| openstackgerrit | Maru Newby proposed a change to openstack/neutron: Exclude .ropeproject from flake8 checks https://review.openstack.org/82676 | 00:01 |
|---|---|---|
| openstackgerrit | Joe Gordon proposed a change to openstack/neutron: Disable oslo.messaging debug logs https://review.openstack.org/82618 | 00:02 |
| *** carlp has joined #openstack-neutron | 00:10 | |
| *** openstack has joined #openstack-neutron | 00:20 | |
| *** matsuhashi has joined #openstack-neutron | 00:26 | |
| *** baoli has joined #openstack-neutron | 00:27 | |
| *** baoli has quit IRC | 00:28 | |
| *** baoli has joined #openstack-neutron | 00:28 | |
| *** sungju has joined #openstack-neutron | 00:29 | |
| *** manishg has quit IRC | 00:30 | |
| *** singhs has quit IRC | 00:32 | |
| *** salv-orlando has joined #openstack-neutron | 00:37 | |
| *** pcm_ has joined #openstack-neutron | 00:39 | |
| pcm_ | nati_ueno: ping | 00:40 |
| *** welldannit has quit IRC | 00:43 | |
| *** _cjones_ has quit IRC | 00:52 | |
| nati_ueno | pcm_: pong | 00:53 |
| nati_ueno | pcm_: sorry still working on fixing ut | 00:53 |
| nati_ueno | pcm_: can we talk tommorow? | 00:53 |
| *** yamahata has joined #openstack-neutron | 00:55 | |
| *** yamahata_ has quit IRC | 00:56 | |
| *** yamahata_ has joined #openstack-neutron | 00:56 | |
| *** tomoe_ has quit IRC | 00:58 | |
| pcm_ | nati_ueno: OK. Let me know as soon as you can. I'll try to email my thoughts. | 00:58 |
| *** armitage81 has joined #openstack-neutron | 01:00 | |
| *** armitage81 has quit IRC | 01:02 | |
| *** krtaylor has quit IRC | 01:07 | |
| *** thuc has quit IRC | 01:10 | |
| *** thuc has joined #openstack-neutron | 01:10 | |
| *** krtaylor has joined #openstack-neutron | 01:11 | |
| *** dave_tucker is now known as dave_tucker_zzz | 01:13 | |
| *** thuc has quit IRC | 01:15 | |
| *** mlavalle has quit IRC | 01:18 | |
| openstackgerrit | Abhishek Raut proposed a change to openstack/neutron: Fix segment allocation tables in Cisco N1kv plugin https://review.openstack.org/78506 | 01:19 |
| *** krtaylor has quit IRC | 01:20 | |
| *** mwagner_lap has joined #openstack-neutron | 01:23 | |
| *** krtaylor has joined #openstack-neutron | 01:23 | |
| *** rwsu has quit IRC | 01:26 | |
| *** SumitNaiksatam has joined #openstack-neutron | 01:29 | |
| *** matsuhashi has quit IRC | 01:30 | |
| *** matsuhashi has joined #openstack-neutron | 01:30 | |
| *** carl_baldwin has joined #openstack-neutron | 01:35 | |
| *** Jianyong has joined #openstack-neutron | 01:36 | |
| *** emagana has quit IRC | 01:41 | |
| *** rwsu has joined #openstack-neutron | 01:42 | |
| *** SumitNaiksatam has quit IRC | 01:45 | |
| *** yamahata_ has quit IRC | 01:48 | |
| openstackgerrit | Nachi Ueno proposed a change to openstack/neutron: Improve vif attributes related with firewalling https://review.openstack.org/21946 | 01:48 |
| *** tomoe_ has joined #openstack-neutron | 01:50 | |
| *** yamahata_ has joined #openstack-neutron | 01:50 | |
| *** BuSerD has quit IRC | 01:50 | |
| openstackgerrit | shihanzhang proposed a change to openstack/neutron: Prevent dhcp port deletion from the API https://review.openstack.org/81768 | 01:52 |
| *** tdx-ram has joined #openstack-neutron | 01:52 | |
| *** SumitNaiksatam has joined #openstack-neutron | 01:52 | |
| *** spandhe has quit IRC | 01:54 | |
| *** BuSerD has joined #openstack-neutron | 01:56 | |
| *** yonglihe_ has quit IRC | 01:56 | |
| *** carl_baldwin has quit IRC | 01:58 | |
| *** yonglihe_ has joined #openstack-neutron | 01:59 | |
| *** Guest3042 has joined #openstack-neutron | 02:00 | |
| Guest3042 | Hello | 02:00 |
| *** baoli has quit IRC | 02:00 | |
| Guest3042 | Trying to track a request through openstack | 02:01 |
| *** xuhanp has joined #openstack-neutron | 02:01 | |
| Guest3042 | I find plenty on older arch but I need better understanding of new flow with neutron | 02:01 |
| *** gongysh has joined #openstack-neutron | 02:11 | |
| *** xianghui has joined #openstack-neutron | 02:12 | |
| *** carl_baldwin has joined #openstack-neutron | 02:14 | |
| *** carl_baldwin has quit IRC | 02:15 | |
| *** beagles has quit IRC | 02:16 | |
| *** armitage81 has joined #openstack-neutron | 02:19 | |
| *** carl_baldwin has joined #openstack-neutron | 02:19 | |
| *** armitage81 has quit IRC | 02:19 | |
| *** armitage81 has joined #openstack-neutron | 02:19 | |
| *** thuc has joined #openstack-neutron | 02:21 | |
| *** xianghui has quit IRC | 02:22 | |
| *** carl_baldwin has quit IRC | 02:23 | |
| *** xianghui has joined #openstack-neutron | 02:25 | |
| *** thuc has quit IRC | 02:26 | |
| *** carl_baldwin has joined #openstack-neutron | 02:27 | |
| *** oda-g_ has joined #openstack-neutron | 02:29 | |
| *** HenryG has joined #openstack-neutron | 02:29 | |
| *** oda-g has quit IRC | 02:30 | |
| *** sbalukoff has quit IRC | 02:31 | |
| *** carl_baldwin has quit IRC | 02:31 | |
| openstackgerrit | shihanzhang proposed a change to openstack/neutron: Prevent dhcp port deletion from the API https://review.openstack.org/81768 | 02:31 |
| *** tomoe_ has quit IRC | 02:33 | |
| *** mestery has joined #openstack-neutron | 02:34 | |
| *** tomoe_ has joined #openstack-neutron | 02:36 | |
| *** Sukhdev has joined #openstack-neutron | 02:40 | |
| *** rkukura has joined #openstack-neutron | 02:43 | |
| *** mestery has quit IRC | 02:46 | |
| *** thuc has joined #openstack-neutron | 02:46 | |
| *** banix has joined #openstack-neutron | 02:46 | |
| *** mestery has joined #openstack-neutron | 02:46 | |
| *** salv-orlando_ has joined #openstack-neutron | 02:49 | |
| *** salv-orlando has quit IRC | 02:50 | |
| *** salv-orlando_ is now known as salv-orlando | 02:50 | |
| *** banix has quit IRC | 02:51 | |
| *** thuc_ has joined #openstack-neutron | 02:51 | |
| *** thuc has quit IRC | 02:54 | |
| *** tomoe_ has quit IRC | 02:58 | |
| *** banix has joined #openstack-neutron | 03:01 | |
| *** carlp has quit IRC | 03:10 | |
| *** matsuhashi has quit IRC | 03:12 | |
| openstackgerrit | Paul Michali proposed a change to openstack/neutron: Cisco VPN driver correct reporting for admin state chg https://review.openstack.org/82306 | 03:15 |
| *** pcm_ has quit IRC | 03:16 | |
| *** devlaps has quit IRC | 03:19 | |
| *** sweston has joined #openstack-neutron | 03:21 | |
| *** oda-g_ has quit IRC | 03:27 | |
| *** oda-g has joined #openstack-neutron | 03:28 | |
| *** matsuhashi has joined #openstack-neutron | 03:30 | |
| *** thuc has joined #openstack-neutron | 03:31 | |
| *** chandankumar_ has joined #openstack-neutron | 03:32 | |
| *** thuc_ has quit IRC | 03:34 | |
| *** armitage81 has quit IRC | 03:34 | |
| *** thuc has quit IRC | 03:35 | |
| *** matsuhashi has quit IRC | 03:36 | |
| *** chandankumar_ has quit IRC | 03:40 | |
| *** Sukhdev has quit IRC | 03:43 | |
| *** sridhar has joined #openstack-neutron | 03:45 | |
| openstackgerrit | shihanzhang proposed a change to openstack/neutron: Prevent dhcp port deletion from the API https://review.openstack.org/81768 | 03:47 |
| *** BuSerD has quit IRC | 03:49 | |
| *** harlowja_ is now known as harlowja_away | 03:49 | |
| *** banix has quit IRC | 03:54 | |
| *** chandankumar_ has joined #openstack-neutron | 03:55 | |
| *** banix has joined #openstack-neutron | 03:56 | |
| *** tomoe_ has joined #openstack-neutron | 04:00 | |
| *** Longgeek has joined #openstack-neutron | 04:10 | |
| *** banix has quit IRC | 04:11 | |
| *** nati_ueno has quit IRC | 04:13 | |
| *** tomoe_ has quit IRC | 04:13 | |
| *** tomoe_ has joined #openstack-neutron | 04:14 | |
| *** thuc has joined #openstack-neutron | 04:18 | |
| *** tomoe_ has quit IRC | 04:18 | |
| *** thuc_ has joined #openstack-neutron | 04:18 | |
| *** devlaps has joined #openstack-neutron | 04:21 | |
| *** tdx-ram has quit IRC | 04:21 | |
| *** thuc has quit IRC | 04:22 | |
| *** dfarrell07 has joined #openstack-neutron | 04:23 | |
| *** Guest3042 has quit IRC | 04:23 | |
| *** spandhe has joined #openstack-neutron | 04:23 | |
| *** matsuhashi has joined #openstack-neutron | 04:23 | |
| *** chandankumar_ has quit IRC | 04:23 | |
| *** spandhe has quit IRC | 04:28 | |
| *** spandhe_ has joined #openstack-neutron | 04:28 | |
| *** tomoe_ has joined #openstack-neutron | 04:29 | |
| *** jecarey has quit IRC | 04:30 | |
| *** Sukhdev has joined #openstack-neutron | 04:32 | |
| *** thuc_ has quit IRC | 04:40 | |
| *** nati_ueno has joined #openstack-neutron | 04:40 | |
| *** thuc has joined #openstack-neutron | 04:41 | |
| *** pasquier-s has quit IRC | 04:41 | |
| *** ramishra_ has joined #openstack-neutron | 04:45 | |
| *** thuc_ has joined #openstack-neutron | 04:45 | |
| *** thuc has quit IRC | 04:45 | |
| *** sphoorti has joined #openstack-neutron | 04:52 | |
| sphoorti | ihrachys: are you around ? | 04:53 |
| *** pasquier-s has joined #openstack-neutron | 04:54 | |
| *** tomoe_ has quit IRC | 04:54 | |
| *** tomoe_ has joined #openstack-neutron | 04:54 | |
| *** yamahata has quit IRC | 04:58 | |
| *** yamahata has joined #openstack-neutron | 04:58 | |
| *** tomoe_ has quit IRC | 04:59 | |
| *** yamahata has quit IRC | 04:59 | |
| *** yamahata has joined #openstack-neutron | 04:59 | |
| openstackgerrit | Mike Kolesnik proposed a change to openstack/neutron: Allow unsharing a network used as gateway https://review.openstack.org/82352 | 05:03 |
| *** thuc has joined #openstack-neutron | 05:03 | |
| *** thuc_ has quit IRC | 05:06 | |
| *** tomoe_ has joined #openstack-neutron | 05:07 | |
| *** thuc has quit IRC | 05:08 | |
| *** yfried has quit IRC | 05:10 | |
| *** matsuhashi has quit IRC | 05:14 | |
| *** pasquier-s has quit IRC | 05:16 | |
| *** matsuhashi has joined #openstack-neutron | 05:18 | |
| *** skraynev_afk is now known as skraynev | 05:19 | |
| *** emagana has joined #openstack-neutron | 05:27 | |
| *** tomoe_ has quit IRC | 05:28 | |
| openstackgerrit | Nachi Ueno proposed a change to openstack/neutron: Improve vif attributes related with firewalling https://review.openstack.org/21946 | 05:30 |
| *** zhipeng has quit IRC | 05:34 | |
| *** dguitarbite has quit IRC | 05:37 | |
| *** dguitarbite has joined #openstack-neutron | 05:41 | |
| *** irenab has joined #openstack-neutron | 05:44 | |
| *** nati_ueno has quit IRC | 05:47 | |
| *** chandan_kumar has quit IRC | 05:48 | |
| *** nati_ueno has joined #openstack-neutron | 05:51 | |
| *** tomoe_ has joined #openstack-neutron | 05:55 | |
| *** otherwiseguy has quit IRC | 05:59 | |
| *** pradipta_away is now known as pradipta | 06:00 | |
| *** Jabadia has joined #openstack-neutron | 06:03 | |
| *** Sukhdev has quit IRC | 06:04 | |
| *** ramishra_ has quit IRC | 06:06 | |
| *** sridhar has quit IRC | 06:11 | |
| *** thuc has joined #openstack-neutron | 06:14 | |
| *** raies has joined #openstack-neutron | 06:15 | |
| *** rotbeard has joined #openstack-neutron | 06:15 | |
| raies | hi | 06:15 |
| raies | I am trying to add firewall create API test in tempest | 06:16 |
| raies | but when I create a firewall it goes in "PENDING_CREATE" status for infinite long | 06:16 |
| raies | I read some help here https://bugs.launchpad.net/neutron/+bug/1223472 | 06:16 |
| raies | I created a router using that tenant with which firewall is being created | 06:17 |
| raies | but still it does not get Active | 06:17 |
| raies | any help on this ?? ^^^ | 06:17 |
| *** mestery has quit IRC | 06:18 | |
| *** thuc has quit IRC | 06:19 | |
| *** spandhe_ has quit IRC | 06:30 | |
| *** yfried has joined #openstack-neutron | 06:31 | |
| *** sungju has quit IRC | 06:32 | |
| openstackgerrit | Jenkins proposed a change to openstack/neutron: Imported Translations from Transifex https://review.openstack.org/82435 | 06:32 |
| SumitNaiksatam | raies: is your l3 agent active? | 06:33 |
| *** djoreilly has joined #openstack-neutron | 06:41 | |
| *** gdubreui has quit IRC | 06:41 | |
| openstackgerrit | Mike Kolesnik proposed a change to openstack/neutron: Allow unsharing a network used as gateway https://review.openstack.org/82352 | 06:42 |
| raies | SumitNaiksatam: Yes I am using devstack and I have set q-l3 in enable services | 06:45 |
| raies | SumitNaiksatam: Actually when I try using netron CLI then status gets ACTIVE. But during implementation in tempest status is PENDING_CREATE | 06:46 |
| raies | SumitNaiksatam: As I read https://bugs.launchpad.net/neutron/+bug/1223472 which tells that one router should be there with associated tenant | 06:46 |
| raies | SumitNaiksatam: I first created a router with that tenant and with same tenant a firewall is created | 06:47 |
| openstackgerrit | shihanzhang proposed a change to openstack/neutron: Prevent dhcp port deletion from the API https://review.openstack.org/81768 | 06:47 |
| *** Longgeek has quit IRC | 06:49 | |
| *** iwamoto has joined #openstack-neutron | 06:50 | |
| *** ramishra has joined #openstack-neutron | 06:51 | |
| *** ramishra_ has joined #openstack-neutron | 06:53 | |
| *** ramishra has quit IRC | 06:54 | |
| *** Longgeek has joined #openstack-neutron | 06:55 | |
| *** ramishra_ has quit IRC | 06:55 | |
| openstackgerrit | Ann Kamyshnikova proposed a change to openstack/neutron: Restore NOT NULL constraint lost by earlier migrations https://review.openstack.org/78591 | 06:57 |
| *** sridhar has joined #openstack-neutron | 06:58 | |
| openstackgerrit | Maru Newby proposed a change to openstack/neutron: Add preliminary support for in-tree api tests https://review.openstack.org/82237 | 07:00 |
| *** jlibosva has joined #openstack-neutron | 07:06 | |
| *** dfarrell07 has quit IRC | 07:06 | |
| *** sphoorti has quit IRC | 07:09 | |
| *** ramishra has joined #openstack-neutron | 07:13 | |
| *** oda-g has quit IRC | 07:15 | |
| openstackgerrit | Ann Kamyshnikova proposed a change to openstack/neutron: Set correct columns' length https://review.openstack.org/80539 | 07:15 |
| raies | SumitNaiksatam: Any answer ?? | 07:15 |
| *** devvesa has joined #openstack-neutron | 07:16 | |
| *** devlaps has quit IRC | 07:17 | |
| *** dvorkinista has joined #openstack-neutron | 07:21 | |
| *** Longgeek has quit IRC | 07:22 | |
| *** saju_m has joined #openstack-neutron | 07:30 | |
| *** Longgeek has joined #openstack-neutron | 07:31 | |
| *** nati_ueno has quit IRC | 07:37 | |
| *** sacharya has joined #openstack-neutron | 07:37 | |
| openstackgerrit | Berezovsky Irena proposed a change to openstack/neutron: Add L2 Agent side handling for non consistent security_group settings https://review.openstack.org/82729 | 07:38 |
| *** luqas has joined #openstack-neutron | 07:45 | |
| *** sungju has joined #openstack-neutron | 07:46 | |
| *** pasquier-s has joined #openstack-neutron | 07:48 | |
| *** rkukura has quit IRC | 07:48 | |
| *** dvorkini_ has joined #openstack-neutron | 07:56 | |
| *** dvorkinista has quit IRC | 07:56 | |
| *** sungju has quit IRC | 08:03 | |
| openstackgerrit | Maru Newby proposed a change to openstack/neutron: Add support for retargetable functional api testing https://review.openstack.org/72585 | 08:06 |
| *** jprovazn has joined #openstack-neutron | 08:10 | |
| *** dvorkini_ has quit IRC | 08:11 | |
| openstackgerrit | enikanorov proposed a change to openstack/neutron: Fix namespace exist() method https://review.openstack.org/81537 | 08:13 |
| *** sphoorti has joined #openstack-neutron | 08:15 | |
| openstackgerrit | Maru Newby proposed a change to openstack/neutron: Add support for retargetable functional api testing https://review.openstack.org/72585 | 08:16 |
| *** sungju has joined #openstack-neutron | 08:20 | |
| *** bashok has joined #openstack-neutron | 08:20 | |
| *** pasquier-s has quit IRC | 08:22 | |
| *** sungju has quit IRC | 08:23 | |
| *** pasquier-s has joined #openstack-neutron | 08:25 | |
| openstackgerrit | Ann Kamyshnikova proposed a change to openstack/neutron: Fix 'server_default' parameter usage in migrations and models https://review.openstack.org/82073 | 08:25 |
| *** bashok_ has joined #openstack-neutron | 08:29 | |
| *** bashok has quit IRC | 08:29 | |
| *** sphoorti has quit IRC | 08:33 | |
| *** jgallard has joined #openstack-neutron | 08:35 | |
| openstackgerrit | Evgeny Fedoruk proposed a change to openstack/neutron: Add 2-leg configuration to Radware LBaaS Driver https://review.openstack.org/69009 | 08:36 |
| openstackgerrit | Evgeny Fedoruk proposed a change to openstack/neutron: Cancelling thread start while unit tests running https://review.openstack.org/81323 | 08:36 |
| *** sphoorti has joined #openstack-neutron | 08:36 | |
| *** sphoorti_ has joined #openstack-neutron | 08:38 | |
| ihrachys | sphoorti: yeah, what's up? | 08:40 |
| *** leseb has joined #openstack-neutron | 08:40 | |
| openstackgerrit | Sphoorti proposed a change to openstack/neutron: Add unit test for add_vxlan in test_linux_ip_lib https://review.openstack.org/80554 | 08:42 |
| *** sphoorti has quit IRC | 08:42 | |
