*** josecastroleon has joined #openstack-nova | 00:00 | |
*** bradjones has joined #openstack-nova | 00:00 | |
*** bradjones has quit IRC | 00:00 | |
*** bradjones has joined #openstack-nova | 00:00 | |
*** josecastroleon has quit IRC | 00:03 | |
*** dims has quit IRC | 00:03 | |
*** dims has joined #openstack-nova | 00:03 | |
*** asettle has joined #openstack-nova | 00:04 | |
*** josecastroleon has joined #openstack-nova | 00:04 | |
*** josecastroleon has quit IRC | 00:06 | |
*** josecastroleon has joined #openstack-nova | 00:07 | |
*** marun has joined #openstack-nova | 00:08 | |
*** josecastroleon has quit IRC | 00:09 | |
*** claudiub has quit IRC | 00:10 | |
*** josecastroleon has joined #openstack-nova | 00:10 | |
*** josecastroleon has quit IRC | 00:12 | |
*** josecastroleon has joined #openstack-nova | 00:13 | |
*** josecastroleon has quit IRC | 00:15 | |
*** josecastroleon has joined #openstack-nova | 00:16 | |
*** josecastroleon has quit IRC | 00:18 | |
*** annashen has quit IRC | 00:18 | |
*** josecastroleon has joined #openstack-nova | 00:19 | |
*** josecastroleon has quit IRC | 00:21 | |
*** tsekiyama has quit IRC | 00:21 | |
*** markvoelker has joined #openstack-nova | 00:21 | |
*** gholler has quit IRC | 00:21 | |
*** gholler_ has quit IRC | 00:21 | |
*** josecastroleon has joined #openstack-nova | 00:22 | |
*** lchen has quit IRC | 00:23 | |
*** josecastroleon has quit IRC | 00:24 | |
*** marcusvrn has quit IRC | 00:24 | |
*** josecastroleon has joined #openstack-nova | 00:25 | |
*** markvoelker has quit IRC | 00:26 | |
*** josecastroleon has quit IRC | 00:27 | |
*** mtanino has quit IRC | 00:27 | |
*** josecastroleon has joined #openstack-nova | 00:28 | |
*** josecastroleon has quit IRC | 00:30 | |
*** josecastroleon has joined #openstack-nova | 00:31 | |
*** josecastroleon has quit IRC | 00:33 | |
openstackgerrit | Alex Xu proposed openstack/nova-specs: Remove v3 from nova code tree https://review.openstack.org/189218 | 00:33 |
---|---|---|
*** josecastroleon has joined #openstack-nova | 00:34 | |
*** josecastroleon has quit IRC | 00:36 | |
*** rfolco has quit IRC | 00:37 | |
*** josecastroleon has joined #openstack-nova | 00:37 | |
*** josecastroleon has quit IRC | 00:39 | |
*** josecastroleon has joined #openstack-nova | 00:40 | |
*** josecastroleon has quit IRC | 00:42 | |
*** josecastroleon has joined #openstack-nova | 00:43 | |
*** josecastroleon has quit IRC | 00:45 | |
*** dprince has joined #openstack-nova | 00:46 | |
*** josecastroleon has joined #openstack-nova | 00:46 | |
*** tjones1 has quit IRC | 00:47 | |
*** zhenguo has joined #openstack-nova | 00:47 | |
*** dprince has quit IRC | 00:48 | |
*** josecastroleon has quit IRC | 00:48 | |
*** josecastroleon has joined #openstack-nova | 00:49 | |
*** josecastroleon has quit IRC | 00:51 | |
*** vilobhmm has quit IRC | 00:52 | |
*** josecastroleon has joined #openstack-nova | 00:52 | |
*** ijw has quit IRC | 00:52 | |
*** josecastroleon has quit IRC | 00:54 | |
*** josecastroleon has joined #openstack-nova | 00:55 | |
openstackgerrit | Ghanshyam Mann proposed openstack/nova: Merge V2 and V2.1 hypervisor functional tests https://review.openstack.org/167082 | 00:55 |
*** josecastroleon has quit IRC | 00:57 | |
*** josecastroleon has joined #openstack-nova | 00:58 | |
*** josecastroleon has quit IRC | 01:00 | |
*** Guest5484 has quit IRC | 01:01 | |
*** josecastroleon has joined #openstack-nova | 01:01 | |
openstackgerrit | wangxiyuan proposed openstack/nova: Remove "rebuild" parameter in migrate_server https://review.openstack.org/184060 | 01:01 |
*** electroc_ has quit IRC | 01:03 | |
*** Sukhdev has quit IRC | 01:03 | |
*** josecastroleon has quit IRC | 01:03 | |
*** josecastroleon has joined #openstack-nova | 01:04 | |
*** josecastroleon has quit IRC | 01:06 | |
*** mdbooth has quit IRC | 01:07 | |
*** josecastroleon has joined #openstack-nova | 01:07 | |
*** josecastroleon has quit IRC | 01:09 | |
*** otter768 has joined #openstack-nova | 01:09 | |
*** josecastroleon has joined #openstack-nova | 01:10 | |
*** josecastroleon has quit IRC | 01:12 | |
*** josecastroleon has joined #openstack-nova | 01:13 | |
*** patrickeast has quit IRC | 01:13 | |
*** mdbooth has joined #openstack-nova | 01:13 | |
*** otter768 has quit IRC | 01:14 | |
*** josecastroleon has quit IRC | 01:15 | |
*** josecastroleon has joined #openstack-nova | 01:16 | |
*** pixelbeat_ has quit IRC | 01:18 | |
*** tobe has joined #openstack-nova | 01:18 | |
*** josecastroleon has quit IRC | 01:18 | |
*** annashen has joined #openstack-nova | 01:19 | |
*** josecastroleon has joined #openstack-nova | 01:19 | |
*** dsanders has quit IRC | 01:20 | |
*** josecastroleon has quit IRC | 01:21 | |
*** josecastroleon has joined #openstack-nova | 01:22 | |
*** markvoelker has joined #openstack-nova | 01:22 | |
*** dims has quit IRC | 01:23 | |
*** josecastroleon has quit IRC | 01:24 | |
*** ociuhandu has quit IRC | 01:24 | |
*** annashen has quit IRC | 01:24 | |
*** josecastroleon has joined #openstack-nova | 01:25 | |
openstackgerrit | Zhenzan Zhou proposed openstack/nova: Bypass ironic server not available issue https://review.openstack.org/173681 | 01:27 |
