18
SridarK_the project-config stuff has always been a mystery to me - i can see some light now :-)20:19
« Wednesday, 2016-11-16 Index Friday, 2016-11-18 »

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