| sphoorti_ | ihrachys: I was having some troubles earlier with tox tests. fixed them Have a look. I have submitted a fresh patch | 08:43 |
| ihrachys | sphoorti_: link? | 08:47 |
| ihrachys | sphoorti_: ah, I see | 08:47 |
| *** matsuhashi has quit IRC | 08:49 | |
| *** jistr has joined #openstack-neutron | 08:49 | |
| *** roeyc has joined #openstack-neutron | 08:50 | |
| *** saju_m has quit IRC | 08:50 | |
| *** roeyc has quit IRC | 08:51 | |
| *** saju_m has joined #openstack-neutron | 08:52 | |
| *** iwamoto has quit IRC | 08:53 | |
| *** jistr is now known as jistr|training | 08:54 | |
| ihrachys | sphoorti_: done | 08:54 |
| *** ygbo has joined #openstack-neutron | 08:55 | |
| *** matsuhashi has joined #openstack-neutron | 08:56 | |
| *** sungju has joined #openstack-neutron | 08:59 | |
| *** sacharya has quit IRC | 09:00 | |
| sphoorti_ | ihrachys: I had got a comment which suggested that I should add docstrings | 09:00 |
| *** jpich has joined #openstack-neutron | 09:02 | |
| sphoorti_ | and about the port numbers, they are passed as two args . have a look https://github.com/openstack/neutron/blob/master/neutron/agent/linux/ip_lib.py#L175 | 09:02 |
| ihrachys | sphoorti_: self-evident docstrings are worthless. If you still think it's a must, you may elaborate on test intention | 09:02 |
| ihrachys | sphoorti_: yes, they are. still, the programmer assumed it should be a tuple of min:max, not a string | 09:02 |
| ihrachys | sphoorti_: as I told before, this may be a bug in add_vxlan too. but we should discuss details in review, not in irc | 09:03 |
| openstackgerrit | Cedric Brandily proposed a change to openstack/neutron: Replace a usage of the deprecated root_helper option https://review.openstack.org/82746 | 09:03 |
| sphoorti_ | ihrachys: about the docstring I think I should remove them, since when I had elaborated them I received comments to modify them | 09:03 |
| *** yfried has quit IRC | 09:04 | |
| sphoorti_ | ihrachys: what do you suggest for the port tuple ? | 09:07 |
| ihrachys | sphoorti_: it's not that you elaborated too much; instead, you could describe the intention, like 'Test add_vxlan function with valid arguments.' But really, I don't even think this is a major issue anyway. | 09:07 |
| *** rkukura has joined #openstack-neutron | 09:07 | |
| ihrachys | sphoorti_: whatever is valid as per add_vlan function? but again, let's discuss details in gerrit. | 09:08 |
| sphoorti_ | ihrachys: sure | 09:09 |
| *** sbalukoff has joined #openstack-neutron | 09:12 | |
| *** chandan_kumar has joined #openstack-neutron | 09:13 | |
| *** matsuhashi has quit IRC | 09:14 | |
| *** safchain has joined #openstack-neutron | 09:15 | |
| *** irenab has quit IRC | 09:17 | |
| *** matsuhashi has joined #openstack-neutron | 09:18 | |
| *** bashok has joined #openstack-neutron | 09:20 | |
| *** bashok_ has quit IRC | 09:20 | |
| *** yfried has joined #openstack-neutron | 09:21 | |
| *** dvorkinista has joined #openstack-neutron | 09:22 | |
| *** pradipta is now known as pradipta_away | 09:24 | |
| *** tomoe_ has quit IRC | 09:24 | |
| *** pradipta_away is now known as pradipta | 09:24 | |
| *** luqas has quit IRC | 09:25 | |
| sphoorti_ | ihrachys: I replied on gerrit. | 09:25 |
| sphoorti_ | however I see them as drafts rather than comments ihrachys | 09:26 |
| ihrachys | sphoorti_: you need to post them first | 09:26 |
| ihrachys | sphoorti_: Review button will do the job | 09:27 |
| openstackgerrit | Oleg Bondarev proposed a change to openstack/neutron: LBaaS: haproxy driver should do undeploy when needed https://review.openstack.org/82749 | 09:27 |
| openstackgerrit | Oleg Bondarev proposed a change to openstack/neutron: LBaaS: make device driver decide whether to deploy instance https://review.openstack.org/82514 | 09:27 |
| *** dvorkinista has quit IRC | 09:27 | |
| sphoorti_ | thanks a lot ihrachys | 09:29 |
| *** luqas has joined #openstack-neutron | 09:29 | |
| *** Longgeek has quit IRC | 09:29 | |
| *** sungju has quit IRC | 09:35 | |
| *** dave_tucker_zzz is now known as dave_tucker | 09:36 | |
| *** saju_m has quit IRC | 09:38 | |
| sphoorti_ | ihrachys: when I declare port as a tuple and pass them as two args, the tox tests work fine. single argument raises AssertionError for passing wrong number of parameteres | 09:38 |
| openstackgerrit | Evgeny Fedoruk proposed a change to openstack/neutron: Add 2-leg configuration to Radware LBaaS Driver https://review.openstack.org/69009 | 09:41 |
| *** rkukura has quit IRC | 09:41 | |
| *** Longgeek has joined #openstack-neutron | 09:43 | |
| *** sungju has joined #openstack-neutron | 09:45 | |
| ihrachys | sphoorti_: that's probably expected, as per add_vxlan code | 09:46 |
| *** bvandenh has quit IRC | 09:47 | |
| sphoorti_ | so should I just resolve the docstring comment ? and also, should the port be a tuple ? ihrachys ? | 09:47 |
| ihrachys | sphoorti_: yes, as per comments in add_vxlan. Have you really checked the function that you're testing? | 09:48 |
| sphoorti_ | yes ihrachys , I posted it in the comments too | 09:49 |
| *** sungju has quit IRC | 09:51 | |
| *** gongysh has quit IRC | 09:55 | |
| *** saju_m has joined #openstack-neutron | 09:56 | |
| openstackgerrit | Sphoorti proposed a change to openstack/neutron: Add unit test for add_vxlan in test_linux_ip_lib https://review.openstack.org/80554 | 09:59 |
| sphoorti_ | ihrachys: made those changes. | 10:00 |
| sphoorti_ | also should checking the port arguments be filed as a bug ? ihrachys ? | 10:01 |
| sphoorti_ | ihrachys: sorry for that one :( | 10:03 |
| ihrachys | sphoorti_: if I were you, I would do it for sure :) and probably I would send a patch for this, this should be easy. [if not isinstance(...): ...] | 10:04 |
| *** pcm__ has joined #openstack-neutron | 10:04 | |
| *** morganfainberg is now known as morganfainberg_Z | 10:05 | |
| openstackgerrit | Berezovsky Irena proposed a change to openstack/neutron: Add L2 Agent side handling for non consistent security_group settings https://review.openstack.org/82729 | 10:06 |
| *** pcm__ has quit IRC | 10:07 | |
| *** pcm__ has joined #openstack-neutron | 10:08 | |
| *** matsuhashi has quit IRC | 10:08 | |
| *** networkstatic has joined #openstack-neutron | 10:09 | |
| *** sungju has joined #openstack-neutron | 10:09 | |
| *** matsuhashi has joined #openstack-neutron | 10:10 | |
| *** yfried has quit IRC | 10:10 | |
| *** bvandenh has joined #openstack-neutron | 10:10 | |
| *** emagana has quit IRC | 10:13 | |
| *** salv-orlando has quit IRC | 10:14 | |
| *** jp_at_hp has joined #openstack-neutron | 10:16 | |
| openstackgerrit | Sphoorti proposed a change to openstack/neutron: Add unit test for add_vxlan in test_linux_ip_lib https://review.openstack.org/80554 | 10:16 |
| sphoorti_ | ihrachys: finally made that last change too. | 10:17 |
| sphoorti_ | also I didnt follow your :- [if not isinstance(...): ...]. | 10:17 |
| ihrachys | sphoorti_: I meant the probable fix for the missing check in add_vxlan() - checking that the port argument is a tuple/list/... | 10:19 |
| *** overlayer has joined #openstack-neutron | 10:21 | |
| sphoorti_ | thank you ihrachys for the +1 and helping me out :) | 10:22 |
| *** xuhanp has quit IRC | 10:23 | |
| sphoorti_ | ihrachys: I have a question though. The parameters of the port tuple could be strings too right ? | 10:24 |
| sphoorti_ | so accessing the port[0] and port[1] is possible. | 10:24 |
| ihrachys | sphoorti_: it's accepted by the current validation code, but I think it was not an intent of the developer who implemented it | 10:24 |
| *** yfried has joined #openstack-neutron | 10:24 | |
| *** sungju has quit IRC | 10:24 | |
| ihrachys | sphoorti_: the problem is that string is also an iterable, with __len__ method, so it passes thru the check. But I guess this is not expected. | 10:25 |
| *** overlayer has quit IRC | 10:26 | |
| *** gdubreui has joined #openstack-neutron | 10:27 | |
| *** overlayer has joined #openstack-neutron | 10:28 | |
| sphoorti_ | ihrachys: true. How can we confirm the expected behavior of the function ? | 10:28 |
| *** overlayer has quit IRC | 10:31 | |
| ihrachys | sphoorti_: what do you mean? | 10:31 |
| *** pradipta is now known as pradipta_away | 10:32 | |
| sphoorti_ | ihrachys: how can the expected behavior of add_vxlan be confirmed? since the current test checks for the add_vxlan function | 10:32 |
| *** overlayer has joined #openstack-neutron | 10:32 | |
| *** sungju has joined #openstack-neutron | 10:33 | |
| ihrachys | sorry, I don't understand your question. your test is designed to check that proper command is passed to the system, right? what else do you need? | 10:33 |
| *** sungju has quit IRC | 10:33 | |
| sphoorti_ | ihrachys: true. sorry I got confused between checking the port argument as tuple/list | 10:35 |
| *** markmcclain has joined #openstack-neutron | 10:39 | |
| *** sungju has joined #openstack-neutron | 10:40 | |
| *** emagana has joined #openstack-neutron | 10:44 | |
| *** Jianyong has quit IRC | 10:49 | |
| *** Longgeek has quit IRC | 10:49 | |
| openstackgerrit | Ann Kamyshnikova proposed a change to openstack/neutron: Set correct nullable parameter for columns https://review.openstack.org/82089 | 10:49 |
| *** sungju has quit IRC | 10:50 | |
| *** emagana has quit IRC | 10:52 | |
| openstackgerrit | Maru Newby proposed a change to openstack/neutron: Refactor plugin setup helpers out of test.base https://review.openstack.org/82775 | 10:58 |
| openstackgerrit | Maru Newby proposed a change to openstack/neutron: Add support for retargetable functional api testing https://review.openstack.org/72585 | 11:01 |
| marun | sphoorti_: there is little value in that type of testing | 11:06 |
| marun | sphoorti_: recommended reading: http://googletesting.blogspot.com/2013/03/testing-on-toilet-testing-state-vs.html | 11:07 |
| *** saju_m has quit IRC | 11:07 | |
| marun | sphoorti_: if you want to do useful testing, consider adding a functional test that actually interacts with the system | 11:07 |
| *** saju_m has joined #openstack-neutron | 11:08 | |
| *** saju_m has quit IRC | 11:09 | |
| marun | sphoorti_: and I'd be happy to assist you in making that happen. | 11:09 |
| *** ramishra has quit IRC | 11:11 | |
| markmcclain | marun: I would argue that the unit test case added is necessary from a UT perspective | 11:11 |
| marun | markmcclain: disagree entirely | 11:11 |
| markmcclain | we also need functional tests too | 11:12 |
| marun | markmcclain: if you can test it in a unittest, great | 11:12 |
| marun | markmcclain: but did you look at the addition? | 11:12 |
| marun | markmcclain: there is zero need to check that a method was called with a given set of arguments | 11:12 |
| *** Longgeek has joined #openstack-neutron | 11:12 | |
| markmcclain | this? https://review.openstack.org/#/c/80554/10/neutron/tests/unit/test_linux_ip_lib.py | 11:12 |
| marun | yes | 11:12 |
| marun | that first test is bs | 11:12 |
| *** mwagner_lap has quit IRC | 11:12 | |
| marun | whenever I see that, I think "someone doesn't know the value of functional testing" | 11:13 |
| markmcclain | verifies that we're constructing the shell correctly | 11:13 |
| marun | it does nothing at all | 11:13 |
| marun | not without a functional test | 11:13 |
| marun | most of our unit testing is like that | 11:13 |
| marun | mock, mock, mock | 11:13 |
| *** gdubreui has quit IRC | 11:13 | |
| markmcclain | so this test validates the logic we have for constructing the system call | 11:13 |
| markmcclain | the functional test validates that the call actually does what we want on the dataplane | 11:14 |
| marun | it duplicates the internal logic of the code under test | 11:14 |
| marun | it's a waste of time | 11:14 |
| marun | if we're wrong, we're just as likely wrong twice | 11:14 |
| marun | double-entry bookkeeping was only useful pre-computer. we're past that | 11:14 |
| marun | you may want to review: http://googletesting.blogspot.com/2013/03/testing-on-toilet-testing-state-vs.html | 11:14 |
| markmcclain | ok then for ip_lib was the proper unit of work boundary? | 11:14 |
| markmcclain | specifically for .add_vxlan()? | 11:15 |
| *** saju_m has joined #openstack-neutron | 11:15 | |
| marun | i'm afraid I don't follow 'proper unit of work boundary'? | 11:15 |
| markmcclain | so unit tests need to test a unit of work | 11:16 |
| markmcclain | is the proper unit not the resulting call to .execute()? | 11:16 |
| marun | unit tests need to test what is reasonable to test in a unit test | 11:16 |
| markmcclain | seems that command construction fits your definition | 11:16 |
| marun | if all the test is going to do is duplicate the logic of the code under test, that's a sign that you have to go up a level | 11:16 |
| marun | validate the actual system interaction | 11:17 |
| marun | if it's possible to test without the duplication, then a unit test is appropriate | 11:17 |
| markmcclain | if we're modifying the dataplane we're no longer a unit test and have moved into functional testing | 11:17 |
| markmcclain | we should be doing both | 11:17 |
| marun | we should be doing both, but they don't have to duplicate each other | 11:17 |
| marun | coverage should be a measure of both functional and unit testing execution | 11:18 |
| marun | test at the lowest level possible, but no lower | 11:18 |
| sphoorti_ | marun: I would like to work on functional tests too. The https://review.openstack.org/#/c/80554/ was my first patch which was a prerequisite for my OPW application. My mentor rossella_ had suggested that I could start with a unit test. Hence the patch | 11:18 |
| *** ramishra has joined #openstack-neutron | 11:19 | |
| marun | sphoorti_: I know you're just following the testing convention already in use. unfortunately it's flawed. | 11:19 |
| *** leseb has quit IRC | 11:19 | |
| markmcclain | oh see I don't see this duplicating the functional test | 11:19 |
| marun | markmcclain: really? | 11:20 |
| markmcclain | the unit test validates the construction of the shell command | 11:20 |
| *** leseb has joined #openstack-neutron | 11:20 | |
| markmcclain | for functional you'll be inspecting the dataplane | 11:20 |
| marun | have you been working on neutron that long? | 11:20 |
| *** xianghui has quit IRC | 11:20 | |
| markmcclain | state and not the system call right? | 11:20 |
| marun | the second test I can see the point of, validating logic in the code | 11:20 |
| markmcclain | so in theory if iputils were to change the command line interface the functional test | 11:20 |
| marun | the first, no | 11:20 |
| markmcclain | would not need to be altered, but the UT would | 11:21 |
| sphoorti_ | but marun , arent unit tests supposed to check the commands are being passed ? | 11:21 |
| marun | I think so, yes. | 11:21 |
| *** rkukura has joined #openstack-neutron | 11:21 | |
| marun | sorry, that was for mark | 11:21 |
| marun | sphoorti_: unit tests are supposed to validate logic, not duplicate it | 11:22 |
| markmcclain | not seeing how this test duplicates logic | 11:22 |
| marun | sphoorti_: as per that google blog post, it's a bad sign if you have to validate interactions | 11:22 |
| markmcclain | if ensures that based on arguments passed the resulting executed is formed in a certain way | 11:22 |
| marun | markmcclain: it duplicates the logic of the call | 11:23 |
| rossella_ | markmcclain: +1 | 11:23 |
| marun | markmcclain: it duplicates exactly the logic of the call | 11:23 |
| marun | markmcclain: have you read that google blog post? | 11:23 |
| marun | rossella_: http://googletesting.blogspot.com/2013/03/testing-on-toilet-testing-state-vs.html | 11:23 |
| marun | this is a classic case of validating interaction | 11:23 |
| marun | because there is no state to validate | 11:23 |
| rossella_ | marun: I am reading it :) | 11:23 |
| *** dave_tucker is now known as dave_tucker_zzz | 11:24 | |
| markmcclain | marun I've read that post before | 11:24 |
| *** leseb has quit IRC | 11:24 | |
| markmcclain | but not sure that fully applies | 11:24 |
| marun | markmcclain: ?! | 11:24 |
| marun | markmcclain: ok, I guess we'll just have to disagree then. | 11:25 |
| markmcclain | in this case .execute does not have a return value | 11:25 |
| marun | markmcclain: I'm not sure how it could be clearer. | 11:25 |
| markmcclain | the system call here is basically throwing the command over the wall and hoping for the best | 11:25 |
| marun | markmcclain: the point is, you can't test the method in question without testing interactions rather than state. | 11:25 |
| marun | markmcclain: so it's a stupid thing to test | 11:25 |
| openstackgerrit | Oleg Bondarev proposed a change to openstack/neutron: LBaaS: haproxy driver should respect vip/pool admin state https://review.openstack.org/82749 | 11:25 |