*** josecastroleon has quit IRC | 01:27 | |
*** markvoelker has quit IRC | 01:27 | |
*** diogogmt has joined #openstack-nova | 01:28 | |
*** josecastroleon has joined #openstack-nova | 01:28 | |
*** josecastroleon has quit IRC | 01:30 | |
*** josecastroleon has joined #openstack-nova | 01:31 | |
*** josecastroleon has quit IRC | 01:33 | |
*** josecastroleon has joined #openstack-nova | 01:35 | |
*** josecastroleon has quit IRC | 01:36 | |
*** haigang has joined #openstack-nova | 01:37 | |
*** josecastroleon has joined #openstack-nova | 01:37 | |
*** dboik has joined #openstack-nova | 01:39 | |
*** josecastroleon has quit IRC | 01:39 | |
*** dboik_ has joined #openstack-nova | 01:39 | |
*** josecastroleon has joined #openstack-nova | 01:41 | |
*** josecastroleon has quit IRC | 01:42 | |
*** dboik has quit IRC | 01:43 | |
*** josecastroleon has joined #openstack-nova | 01:44 | |
*** tjones1 has joined #openstack-nova | 01:44 | |
*** vilobhmm has joined #openstack-nova | 01:45 | |
*** josecastroleon has quit IRC | 01:45 | |
*** haigang has quit IRC | 01:47 | |
*** josecastroleon has joined #openstack-nova | 01:47 | |
*** haigang has joined #openstack-nova | 01:48 | |
*** josecastroleon has quit IRC | 01:49 | |
*** josecastroleon has joined #openstack-nova | 01:50 | |
*** josecastroleon has quit IRC | 01:52 | |
*** josecastroleon has joined #openstack-nova | 01:53 | |
*** josecastroleon has quit IRC | 01:55 | |
*** josecastroleon has joined #openstack-nova | 01:56 | |
*** josecastroleon has quit IRC | 01:58 | |
*** josecastroleon has joined #openstack-nova | 01:59 | |
*** patrickeast has joined #openstack-nova | 02:00 | |
*** josecastroleon has quit IRC | 02:01 | |
*** vilobhmm has quit IRC | 02:02 | |
*** josecastroleon has joined #openstack-nova | 02:02 | |
*** josecastroleon has quit IRC | 02:04 | |
*** asettle has quit IRC | 02:04 | |
*** josecastroleon has joined #openstack-nova | 02:05 | |
*** dan_ has joined #openstack-nova | 02:05 | |
*** dan_ is now known as Guest19563 | 02:05 | |
*** josecastroleon has quit IRC | 02:07 | |
*** josecastroleon has joined #openstack-nova | 02:08 | |
*** yamahata has quit IRC | 02:09 | |
*** josecastroleon has quit IRC | 02:10 | |
*** VW has joined #openstack-nova | 02:10 | |
*** ajayaa has joined #openstack-nova | 02:10 | |
*** josecastroleon has joined #openstack-nova | 02:11 | |
*** boris-42 has quit IRC | 02:12 | |
*** dims has joined #openstack-nova | 02:12 | |
*** josecastroleon has quit IRC | 02:13 | |
*** josecastroleon has joined #openstack-nova | 02:14 | |
*** josecastroleon has quit IRC | 02:17 | |
*** josecastroleon has joined #openstack-nova | 02:18 | |
*** josecastroleon has quit IRC | 02:20 | |
*** park_heijlong has joined #openstack-nova | 02:20 | |
*** josecastroleon has joined #openstack-nova | 02:21 | |
*** dsanders has joined #openstack-nova | 02:22 | |
*** park_heijlong has quit IRC | 02:22 | |
*** josecastroleon has quit IRC | 02:23 | |
*** dims has quit IRC | 02:23 | |
*** josecastroleon has joined #openstack-nova | 02:24 | |
*** josecastroleon has quit IRC | 02:26 | |
*** josecastroleon has joined #openstack-nova | 02:27 | |
openstackgerrit | Lianhao Lu proposed openstack/nova: Fixed some misspellings https://review.openstack.org/173153 | 02:29 |
*** josecastroleon has quit IRC | 02:29 | |
*** unicell has quit IRC | 02:30 | |
*** josecastroleon has joined #openstack-nova | 02:30 | |
*** haigang has quit IRC | 02:31 | |
*** josecastroleon has quit IRC | 02:32 | |
*** josecastroleon has joined #openstack-nova | 02:33 | |
*** josecastroleon has quit IRC | 02:35 | |
*** josecastroleon has joined #openstack-nova | 02:36 | |
*** josecastroleon has quit IRC | 02:38 | |
*** VW has quit IRC | 02:39 | |
*** josecastroleon has joined #openstack-nova | 02:39 | |
*** josecastroleon has quit IRC | 02:41 | |
*** gyee is now known as operator99 | 02:42 | |
*** josecastroleon has joined #openstack-nova | 02:43 | |
*** josecastroleon has quit IRC | 02:44 | |
*** josecastroleon has joined #openstack-nova | 02:46 | |
*** varya has joined #openstack-nova | 02:46 | |
*** tjones1 has quit IRC | 02:47 | |
openstackgerrit | heijlong proposed openstack/nova: Merge quota_sets functional tests of v2 and v2.1 https://review.openstack.org/167899 | 02:47 |
*** josecastroleon has quit IRC | 02:47 | |
*** asettle has joined #openstack-nova | 02:48 | |
*** josecastroleon has joined #openstack-nova | 02:49 | |
*** asettle has quit IRC | 02:49 | |
*** asettle has joined #openstack-nova | 02:50 | |
*** josecastroleon has quit IRC | 02:50 | |
*** bradjones has quit IRC | 02:52 | |
*** josecastroleon has joined #openstack-nova | 02:52 | |
*** bradjones has joined #openstack-nova | 02:53 | |
*** bradjones has quit IRC | 02:53 | |
*** bradjones has joined #openstack-nova | 02:53 | |
*** josecastroleon has quit IRC | 02:53 | |
*** alex_klimov has quit IRC | 02:54 | |
*** josecastroleon has joined #openstack-nova | 02:55 | |
*** varya_ has joined #openstack-nova | 02:55 | |
*** unicell has joined #openstack-nova | 02:55 | |
*** jecarey has joined #openstack-nova | 02:56 | |
*** varya has quit IRC | 02:56 | |
*** josecastroleon has quit IRC | 02:56 | |
*** josecastroleon has joined #openstack-nova | 02:58 | |
*** josecastroleon has quit IRC | 02:59 | |
*** josecastroleon has joined #openstack-nova | 03:01 | |
*** kiran-r has joined #openstack-nova | 03:02 | |
*** josecastroleon has quit IRC | 03:03 | |
*** kiran-r has quit IRC | 03:04 | |
*** josecastroleon has joined #openstack-nova | 03:04 | |
*** kiran-r has joined #openstack-nova | 03:04 | |
*** asettle has quit IRC | 03:05 | |
*** rushiagr_away is now known as rushiagr | 03:05 | |
*** asettle has joined #openstack-nova | 03:05 | |
*** josecastroleon has quit IRC | 03:06 | |
*** ajayaa has quit IRC | 03:07 | |
*** josecastroleon has joined #openstack-nova | 03:07 | |
*** josecastroleon has quit IRC | 03:09 | |
*** changbl_ has joined #openstack-nova | 03:09 | |
*** otter768 has joined #openstack-nova | 03:10 | |
*** josecastroleon has joined #openstack-nova | 03:10 | |
*** markvoelker has joined #openstack-nova | 03:11 | |
*** josecastroleon has quit IRC | 03:12 | |
*** asettle has quit IRC | 03:13 | |
*** josecastroleon has joined #openstack-nova | 03:13 | |
*** asettle has joined #openstack-nova | 03:14 | |
*** ssurana has left #openstack-nova | 03:14 | |
*** otter768 has quit IRC | 03:15 | |
*** josecastroleon has quit IRC | 03:15 | |
*** markvoelker has quit IRC | 03:16 | |
*** josecastroleon has joined #openstack-nova | 03:16 | |
*** josecastroleon has quit IRC | 03:18 | |
*** svalleru has joined #openstack-nova | 03:19 | |
*** josecastroleon has joined #openstack-nova | 03:19 | |
*** josecastroleon has quit IRC | 03:21 | |
*** annashen has joined #openstack-nova | 03:21 | |
*** josecastroleon has joined #openstack-nova | 03:22 | |
*** kiran-r has quit IRC | 03:23 | |
*** josecastroleon has quit IRC | 03:24 | |
*** rushiagr is now known as rushiagr_away | 03:25 | |
*** josecastroleon has joined #openstack-nova | 03:25 | |
*** annashen has quit IRC | 03:26 | |
*** josecastroleon has quit IRC | 03:27 | |
*** mmedvede has quit IRC | 03:28 | |
*** josecastroleon has joined #openstack-nova | 03:28 | |
*** Nikolay_St has quit IRC | 03:28 | |
*** asselin_ has joined #openstack-nova | 03:29 | |
asselin_ | anyone know how to set this up? https://github.com/openstack/nova/blob/f75aba12a932f7b445631e2cdf6c4005d8b4c205/nova/virt/firewall.py#L269 | 03:29 |
asselin_ | # Pass through provider-wide drops | 03:29 |
openstackgerrit | Yuuichi Fujioka proposed openstack/nova: pass environment variables of proxy to tox https://review.openstack.org/189569 | 03:31 |
*** josecastroleon has quit IRC | 03:33 | |
*** josecastroleon has joined #openstack-nova | 03:34 | |
*** vilobhmm has joined #openstack-nova | 03:34 | |
*** mmedvede has joined #openstack-nova | 03:35 | |
*** josecastroleon has quit IRC | 03:36 | |
*** josecastroleon has joined #openstack-nova | 03:37 | |
openstackgerrit | Alex Xu proposed openstack/nova: Remove db layer hard-code permission checks for quota_get_all_* https://review.openstack.org/160202 | 03:37 |
openstackgerrit | Alex Xu proposed openstack/nova: Remove db layer hard-code permission checks for quota_usage_update https://review.openstack.org/160215 | 03:37 |
*** josecastroleon has quit IRC | 03:39 | |
asselin_ | I found this wiki: https://wiki.openstack.org/wiki/Spec-provider-firewall | 03:39 |
*** josecastroleon has joined #openstack-nova | 03:40 | |
*** coolsvap|afk is now known as coolsvap | 03:42 | |
*** josecastroleon has quit IRC | 03:42 | |
*** josecastroleon has joined #openstack-nova | 03:43 | |
*** dsanders has quit IRC | 03:44 | |
*** diogogmt has quit IRC | 03:45 | |
*** josecastroleon has quit IRC | 03:45 | |
*** josecastroleon has joined #openstack-nova | 03:46 | |
*** josecastroleon has quit IRC | 03:48 | |
*** josecastroleon has joined #openstack-nova | 03:49 | |
openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova: Add JSON-Home document methods https://review.openstack.org/189551 | 03:50 |
*** josecastroleon has quit IRC | 03:51 | |
*** Sukhdev has joined #openstack-nova | 03:51 | |
*** deepthi has joined #openstack-nova | 03:52 | |
*** josecastroleon has joined #openstack-nova | 03:52 | |
*** josecastroleon has quit IRC | 03:54 | |
*** josecastroleon has joined #openstack-nova | 03:55 | |
*** josecastroleon has quit IRC | 03:57 | |
*** dsanders has joined #openstack-nova | 03:58 | |
*** josecastroleon has joined #openstack-nova | 03:58 | |
*** josecastroleon has quit IRC | 04:00 | |
*** josecastroleon has joined #openstack-nova | 04:01 | |
*** josecastroleon has quit IRC | 04:03 | |
*** josecastroleon has joined #openstack-nova | 04:04 | |
*** rushiagr_away is now known as rushiagr | 04:05 | |
*** rushiagr is now known as rushiagr_away | 04:05 | |
*** rushiagr_away is now known as rushiagr | 04:06 | |
*** josecastroleon has quit IRC | 04:06 | |
*** josecastroleon has joined #openstack-nova | 04:07 | |
*** josecastroleon has quit IRC | 04:09 | |
*** josecastroleon has joined #openstack-nova | 04:10 | |
*** garyk has quit IRC | 04:12 | |
*** josecastroleon has quit IRC | 04:12 | |
*** josecastroleon has joined #openstack-nova | 04:13 | |
*** rajesht has joined #openstack-nova | 04:16 | |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: virt: introduce libosinfo library to set hardware policy https://review.openstack.org/149625 | 04:25 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: use osinfo when configuring network model https://review.openstack.org/149627 | 04:25 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: adding libosinfo configuration https://review.openstack.org/149626 | 04:25 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: use osinfo when configuring the disk bus https://review.openstack.org/149628 | 04:25 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: adjust os_command_empty test to ImageMetaProps object https://review.openstack.org/190017 | 04:25 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: adjust model_kvm_bogus to ImageMetaProps object https://review.openstack.org/190018 | 04:25 |
*** arnaud____ has joined #openstack-nova | 04:27 | |
*** svalleru has quit IRC | 04:28 | |
*** nelsnelson has joined #openstack-nova | 04:30 | |
*** amotoki_ has joined #openstack-nova | 04:40 | |
*** josecastroleon has quit IRC | 04:43 | |
*** nelsnelson has quit IRC | 04:43 | |
*** nelsnelson has joined #openstack-nova | 04:44 | |
*** josecastroleon has joined #openstack-nova | 04:44 | |
*** amotoki_ has quit IRC | 04:46 | |
*** josecastroleon has quit IRC | 04:46 | |
*** josecastroleon has joined #openstack-nova | 04:47 | |
*** josecastroleon has quit IRC | 04:49 | |
*** tojuvone has joined #openstack-nova | 04:49 | |
*** josecastroleon has joined #openstack-nova | 04:51 | |
*** josecastroleon has quit IRC | 04:52 | |
*** patrickeast has quit IRC | 04:52 | |
*** josecastroleon has joined #openstack-nova | 04:54 | |
*** josecastroleon has quit IRC | 04:55 | |
*** rushiagr is now known as rushiagr_away | 04:55 | |
*** josecastroleon has joined #openstack-nova | 04:56 | |
*** arnaud____ has quit IRC | 04:57 | |
*** achanda has joined #openstack-nova | 04:58 | |
*** josecastroleon has quit IRC | 04:58 | |
*** josecastroleon has joined #openstack-nova | 05:00 | |
*** markvoelker has joined #openstack-nova | 05:00 | |
openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova: Add jsonschema validation for creating secgroup https://review.openstack.org/189544 | 05:00 |
*** josecastroleon has quit IRC | 05:03 | |
*** josecastroleon has joined #openstack-nova | 05:04 | |
*** markvoelker has quit IRC | 05:04 | |
*** arnaud____ has joined #openstack-nova | 05:05 | |
*** josecastroleon has quit IRC | 05:06 | |
*** jecarey has quit IRC | 05:07 | |
*** fawadkhaliq has joined #openstack-nova | 05:07 | |
*** josecastroleon has joined #openstack-nova | 05:07 | |
*** josecastroleon has quit IRC | 05:09 | |
*** josecastroleon has joined #openstack-nova | 05:10 | |
*** ajayaa has joined #openstack-nova | 05:10 | |
*** otter768 has joined #openstack-nova | 05:11 | |
*** vilobhmm has quit IRC | 05:11 | |
*** josecastroleon has quit IRC | 05:12 | |
*** josecastroleon has joined #openstack-nova | 05:13 | |
*** josecastroleon has quit IRC | 05:15 | |
*** otter768 has quit IRC | 05:15 | |
*** josecastroleon has joined #openstack-nova | 05:16 | |
*** josecastroleon has quit IRC | 05:18 | |
*** josecastroleon has joined #openstack-nova | 05:19 | |
*** josecastroleon has quit IRC | 05:21 | |
*** sudipto has joined #openstack-nova | 05:22 | |
*** josecastroleon has joined #openstack-nova | 05:22 | |
*** prashup has joined #openstack-nova | 05:23 | |
*** pradipta has joined #openstack-nova | 05:23 | |
*** josecastroleon has quit IRC | 05:24 | |
*** josecastroleon has joined #openstack-nova | 05:25 | |
*** rushiagr_away is now known as rushiagr | 05:25 | |
*** josecastroleon has quit IRC | 05:27 | |
*** kiran-r has joined #openstack-nova | 05:27 | |
*** josecastroleon has joined #openstack-nova | 05:28 | |
openstackgerrit | Eli Qiao proposed openstack/nova: Clean up Fake_Url for unit test of flavor_access https://review.openstack.org/162869 | 05:29 |
*** josecastroleon has quit IRC | 05:30 | |
*** xiaohui has joined #openstack-nova | 05:31 | |
*** josecastroleon has joined #openstack-nova | 05:31 | |
*** nihilifer_ has joined #openstack-nova | 05:32 | |
*** nkrinner has joined #openstack-nova | 05:32 | |
*** achanda has quit IRC | 05:32 | |
*** josecastroleon has quit IRC | 05:33 | |
*** apuimedo has joined #openstack-nova | 05:33 | |
*** josecastroleon has joined #openstack-nova | 05:34 | |
*** josecastroleon has quit IRC | 05:36 | |
*** josecastroleon has joined #openstack-nova | 05:37 | |
*** josecastroleon has quit IRC | 05:39 | |
*** rook has quit IRC | 05:39 | |
*** josecastroleon has joined #openstack-nova | 05:40 | |
*** Murali has joined #openstack-nova | 05:40 | |
*** amotoki_ has joined #openstack-nova | 05:41 | |
*** unicell1 has joined #openstack-nova | 05:41 | |
*** josecastroleon has quit IRC | 05:42 | |
*** unicell has quit IRC | 05:43 | |
*** josecastroleon has joined #openstack-nova | 05:43 | |
*** rajesht has quit IRC | 05:45 | |
*** josecastroleon has quit IRC | 05:45 | |
*** amotoki_ has quit IRC | 05:46 | |
*** josecastroleon has joined #openstack-nova | 05:46 | |
*** ankit_ag has joined #openstack-nova | 05:47 | |
*** josecastroleon has quit IRC | 05:48 | |
*** eliqiao has joined #openstack-nova | 05:48 | |
*** josecastroleon has joined #openstack-nova | 05:49 | |
*** josecastroleon has quit IRC | 05:51 | |
*** Sukhdev has quit IRC | 05:51 | |
*** josecastroleon has joined #openstack-nova | 05:52 | |
*** haomaiwang has quit IRC | 05:54 | |
*** pilgrimstack has joined #openstack-nova | 05:54 | |
*** pilgrimstack has quit IRC | 06:01 | |
*** belmoreira has joined #openstack-nova | 06:01 | |
*** pilgrimstack has joined #openstack-nova | 06:02 | |
*** marun has quit IRC | 06:03 | |
*** garyk has joined #openstack-nova | 06:05 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Imported Translations from Transifex