| marun | we can test state in a functional test, so it should be done there. | 11:25 |
| markmcclain | marun: but you can validate that we construct the command correctly | 11:25 |
| marun | markmcclain: ?! | 11:25 |
| sphoorti_ | marun I dont get how exactly the blog post applies to this patch. | 11:26 |
| marun | forest for the trees man, neutron is so full of this crap. | 11:26 |
| marun | you're validating that a method was called correctly | 11:26 |
| marun | without actually calling the method | 11:26 |
| markmcclain | marun: it is a shell command | 11:26 |
| marun | ffs | 11:26 |
| marun | "Testing interactions means you're verifying that the code under test calls certain methods properly." | 11:26 |
| marun | how is that not what's happening? | 11:26 |
| markmcclain | Unit test should test a unit of work | 11:27 |
| markmcclain | command construction is a unit work | 11:27 |
| marun | -> construct the command correctly | 11:27 |
| marun | *sigh* | 11:27 |
| markmcclain | https://github.com/openstack/neutron/blob/master/neutron/agent/linux/ip_lib.py#L158 | 11:27 |
| marun | ok, I give up. | 11:27 |
| marun | You can keep you unit tests. | 11:27 |
| markmcclain | looking at the test and the actual method | 11:27 |
| markmcclain | I'm failing to see the duplicated logic | 11:27 |
| markmcclain | an argument could made that the call to as root should be checked instead of execute | 11:27 |
| *** dave_tucker_zzz is now known as dave_tucker | 11:28 | |
| marun | sure, whatever | 11:28 |
| markmcclain | and that the return value of add_vxlan() is an IPDevice right? | 11:28 |
| sphoorti_ | yep | 11:28 |
| marun | if you don't think that's not duplication, our respective meanings of 'duplicate' clearly differ and we should just stop | 11:28 |
| marun | the unit tests are full of this crap and I think it's pointless | 11:29 |
| marun | but if you think it's valuable, well, you're PTL. | 11:29 |
| markmcclain | I'm just trying to gain a better understand of where the problem lies | 11:29 |
| markmcclain | unit test should test the add_vxlan() method in isolation right? | 11:29 |
| markmcclain | functional test should examine all the way down the data plane right? | 11:30 |
| *** rotbeard has quit IRC | 11:30 | |
| marun | if possible. but the equivalent functional test does what the unit test does and one better - actual system itneraction | 11:30 |
| marun | ergo, don't bother with the unit test. | 11:30 |
| marun | writing unit tests - or any type of test - in isolation is the devils work | 11:30 |
| marun | considering how to test a feature really should require considering all the different ways it can be done. | 11:31 |
| marun | rather than do everything in unit, then everything in functional then everything in integration... | 11:31 |
| openstackgerrit | Yuriy Taraday proposed a change to openstack/neutron: Abstract out root_helper calls to classes https://review.openstack.org/82787 | 11:31 |
| marun | because we can't. we don't have the resources. | 11:31 |
| rossella_ | marun: so your point is that this kind of unit tests are an useless extra check if we have functional tests that test the same | 11:31 |
| marun | rossella_: correct. | 11:31 |
| marun | rossella_: if I can write a unit test for not much more cost, then it's worth skipping the unit test | 11:32 |
| marun | rossella_: less cost to write, less cost to maintain | 11:32 |
| marun | oops | 11:32 |
| marun | write a unit test -> functional test | 11:32 |
| markmcclain | marun: functional tests have cost that is always greater than unit test | 11:32 |
| rossella_ | yes | 11:33 |
| marun | markmcclain: agreed | 11:33 |
| rossella_ | also ramping up cost is higher | 11:33 |
| marun | but we _need_ the functional test | 11:33 |
| rossella_ | marun: agreed | 11:33 |
| marun | so, unit test < functional test < unit test + functional test | 11:33 |
| rossella_ | :) | 11:33 |
| marun | we get away with the least cost item | 11:33 |
| markmcclain | marun: we agree that we need both, but adding the UT is a good first step | 11:34 |
| marun | in the case of system interaction, functional tests should be the default | 11:34 |
| marun | because we need them | 11:34 |
| rossella_ | do we agree that for somebody that has just started looking into neutron a unit test is an achievable target while a functional test is kind of high? | 11:34 |
| marun | unit tests are often of questionable value for code that does system interaction | 11:34 |
| rossella_ | right now we request a UT for a patch that adds something, would you request a functional test too in future? | 11:34 |
| marun | rossella_: I agree only if the unit test isn't pointless | 11:34 |
| marun | rossella_: the test for the error -> not pointless | 11:35 |
| marun | rossella_: the test that validates construction of the command -> pointless | 11:35 |
| marun | rossella_: (though you wouldn't know it from the existing tests) | 11:35 |
| markmcclain | marun: I think part of the problem is that our unit tests take too long run | 11:35 |
| markmcclain | unit testing should be the first bar | 11:36 |
| marun | markmcclain: uh | 11:36 |
| markmcclain | functional the next one | 11:36 |
| markmcclain | and integration the final hurdle | 11:36 |
| marun | markmcclain: our unit tests take too long because they are all fscking functional | 11:36 |
| markmcclain | that we all agree on | 11:36 |
| marun | markmcclain: and pretending we can keep on writing stupid tests and the problem will go away doesn't really solve anything | 11:36 |
| markmcclain | they are also 70-80% duplicative | 11:36 |
| rossella_ | marun markmcclain : let's start writing the functional test then we can prune the useless UT | 11:37 |
| marun | +1 | 11:37 |
| markmcclain | rossella_: that I agree with | 11:38 |
| rossella_ | :) | 11:38 |
| *** saju_m has quit IRC | 11:38 | |
| marun | rossella_: tests in the tests/functional tree can interact with the system as configured by devstack, and can use sudo | 11:38 |
| markmcclain | I just think we'll end up keeping this test because it does have unit value | 11:38 |
| *** saju_m has joined #openstack-neutron | 11:38 | |
| marun | markmcclain: 'unit value'? | 11:38 |
| marun | is that some kind of negative quantity? :p | 11:38 |
| markmcclain | yes as a valid unit tests it guards against regressions | 11:38 |
| marun | bs | 11:39 |
| marun | regressions from what? | 11:39 |
| markmcclain | so let's say that we find out we're passing the wrong args to iputils | 11:39 |
| markmcclain | we update this test to the proper command | 11:39 |
| markmcclain | and then update ip_lib to construct the command properly | 11:39 |
| *** jistr|training has quit IRC | 11:40 | |
| marun | again, functional test > unit test in this case | 11:40 |
| *** irenab has joined #openstack-neutron | 11:40 | |
| markmcclain | no the bug is in command construction | 11:40 |
| marun | the interaction is all with iputils, so test there | 11:40 |
| marun | i'm starting to hate on mocking | 11:40 |
| markmcclain | mocking is necessary when you start inspecting the data plane you are testing more than one level at once | 11:41 |
| marun | you're using big words like 'data plane' | 11:41 |
| markmcclain | multiple levels are not unit tests | 11:41 |
| *** jistr has joined #openstack-neutron | 11:41 | |
| markmcclain | iputils modifies the data plane does it not? | 11:41 |
| *** jistr is now known as jistr|training | 11:41 | |
| marun | it doesn't if it's mocked out | 11:41 |
| marun | and if all that's being done is constructing a command.... | 11:41 |
| rossella_ | but if you construct it in the wrong way...how long would you need to catch this trivial flaw in a functional test and how long in a stupid, straightforward UT? | 11:43 |
| marun | again, if you think validating trivial command construction is somehow not duplicative, we're simply not on the same page | 11:43 |
| marun | 'trivial flaw'? | 11:43 |
| markmcclain | well the functional test isn't going to care how the command is constructed | 11:43 |
| marun | if there's a flaw, how can you validate it _except_ against the damn command? | 11:43 |
| markmcclain | only that the device is created and configured correctly | 11:44 |
| marun | doing it with mocks is utterly pointless | 11:44 |
| marun | markmcclain: isn't that the only thing that matters? | 11:44 |
| marun | -> results? | 11:44 |
| marun | if there's an error or the results are unexpected, then something is wrong... | 11:44 |
| marun | if things are mocked out, what are we actually testing? | 11:45 |
| rossella_ | marun: in the end yes, results matter. But time matters too...detecting what's wrong if a functional test fails is not so trivial | 11:45 |
| markmcclain | ++ | 11:45 |
| marun | rossella_: what am I missing here? | 11:45 |
| marun | you are both telling me that a flaw in creating the command is easier to detect in a unit test, when the unit test doesn't have any ability to tell you that _something is actually wrong_ | 11:45 |
| marun | it's an echo chamber | 11:46 |
| marun | 'i did it wrong according to my internal model' | 11:46 |
| marun | not 'i did it wrong according to the world' | 11:46 |
| marun | if we were talking about something non-trivial, things would be different | 11:46 |
| marun | but iplib / ovslib are just thin wrappers over cli | 11:46 |
| marun | they don't deserve much in the way of unit testing, they deserve functional testing | 11:46 |
| rossella_ | maybe you are just smarter than me...sometimes I do stupid stuff and then I run a trivial, dummy UT and I discover the problem. If I don't have a stupid UT I have to run a complex functional test and believe me it will take me hours to find out that I am just passing the wrong parameters | 11:47 |
| marun | i *like* unit tests | 11:47 |
| marun | unit tests are awesome | 11:47 |
| marun | but they aren't for everything | 11:47 |
| marun | if I can't understand an error in a functional test, it's really nice to have a unit test to isolate the problem | 11:47 |
| marun | either write it or have it exist | 11:47 |
| marun | but in this case, it doesn't buy us anything. | 11:47 |
| marun | if there's an error in constructing the command, we have just as much ability to look at that in a functional tests | 11:48 |
| marun | test | 11:48 |
| marun | so, in general I completely agree with having test coverage at as low a level as possible to make detection and reproduction of failure easier | 11:49 |
| marun | but sometimes it doesn't make sense, and this is one of those cases. | 11:49 |
| * marun thinks we should just agree to disagree, in any case | 11:49 | |
| rossella_ | marun: I think we agree generally, for this specific case we don't agree | 11:49 |
| marun | rossella_: i've maintained way too much legacy code to want to waste my time and the time of others on pointless testing | 11:50 |
| marun | rossella_: but hey, whatever floats your boat | 11:50 |
| * rossella_ thinks that she will buy a candy to marun just to bribe him | 11:50 | |
| *** leseb has joined #openstack-neutron | 11:50 | |
| *** vbellur has joined #openstack-neutron | 11:51 | |
| *** jgallard has quit IRC | 11:51 | |
| sphoorti_ | marun: markmcclain rossella_ I am a bit lost in the conversation now | 11:52 |
| sphoorti_ | and marun I am eager to learn about functional tests too :) | 11:53 |
| marun | sphoorti_: just ignore my ranting | 11:53 |
| rossella_ | marun: I mean, it's one test, we can agree to disagree. But it's important that we agree on some general principle. When we will have the functional tests and we will prune the current UT we don't want to discuss on any single test | 11:53 |
| marun | i'm fighting for the principle because maintaining tests isn't free | 11:54 |
| marun | most of the unit tests in neutron that validate system interaction are simply pointless | 11:54 |
| marun | but they have a cost nonetheless | 11:55 |
| rossella_ | I know. So maybe we should extend the discussion to the community and write some general rules | 11:55 |
| marun | agree completely... | 11:55 |
| rossella_ | :) | 11:55 |
| marun | https://blueprints.launchpad.net/neutron/+spec/neutron-testing-refactor | 11:55 |
| *** leseb has quit IRC | 11:55 | |
| marun | https://etherpad.openstack.org/p/neutron-testing-refactor | 11:55 |
| marun | the first step was getting a functional job running, which it is | 11:55 |
| rossella_ | great! sorry I missed that blueprint | 11:55 |
| marun | way down that list is developer education | 11:55 |
| marun | i'm working on getting examples in-tree of how to do different types of testing so there'll framework for people to work from | 11:56 |
| rossella_ | marun: great work! | 11:56 |
| marun | if you want to contribute in any way, please don't hesitate to dive in | 11:56 |
| marun | the section at the top is to track individual effort so we don't duplicate | 11:56 |
| rossella_ | I am interested, I will try to contribute | 11:57 |
| marun | ok, I must sleep. i think i've reached my limit on crankiness, thanks for bearing with me. | 11:58 |
| *** dave_tucker is now known as dave_tucker_zzz | 11:58 | |
| marun | rossella_: also: https://review.openstack.org/#/c/72585/ | 11:58 |
| rossella_ | marun: it was a pleasure talking with you...I like people with strong opinions. Get some rest! | 11:59 |
| marun | :) | 11:59 |
| *** networkstatic is now known as networkstatic_zZ | 12:01 | |
| *** skraynev is now known as skraynev_afk | 12:01 | |
| *** baoli has joined #openstack-neutron | 12:02 | |
| *** harlowja_away has quit IRC | 12:03 | |
| *** dave_tucker_zzz is now known as dave_tucker | 12:03 | |
| rossella_ | markmcclain: before this conversation about unit testing started I just wanted to say that sphoorti_ is an applicant of the Outreach Program for Women. She'd like to contribute to Neutron. That UT was her first patch and was required to apply for the internship. Welcome sphoorti_ ! | 12:07 |
| markmcclain | sphoorti_: Welcome! | 12:08 |
| markmcclain | rossella_: awesome | 12:08 |
| anteaya | sphoorti_: you will be lost for a good while, don't take feeling lost as a metric | 12:08 |
| anteaya | sphoorti_: show up everyday and ask questions | 12:08 |
| anteaya | sphoorti_: you will learn more than you think just be doing that | 12:08 |
| sphoorti_ | thank you markmcclain rossella_ anteaya :) | 12:09 |
| sphoorti_ | markmcclain: I have some doubts regarding your comment, would you be around if I come after sometime ? | 12:11 |
| markmcclain | sphoorti_: yes feel free to reach out on IRC | 12:12 |
| sphoorti_ | and anteaya agreed with you. being on this channel did help me understand some concepts regarding unit tests. Initially my unit tests used to fail many a times. But once I followed a conversation on the channel and that helped me solve my doubt. | 12:13 |
| sphoorti_ | :) | 12:13 |
| sphoorti_ | markmcclain: sure :) | 12:13 |
| *** julim has joined #openstack-neutron | 12:13 | |
| *** matsuhashi has quit IRC | 12:16 | |
| *** sphoorti_ has quit IRC | 12:19 | |
| *** jgallard has joined #openstack-neutron | 12:19 | |
| *** matsuhashi has joined #openstack-neutron | 12:23 | |
| *** markmcclain has quit IRC | 12:24 | |
| *** markmcclain has joined #openstack-neutron | 12:24 | |
| *** markmcclain1 has joined #openstack-neutron | 12:25 | |
| *** tomoe_ has joined #openstack-neutron | 12:26 | |
| *** yamahata has quit IRC | 12:27 | |
| *** markmcclain2 has joined #openstack-neutron | 12:27 | |
| *** markmcclain2 has quit IRC | 12:27 | |
| *** markmcclain2 has joined #openstack-neutron | 12:28 | |
| *** markmcclain has quit IRC | 12:29 | |
| *** markmcclain1 has quit IRC | 12:29 | |
| *** markmcclain2 has quit IRC | 12:30 | |
| *** markmcclain has joined #openstack-neutron | 12:31 | |
| *** b3nt_pin has joined #openstack-neutron | 12:31 | |
| *** b3nt_pin is now known as beagles | 12:31 | |
| *** pba2 has quit IRC | 12:32 | |
| *** pba2 has joined #openstack-neutron | 12:32 | |
| *** doude_ has joined #openstack-neutron | 12:33 | |
| *** doude has quit IRC | 12:33 | |
| *** saju_m has quit IRC | 12:41 | |
| *** roeyc has joined #openstack-neutron | 12:42 | |
| *** leseb has joined #openstack-neutron | 12:43 | |
| *** mwagner_lap has joined #openstack-neutron | 12:46 | |
| enikanorov_ | markmcclain: hi. could you please explain negative scoring on https://review.openstack.org/#/c/81537/ ? latest patchset adds bug-id to commit message that we believe is fixed with the patch | 12:47 |
| *** dims_ has quit IRC | 12:52 | |
| *** irenab has quit IRC | 12:52 | |
| *** luqas has quit IRC | 12:52 | |
| *** saju_m has joined #openstack-neutron | 12:55 | |
| *** tomoe_ has quit IRC | 12:55 | |
| *** tomoe_ has joined #openstack-neutron | 12:56 | |
| *** dguitarbite has quit IRC | 12:56 | |