https://review.openstack.org/190034 | 06:07 |
*** jcoufal has joined #openstack-nova | 06:08 | |
*** oro has joined #openstack-nova | 06:10 | |
*** pbandzi has joined #openstack-nova | 06:11 | |
*** asettle has quit IRC | 06:11 | |
*** destiny has joined #openstack-nova | 06:14 | |
*** eglynn_ has joined #openstack-nova | 06:15 | |
*** arnaud____ has quit IRC | 06:17 | |
*** Sukhdev has joined #openstack-nova | 06:19 | |
*** sahid has joined #openstack-nova | 06:22 | |
*** annashen has joined #openstack-nova | 06:23 | |
*** oro has quit IRC | 06:25 | |
*** achanda has joined #openstack-nova | 06:26 | |
*** lsmola has joined #openstack-nova | 06:27 | |
*** annashen has quit IRC | 06:29 | |
*** Sukhdev has quit IRC | 06:33 | |
*** rook has joined #openstack-nova | 06:34 | |
*** pilgrimstack has quit IRC | 06:35 | |
*** mdbooth has quit IRC | 06:39 | |
*** amotoki_ has joined #openstack-nova | 06:41 | |
*** moshele has joined #openstack-nova | 06:44 | |
*** mdbooth has joined #openstack-nova | 06:45 | |
*** ildikov_ has quit IRC | 06:45 | |
*** jwcroppe_ has quit IRC | 06:46 | |
*** amotoki_ has quit IRC | 06:47 | |
*** markvoelker has joined #openstack-nova | 06:49 | |
*** romainh has joined #openstack-nova | 06:49 | |
*** browne has quit IRC | 06:50 | |
*** markvoelker has quit IRC | 06:53 | |
*** lpetrut has joined #openstack-nova | 06:58 | |
*** ildikov_ has joined #openstack-nova | 06:58 | |
*** ssurana has joined #openstack-nova | 06:59 | |
openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova: Add jsonschema validation for creating secgroup https://review.openstack.org/189544 | 07:00 |
*** sahid has quit IRC | 07:00 | |
*** marun has joined #openstack-nova | 07:01 | |
*** sahid has joined #openstack-nova | 07:01 | |
*** kiran-r has quit IRC | 07:03 | |
*** kiran-r has joined #openstack-nova | 07:04 | |
*** kiran-r has quit IRC | 07:04 | |
*** kiranr has joined #openstack-nova | 07:04 | |
*** kiranr has quit IRC | 07:04 | |
*** kiranr has joined #openstack-nova | 07:05 | |
*** marun has quit IRC | 07:05 | |
*** kiranr is now known as kiran-r | 07:05 | |
openstackgerrit | garyk proposed openstack/nova: Compute: enable admin to decide action for conflicting states https://review.openstack.org/190047 | 07:05 |
*** mwagner_afk has quit IRC | 07:06 | |
*** slong has quit IRC | 07:07 | |
xiaohui | x123 | 07:09 |
xiaohui | lllllll | 07:09 |
*** xiaohui has quit IRC | 07:09 | |
*** otter768 has joined #openstack-nova | 07:12 | |
*** abhishekk has joined #openstack-nova | 07:12 | |
*** dsanders has quit IRC | 07:13 | |
*** otter768 has quit IRC | 07:16 | |
*** p_deore has joined #openstack-nova | 07:16 | |
*** arnaud____ has joined #openstack-nova | 07:18 | |
*** mwagner_afk has joined #openstack-nova | 07:18 | |
*** achanda has quit IRC | 07:20 | |
*** haypo has joined #openstack-nova | 07:22 | |
*** haypo has quit IRC | 07:22 | |
*** haypo has joined #openstack-nova | 07:22 | |
*** rajesht has joined #openstack-nova | 07:22 | |
*** arnaud____ has quit IRC | 07:22 | |
*** armax has quit IRC | 07:26 | |
*** mpaolino has joined #openstack-nova | 07:27 | |
*** eglynn_ has quit IRC | 07:30 | |
*** fawadkhaliq has quit IRC | 07:32 | |
*** doude has joined #openstack-nova | 07:33 | |
*** mmedvede has quit IRC | 07:37 | |
*** jistr has joined #openstack-nova | 07:37 | |
*** park_heijlong has joined #openstack-nova | 07:38 | |
*** jlanoux has joined #openstack-nova | 07:38 | |
*** dsanders has joined #openstack-nova | 07:40 | |
*** amotoki_ has joined #openstack-nova | 07:43 | |
*** achanda has joined #openstack-nova | 07:44 | |
*** yingjun has joined #openstack-nova | 07:46 | |
*** amotoki_ has quit IRC | 07:48 | |
*** cfriesen has quit IRC | 07:55 | |
*** haomaiwang has joined #openstack-nova | 07:56 | |
*** park_heijlong has quit IRC | 07:57 | |
*** oro has joined #openstack-nova | 07:57 | |
*** rotbeard has joined #openstack-nova | 07:57 | |
openstackgerrit | Rui Chen proposed openstack/nova: Improve functional test base for microversion https://review.openstack.org/190063 | 07:58 |
*** leopoldj has joined #openstack-nova | 08:00 | |
*** ZZelle has quit IRC | 08:01 | |
*** ijw has joined #openstack-nova | 08:01 | |
*** ZZelle has joined #openstack-nova | 08:02 | |
*** fawadkhaliq has joined #openstack-nova | 08:03 | |
*** lucasagomes has joined #openstack-nova | 08:07 | |
*** park_heijlong has joined #openstack-nova | 08:08 | |
*** atuvenie has joined #openstack-nova | 08:12 | |
openstackgerrit | Alex Xu proposed openstack/nova: The devref for Nova stable API https://review.openstack.org/162912 | 08:13 |
*** dsanders has quit IRC | 08:13 | |
openstackgerrit | Rui Chen proposed openstack/nova: CreateVM supports subnet specified https://review.openstack.org/182969 | 08:14 |
openstackgerrit | Yongli He proposed openstack/nova-specs: Allocate least feature device to user https://review.openstack.org/190065 | 08:14 |