| *** dguitarbite has joined #openstack-neutron | 12:56 | |
| *** thuc has joined #openstack-neutron | 12:57 | |
| *** xuhanp has joined #openstack-neutron | 12:58 | |
| *** thuc_ has joined #openstack-neutron | 12:58 | |
| *** saju_m has quit IRC | 12:59 | |
| baoli | #startmeeting PCI Passthrough | 13:00 |
| openstack | Meeting started Tue Mar 25 13:00:11 2014 UTC and is due to finish in 60 minutes. The chair is baoli. Information about MeetBot at http://wiki.debian.org/MeetBot. | 13:00 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 13:00 |
| openstack | The meeting name has been set to 'pci_passthrough' | 13:00 |
| *** Gil_McGrath has joined #openstack-neutron | 13:00 | |
| *** saju_m has joined #openstack-neutron | 13:00 | |
| baoli | Hi | 13:00 |
| *** vbellur has quit IRC | 13:01 | |
| *** thuc has quit IRC | 13:01 | |
| *** heyongli has joined #openstack-neutron | 13:02 | |
| *** itzikb has joined #openstack-neutron | 13:02 | |
| *** dims_ has joined #openstack-neutron | 13:02 | |
| enikanorov_ | baoli: hi | 13:02 |
| enikanorov_ | it's openstack-neutron, are you sure the meeting is going to be here? | 13:03 |
| *** matsuhashi has quit IRC | 13:03 | |
| baoli | enikanorov_, we have weekly meeting at utc 1300 wednesday | 13:04 |
| enikanorov_ | at which channel? | 13:04 |
| baoli | sorry, wrong channel | 13:04 |
| enikanorov_ | :) | 13:04 |
| baoli | my bad | 13:04 |
| *** saju_m has quit IRC | 13:04 | |
| anteaya | baoli: can you end the meeting please? | 13:04 |
| *** saju_m has joined #openstack-neutron | 13:05 | |
| baoli | #endmeeting | 13:05 |
| openstack | Meeting ended Tue Mar 25 13:05:24 2014 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 13:05 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/pci_passthrough/2014/pci_passthrough.2014-03-25-13.00.html | 13:05 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/pci_passthrough/2014/pci_passthrough.2014-03-25-13.00.txt | 13:05 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/pci_passthrough/2014/pci_passthrough.2014-03-25-13.00.log.html | 13:05 |
| baoli | Sorry, guys. A little rusty in the morning | 13:05 |
| anteaya | baoli: np | 13:05 |
| *** heyongli has quit IRC | 13:07 | |
| *** heyongli has joined #openstack-neutron | 13:07 | |
| xuhanp | safchain, ping | 13:07 |
| safchain | xuhanp, pong | 13:08 |
| xuhanp | I have a question about L3 HA if you have a minute | 13:08 |
| *** irenab has joined #openstack-neutron | 13:08 | |
| safchain | xuhanp, sure | 13:08 |
| xuhanp | I wanted to try your patch out today and I get stuck with a problem | 13:08 |
| xuhanp | so I created a route today and I haven't attached any subnets to it yet. | 13:08 |
| xuhanp | I can find the HA ports created on both my 2 L3 agent nodes | 13:09 |
| *** saju_m has quit IRC | 13:09 | |
| xuhanp | but I am having trouble to ping from one port's IP to the other from their namespace | 13:09 |
| xuhanp | I was worried about keepalived won't work under this situation. | 13:10 |
| *** saju_m has joined #openstack-neutron | 13:10 | |
| xuhanp | Have you find this problem before? any clue? | 13:10 |
| safchain | xuhanp, no, but I can try to redo the same test | 13:12 |
| xuhanp | I am using GRE so it may related to my GRE setup problem. | 13:12 |
| safchain | xuhanp, yes could you test the GRE setup | 13:12 |
| xuhanp | safchain, any clue about how to test GRE? | 13:13 |
| *** aveiga has joined #openstack-neutron | 13:13 | |
| safchain | xuhanp, maybe without the HA patches, test to spawn 2 routers and try to do the same test | 13:14 |
| safchain | xuhanp, maybe between the dhcp namespace and a router namespace | 13:14 |
| xuhanp | you mean ping from the qr namespace to the qr namespace on the other node? | 13:15 |
| xuhanp | and do the same for dhcp namespace? | 13:15 |
| *** ChanServ changes topic to "the gerrit event stream is currently hung, blocking all testing. troubleshooting is in progress (next update at 14:00 utc)" | 13:21 | |
| *** changbl has quit IRC | 13:22 | |
| *** devvesa has quit IRC | 13:27 | |
| *** mestery has joined #openstack-neutron | 13:28 | |
| *** mestery has quit IRC | 13:28 | |
| *** mestery has joined #openstack-neutron | 13:29 | |
| *** ChanServ changes topic to "Discussion of OpenStack Networking || for support join #openstack" | 13:30 | |
| *** ramishra has quit IRC | 13:30 | |
| *** banix has joined #openstack-neutron | 13:31 | |
| *** markmcclain has quit IRC | 13:34 | |
| *** yamahata has joined #openstack-neutron | 13:38 | |
| *** nati_ueno has joined #openstack-neutron | 13:40 | |
| *** thuc_ has quit IRC | 13:41 | |
| *** luqas has joined #openstack-neutron | 13:42 | |
| *** thuc has joined #openstack-neutron | 13:42 | |
| *** thuc has quit IRC | 13:47 | |
| openstackgerrit | Nachi Ueno proposed a change to openstack/neutron: Improve vif attributes related with firewalling https://review.openstack.org/21946 | 13:49 |
| *** Gil_McGrath has quit IRC | 13:51 | |
| anteaya | I'm going to test something with meetbot/openstack, please ignore me | 13:52 |
| anteaya | #startmeeting test_lurk | 13:52 |
| openstack | Meeting started Tue Mar 25 13:52:22 2014 UTC and is due to finish in 60 minutes. The chair is anteaya. Information about MeetBot at http://wiki.debian.org/MeetBot. | 13:52 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 13:52 |
| *** jecarey has joined #openstack-neutron | 13:52 | |
| openstack | The meeting name has been set to 'test_lurk' | 13:52 |
| anteaya | #endmeeting | 13:52 |
| openstack | Meeting ended Tue Mar 25 13:52:33 2014 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 13:52 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/test_lurk/2014/test_lurk.2014-03-25-13.52.html | 13:52 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/test_lurk/2014/test_lurk.2014-03-25-13.52.txt | 13:52 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/test_lurk/2014/test_lurk.2014-03-25-13.52.log.html | 13:52 |
| anteaya | #lurk | 13:52 |
| anteaya | #startmeeting test_lurk | 13:53 |
| openstack | Meeting started Tue Mar 25 13:53:21 2014 UTC and is due to finish in 60 minutes. The chair is anteaya. Information about MeetBot at http://wiki.debian.org/MeetBot. | 13:53 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 13:53 |
| openstack | The meeting name has been set to 'test_lurk' | 13:53 |
| anteaya | #endmeeting | 13:53 |
| openstack | Meeting ended Tue Mar 25 13:53:33 2014 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 13:53 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/test_lurk/2014/test_lurk.2014-03-25-13.53.html | 13:53 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/test_lurk/2014/test_lurk.2014-03-25-13.53.txt | 13:53 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/test_lurk/2014/test_lurk.2014-03-25-13.53.log.html | 13:53 |
| anteaya | #unlurk | 13:53 |
| *** dguitarbite has quit IRC | 13:54 | |
| *** nati_ueno has quit IRC | 13:59 | |
| *** tomoe_ has quit IRC | 14:00 | |
| *** ramishra has joined #openstack-neutron | 14:00 | |
| *** tomoe_ has joined #openstack-neutron | 14:00 | |
| *** ramishra has quit IRC | 14:02 | |
| *** heyongli has quit IRC | 14:02 | |
| *** ramishra has joined #openstack-neutron | 14:04 | |
| *** peristeri has joined #openstack-neutron | 14:04 | |
| *** tomoe_ has quit IRC | 14:05 | |
| *** dvorkinista has joined #openstack-neutron | 14:06 | |
| *** nplanel has quit IRC | 14:14 | |
| *** dvorkinista has quit IRC | 14:19 | |
| *** flwang has joined #openstack-neutron | 14:20 | |
| openstackgerrit | Armando Migliaccio proposed a change to openstack/python-neutronclient: Show the unknown auth stratey in neutron client https://review.openstack.org/82834 | 14:20 |
| openstackgerrit | Oleg Bondarev proposed a change to openstack/neutron: Delete disassociated floating ips on external network deletion https://review.openstack.org/53364 | 14:21 |
| *** mestery has quit IRC | 14:24 | |
| *** dvorkinista has joined #openstack-neutron | 14:26 | |
| *** alexpilotti has quit IRC | 14:28 | |
| *** dvorkinista has quit IRC | 14:29 | |
| *** tomoe_ has joined #openstack-neutron | 14:29 | |
| *** nplanel has joined #openstack-neutron | 14:30 | |
| openstackgerrit | Yuriy Taraday proposed a change to openstack/neutron: Abstract out root_helper calls to classes https://review.openstack.org/82787 | 14:30 |
| *** markmcclain has joined #openstack-neutron | 14:32 | |
| *** markmcclain has quit IRC | 14:32 | |
| *** mestery has joined #openstack-neutron | 14:33 | |
| *** markmcclain has joined #openstack-neutron | 14:33 | |
| enikanorov_ | markmcclain: ping | 14:35 |
| pcm__ | ANYONE: I'm getting a Jenkins fail on py2.6, but don't see to see the failure in log. Tried to run on my machine, but have issue w/tox saying zlib import fail. Are there instructions on setup of tox for py2.6? | 14:37 |
| *** otherwiseguy has joined #openstack-neutron | 14:37 | |
| *** irenab has quit IRC | 14:38 | |
| *** leseb has quit IRC | 14:38 | |
| *** dvorkinista has joined #openstack-neutron | 14:42 | |
| openstackgerrit | Roey Chen proposed a change to openstack/neutron: Fixed TypeError when creating MlnxException https://review.openstack.org/82837 | 14:44 |
| xuhanp | amotoki, ping | 14:46 |
| enikanorov_ | pcm__: pls give a link to jankins failure | 14:47 |
| enikanorov_ | *e | 14:48 |
| *** dvorkinista has quit IRC | 14:48 | |
| *** armax has joined #openstack-neutron | 14:49 | |
| enikanorov_ | pcm__: if you're talking about this one https://review.openstack.org/#/c/82306/ | 14:50 |
| enikanorov_ | pcm__: py26 job exceeded 1 hour in duration | 14:50 |
| enikanorov_ | nothing specific to fix. it happens from time to time | 14:50 |
| *** BillTheKat has joined #openstack-neutron | 14:52 | |
| markmcclain | enikanorov_: updated the review. The problem is the linked bugs and the actual change do not agree. | 14:53 |
| *** BillTheKat has quit IRC | 14:53 | |
| enikanorov_ | markmcclain: i don't quite understand why so. the change gets rid of running namespace listing with a root and within the namespace. also, we found out that the issue (existing behaviour causes exception on lbaas agent and respawning the haproxy) is responsible for scenario test failure | 14:54 |
| *** thedodd has joined #openstack-neutron | 14:54 | |
| markmcclain | enikanorov_: because first bug only lowers the priveledges | 14:55 |
| enikanorov_ | markmcclain: but you can't do 'ip netns exec namespace' without root, so you need to get rid of both sudo and 'exec' | 14:55 |
| *** thuc_ has joined #openstack-neutron | 14:57 | |
| *** tvardeman has joined #openstack-neutron | 14:58 | |
| *** sacharya has joined #openstack-neutron | 14:58 | |
| *** baoli has quit IRC | 14:59 | |
| openstackgerrit | rcurran proposed a change to openstack/neutron: ML2 Cisco Nexus MD: Remove workaround for bug 1276395 https://review.openstack.org/82839 | 15:01 |
| *** carl_baldwin has joined #openstack-neutron | 15:02 | |
| *** luqas has quit IRC | 15:03 | |
| *** aveiga has quit IRC | 15:03 | |
| enikanorov_ | markmcclain: so would you suggest to file a new bug for that change? or change the commit message? | 15:03 |
| *** networkstatic_zZ has quit IRC | 15:03 | |
| *** devlaps has joined #openstack-neutron | 15:04 | |
| *** jobewan has joined #openstack-neutron | 15:05 | |
| *** baoli has joined #openstack-neutron | 15:05 | |
| *** leseb has joined #openstack-neutron | 15:05 | |
| *** ygbo has quit IRC | 15:06 | |
| *** carlp has joined #openstack-neutron | 15:07 | |
| pcm__ | enikanorov_: Thanks. I couldn't figure out what the issue was, seemed like a time out. Ran again and got same result. | 15:09 |
| pcm__ | enikanorov_: Do you know how to setup environment to run tox with py26? | 15:09 |
| openstackgerrit | Miguel Angel Ajo proposed a change to openstack/neutron: fixes broken neutron-netns-cleanup https://review.openstack.org/80261 | 15:10 |
| *** ygbo has joined #openstack-neutron | 15:11 | |
| *** saju_m has quit IRC | 15:13 | |
| *** humbolt1 has joined #openstack-neutron | 15:14 | |
| *** networkstatic has joined #openstack-neutron | 15:14 | |
| *** humbolt1 has quit IRC | 15:14 | |
| *** packet has joined #openstack-neutron | 15:14 | |
| *** humbolt has quit IRC | 15:14 | |
| *** nplanel has quit IRC | 15:14 | |
| *** humbolt has joined #openstack-neutron | 15:15 | |
| *** humbolt has quit IRC | 15:15 | |
| *** humbolt has joined #openstack-neutron | 15:16 | |
| *** humbolt has quit IRC | 15:16 | |
| *** nplanel has joined #openstack-neutron | 15:17 | |
| *** humbolt has joined #openstack-neutron | 15:17 | |
| *** humbolt has quit IRC | 15:17 | |
| *** pcm__ has quit IRC | 15:18 | |
| *** tvardeman has quit IRC | 15:18 | |
| *** humbolt has joined #openstack-neutron | 15:18 | |
| *** humbolt has quit IRC | 15:18 | |
| *** pcm_ has joined #openstack-neutron | 15:19 | |
| *** humbolt has joined #openstack-neutron | 15:19 | |
| *** humbolt has quit IRC | 15:19 | |
| *** humbolt has joined #openstack-neutron | 15:20 | |
| *** humbolt has quit IRC | 15:20 | |
| *** changbl has joined #openstack-neutron | 15:20 | |
| *** pcm___ has joined #openstack-neutron | 15:20 | |
| *** humbolt has joined #openstack-neutron | 15:21 | |
| *** xuhanp has quit IRC | 15:21 | |
| *** humbolt has quit IRC | 15:21 | |
| *** humbolt has joined #openstack-neutron | 15:21 | |
| *** devvesa has joined #openstack-neutron | 15:21 | |
| *** humbolt has quit IRC | 15:21 | |
| *** tvardeman has joined #openstack-neutron | 15:22 | |
| *** humbolt has joined #openstack-neutron | 15:22 | |
| anteaya | humbolt: please check your client | 15:22 |
| *** humbolt has quit IRC | 15:22 | |
| *** rkukura has quit IRC | 15:23 | |
| *** humbolt has joined #openstack-neutron | 15:23 | |
| *** humbolt has quit IRC | 15:23 | |
| *** pcm_ has quit IRC | 15:23 | |
| *** humbolt has joined #openstack-neutron | 15:24 | |
| *** humbolt has quit IRC | 15:24 | |
| *** mestery has quit IRC | 15:24 | |
| jaypipes | marun: ready for a review on https://review.openstack.org/#/c/72585/ yet? (I know you've been pushing a lot of patchsets on that, just want to make sure you're ready for a review :) | 15:25 |
| *** humbolt has joined #openstack-neutron | 15:25 | |
| *** humbolt has quit IRC | 15:25 | |
| *** humbolt has joined #openstack-neutron | 15:26 | |
| *** humbolt has quit IRC | 15:26 | |
| *** humbolt has joined #openstack-neutron | 15:27 | |
| *** humbolt has quit IRC | 15:27 | |
| openstackgerrit | Yuriy Taraday proposed a change to openstack/neutron: Abstract out root_helper calls to classes https://review.openstack.org/82787 | 15:27 |
| *** humbolt has joined #openstack-neutron | 15:29 | |
| *** humbolt has quit IRC | 15:29 | |
| *** humbolt has joined #openstack-neutron | 15:30 | |
| *** dave_tucker is now known as dave_tucker_zzz | 15:30 | |
| *** otherwiseguy has quit IRC | 15:32 | |
| *** itzikb has quit IRC | 15:35 | |
| *** emagana has joined #openstack-neutron | 15:38 | |
| *** feleouet has quit IRC | 15:38 | |
| *** krtaylor has quit IRC | 15:42 | |
| *** BillTheKat has joined #openstack-neutron | 15:42 | |
| *** rossella_ has quit IRC | 15:44 | |
| *** alexpilotti has joined #openstack-neutron | 15:44 | |
| *** rossella_ has joined #openstack-neutron | 15:45 | |
| *** bashok has quit IRC | 15:45 | |
| *** alexpilotti has quit IRC | 15:46 | |
| *** markwash has joined #openstack-neutron | 15:46 | |
| ihrachys | anyone with stable +2 rights? I have some patches that I would like to see in stable branch before the next release cut off? | 15:47 |
| ihrachys | s/\?$// | 15:47 |
| markmcclain | ihrachys: stable gating for Neutron is not stable | 15:47 |
| *** SumitNaiksatam has quit IRC | 15:47 | |
| markmcclain | +2ing anything is likely to result in a gate failure | 15:47 |
| *** alagalah has joined #openstack-neutron | 15:47 | |
| enikanorov__ | ^^ sounds funny :) | 15:48 |
| *** _cjones_ has joined #openstack-neutron | 15:48 | |
| *** _cjones__ has joined #openstack-neutron | 15:49 | |
| ihrachys | markmcclain: hm, I didn't know that. The last time I've asked apevec about stable gate state, he said everything is fine now. :| | 15:49 |
| markmcclain | the kernel bug still shows up in the gate | 15:50 |
| ihrachys | markmcclain: is it neutron only? | 15:50 |
| *** krtaylor has joined #openstack-neutron | 15:50 | |
| markmcclain | so anything that is approved for Neutron has decent chance of hitting the bug | 15:50 |
| *** thuc_ has quit IRC | 15:51 | |
| *** luqas has joined #openstack-neutron | 15:51 | |
| *** thuc has joined #openstack-neutron | 15:51 | |
| *** _cjones_ has quit IRC | 15:52 | |
| *** alagalah has quit IRC | 15:52 | |
| openstackgerrit | Sean M. Collins proposed a change to openstack/neutron: QoS API and DB models https://review.openstack.org/28313 | 15:52 |
| openstackgerrit | Sean M. Collins proposed a change to openstack/neutron: Quality of Service API extension - RPC & Driver support https://review.openstack.org/59970 | 15:52 |
| ihrachys | markmcclain: is there any list of known issues that I can start to work on? | 15:53 |
| markmcclain | so have to work altering the stable tests | 15:54 |
| markmcclain | to disable key injection | 15:54 |
| *** apevec has joined #openstack-neutron | 15:56 | |
| *** thuc has quit IRC | 15:56 | |
| ihrachys | apevec: markmcclain claims there are still issues with stable gating, related to key injection triggering kernel oops | 15:56 |
| apevec | markmcclain, hi - ihrachys says there are still stable neutron gate issues, what are examples? | 15:56 |
| *** amotoki_ has joined #openstack-neutron | 15:56 | |
| apevec | wasn't key injection disabled or somehting? | 15:56 |
| markmcclain | apevec: only in master | 15:56 |
| markmcclain | still running for some stable tests | 15:57 |
| apevec | damn | 15:57 |
| *** packet has quit IRC | 15:57 | |
| apevec | what prevents us to do it on stable/havana too? | 15:57 |
| *** yfried has quit IRC | 15:57 | |
| *** devvesa has quit IRC | 15:58 | |
| markmcclain | apevec: really just have to find the test cases where it is still being activated | 15:58 |
| markmcclain | I thought I had tracked them all down, but was noticing failures still | 15:58 |
| *** dvorkinista has joined #openstack-neutron | 15:59 | |
| *** alagalah has joined #openstack-neutron | 15:59 | |
| *** packet has joined #openstack-neutron | 15:59 | |
| *** mestery has joined #openstack-neutron | 15:59 | |
| openstackgerrit | Yuriy Taraday proposed a change to openstack/neutron: Abstract out root_helper calls to classes https://review.openstack.org/82787 | 16:01 |
| *** alagalah has quit IRC | 16:01 | |
| *** Sukhdev has joined #openstack-neutron | 16:02 | |
| *** devvesa has joined #openstack-neutron | 16:04 | |
| *** SumitNaiksatam has joined #openstack-neutron | 16:04 | |
| *** dvorkinista has quit IRC | 16:04 | |
| *** Sukhdev has quit IRC | 16:06 | |
| apevec | markmcclain, how can I help tracking those down, do you have an example what to look for in tempest? | 16:07 |
| apevec | we're supposed to freeze stable/havana for 2013.2.3 this Thursday | 16:08 |
| *** baoli has quit IRC | 16:08 | |
| markmcclain | apevec: it is not a tempest thing | 16:08 |
| markmcclain | the error is in how the devstack instance is configured | 16:09 |
| markmcclain | I'm trying to find an example review | 16:09 |
| *** mestery has quit IRC | 16:09 | |
| *** SumitNaiksatam has quit IRC | 16:10 | |
| *** mestery has joined #openstack-neutron | 16:10 | |
| *** SumitNaiksatam has joined #openstack-neutron | 16:10 | |
| *** baoli has joined #openstack-neutron | 16:10 | |
| *** Jabadia has quit IRC | 16:12 | |
| *** Jabadia has joined #openstack-neutron | 16:12 | |
| *** dave_tucker_zzz is now known as dave_tucker | 16:13 | |
| *** skraynev_afk is now known as skraynev | 16:14 | |
| *** rkukura has joined #openstack-neutron | 16:16 | |
| *** nati_ueno has joined #openstack-neutron | 16:17 | |
| *** Jabadia has quit IRC | 16:17 | |
| *** blogan has joined #openstack-neutron | 16:19 | |
| *** markwash has quit IRC | 16:20 | |
| openstackgerrit | Yuriy Taraday proposed a change to openstack/neutron: Abstract out root_helper calls to classes https://review.openstack.org/82787 | 16:22 |
| *** markwash has joined #openstack-neutron | 16:23 | |
| openstackgerrit | Aaron Rosen proposed a change to openstack/neutron: Who's CI is lying! :) https://review.openstack.org/75304 | 16:24 |
| *** spandhe has joined #openstack-neutron | 16:25 | |
| *** thuc has joined #openstack-neutron | 16:25 | |
| *** thuc has quit IRC | 16:26 | |
| *** thuc has joined #openstack-neutron | 16:26 | |
| *** alagalah has joined #openstack-neutron | 16:27 | |
| *** nati_ueno has quit IRC | 16:30 | |
| *** catohornet__ has joined #openstack-neutron | 16:31 | |
| *** alagalah has quit IRC | 16:34 | |
| *** Longgeek has quit IRC | 16:35 | |
| *** Longgeek has joined #openstack-neutron | 16:35 | |
| *** tomoe_ has quit IRC | 16:38 | |
| _cjones__ | kevinbenton, You around today? | 16:38 |
| *** Longgeek has quit IRC | 16:42 | |
| *** spandhe has quit IRC | 16:43 | |
| markmcclain | apevec: http://logstash.openstack.org/#eyJzZWFyY2giOiJcIkZpeGluZyByZWN1cnNpdmUgZmF1bHQgYnV0IHJlYm9vdCBpcyBuZWVkZWRcIiBBTkQgYnVpbGRfYnJhbmNoOlwic3RhYmxlL2hhdmFuYVwiIEFORCBidWlsZF9xdWV1ZTpcImdhdGVcIiIsImZpZWxkcyI6W10sIm9mZnNldCI6MCwidGltZWZyYW1lIjoiYWxsIiwiZ3JhcGhtb2RlIjoiY291bnQiLCJ0aW1lIjp7InVzZXJfaW50ZXJ2YWwiOjB9LCJtb2RlIjoiIiwiYW5hbHl6ZV9maWVsZCI6IiIsInN0YW1wIjoxMzk1NzY1ODk0MzEyfQ== | 16:45 |
| *** markwash_ has joined #openstack-neutron | 16:48 | |
| *** markwash has quit IRC | 16:49 | |
| *** markwash_ is now known as markwash | 16:49 | |
| openstackgerrit | Yuriy Taraday proposed a change to openstack/neutron: Abstract out root_helper calls to classes https://review.openstack.org/82787 | 16:49 |
| kevinbenton | _cjones_: yep | 16:53 |
| *** rossella_s has joined #openstack-neutron | 16:54 | |
| *** spandhe has joined #openstack-neutron | 16:57 | |
| *** jistr|training has quit IRC | 16:58 | |
| *** dave_tucker is now known as dave_tucker_zzz | 17:00 | |
| *** markmcclain1 has joined #openstack-neutron | 17:01 | |
| *** markmcclain1 has quit IRC | 17:01 | |
| *** tomoe_ has joined #openstack-neutron | 17:01 | |
| *** markmcclain1 has joined #openstack-neutron | 17:02 | |
| *** markmcclain has quit IRC | 17:02 | |
| *** nati_ueno has joined #openstack-neutron | 17:02 | |
| *** roeyc has quit IRC | 17:02 | |
| *** harlowja has joined #openstack-neutron | 17:03 | |
| *** rossella_ has quit IRC | 17:04 | |
| *** morganfainberg_Z is now known as morganfainberg | 17:04 | |
| *** _cjones__ has quit IRC | 17:05 | |
| *** sacharya has quit IRC | 17:05 | |
| *** _cjones_ has joined #openstack-neutron | 17:05 | |
| *** zzelle has joined #openstack-neutron | 17:06 | |
| *** claudiub has joined #openstack-neutron | 17:06 | |
| _cjones_ | kevinbenton: I think I worked around my tox issues. HFS+ on Mac seemed to be the cause of a bunch of issues. Out of curiousity how do you go abouts debugging your unit tests? Is there a standard procedure for this in Neutron (ML2) | 17:07 |
| zzelle | markmcclain1, hi | 17:07 |
| markmcclain1 | zzelle: hi | 17:10 |
| *** manishg has joined #openstack-neutron | 17:10 | |
| *** ramishra has quit IRC | 17:11 | |
| *** safchain has quit IRC | 17:11 | |
| *** peristeri has quit IRC | 17:11 | |
| zzelle | markmcclain1, i corrected a small bug about L3 agent using deprecated root_helper option, i think it could be added to the RC | 17:11 |
| zzelle | markmcclain1, https://review.openstack.org/82746/ | 17:12 |
| zzelle | markmcclain1, without the correction extraroute could be not applied | 17:12 |
| *** beagles is now known as beagles_brb | 17:12 | |
| apevec | markmcclain1, so this is bug 1273386 which is worked around in master by switching to config drive but stable/havana devstack is missing support for that ? | 17:13 |
| *** ramishra has joined #openstack-neutron | 17:13 | |
| openstackgerrit | Kevin Benton proposed a change to openstack/neutron: (WIP) BigSwitch: hybrid routing model for plugin https://review.openstack.org/82873 | 17:14 |
| *** ygbo has quit IRC | 17:15 | |
| kevinbenton | _cjones_: is it to the point where all of the tests seem to run now? | 17:15 |
| apevec | devstack-gate (which has master only) sets FORCE_CONFIG_DRIVE by default but that envvar is not present in devstack stable/havana, so ignored there | 17:15 |
| _cjones_ | kevinbenton: Have not tried all no. Is that worth a road going down? | 17:15 |
| kevinbenton | _cjones_: probably not if you already have one unit test that fails. I start by running the unit test module that fails. for example. tox -v -epy27 tests.unit.ml2.drivers.test_bigswitch_mech will run the tests in neutron/tests/unit/ml2/drivers/test_bigswitch_mech.py | 17:17 |
| apevec | markmcclain1, I'll propose https://review.openstack.org/54746 for devstack stable/havana | 17:18 |
| kevinbenton | _cjones_: once you get failures, you should get some useful information about the failure that will tell you what your change did wrong or that the unit tests will require an update to match your change | 17:18 |
| *** leseb has quit IRC | 17:21 | |
| *** catohornet__ has quit IRC | 17:24 | |
| *** zzelle has quit IRC | 17:25 | |
| _cjones_ | kevinbenton: Fair enough. I'll ensure the others are passing. I think the main issue is in the tests for the driver I wrote which is why I was hoping for a debug procedure. | 17:28 |
| *** amotoki_ has quit IRC | 17:29 | |
| *** salv-orlando has joined #openstack-neutron | 17:30 | |
| openstackgerrit | Sean M. Collins proposed a change to openstack/neutron: Ml2 QoS API extension support https://review.openstack.org/59971 | 17:31 |
| *** skraynev is now known as skraynev_afk | 17:32 | |
| *** jprovazn has quit IRC | 17:32 | |
| openstackgerrit | Maru Newby proposed a change to openstack/neutron: Refactor plugin setup helpers out of test.base https://review.openstack.org/82775 | 17:34 |
| *** Sukhdev has joined #openstack-neutron | 17:35 | |
| _cjones_ | kevinbenton: Here's the output when running tox -v -epy27 tests.unit.ml2: | 17:35 |
| *** tomoe_ has quit IRC | 17:35 | |
| _cjones_ | https://gist.github.com/cjones-/bcc6fab369861b12c00a | 17:36 |
| _cjones_ | Not really much to determine why I got those 2 failures. | 17:36 |
| *** markmcclain1 has quit IRC | 17:36 | |
| *** spandhe has quit IRC | 17:37 | |
| *** devvesa has quit IRC | 17:39 | |
| openstackgerrit | Kevin Benton proposed a change to openstack/neutron: Disable XML tests on Py26 https://review.openstack.org/81865 | 17:41 |
| *** devvesa has joined #openstack-neutron | 17:41 | |
| *** alagalah has joined #openstack-neutron | 17:41 | |
| kevinbenton | _cjones_: yeah, that's not very helpful. anything useful in .testrepository/failing ? | 17:44 |
| claudiub | hello guys, could you take a look at a backport bug fix? Already has plenty of +1s. https://review.openstack.org/#/c/63775/ | 17:44 |
| _cjones_ | kevinbenton: Yup. https://gist.github.com/cjones-/f3fac6308462babf2b98 | 17:45 |
| _cjones_ | Pretty sure that has nothing to do with me, but I'm trying again w/o any of my changes. | 17:45 |
| *** alagalah has quit IRC | 17:45 | |
| kevinbenton | _cjones_: another thing you can do is see if that test fails by itself. tox -v -epy27 neutron.tests.unit.ml2.drivers.test_cisco_mech.TestCiscoSubnetsV2.test_create_subnet_with_two_host_routes | 17:46 |
| *** markmcclain has joined #openstack-neutron | 17:46 | |
| _cjones_ | Good call | 17:46 |
| *** SumitNaiksatam has quit IRC | 17:47 | |
| _cjones_ | Hah. Passed! | 17:47 |
| _cjones_ | :| That's a little disconcerting. | 17:48 |
| *** mestery_ has joined #openstack-neutron | 17:48 | |
| markmcclain | apevec: yes need to enable config drive or metadata for test | 17:48 |
| *** mestery has quit IRC | 17:48 | |
| markmcclain | zzelle: approved | 17:48 |
| *** spandhe has joined #openstack-neutron | 17:49 | |
| kevinbenton | _cjones_: maybe try a re-run. the 'killed' part seems like something intervened but I'm not sure why | 17:49 |
| *** SumitNaiksatam has joined #openstack-neutron | 17:50 | |
| *** mestery_ is now known as mestery | 17:50 | |
| *** armitage81 has joined #openstack-neutron | 17:50 | |
| _cjones_ | kevinbenton: I'm wondering if my vagrant/VBox combo is too slow and something is timing out causing subsequent failures. | 17:50 |
| *** armitage81 has quit IRC | 17:51 | |
| *** armitage81 has joined #openstack-neutron | 17:51 | |
| kevinbenton | _cjones_: which os are the UTs running on? | 17:51 |
| *** mlavalle has joined #openstack-neutron | 17:51 | |
| _cjones_ | kevinbenton: Ubuntu Server 12.04 | 17:52 |
| *** sphoorti has joined #openstack-neutron | 17:52 | |
| *** spandhe has quit IRC | 17:52 | |
| *** spandhe has joined #openstack-neutron | 17:54 | |
| *** jgallard has quit IRC | 17:56 | |
| kevinbenton | _cjones_: same set of tests passed in my dev VM in 240 seconds, so I don't think that's the problem | 17:56 |
| kevinbenton | _cjones_: have the unit tests in that folder always been run from the same VM/OS? If so, it might be worth wiping out the .tox and .testrepository directories so it starts out clean | 17:58 |
| *** markmcclain has quit IRC | 17:58 | |
| openstackgerrit | Yuriy Taraday proposed a change to openstack/neutron: Abstract out root_helper calls to classes https://review.openstack.org/82787 | 17:59 |
| nati_ueno | sdague: markmcclain: around? | 18:00 |
| _cjones_ | kevinbenton: Yeah. I tried with a fresh clone (no changes by me). I get a different set in ./testrepository/failing. I think this may be a timing thing. I'll see what I can glean. I've got enough insight to the process to keep me busy for a bit. Thanks :) | 18:00 |
| sdague | looks like we just lost mark | 18:00 |
| sdague | however there are probably other neutron cores | 18:00 |
| sdague | https://review.openstack.org/#/c/82880/ - is a tempest proposed patch to drop a neutron test | 18:00 |
| *** changbl has quit IRC | 18:00 | |
| *** thedodd has quit IRC | 18:00 | |
| salv-orlando | sdague: I can stand in for mark | 18:00 |
| sdague | which is required by this https://review.openstack.org/#/c/21946/ neutron test | 18:00 |
| sdague | in order to drop a tempest test, we have a policy of requiring a +2 on the dependant change | 18:01 |
| sdague | to make sure it's not some crazy drive by | 18:01 |
| sdague | but actually a change the team in question wants | 18:01 |
| *** catohornet__ has joined #openstack-neutron | 18:01 | |
| salv-orlando | nati-ueno, sdague: we know the blocker here might be this: https://review.openstack.org/#/c/44596/ | 18:01 |
| sdague | so if we could get a +2 on https://review.openstack.org/#/c/21946/ | 18:01 |
| sdague | salv-orlando: so we also need a +2 there then by nova | 18:02 |
| salv-orlando | sdague: +2'ind that will solve the tempest dependency (maybe) but that very patch should not be merged if we not have a committment from nova on 44596 | 18:02 |
| sdague | ok, sure | 18:02 |
| sdague | so I guess someone needs to chase a +2 on the nova patch as well | 18:02 |
| salv-orlando | in a nutshell: if it does not make nova-rc1 then there's no game | 18:02 |
| sdague | ok, is this something russell is tracking for rc1? | 18:03 |
| *** beagles_brb is now known as beagles | 18:03 | |
| nati_ueno | I'll discuss with russel | 18:04 |
| nati_ueno | I'm also fixing 44596 now | 18:04 |
| openstackgerrit | Maru Newby proposed a change to openstack/neutron: Add support for retargetable functional api testing https://review.openstack.org/72585 | 18:05 |
| *** Matt2 has left #openstack-neutron | 18:06 | |
| *** packet has quit IRC | 18:07 | |
| *** banix has quit IRC | 18:08 | |
| *** peristeri has joined #openstack-neutron | 18:10 | |
| *** packet has joined #openstack-neutron | 18:11 | |
| *** emagana has quit IRC | 18:12 | |
| *** emagana has joined #openstack-neutron | 18:12 | |
| *** banix has joined #openstack-neutron | 18:12 | |
| *** yfried has joined #openstack-neutron | 18:13 | |
| *** catohornet__ has quit IRC | 18:15 | |
| *** Matt1 has joined #openstack-neutron | 18:16 | |
| *** changbl has joined #openstack-neutron | 18:19 | |
| *** Jabadia has joined #openstack-neutron | 18:22 | |
| *** Jabadia has quit IRC | 18:23 | |
| *** spandhe has quit IRC | 18:23 | |
| *** Jabadia has joined #openstack-neutron | 18:23 | |
| *** shakamunyi has joined #openstack-neutron | 18:25 | |
| *** shakamunyi has quit IRC | 18:25 | |
| *** shakamunyi has joined #openstack-neutron | 18:26 | |
| *** shakamunyi has quit IRC | 18:26 | |
| *** catohornet__ has joined #openstack-neutron | 18:27 | |
| *** spandhe has joined #openstack-neutron | 18:30 | |
| *** manishg_ has joined #openstack-neutron | 18:33 | |
| *** mlavalle_ has joined #openstack-neutron | 18:36 | |
| *** manishg has quit IRC | 18:37 | |
| *** manishg_ is now known as manishg | 18:37 | |
| *** sphoorti has quit IRC | 18:37 | |
| *** mlavalle has quit IRC | 18:37 | |
| *** mlavalle_ is now known as mlavalle | 18:37 | |
| *** alagalah has joined #openstack-neutron | 18:41 | |
| *** overlayer has quit IRC | 18:44 | |
| *** linuxgeek_ has joined #openstack-neutron | 18:44 | |
| *** devlaps1 has joined #openstack-neutron | 18:44 | |
| *** yfried has quit IRC | 18:45 | |
| *** yfried has joined #openstack-neutron | 18:45 | |
| *** alagalah has quit IRC | 18:46 | |
| *** devlaps has quit IRC | 18:46 | |
| *** peristeri has quit IRC | 18:46 | |
| *** crc32 has joined #openstack-neutron | 18:47 | |
| *** BillTheKat has joined #openstack-neutron | 18:47 | |
| *** BillTheKat is now known as Gil_McGrath | 18:48 | |
| *** thuc has quit IRC | 18:48 | |