*** eglynn_ has joined #openstack-nova | 08:14 | |
*** markus_z has joined #openstack-nova | 08:16 | |
*** fawadkhaliq has quit IRC | 08:17 | |
*** yingjun has quit IRC | 08:17 | |
*** arnaud____ has joined #openstack-nova | 08:18 | |
*** Nikolay_St has joined #openstack-nova | 08:19 | |
*** yingjun has joined #openstack-nova | 08:20 | |
*** afazekas has joined #openstack-nova | 08:20 | |
*** yassine has joined #openstack-nova | 08:23 | |
*** ssurana has quit IRC | 08:23 | |
*** arnaud____ has quit IRC | 08:23 | |
*** danpb has joined #openstack-nova | 08:25 | |
*** lsmola has quit IRC | 08:27 | |
*** aarefiev has joined #openstack-nova | 08:28 | |
*** lsmola has joined #openstack-nova | 08:34 | |
*** BobBall_AWOL is now known as BobBall | 08:34 | |
*** markvoelker has joined #openstack-nova | 08:37 | |
*** bradjones has quit IRC | 08:40 | |
openstackgerrit | garyk proposed openstack/nova: DB: add support for volume_type by BDMS https://review.openstack.org/139992 | 08:41 |
openstackgerrit | garyk proposed openstack/nova: Object: add volume_type_id to block_device objects https://review.openstack.org/139994 | 08:41 |
*** alex_klimov has joined #openstack-nova | 08:41 | |
*** markvoelker has quit IRC | 08:42 | |
*** bradjones has joined #openstack-nova | 08:42 | |
*** bradjones has quit IRC | 08:42 | |
*** bradjones has joined #openstack-nova | 08:42 | |
*** lpetrut has quit IRC | 08:43 | |
*** amotoki_ has joined #openstack-nova | 08:44 | |
*** park_heijlong has quit IRC | 08:44 | |
*** amotoki_ has quit IRC | 08:48 | |
*** vishwanathj has quit IRC | 08:51 | |
openstackgerrit | garyk proposed openstack/nova: Compute: enable admin to decide action for conflicting states https://review.openstack.org/190047 | 08:51 |
*** park_heijlong has joined #openstack-nova | 08:56 | |
*** abhishekk has quit IRC | 08:56 | |
*** apuimedo has quit IRC | 08:56 | |
*** ijw has quit IRC | 08:57 | |
*** fawadkhaliq has joined #openstack-nova | 08:59 | |
*** palexster has joined #openstack-nova | 08:59 | |
*** arnaud____ has joined #openstack-nova | 09:00 | |
*** apuimedo has joined #openstack-nova | 09:02 | |
*** arnaud____ has quit IRC | 09:06 | |
*** yamahata has joined #openstack-nova | 09:07 | |
*** otter768 has joined #openstack-nova | 09:13 | |
*** abhishekk has joined #openstack-nova | 09:16 | |
*** otter768 has quit IRC | 09:18 | |
*** lpetrut has joined #openstack-nova | 09:18 | |
*** apuimedo has quit IRC | 09:19 | |
*** apuimedo has joined #openstack-nova | 09:20 | |
*** dims has joined #openstack-nova | 09:27 | |
*** dims has quit IRC | 09:31 | |
*** aix has joined #openstack-nova | 09:34 | |
gilliard | jogo: What was the test which was failing frequently on the multinode job? | 09:35 |
gilliard | I'm doing some logstash-diving. | 09:35 |
gilliard | Looks like the live-migration test I added fails sometimes, but only a dozen times in the last 48h | 09:35 |
gilliard | IIRC there was another one you mentioned at the summit? | 09:35 |
jogo | gilliard: nothing is failing very often | 09:38 |
jogo | its just a few infrequent ones | 09:38 |
jogo | https://bugs.launchpad.net/nova/+bug/1462305 | 09:38 |
openstack | Launchpad bug 1462305 in OpenStack Compute (nova) "multi-node test causes nova-compute to lockup" [Undecided,Incomplete] - Assigned to Joe Gordon (jogo) | 09:38 |
jogo | gilliard: and https://bugs.launchpad.net/nova/+bug/1445569 | 09:39 |
openstack | Launchpad bug 1445569 in OpenStack Compute (nova) "No dhcp lease after shelve unshelve" [High,Confirmed] | 09:39 |
*** toan-tran has joined #openstack-nova | 09:39 | |
jogo | gilliard: http://jogo.github.io/gate/multinode.html paints a good picture | 09:39 |
gilliard | ahh yes shelve/unshelve is the one I was thinking of. | 09:39 |
jogo | although that spike looks really bad | 09:39 |
jogo | gilliard: yup unshelve/shelve and n-cpu hanging are the two that I know of | 09:39 |
*** yingjun has quit IRC | 09:40 | |
*** varya_ has quit IRC | 09:40 | |
gilliard | gilliard: and https://bugs.launchpad.net/nova/+bug/1445569 | 09:41 |
openstack | Launchpad bug 1445569 in OpenStack Compute (nova) "No dhcp lease after shelve unshelve" [High,Confirmed] | 09:41 |
gilliard | oh, c&p error, sorry. | 09:41 |
*** xek has joined #openstack-nova | 09:42 | |
*** varya_ has joined #openstack-nova | 09:42 | |
jogo | dansmith: I revved your partial n-net patch | 09:43 |
jogo | gilliard: I think this failure is new http://logs.openstack.org/92/139992/31/check/check-tempest-dsvm-multinode-full/59bcd9b//console.html#_2015-06-10_09_35_54_675 | 09:45 |
gilliard | Yeah I was looking at that one. | 09:45 |
*** prashup has quit IRC | 09:45 | |
gilliard | ~12 hits in 48h | 09:45 |
jogo | time to check the nova logs | 09:46 |
gilliard | http://logstash.openstack.org/#eyJzZWFyY2giOiJcIkZpbGUgXCJ0ZW1wZXN0L2FwaS9jb21wdXRlL2FkbWluL3Rlc3RfbGl2ZV9taWdyYXRpb24ucHlcIiwgbGluZSAxMTYsIGluIHRlc3RfbGl2ZV9ibG9ja19taWdyYXRpb25cIiAgQU5EIGJ1aWxkX25hbWU6XCJjaGVjay10ZW1wZXN0LWRzdm0tbXVsdGlub2RlLWZ1bGxcIiBBTkQgcHJvamVjdDpcIm9wZW5zdGFjay9ub3ZhXCIiLCJmaWVsZHMiOltdLCJvZmZzZXQiOjAsInRpbWVmcmFtZSI6IjE3MjgwMCIsImdyYXBobW9kZSI6ImNvdW50IiwidGltZSI6eyJ1c2VyX2lud | 09:46 |
gilliard | GVydmFsIjowfSwic3RhbXAiOjE0MzM5Mjg1MDE4ODIsIm1vZGUiOiIiLCJhbmFseXplX2ZpZWxkIjoiIn0= | 09:46 |
gilliard | :( long url. "File "tempest/api/compute/admin/test_live_migration.py", line 116, in test_live_block_migration" AND build_name:"check-tempest-dsvm-multinode-full" AND project:"openstack/nova" | 09:46 |
jogo | thatis a lot of red http://logs.openstack.org/92/139992/31/check/check-tempest-dsvm-multinode-full/59bcd9b/logs/screen-n-cpu.txt.gz?level=TRACE | 09:46 |
gilliard | Worth logging a bug for the failure in test_live_block_migration? | 09:47 |
jogo | gilliard: yes for this stacktrace | 09:47 |