| *** thuc has joined #openstack-neutron | 18:49 | |
| *** apevec has quit IRC | 18:49 | |
| *** jpich has quit IRC | 18:49 | |
| *** dfarrell07 has joined #openstack-neutron | 18:50 | |
| *** shakayumi has joined #openstack-neutron | 18:50 | |
| *** shakayumi has quit IRC | 18:50 | |
| *** thuc has quit IRC | 18:51 | |
| *** thuc_ has joined #openstack-neutron | 18:51 | |
| *** shakayumi has joined #openstack-neutron | 18:51 | |
| *** dave_tucker_zzz is now known as dave_tucker | 18:52 | |
| *** BuSerD has joined #openstack-neutron | 18:53 | |
| marun | nati_ueno: is there a reason the proposed patch is so complicated? | 18:54 |
| nati_ueno | marun: vif security one? | 18:55 |
| *** zzelle has joined #openstack-neutron | 18:55 | |
| marun | nati_ueno: yes | 18:55 |
| *** ramishra has quit IRC | 18:55 | |
| marun | nati_ueno: why isn't it just fixing the regression? | 18:55 |
| marun | adding new capabilities would appear to just have slowed down acceptance. | 18:55 |
| *** ramishra has joined #openstack-neutron | 18:55 | |
| marun | especially on the nova side | 18:55 |
| nati_ueno | marun: you are right | 18:55 |
| *** singhs has joined #openstack-neutron | 18:55 | |
| nati_ueno | marun: let's fix back to previous patch | 18:56 |
| marun | so is it possible to backtrack and just submit the simplest thing that could possibly work? | 18:56 |
| nati_ueno | marun: let's me check when this bug comes in | 18:56 |
| marun | if we engage the nova guys, maybe we can get this merged today | 18:56 |
| marun | ok | 18:56 |
| marun | dansmith: ^^ | 18:57 |
| linuxgeek_ | hi when i execute neutron net-list or any other neutron command i get authentication required | 18:57 |
| linuxgeek_ | and the neutron-server.log has WARNING keystoneclient.middleware.auth_token [-] Authorization failed for token xxxxxx | 18:57 |
| linuxgeek_ | i have verified the conf files of neutron and nova, it seems ok | 18:58 |
| linuxgeek_ | any help/pointers pls | 18:58 |
| dansmith | marun: if it's not targeted to -rc1 it's not getting looked at right now | 18:58 |
| dansmith | marun: ping tjones in -nova with the bug to get it targeted if it's important for -rc1 | 18:59 |
| *** jistr has joined #openstack-neutron | 18:59 | |
| *** thedodd has joined #openstack-neutron | 18:59 | |
| *** ramishra has quit IRC | 18:59 | |
| marun | dansmith: it's important - neutron security groups are still broken | 18:59 |
| pcm___ | nati_ueno: ping | 19:00 |
| *** jlibosva has quit IRC | 19:00 | |
| marun | nati_ueno: who have you been engaging with on the nova side of this? | 19:00 |
| dansmith | marun: if it's a regression from havana and the fix is reasonable, then it should be doable | 19:00 |
| dansmith | marun: if it's not a regression, then I think it might as well wait at this point | 19:00 |
| marun | dansmith: yes, it's a regression. we'll work on getting a reasonable fix up in short order | 19:00 |
| dansmith | okay | 19:00 |
| nati_ueno | pcm___: sorry i'm busy now.. | 19:00 |
| *** gdubreui has joined #openstack-neutron | 19:00 | |
| pcm___ | nati_ueno: Can you review these two... https://review.openstack.org/#/c/81124/ and https://review.openstack.org/#/c/82306/ when you can? | 19:01 |
| nati_ueno | pcm___: I'll ping you back | 19:01 |
| nati_ueno | marun: This is the patch https://review.openstack.org/#/c/49660/5 | 19:01 |
| *** jprovazn has joined #openstack-neutron | 19:02 | |
| nati_ueno | marun: if we revert this one, we could have a valid set of configraion. But I'm not sure these drivers still works | 19:02 |
| nati_ueno | marun: Let's me test this | 19:02 |
| marun | nati_ueno: ok. | 19:02 |
| marun | let me know how I can help | 19:02 |
| marun | i'll talk to tjones in #nova | 19:02 |
| nati_ueno | marun: Thanks! | 19:03 |
| *** shakayumi has quit IRC | 19:05 | |
| claudiub | When you have a little time, can you guys review this one? :) https://review.openstack.org/#/c/63775/ | 19:07 |
| *** markwash has quit IRC | 19:08 | |
| openstackgerrit | Yuriy Taraday proposed a change to openstack/neutron: Abstract out root_helper calls to classes https://review.openstack.org/82787 | 19:08 |
| marun | nati_ueno: what about merging the complex patch on the neutron side but only a minimal patch on the nova side? | 19:08 |
| marun | nati_ueno: i.e. pass all the details in to nova, but nova only uses it for firewall yes/no | 19:09 |
| linuxgeek_ | marun, neutron security groups are still broken - i'm seeing this on i3 env. is there a bug for this? | 19:09 |
| marun | salv-orlando: ^ | 19:09 |
| nati_ueno | marun: if you take the way, I think current patch is minimal https://review.openstack.org/#/c/44596/ | 19:09 |
| marun | linuxgeek_: we're talking about it right now. | 19:09 |
| nati_ueno | marun: it is just 141 ilne | 19:09 |
| marun | 'just' | 19:09 |
| nati_ueno | marun: haha sorry | 19:09 |
| marun | nati_ueno: this isn't working | 19:09 |
| marun | nati_ueno: what's the simplest patch that could possibly work on the nova side? | 19:10 |
| marun | nati_ueno: you can submit a follow-up in juno to make more use of the details passed in by neutron | 19:10 |
| marun | nati_ueno: (or maybe even this cycle) | 19:10 |
| linuxgeek_ | marun, thanks. btw do you have any thoughts on the neutron auth required issue? | 19:10 |
| marun | nati_ueno: but the priority is the regression | 19:10 |
| nati_ueno | marun: I think simplest is reverting original one | 19:10 |
| marun | nati_ueno: which one? | 19:10 |
| nati_ueno | marun: https://review.openstack.org/#/c/49660/5 | 19:10 |
| *** SumitNaiksatam has quit IRC | 19:11 | |
| marun | nati_ueno: so is the answer that without the old vif code we need a complex solution? | 19:12 |
| nati_ueno | marun: https://review.openstack.org/#/c/44596/9/nova/virt/libvirt/vif.py | 19:12 |
| beagles | nati_ueno, what about taking the detail that tells nova that we need to support firewalling and push that along instead of pulling in the other details | 19:12 |
| nati_ueno | marun: so what's we really needs is fixing few lines | 19:12 |
| beagles | nati_ueno, ie. just a flag down in the model and the vif driver? | 19:12 |
| nati_ueno | beagles: Ah That's one way | 19:12 |
| *** SumitNaiksatam has joined #openstack-neutron | 19:13 | |
| marun | nati_ueno: that sounds like the way to go! | 19:13 |
| marun | nati_ueno: -> nova inspects the port and decides firewall yes/no | 19:14 |
| nati_ueno | marun: sorry, let me make it clear | 19:14 |
| beagles | and I guess we could do that neutronv2/api.py instead of pushing neutron details down into the nova code | 19:14 |
| nati_ueno | marun: (option1) current patch (neturon + nova) | 19:14 |
| nati_ueno | (option2) nova only with flag | 19:14 |
| *** yfried has quit IRC | 19:14 | |
| marun | nati_ueno: option2++ | 19:14 |
| nati_ueno | gotcha, I think I can push it in 2-3 hours | 19:15 |
| marun | nati_ueno: wait, what do you mean by flag? | 19:15 |
| marun | like a config option? | 19:15 |
| nati_ueno | marun: yes | 19:15 |
| marun | or can nova figure out what to do from port details?? | 19:15 |
| marun | why is this hard? | 19:15 |
| nati_ueno | marun: Actually, that is option1 | 19:15 |
| marun | nati_ueno: so simplify option 1 | 19:15 |
| rkukura | nati_ueno: are you still planning to have nova pass vif_details though to the VIF driver? | 19:15 |
| marun | nati_ueno: provide enough detail in the port, we'll get the damn patch merged asap | 19:16 |
| marun | rkukura: we need something simpler | 19:16 |
| marun | rkukura: we need something that can _merge_ | 19:16 |
| rkukura | that’s as simple as possible | 19:16 |
| marun | rkukura: the current approach of adding complexity not needed to fix the regression != simple as possible | 19:16 |
| nati_ueno | marun: so how we can simplify the patch? | 19:16 |
| marun | nati_ueno: pass in port details -> use_firewall = True | 19:16 |
| marun | nova checks port details, use_firewall = port_details.get('use_firewall, False) | 19:17 |
| marun | would that work? | 19:17 |
| nati_ueno | marun: Isn't is simple as current solution? | 19:17 |
| nati_ueno | marun: it is almost same | 19:17 |
| marun | nati_ueno: 141 lines? | 19:17 |
| marun | in review for a year? | 19:17 |
| marun | reallly? | 19:17 |
| nati_ueno | marun: yes. Including test code | 19:17 |
| *** claudiub has quit IRC | 19:17 | |
| beagles | nati_ueno, I think the difference is in the minimalistic changes to the data model | 19:17 |
| nati_ueno | marun: I believe it will be 120 line or something like that | 19:17 |
| nati_ueno | marun: actulal changes are really few lines | 19:18 |
| *** openstackgerrit has quit IRC | 19:18 | |
| marun | nati_ueno: if this were simple it would be merged already | 19:18 |
| marun | nati_ueno: and we wouldn't be having this discussion | 19:18 |
| beagles | nati_ueno, so instead of several options, you only add one.. the purpose of which is pretty clear | 19:18 |
| rkukura | what if we kept the current neutron code where vif_details contains just the single item, and pass the vif_details through nova to the driver, and have the nova VIF driver just use that one flag? | 19:18 |
| marun | rkukura: +1 | 19:18 |
| *** openstackgerrit has joined #openstack-neutron | 19:18 | |
| marun | rkukura: that's what i'm suggesting (I think) | 19:18 |
| nati_ueno | marun: rkukura: I'm taking that option now | 19:18 |
| nati_ueno | marun: sorry it may confising, current code is old, I'm testing new version with Bob's sugestion localy now | 19:19 |
| *** sphoorti has joined #openstack-neutron | 19:19 | |
| beagles | we should determine if storing the vif_details for this purpose is contentious and would be an obstacle to mergin | 19:19 |
| rkukura | beagles: Any reason to think it would be contentious? | 19:19 |
| marun | rkukura: because we're in rc? | 19:20 |
| marun | rkukura: the patch on the nova side isn't even on nova's radar right now! | 19:20 |
| marun | rkukura: if we're doing anything more than required to fix the security group regression, we're going to have a harder time than necessary | 19:20 |
| rkukura | beagles: Passing vif_details though as part of the VIF object seems to require minimal change to most of the nova code, but still lets vif_details be used for finer grained security control, and/or SR-IOV in the future | 19:21 |
| beagles | rkukura, I think it might be... does it pose migration/upgrade issues, etc. it would be fine for opaque data maybe, but this isn't.. .it is being used by nova itself | 19:21 |
| marun | rkukura: fsck finger grained control in the future | 19:21 |
| marun | rkukura: the time for adding stuff was pre-rc | 19:21 |
| marun | rkukura: even if we could get away with it on the neutron side, we have the nova side to content with | 19:21 |
| beagles | rkukura, yeah... I know we discussed this in the past and it seemed good to me at the time, but when I thought through the implications (not to mention got a good slap upside the head) I think we need to be more cautious with that kind of approach | 19:22 |
| rkukura | Is the VIF object persisted by nova, or is this just used locally in the compute node? | 19:22 |
| nati_ueno | rkukura: it is locally | 19:22 |
| *** linuxgeek_ has left #openstack-neutron | 19:22 | |
| beagles | rkukura, the information we need right now is whether or not to enable the ovs hybrid bridge | 19:22 |
| rkukura | Seems we need to add something to the VIF object - so I don’t see any different risk in adding the vif_details dict vs. adding some special-purpose flag. | 19:22 |
| nati_ueno | I agree | 19:23 |
| marun | rkukura: I see a risk | 19:23 |
| marun | a risk that it will complicate review | 19:23 |
| marun | we don't add things now | 19:23 |
| *** claudiub has joined #openstack-neutron | 19:23 | |
| nati_ueno | marun: so IMO, revering back is a good way for now | 19:23 |
| nati_ueno | marun: I'm not sure why,,, but this fix get fuge discussions... | 19:24 |
| marun | nati_ueno: so cut out _everything_ that complicates it | 19:24 |
| nati_ueno | marun: ok so let's get back to previous vif driver if it is working | 19:24 |
| nati_ueno | marun: no generalizeation | 19:25 |
| nati_ueno | marun: the revering is safe, because it is adding drivers | 19:25 |
| rkukura | seriously, is passing the vif_details as part of the VIF object and having the OVS vif driver look at vif_details[CAP_PORT_FILTER] that risky? | 19:25 |
| marun | nati_ueno: that's not safe | 19:26 |
| marun | nati_ueno: beagles tells me that politically on the nova side reverting that change would be an uphill battle | 19:26 |
| beagles | it *could* be a battle :) | 19:26 |
| *** alagalah has joined #openstack-neutron | 19:26 | |
| *** catohornet__ has quit IRC | 19:26 | |
| nati_ueno | marun: In Japanese sayings, there is a tigar in front of me, and there is a wolf in the back of me.. | 19:27 |
| rkukura | marun: I’m not clear on what you are saying isn’t safe? | 19:27 |
| beagles | rkukura, I think the issue isn't how difficult or complex the test it is the implications of storing this data... say the dictionary changes in Juno | 19:27 |
| *** jp_at_hp has quit IRC | 19:27 | |
| beagles | rkukura, then we need some way to transform that data... in nova... isn't that weird? | 19:27 |
| *** BuSerD has quit IRC | 19:28 | |
| rkukura | that’s the whole point of the approach - to be less brittle, easier to evolve, etc. | 19:28 |
| beagles | but but but but | 19:28 |
| rkukura | why is anything but null transform needed? | 19:28 |
| beagles | who transforms the data? | 19:28 |
| nati_ueno | IF revering is a risk | 19:28 |
| beagles | all the old vif details would be .. broken? | 19:28 |
| nati_ueno | simplest is | 19:28 |
| nati_ueno | using flag(config) | 19:28 |
| nati_ueno | This needs only nova side fix | 19:28 |
| rkukura | transforming the data is what makes this so difficult - that’s why I’ve been advocating just passing vif_details through untouched end-to-end | 19:29 |
| marun | nati_ueno: find a champion on the nova side to push that and *maybe* it's an option | 19:29 |
| beagles | rkukura, let's not forget that nova is using that info.. but let's back up | 19:29 |
| marun | nati_ueno: also not free/easy from a deployment perspective | 19:29 |
| marun | nati_ueno: easy for a developer using devstack, sure | 19:29 |
| beagles | we need to get the security group functionality working.. how are we going to do it? | 19:30 |
| nati_ueno | marun: anyway, we need to have a discussion with nova guys | 19:30 |
| marun | nati_ueno: so, do you have someone on that side you've been working with? | 19:30 |
| nati_ueno | marun: yes. That's the intension of the orignal fix | 19:30 |
| marun | nati_ueno: the original intention, I'll believe | 19:30 |
| beagles | actually nati_ueno, if you have a way you have a good feeling will get through.. man you should do it. | 19:31 |
| nati_ueno | marun: Brian Haley reviewed this recently | 19:31 |
| marun | nati_ueno: is he core? | 19:31 |
| beagles | I'm coming at this kind of from left field... so I could be off-base and the last thing I want to do is lead folks astray or in circles | 19:31 |
| nati_ueno | marun: I'm not sure.. do you know some cores who may review this? | 19:32 |
| marun | nati_ueno: *sigh* | 19:32 |
| *** luqas has quit IRC | 19:32 | |
| marun | nati_ueno: the first step in merging a patch in nova should be engaging one or more nova core | 19:33 |
| marun | nati_ueno: right now getting attention is all but impossible | 19:33 |
| marun | nati_ueno: i.e. if you were serious about merging this the time to engage with nova cores was pre-rc | 19:33 |
| marun | nati_ueno: is there anyone you were talking to at least? | 19:33 |
| nati_ueno | marun: sorry I was focusing on neutron side. This nova fix depends on it | 19:34 |
| nati_ueno | marun: Without neutron fix, tempest won't success on nova side | 19:34 |
| nati_ueno | marun: even if neturon side isn't merged yet | 19:34 |
| marun | nati_ueno: ok | 19:34 |
| *** baoli has quit IRC | 19:35 | |
| *** Sukhdev has quit IRC | 19:35 | |
| nati_ueno | marun: I'm keep working on this, and requested ton's of mails.. | 19:35 |
| nati_ueno | but it causes new discussion.. | 19:35 |
| *** BuSerD has joined #openstack-neutron | 19:35 | |
| nati_ueno | so reverting the patch is only realistic way for me | 19:37 |
| *** jecarey has quit IRC | 19:37 | |
| marun | nati_ueno: reverting/adding a new flag | 19:37 |
| marun | nati_ueno: so we need nova buy-in | 19:37 |
| nati_ueno | I'm not sure the patch will be merged even if it is neutron side | 19:37 |
| nati_ueno | ok | 19:38 |
| nati_ueno | russellb: Can we have a quick discussion? | 19:38 |
| *** baoli has joined #openstack-neutron | 19:38 | |
| marun | russellb: double plus important | 19:38 |
| * russellb perks up | 19:39 | |
| russellb | o/ | 19:39 |