*** fesp has joined #openstack-nova | 09:48 | |
*** ildikov_ has quit IRC | 09:48 | |
*** rushiagr is now known as rushiagr_away | 09:49 | |
*** artom has joined #openstack-nova | 09:49 | |
*** prashup has joined #openstack-nova | 09:49 | |
gilliard | https://bugs.launchpad.net/nova/+bug/1463747 | 09:50 |
openstack | Launchpad bug 1463747 in OpenStack Compute (nova) "multinode job failing in test_live_block_migration" [Undecided,New] | 09:50 |
*** dims has joined #openstack-nova | 09:51 | |
*** dims_ has joined #openstack-nova | 09:52 | |
jogo | gilliard: this is a 100% breakage | 09:54 |
jogo | that is the good news | 09:54 |
*** fesp has quit IRC | 09:55 | |
jogo | and this is the offending patch: https://review.openstack.org/#/c/177437/ | 09:56 |
*** dims has quit IRC | 09:56 | |
gilliard | I have 2 questions. 1/ how do you know it's 100% breakage? I really struggle to get that kind of info out of logstash | 09:57 |
*** park_heijlong has quit IRC | 09:57 | |
gilliard | 2/ How did you find that patch so fast?! | 09:57 |
jogo | gilliard: great questions. | 09:58 |
jogo | 1) with this query message:"Finished: SUCCESS" AND build_name:"check-tempest-dsvm-multinode-full" | 09:58 |
jogo | so that will show you the last time multinode finished successfully | 09:58 |
jogo | first hint that it was 100% failure was my graphite page http://jogo.github.io/gate/multinode.html | 09:59 |
jogo | that is a very steep curve (moving average) | 09:59 |
jogo | 2) I went into n-cpu and looked for stacktraces | 09:59 |
jogo | http://logs.openstack.org/92/139992/31/check/check-tempest-dsvm-multinode-full/59bcd9b/logs/screen-n-cpu.txt.gz?level=TRACE | 09:59 |
jogo | and found this fingerprint: message:"TypeError: string indices must be integers" AND tags:"screen-n-cpu.txt" | 09:59 |
jogo | then looked for the first few hits for it | 10:00 |
garyk | BobBall: any idea why citrix CI is broken - can we stop the -1 vote until it is fixed. My world is red (and I am not a liverpool fan) | 10:00 |
jogo | the first hits are from two separate runs on the same patch | 10:00 |
jogo | (build_change in logstash) | 10:00 |
BobBall | I've already turnd it off | 10:01 |
BobBall | no idea what's going on though | 10:01 |
BobBall | trying to investigate | 10:01 |
BobBall | Just one more job running that'll get a -1 vote (189576) - but no more -1s will be posted | 10:01 |
BobBall | Don't suppose you've got an idea why requests might disappear? http://dd6b71949550285df7dc-dda4e480e005aaa13ec303551d2d8155.r49.cf1.rackcdn.com/63/190063/1/581/logs/devstacklog.txt.gz | 10:02 |
jogo | gilliard: it would be great of you try what I just did and hopefully come to the same conclusion | 10:02 |
gilliard | jogo: I'm just working through it... | 10:02 |
jogo | gilliard: thanks! | 10:03 |
gilliard | thanks for the tutorial :) | 10:03 |
BobBall | Ok... It's hitting https://git.openstack.org/cgit/openstack-dev/devstack/commit/?id=1bbe71c856916a81e2a4aebcf380235cf0fb819b hmmmmz | 10:03 |
jogo | gilliard: I am happy to spread the know how around | 10:03 |
*** Kennan2 has joined #openstack-nova | 10:03 | |
*** Kennan has quit IRC | 10:04 | |
*** achanda has quit IRC | 10:04 | |
BobBall | Looks like devstack is broken for some systems now! :) yay. | 10:04 |
dims_ | ndipanov: danpb: thanks for the py34 review yday. i got it working (https://review.openstack.org/#/c/187785/) | 10:06 |
openstackgerrit | Joe Gordon proposed openstack/nova: Revert "libvirt: Remove unnecessary JSON conversions" https://review.openstack.org/190098 | 10:06 |
jogo | BobBall: if it makes you feel better, someone broke live migration as well | 10:07 |
*** achanda has joined #openstack-nova | 10:07 | |
BobBall | heh :) Wasn't me! | 10:08 |
jogo | dims_: isn't it 6AM by you? | 10:08 |
lifeless | of course they did | 10:08 |
dims_ | ndipanov: danpb: py34 does not support unbound methods and ismethod was not working, so had to add isfunction | 10:08 |
dims_ | jogo: hehe, yes :) | 10:08 |
jogo | lifeless: well we did catch the issue in <6 hours | 10:08 |
dims_ | jogo: and you are travelling? :) | 10:08 |
jogo | dims_: just checking | 10:08 |
jogo | dims_: yup, in Israel (and Rome next week) | 10:09 |
*** boris-42 has joined #openstack-nova | 10:09 | |
jogo | gilliard: any other logstash.o.o questions I can help with? | 10:09 |
dims_ | safe travels jogo! | 10:09 |
jogo | dims_: thanks | 10:09 |
gilliard | I'm sure I'll come up with some more! | 10:09 |
*** achanda has quit IRC | 10:13 | |
*** gszasz has joined #openstack-nova | 10:14 | |
*** matrohon has joined #openstack-nova | 10:16 | |
*** pixelbeat_ has joined #openstack-nova | 10:20 | |
gilliard | jogo: https://gist.github.com/mjg123/cb544f2f6853df8d866c | 10:22 |
*** markvoelker has joined #openstack-nova | 10:23 | |
jogo | gilliard: yup | 10:23 |
jogo | note logstash.openstack.org has 10 days worth of logs | 10:23 |
*** wuhg has quit IRC | 10:24 | |
*** romainh has quit IRC | 10:27 | |
*** annashen has joined #openstack-nova | 10:27 | |
*** markvoelker has quit IRC | 10:27 | |
*** ociuhandu has joined #openstack-nova | 10:28 | |
*** annashen has quit IRC | 10:31 | |
*** mpaolino has quit IRC | 10:32 | |
jogo | danpb: want to post non-revert fix for https://review.openstack.org/#/c/190098/ ? | 10:33 |
*** cbader has quit IRC | 10:34 | |
ndipanov | dims_, cool - will take a look | 10:37 |