| nati_ueno | russellb: Thank you for taking time for this such a busy time | 19:39 |
| nati_ueno | russellb: sorry I should inform you this more earlier, but actually security group fro neturon + nova + ovs | 19:39 |
| nati_ueno | is broken now | 19:39 |
| russellb | i saw the patches just a couple minutes ago | 19:40 |
| russellb | looked like they started way back in Havana dev | 19:40 |
| russellb | has it been broken that long (or longer) ? | 19:40 |
| marun | russellb: it's been broken as of icehouse (removal of hybrid vif driver) | 19:40 |
| nati_ueno | russellb: yes. We are working on fixing neutron side fix, and it takes long time | 19:40 |
| nati_ueno | russellb: neutron side fix isn't merged yet | 19:40 |
| marun | russellb: we're so late in the cycle that fixing things properly (neutron passing in vif details to determine firewall yes/no) seems like a hard sell | 19:40 |
| russellb | yes, it's very late :-) | 19:41 |
| marun | russellb: the dirty fix is - nova configured to use/not use firewall | 19:41 |
| marun | russellb: either by reverting removal of hybrid vif driver | 19:41 |
| nati_ueno | or revert the removal of hybrid vif driver | 19:41 |
| marun | russellb: ...or setting an explicit flag that has the same effect | 19:41 |
| russellb | can you expand on what you mean by "nova configured to use/not use firewall" ? | 19:41 |
| russellb | you mean, you just have to turn it on manually? | 19:41 |
| russellb | instead of determining it via info neutron passes in? | 19:42 |
| marun | russellb: that's the way it used to work, per-compute host configuration | 19:42 |
| nati_ueno | russellb: so we should select hybrid vif plugging method when we are using OVS | 19:42 |
| rkukura | russellb: note that neutron is already including a flag in the vif_details port attribute that nova could use | 19:42 |
| nati_ueno | russellb: Original intension was passing that information from neturon to nova | 19:42 |
| nati_ueno | russellb: but it is too late, so another option is configuring it by confi | 19:42 |
| marun | rkukura: and we've definitively worked out the details of how that will be persisted/upgraded/etc to merge so late in the cycle? | 19:42 |
| russellb | how is this just coming up? | 19:42 |
| russellb | this isn't tested? | 19:43 |
| marun | russellb: we're asking ourselves the same question. :( | 19:43 |
| marun | russellb: somehow this kept slipping and slipping and slipping... | 19:43 |
| marun | russellb: trying to agree on the 'right' fix, and the fact that the regression should have priority was lost in translation | 19:43 |
| russellb | a config option sounds reasonable i guess, if that's what we have to do | 19:43 |
| nati_ueno | russellb: sure I'll send the patch in 2-3 hours. I'll ping you when I push it | 19:44 |
| russellb | but honestly, i'm not willing to block on this | 19:44 |
| russellb | OK | 19:44 |
| russellb | if it can be ASAP we can probably get it in | 19:44 |
| russellb | make it as simple as possible to ease review | 19:44 |
| nati_ueno | russellb: I got it | 19:44 |
| marun | russellb: we'll do everything we need to to make it happen | 19:44 |
| russellb | ok | 19:44 |
| marun | russellb: it's top priority on the neutron side | 19:44 |
| *** shakayumi has joined #openstack-neutron | 19:45 | |
| russellb | OK | 19:45 |
| nati_ueno | russellb: Thank you for your talking time on this | 19:45 |
| marun | russellb: danke | 19:45 |
| russellb | sure, thanks for working on a quick fix for it | 19:45 |
| *** shakayumi has quit IRC | 19:45 | |
| zzelle | marun, rkukura, hi | 19:45 |
| marun | nati_ueno: so what needs to happen to get this fix in? | 19:46 |
| marun | nati_ueno: nova patch, check. will a devstack patch be required? | 19:46 |
| marun | zzelle: hi | 19:46 |
| *** jecarey has joined #openstack-neutron | 19:46 | |
| nati_ueno | marun: In order to get tempest success, it is enough if we have nova side patch only.. (This sounds not good, but it is reality) | 19:47 |
| nati_ueno | I just take a look code again, and option2' | 19:47 |
| marun | nati_ueno: because negative tests couldn't merge so long as things were broken? | 19:47 |
| nati_ueno | is coming up to me | 19:47 |
| nati_ueno | marun: yep | 19:47 |
| marun | nati_ueno: ah, okk. | 19:48 |
| zzelle | marun, could you re-approve https://review.openstack.org/#/c/82746/ which has failed gate check, a priori because of a false negative test | 19:48 |
| nati_ueno | option2' is add new vif driver inherits require_iptables_option | 19:48 |
| marun | nati_ueno: then we can coordinate with yfried to get the tempest tests merged as soon as the nova patch does | 19:48 |
| nati_ueno | option2' is add new vif driver inherits LibvirtGenericVIFDriver | 19:48 |
| marun | ? | 19:48 |
| nati_ueno | In this case, there is no need to add new flag | 19:48 |
| *** sphoorti has quit IRC | 19:48 | |
| nati_ueno | so code will be more simple | 19:48 |
| zzelle | marun, the change is only replacing the deprecated root_helper by the "next" one and passed others tests | 19:49 |
| marun | nati_ueno: so we go back to the old way of specifying a vif driver? | 19:50 |
| zzelle | s/next/new/ | 19:50 |
| marun | nati_ueno: I think that makes sense. | 19:50 |
| marun | zzelle: right, test timeout. i'll reapprove | 19:50 |
| nati_ueno | marun: I'll push wip code soon, let's discussion on code | 19:50 |
| zzelle | marun, thanks :) | 19:51 |
| marun | nati_ueno: ok i'll be online until 2:30 pst, then back at 4:30 | 19:52 |
| *** shakayumi has joined #openstack-neutron | 19:54 | |
| *** emagana has quit IRC | 19:55 | |
| *** ramishra has joined #openstack-neutron | 19:56 | |
| *** shakamunyi has joined #openstack-neutron | 19:56 | |
| *** shakayumi has quit IRC | 19:58 | |
| *** leseb has joined #openstack-neutron | 19:58 | |
| nati_ueno | marun: do you like this change? https://review.openstack.org/#/c/82904/1/nova/virt/libvirt/vif.py | 19:59 |
| nati_ueno | marun: my bad give me some sec | 20:00 |
| *** changbl has quit IRC | 20:00 | |
| *** armax has quit IRC | 20:00 | |
| *** armax has joined #openstack-neutron | 20:01 | |
| *** alagalah has quit IRC | 20:01 | |
| nati_ueno | marun: https://review.openstack.org/#/c/82904/3/nova/virt/libvirt/vif.py | 20:02 |
| nati_ueno | marun: so 12L | 20:02 |
| *** leseb has quit IRC | 20:03 | |
| openstackgerrit | A change was merged to openstack/neutron: Add enable_security_group to BigSwitch and OneConvergence ini files https://review.openstack.org/82287 | 20:03 |
| *** shakamunyi has quit IRC | 20:03 | |
| openstackgerrit | A change was merged to openstack/neutron: ML2 Cisco Nexus MD: Remove workaround for bug 1276395 https://review.openstack.org/82839 | 20:03 |
| beagles | um | 20:03 |
| beagles | what | 20:03 |
| *** dave_tucker is now known as dave_tucker_zzz | 20:03 | |
| *** baoli has quit IRC | 20:04 | |
| beagles | nati_ueno, isn't it the case that you what the ovs_hybrid bridges to get configured... so what you want to happen is to get get_firewall_required returning true | 20:04 |
| beagles | nati_ueno, the nwfilter stuff is arp-spoofing, etc, which did not exist in previous versions... unless I am mistaken it is the hybrid bridges that need to get configured | 20:05 |
| nati_ueno | beagles: without nwfilter stuff change, it will be stuck | 20:05 |
| nati_ueno | beagles: because we don't initialize basic nw filter | 20:05 |
| nati_ueno | beagles: you mean L164? | 20:06 |
| *** jistr has quit IRC | 20:06 | |
| *** shakamunyi has joined #openstack-neutron | 20:06 | |
| beagles | nati_ueno, mrmmmmm... okay let's consider one of your other proposals... that is reverting the removal of LibvirtHybridOVSBridgeDriver | 20:07 |
| nati_ueno | https://review.openstack.org/#/c/82904/4/nova/virt/libvirt/vif.py sorry I found typo. | 20:08 |
| nati_ueno | beagles: yes | 20:08 |
| beagles | nati_ueno, so .. it calls plug_ovs_hybrid() | 20:08 |
| nati_ueno | beagles: yes | 20:08 |
| beagles | nati_ueno, so the generic driver will only do this if get_firewall_required returns true, right? | 20:08 |
| nati_ueno | beagles: yes | 20:09 |
| beagles | nati_ueno, but for neutron to work, nova's firewall driver must be configured to be nova.virt.firewall.NoopFirewallDriver | 20:09 |
| beagles | nati_ueno: so this returns false | 20:09 |
| nati_ueno | beagles: and conf.filtername should be None (see L180) | 20:09 |
| beagles | nati_ueno, oh bugger...sorry missed the override below | 20:10 |
| beagles | for the firewall thing | 20:11 |
| *** spandhe has quit IRC | 20:11 | |
| nati_ueno | beagles: yes. This combination is needed | 20:11 |
| *** spandhe has joined #openstack-neutron | 20:13 | |
| *** jprovazn has quit IRC | 20:15 | |
| *** dfarrell07 has quit IRC | 20:19 | |
| *** itzikb has joined #openstack-neutron | 20:21 | |
| *** yfried has joined #openstack-neutron | 20:23 | |
| marun | nati_ueno: looking | 20:25 |
| nati_ueno | marun: Thanks | 20:25 |
| *** banix has quit IRC | 20:26 | |
| *** spandhe has quit IRC | 20:26 | |
| *** alagalah has joined #openstack-neutron | 20:27 | |
| marun | nati_ueno: quibble, but should the name be 'LibvirtNeutronFirewallVifDriver'? | 20:27 |
| marun | nati_ueno: so get_firewall_required is used elsewhere, necessitating the addition of get_nw_filter_required? | 20:28 |
| nati_ueno | marun: OK give me a sec. I'll change the name | 20:28 |
| nati_ueno | marun: yes. get_firewall_required is used | 20:28 |
| nati_ueno | marun: the problem is get_firewall_required used for both of nwfilter selection and vif method selection | 20:29 |
| nati_ueno | marun: For neturon case, it is not same | 20:29 |
| marun | nati_ueno: so get_firewall_required -> use hybrid plugging | 20:29 |
| marun | nati_ueno: get_nw_filter_required -> ?? | 20:30 |
| nati_ueno | marun: yes | 20:30 |
| marun | nati_ueno: I don't know why they didn't just name it 'use_hybrid plugging' ;) | 20:30 |
| nati_ueno | marun: it set conf.filtername | 20:30 |
| nati_ueno | marun: if conf.filtername isn't none, it won't work with NoopFirewallDriver which neutron using | 20:30 |
| nati_ueno | so we need get_nw_filter_required to be set | 20:31 |
| nati_ueno | For neutron case, it should be get_nw_filter_required=False get_firewall_required=True | 20:31 |
| marun | nati_ueno: the override in the new vif driver mispells the method name | 20:31 |
| marun | misspells | 20:31 |
| marun | nati_ueno: get_filter_required -> get_nwfilter_required | 20:32 |
| nati_ueno | marun: ah sorry that' one is fixed in new patch set | 20:32 |
| nati_ueno | marun: https://review.openstack.org/#/c/82904/4/nova/virt/libvirt/vif.py | 20:32 |
| marun | ok | 20:32 |
| marun | nati_ueno: so this looks fine to me. let's get a nova core to take a look. | 20:33 |
| nati_ueno | marun: I'm adding UT, so could you give me few sec? | 20:33 |
| marun | dansmith: could use your eyes for a a couple minutes. no longer, promise! https://review.openstack.org/#/c/82904/4/nova/virt/libvirt/vif.py | 20:33 |
| marun | nati_ueno: I'd like to make sure we're good on the approach before UT are required | 20:34 |
| salv-orlando | marun, nati_ueno: so are we cool with both patches? | 20:34 |
| salv-orlando | marun: ok, then | 20:34 |
| marun | salv-orlando: we're reverting | 20:34 |
| marun | salv-orlando: or what did you mean by 'both patches'? | 20:34 |
| nati_ueno | marun: no, we are selecting small patch in nova side only in Icehouse | 20:34 |
| salv-orlando | ok. I missed a part of the conversation as I was away for dinner. | 20:35 |
| nati_ueno | salv-orlando: I gave up to merge original one, and I have a discussion with Russel | 20:35 |
| marun | salv-orlando: too late in cycle to push nova-side change through | 20:35 |
| salv-orlando | so it looks like we're going to push for a small change that selects the appropriate VIF driver if neutron is going to do sec groups? | 20:35 |
| marun | salv-orlando: so we simplify - go back to old way of requiring explicit vif driver | 20:35 |
| salv-orlando | marun: I see. I'm not sure this will go down well in nova, if we're moving away from the generic driver | 20:36 |
| marun | salv-orlando: if by 'select' you mean 'nova has to be configured to use the appropriate driver' | 20:36 |
| marun | salv-orlando: it's either this or we add a config flag that has the same effect in the generic vif driver | 20:36 |
| nati_ueno | salv-orlando: no we aren't moving away from generic one, we have generic one' | 20:36 |
| marun | salv-orlando: the effect is the same - explicit config | 20:36 |
| nati_ueno | salv-orlando: please take a look https://review.openstack.org/#/c/82904/5/nova/virt/libvirt/vif.py | 20:36 |
| dansmith | marun: ...and we swap this subclass in how/when? | 20:37 |
| *** spandhe has joined #openstack-neutron | 20:37 | |
| openstackgerrit | Hareesh Puthalath proposed a change to openstack/neutron: Include cisco plugin in migration plugins with ovs https://review.openstack.org/80406 | 20:37 |
| marun | dansmith: i'm presuming it has to be set per compute node in nova.conf, same as the old hybrid vif driver needed to be configured | 20:37 |
| marun | dansmith: if that's not acceptable, the alternative is adding a new config option that has the same effect on the generic driver | 20:37 |
| marun | dansmith: i.e. if set, control whether hybrid vif plugging is used. | 20:38 |
| dansmith | marun: I'm pretty sure we're not merging any new config variables at this point :) | 20:38 |
| marun | dansmith: so is the new vif driver an option? | 20:38 |
| nati_ueno | dansmith: we are not going add new config option. Adding new driver | 20:38 |
| marun | russelb agreed in principle to the idea of explicit configuration, but we didn't get into the specifics | 20:38 |
| *** otherwiseguy has joined #openstack-neutron | 20:39 | |
| nati_ueno | dansmith: could you take a look https://review.openstack.org/#/c/82904/5 ? | 20:39 |
| russellb | in meeting so hard to talk detail now | 20:39 |
| marun | dansmith: you might be able to scroll back or eavesdrop | 20:39 |
| marun | russellb: no worries, I'm invoking your name in vain | 20:39 |
| russellb | but basically what i want is: 1) not the big thing where we're trying to read data from neutron, too late. | 20:39 |
| russellb | and 2) as simple as possible to fix the regression | 20:39 |
| russellb | whatever that is | 20:39 |
| dansmith | nati_ueno: yeah, I'm looking at it right now | 20:39 |
| nati_ueno | dansmith: Thanks | 20:40 |
| marun | russellb: https://review.openstack.org/#/c/82904/5/nova/virt/libvirt/vif.py | 20:40 |
| nati_ueno | I'm still testing it with devstack, but hopuly I can get direction now | 20:40 |
| marun | nati_ueno: you could probably drop the Libvirt prefix off the name of the subclass since in config the fully qualified path has to be specified | 20:41 |
| marun | i.e. nova.virt.libvirt.vif.NeutronFirewallVifDriver | 20:41 |
| nati_ueno | marun: sure! I'll do it right now | 20:41 |
| beagles | um... this is still something that needs tob e configured though right? | 20:41 |
| beagles | ie in nova.conf | 20:41 |
| salv-orlando | But if nova sets as firewall driver the NoOp driver, isthe hybrid driver used and sec groups correctly work with neutron? | 20:41 |
| salv-orlando | or is there something else? | 20:41 |
| marun | beagles: it used to, back when the hybrid vif driver still existed | 20:42 |
| marun | oops, salv-orlando ^ | 20:42 |
| dansmith | oh, I see vif driver is pluggable right now, | 20:42 |
| dansmith | but I don't know why | 20:42 |
| marun | beagles: is it still possible to specify vif driver? | 20:42 |
| salv-orlando | my understanding of the logic is that if nova specifies the Noopdriver | 20:42 |
| beagles | oic .. we are adding a value not a variable | 20:43 |
| marun | salv-orlando: easy enough to check | 20:43 |
| salv-orlando | then the generic vif driver does hybrid plugging | 20:43 |
| beagles | marun % | 20:43 |
| marun | salv-orlando: if that was the case, why are security groups borked? | 20:43 |
| beagles | marun sorry that was supposed to be ^ | 20:43 |
| salv-orlando | marun: that's what I've done again now, but I'm usually bad in that case | 20:43 |
| salv-orlando | so, I remember from what I discussed in yfried that a devstack patch restored the default value of firewall_driver to iptables | 20:43 |
| salv-orlando | regardless of whether neutron was in place or not. | 20:44 |
| marun | hmmm | 20:44 |
| salv-orlando | But this looks to easy to be true | 20:44 |
| nati_ueno | salv-orlando: The issue is L180. Neutron don't want to config.filtername | 20:44 |
| dansmith | marun: nati_ueno can we not choose to do the right thing if is_neutron() is true? | 20:44 |
| salv-orlando | meaning that if it's there everthing is broken? | 20:44 |
| nati_ueno | salv-orlando: if we set require_firewall=True, current code set config.filtername | 20:44 |
| yfried | salv-orlando: ? | 20:44 |
| nati_ueno | dansmith: It will break VMWare plugin | 20:44 |
| salv-orlando | yfried: the thing about security groups not enforced in the gate | 20:44 |
| nati_ueno | yfried: We are lack of negative test | 20:45 |
| dansmith | nati_ueno: how? | 20:45 |
| nati_ueno | dansmith: so VMWare plugin don't need iptables | 20:45 |
| marun | yfried: was there a devstack-only fix for that or are your test additions still awaiting merge? | 20:45 |
| dansmith | nati_ueno: use more words | 20:45 |
| nati_ueno | dansmith: and they need to use NoopDriver | 20:45 |
| marun | dansmith: it needs to be configurable depending on the selected plugin | 20:45 |
| marun | dansmith: not all plugins want hybrid plugging | 20:46 |
| yfried | salv-orlando: nati_ueno: I have a patch ready for this. https://review.openstack.org/#/c/62702/ | 20:46 |
| nati_ueno | dansmith: ok, so vmware isn't using iptables for realizing security group functionalists. They are using OVS for that. They don't wan't to use Nova side firewall too. | 20:46 |
| dansmith | I don't even know what hybrid plugging means | 20:46 |
| yfried | marun: this should verify secgroup enforcemnt | 20:46 |
| marun | dansmith: hybrid plugging is when there is a linux bridge between vif and ovs so that iptables filtering can be done (e.g. for security groups) | 20:46 |
| yfried | marun: nati_ueno: salv-orlando: ^ is that what you were looking for? | 20:46 |
| nati_ueno | dansmith: hybrid plugging is needed to use iptables + OVS. Basically, OVS didn't support iptables. so for that case, we are using linuxbridge and OVS combination. And This is called Hybridplugging | 20:47 |
| salv-orlando | yfried: yes. | 20:47 |
| marun | yfried: we were debating whether there was a devstack workaround to enable security groups, but I'm assuming a real fix is needed before we can merge your patch | 20:47 |
| nati_ueno | yfried: nice! | 20:47 |
| nati_ueno | marun: +1 | 20:47 |
| dansmith | so, I'm not sure I can reasonably grok all of this in time to make a good decision, personally | 20:47 |
| *** ramishra has quit IRC | 20:47 | |
| dansmith | russellb: ^ | 20:47 |
| yfried | marun: "devstack workaround to enable security groups"? | 20:47 |
| yfried | marun: shouldn't they be on by default? | 20:48 |
| *** ramishra has joined #openstack-neutron | 20:48 | |
| marun | dansmith: It's pretty simple actually, but it's a hectic time. | 20:48 |
| russellb | should we jump on the phone to talk through it? | 20:48 |
| russellb | maybe tomorrow better for me though | 20:48 |
| dansmith | yes, today is not good for me | 20:48 |
| marun | russellb: +1 | 20:48 |
| russellb | k, let's sync tomorrow then | 20:48 |
| marun | russellb: ok. | 20:49 |
| marun | nati_ueno: you're up for that? | 20:49 |
| russellb | maybe proposed workaround will be settled by then | 20:49 |
| *** markwash has joined #openstack-neutron | 20:49 | |
| nati_ueno | yeah, I can join | 20:49 |
| marun | nati_ueno: I think you can probably add UT and test with devstack prior. I think the proposed solution is about as simple as we're going to get. | 20:49 |
| nati_ueno | marun: Sure | 20:49 |
| marun | russellb, dansmith: I think the only alternative is adding a config variable that has the same effect as a new vif driver, but I understand we're too late for that. | 20:50 |
| marun | we can discuss tomorrow | 20:50 |
| salv-orlando | dansmith, russellb: if the proposed vif driver workaround is fine for you, I'm fine either (obviously) I raised the concern just because I recall the direction is to use Hybrid Driver for everything | 20:50 |
| salv-orlando | sorry Generic | 20:50 |
| dansmith | right, I'm not even sure why we have this plug point | 20:51 |
| marun | dansmith: fortuitous luck? :) | 20:51 |
| *** changbl has joined #openstack-neutron | 20:51 | |
| dansmith | marun: no, definitely not that. | 20:51 |
| russellb | dansmith: pretty sure it should be set on fire | 20:51 |
| dansmith | yes | 20:52 |
| nati_ueno | salv-orlando: dansmith: I agree. But we need some workaround on this | 20:52 |
| russellb | and all automatic based on talking to neutron | 20:52 |
| russellb | i think that was the point of the generic driver right? | 20:52 |
| *** ramishra has quit IRC | 20:52 | |
| marun | agreed that is the way forward | 20:52 |
| russellb | to move away from having to configure this crap | 20:52 |
| nati_ueno | russellb: right | 20:52 |
| dansmith | yes | 20:52 |
| marun | but in the absence of support for dynamic config, we need security groups to work. | 20:52 |
| russellb | and there's something nova doesn't know from neutron right now? | 20:52 |
| salv-orlando | nati_ueno: I have said that if the nova team takes https://review.openstack.org/#/c/82904 as an acceptable workaround I am fine | 20:53 |
| russellb | what is it we're enabling or not on the nova side | 20:53 |
| nati_ueno | salv-orlando: gotcha | 20:53 |
| nati_ueno | russellb: security group with neturon + ovs | 20:53 |
| salv-orlando | russellb: nati-ueno has patch 21946 where he has a few new binding attributes that would let neutron tell nova how security should be configured on a port... | 20:54 |
| russellb | OK | 20:54 |
| russellb | so, what drivers from neutron need this | 20:54 |
| marun | russellb: it's simply too late to get it ready to merge for icehouse | 20:54 |
| russellb | just OVS? | 20:54 |
| salv-orlando | I was more of the opinion that neutron should just tell straight nova how to plug a port, without too many attributes | 20:54 |
| marun | russellb: a whole bunch of plugins use the same path | 20:54 |
| nati_ueno | russellb: yes just OVS | 20:54 |
| marun | russellb: it's not just one | 20:54 |
| salv-orlando | and assume that if neutron is there, neutron does security groups. | 20:54 |
| marun | russellb: ah, sorry. yes. just the ovs vif driver. | 20:54 |
| russellb | what's broken, just OVS? | 20:54 |
| russellb | ok. | 20:54 |
| russellb | does nova know if ovs is used on the neutron side? | 20:54 |
| russellb | if so, can nova just "do the right thing" ? | 20:55 |
| russellb | instead of having to configure junk? | 20:55 |
| marun | no | 20:55 |
| russellb | even if it's special casing OVS | 20:55 |
| russellb | ah, boo. | 20:55 |
| salv-orlando | russellb: do you know the trick the ovs agent does for making sec groups work on ovs? the "shim bridge | 20:55 |
| russellb | no | 20:55 |
| marun | that's why we need to pass vif details with the port to configure dynamically | 20:55 |
| marun | (or failing that, statically config) | 20:56 |
| salv-orlando | basically when using the ovs agent (either ovs or ml2 plugin), neutron actually needs a linux bridge to attach iptables rules, otheriwse they're not enforced | 20:56 |
| nati_ueno | russellb: so ML2+OVSDriver is using iptables + OVS. VMWare plugin is using OVS only | 20:56 |
| salv-orlando | as ovs does not send packets to netfilter for processing when they come in/out of their interface | 20:56 |
| marun | long term solution is to filter in ovs directly... | 20:57 |
| marun | but until then we're stuck with a hybrid solution | 20:57 |
| salv-orlando | to address that a tap from a VM is actually plugged into a LB instance, and that LB instance then has another port on OVS | 20:57 |
| salv-orlando | russellb: I hope you're not puking after hearing this ;) | 20:57 |
| russellb | no, in IRC meeting at the same time, so behind | 20:57 |
| kevinbenton | salv-orlando: isn't this the same limitation that nova security groups have been dealing with as well? | 20:59 |
| salv-orlando | oh I think you're in the TC meeting now. I'll be around later if you have time | 20:59 |
| kevinbenton | russellb: i.e. why would russelb be puking? :-) | 20:59 |
| salv-orlando | kevinbenton: I've not been following nova network development. I recall it always used bridges | 21:00 |
| russellb | i'm mostly puking that this has been broken so long and we're hearing about it 2 days before cutting a RC | 21:00 |
| *** manishg has quit IRC | 21:00 | |
| *** spandhe has quit IRC | 21:01 | |
| salv-orlando | russellb: this is not one of the bugs I've kept closely on my radar. to me it was always a regression introduced by a devstack patch; but only recently I'm hearing it's a deeper problem. | 21:01 |
| *** singhs_ has joined #openstack-neutron | 21:02 | |
| kevinbenton | salv-orlando: is the LB strategy the same as this one? https://github.com/openstack/nova/blob/stable/havana/nova/virt/libvirt/vif.py#L410 | 21:02 |
| dansmith | that makes me less willing to take a ninja patch, personally :/ | 21:02 |
| *** singhs has quit IRC | 21:02 | |
| kevinbenton | salv-orlando: veth with ovs | 21:02 |
| *** singhs_ is now known as singhs | 21:02 | |
| *** spandhe has joined #openstack-neutron | 21:03 | |
| *** manishg has joined #openstack-neutron | 21:04 | |
| *** banix has joined #openstack-neutron | 21:04 | |
| marun | dansmith: ninja patch? | 21:05 |
| dansmith | yeah | 21:05 |
| marun | dansmith: What qualifies as a ninja patch? | 21:05 |
| dansmith | marun: fast, last minute, minimal consideration, under-duress | 21:06 |
| marun | dansmith: some of those are true | 21:06 |
| marun | dansmith: but the solution is basically to use the old proven way of configuring nova to use hybrid plugging | 21:06 |
| marun | dansmith: it's a regression from moving towards entirely dynamic configuration | 21:07 |
| marun | dansmith: but it isn't doing anything new - this approach as pretty much proven (if cumbersome) | 21:07 |
| dansmith | weren't we going to table this until tomorrow when we can all concentrate on the proposal?> | 21:07 |
| marun | dansmith: fair enough :) | 21:07 |
| salv-orlando | yeah let's get back at this tomorrow. It's also late my time, and tomorrow we'll have also markmcclain around | 21:07 |
| *** leseb has joined #openstack-neutron | 21:08 | |
| *** shakayumi has joined #openstack-neutron | 21:08 | |
| *** mlavalle has quit IRC | 21:09 | |
| *** shakayumi has quit IRC | 21:09 | |
| nati_ueno | so which time we will have the meeting? | 21:09 |
| *** shakamunyi has quit IRC | 21:11 | |
| *** sridhar has quit IRC | 21:12 | |
| marun | nati_ueno: I think the timing was indeterminate | 21:13 |
| nati_ueno | marun: indeterminate ? | 21:14 |
| *** tvardeman has quit IRC | 21:15 | |
| salv-orlando | yeah, like in the heisenberg principle ;) | 21:17 |
| *** dave_tucker_zzz is now known as dave_tucker | 21:19 | |
| nati_ueno | russellb: salv-orlando: marun: dansmith: kevinbenton: about UTC 17:00? | 21:20 |
| kevinbenton | nati_ueno: i'm just an innocent bystander :-) | 21:20 |
| nati_ueno | kevinbenton: sure :) | 21:20 |
| dansmith | that should be okay for me | 21:21 |
| kevinbenton | nati_ueno: which channel is this meeting in? | 21:21 |
| nati_ueno | kevinbenton: #openstack-neturon | 21:22 |
| nati_ueno | kevinbenton: I'll post conf call url on here | 21:22 |
| kevinbenton | nati_ueno: ok | 21:22 |
| nati_ueno | kevinbenton: Thanks | 21:22 |
| salv-orlando | nati_ueno: this might sound silly but I have to take the cat to the vet | 21:23 |
| salv-orlando | go ahed without me! | 21:23 |
| nati_ueno | salv-orlando: sure! Take care your cat :) | 21:24 |
| *** carlp has quit IRC | 21:28 | |
| marun | nati_ueno: wfm | 21:34 |
| *** banix has quit IRC | 21:35 | |
| *** Jabadia has quit IRC | 21:35 | |
| nati_ueno | marun: ? | 21:36 |
| marun | 17:00utc - works for me | 21:36 |
| nati_ueno | marun: sure | 21:36 |
| *** BuSerD has quit IRC | 21:36 | |
| *** Apsu has quit IRC | 21:37 | |
| *** dave_tucker is now known as dave_tucker_zzz | 21:39 | |
| *** carlp has joined #openstack-neutron | 21:41 | |
| *** humbolt has quit IRC | 21:42 | |
| *** alexpilotti has joined #openstack-neutron | 21:44 | |
| *** Diopter has joined #openstack-neutron | 21:45 | |
| *** russellb has quit IRC | 21:47 | |
| *** metral has quit IRC | 21:47 | |
| *** Diopter has quit IRC | 21:47 | |
| *** zigo has quit IRC | 21:47 | |
| *** wendar has quit IRC | 21:47 | |
| *** russellb has joined #openstack-neutron | 21:48 | |
| *** Apsu has joined #openstack-neutron | 21:48 | |
| *** mwagner_lap has quit IRC | 21:49 | |
| *** metral has joined #openstack-neutron | 21:49 | |
| *** wendar has joined #openstack-neutron | 21:49 | |
| *** zigo has joined #openstack-neutron | 21:49 | |
| *** catohornet__ has joined #openstack-neutron | 21:50 | |
| *** BuSerD has joined #openstack-neutron | 21:50 | |
| *** Gil_McGrath has quit IRC | 21:50 | |
| *** dave_tucker_zzz is now known as dave_tucker | 21:54 | |
| *** claudiub has quit IRC | 21:58 | |
| *** spandhe has quit IRC | 21:58 | |
| *** dave_tucker is now known as dave_tucker_zzz | 22:00 | |
| *** djoreilly has quit IRC | 22:01 | |
| *** leseb has quit IRC | 22:02 | |
| *** spandhe has joined #openstack-neutron | 22:02 | |
| *** networkstatic is now known as networkstatic_zZ | 22:04 | |
| *** armax has quit IRC | 22:05 | |
| openstackgerrit | Kyle Mestery proposed a change to openstack/neutron: Correct OVS VXLAN version check https://review.openstack.org/82931 | 22:10 |
| *** catohornet___ has joined #openstack-neutron | 22:11 | |
| *** markwash_ has joined #openstack-neutron | 22:12 | |
| *** markwash has quit IRC | 22:14 | |
| *** markwash_ is now known as markwash | 22:14 | |
| *** catohornet__ has quit IRC | 22:15 | |
| *** devlaps1 has quit IRC | 22:17 | |
| *** changbl has quit IRC | 22:17 | |
| *** dave_tucker_zzz is now known as dave_tucker | 22:17 | |
| *** devlaps has joined #openstack-neutron | 22:17 | |
| *** dims_ has quit IRC | 22:18 | |
| *** catohornet____ has joined #openstack-neutron | 22:19 | |
| *** catohornet____ is now known as catohornet__ | 22:19 | |
| *** catohornet___ has quit IRC | 22:21 | |
| *** itzikb has quit IRC | 22:23 | |
| *** jobewan has quit IRC | 22:23 | |
| *** catohornet__ has quit IRC | 22:32 | |
| *** dims_ has joined #openstack-neutron | 22:34 | |
| *** banix has joined #openstack-neutron | 22:54 | |
| *** WackoRobie has joined #openstack-neutron | 22:57 | |
| openstackgerrit | Kyle Mestery proposed a change to openstack/neutron: Correct OVS VXLAN version check https://review.openstack.org/82931 | 22:58 |
| *** _1_overlayer has joined #openstack-neutron | 23:01 | |
| *** mwagner_lap has joined #openstack-neutron | 23:02 | |
| openstackgerrit | A change was merged to openstack/neutron: Replace a usage of the deprecated root_helper option https://review.openstack.org/82746 | 23:02 |
| *** _1_overlayer has quit IRC | 23:04 | |
| *** blogan has quit IRC | 23:10 | |
| *** markmcclain has joined #openstack-neutron | 23:10 | |
| *** overlayer has joined #openstack-neutron | 23:10 | |
| *** markmcclain1 has joined #openstack-neutron | 23:11 | |
| *** overlayer has quit IRC | 23:12 | |
| *** otherwiseguy has quit IRC | 23:13 | |
| *** _cjones_ has quit IRC | 23:14 | |
| *** _cjones_ has joined #openstack-neutron | 23:14 | |
| *** markmcclain has quit IRC | 23:15 | |
| *** thedodd has quit IRC | 23:17 | |
| openstackgerrit | Evgeny Fedoruk proposed a change to openstack/neutron: Cancelling thread start while unit tests running https://review.openstack.org/81323 | 23:18 |
| *** thuc_ has quit IRC | 23:18 | |
| *** thuc has joined #openstack-neutron | 23:18 | |
| *** thuc has quit IRC | 23:23 | |
| *** krtaylor has quit IRC | 23:23 | |
| *** jgrimm has quit IRC | 23:25 | |
| *** networkstatic_zZ is now known as networkstatic | 23:25 | |
| *** gdubreui has quit IRC | 23:27 | |
| *** overlayer has joined #openstack-neutron | 23:28 | |
| *** julim has quit IRC | 23:30 | |
| *** markmcclain1 has quit IRC | 23:30 | |
| *** overlayer has quit IRC | 23:31 | |
| *** overlayer has joined #openstack-neutron | 23:31 | |
| *** banix has quit IRC | 23:33 | |
| *** spandhe has quit IRC | 23:33 | |
| *** overlayer has quit IRC | 23:34 | |
| *** dave_tucker is now known as dave_tucker_zzz | 23:36 | |
| *** spandhe has joined #openstack-neutron | 23:39 | |
| openstackgerrit | Robert Kukura proposed a change to openstack/neutron: ML2: Bind ports outside transactions https://review.openstack.org/82945 | 23:40 |
| *** zzelle has quit IRC | 23:44 | |
| *** baoli has joined #openstack-neutron | 23:46 | |
| *** pcm___ has quit IRC | 23:47 | |
| *** devlaps1 has joined #openstack-neutron | 23:50 | |
| *** devlaps has quit IRC | 23:53 | |
| *** mwagner_ has joined #openstack-neutron | 23:54 | |
| *** carlp has quit IRC | 23:55 | |
| openstackgerrit | Daniel Gollub proposed a change to openstack/neutron: Validate local_ip when ovs-agent starts https://review.openstack.org/77100 | 23:56 |
| *** carl_baldwin has quit IRC | 23:57 | |
| alexpilotti | rkukura: hi | 23:58 |
| rkukura | alexpilotti: hi | 23:59 |
| alexpilotti | rkukura: I have a question on a backporting procedure | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!