*** ildikov_ has joined #openstack-nova | 10:37 | |
*** marun has joined #openstack-nova | 10:38 | |
*** derekh has joined #openstack-nova | 10:38 | |
openstackgerrit | Petrut Lucian proposed openstack/nova-specs: Hyper-V: Add Fibre Channel support https://review.openstack.org/190107 | 10:39 |
*** ociuhandu has quit IRC | 10:40 | |
openstackgerrit | Petrut Lucian proposed openstack/nova-specs: Hyper-V: Add Fibre Channel support https://review.openstack.org/190107 | 10:40 |
*** subscope has joined #openstack-nova | 10:41 | |
*** varya_ has quit IRC | 10:42 | |
*** Kennan2 is now known as Kennan | 10:42 | |
danpb | jogo: yes, i'm just doing that now | 10:42 |
*** marun has quit IRC | 10:43 | |
jogo | danpb: thanks | 10:43 |
*** kmartin has quit IRC | 10:44 | |
*** mpaolino has joined #openstack-nova | 10:44 | |
*** amotoki_ has joined #openstack-nova | 10:45 | |
*** coolsvap is now known as coolsvap|afk | 10:45 | |
openstackgerrit | garyk proposed openstack/nova: VMware: improve instance names on VC https://review.openstack.org/166608 | 10:47 |
openstackgerrit | garyk proposed openstack/nova: VMware: add in folder support on VC https://review.openstack.org/165060 | 10:47 |
*** fawadkhaliq has quit IRC | 10:48 | |
*** fawadkhaliq has joined #openstack-nova | 10:48 | |
*** arnaud____ has joined #openstack-nova | 10:50 | |
*** amotoki_ has quit IRC | 10:50 | |
*** asselin has quit IRC | 10:50 | |
lpetrut | jogo: Hi. We'd like to add support for nova assisted volume snapshots using a similar workflow as the one used by the Libvirt driver. Considering the fact that this is a parity feature, would it require a spec or is a blueprint enough? | 10:51 |
lpetrut | jogo: this is targeting the Hyper-V driver | 10:51 |
jogo | lpetrut: not sure, a blueprint at the very least. You can add the blueprint to the next nova meeting schedule to see if needs a spec or not | 10:52 |
jogo | lpetrut: I am inclined to say a blueprint is enough though | 10:52 |
openstackgerrit | Petrut Lucian proposed openstack/nova-specs: Hyper-V: Add Fibre Channel support https://review.openstack.org/190107 | 10:53 |
lpetrut | jogo: great, thanks! | 10:53 |
openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: fix live migration handling of disk_info https://review.openstack.org/190110 | 10:53 |
*** arnaud____ has quit IRC | 10:54 | |
jogo | danpb: thanks | 10:58 |
danpb | the sooner we objectify the rest of the nova/virt/driver.py API and kill these damn dicts and json strings the better | 10:59 |
openstackgerrit | Ritesh proposed openstack/python-novaclient: Enhance CLI commands for more usability https://review.openstack.org/190111 | 11:00 |
jogo | danpb: amen | 11:00 |
jogo | danpb: good news is the revert worked, so hopefully your patch will work too | 11:00 |
*** yingjun has joined #openstack-nova | 11:02 | |
*** eglynn_ has quit IRC | 11:07 | |
*** prashup has quit IRC | 11:07 | |
*** alexpilotti has joined #openstack-nova | 11:08 | |
*** prashup has joined #openstack-nova | 11:10 | |
*** mwagner_afk has quit IRC | 11:12 | |
*** tojuvone has quit IRC | 11:13 | |
*** otter768 has joined #openstack-nova | 11:14 | |
*** lucasagomes is now known as lucas-hungry | 11:18 | |
alexpilotti | johnthetubaguy: morning, about the supports_migrate_to_same_host, it’s a bit blocking for Hyper-V: https://review.openstack.org/#/c/189798/ | 11:18 |
*** otter768 has quit IRC | 11:18 | |
alexpilotti | johnthetubaguy: in testing resizes on single host | 11:18 |
johnthetubaguy | alexpilotti: sorry, fighting an internal fire right now | 11:19 |
johnthetubaguy | alexpilotti: is that top of your hyper-v ready to merge list? | 11:19 |
alexpilotti | johnthetubaguy: np :-) | 11:20 |
johnthetubaguy | other folks should spot it there | 11:20 |
danpb | alexpilotti: jogo can probably give you the second +2 | 11:20 |
alexpilotti | johnthetubaguy: well, it’s a 6 lines patch which just completes zhangtralon’s patch | 11:20 |
alexpilotti | johnthetubaguy: basically a 0 review effort :-) | 11:21 |
johnthetubaguy | alexpilotti: why has the hyperv CI failed on that? | 11:21 |
johnthetubaguy | ping we when its passed, and I can take a peak | 11:21 |
alexpilotti | johnthetubaguy: the CI has 2 Hyper-V nodes (for live migration) | 11:21 |
danpb | alexpilotti: ha, those words are sure to trigger murphy's law | 11:21 |
* danpb has been burnt too often by patches which are "trivial" :-) | 11:21 | |
johnthetubaguy | alexpilotti: sill would expect that test to be green though, if it doesn't break existing stuff | 11:22 |
johnthetubaguy | danpb: me too | 11:22 |
*** Longgeek has joined #openstack-nova | 11:22 | |
alexpilotti | johnthetubaguy: this patch allows to do resizes on *single* nodes | 11:22 |
johnthetubaguy | alexpilotti: it should still be green | 11:22 |
alexpilotti | johnthetubaguy: I misunderstood your question and good point | 11:22 |
*** prashup has quit IRC | 11:22 | |
alexpilotti | johnthetubaguy: ociuhandu ^ | 11:22 |
alexpilotti | danpb: looks like murphy’s laws have ben triggered retroactively :-) | 11:23 |
*** markvoelker has joined #openstack-nova | 11:24 | |
*** Longgeek_ has joined #openstack-nova | 11:25 | |
*** deepthi has quit IRC | 11:26 | |
*** markvoelker has quit IRC | 11:28 | |
*** Longgeek has quit IRC | 11:28 | |
*** eglynn_ has joined #openstack-nova | 11:30 | |
*** oro has quit IRC | 11:32 | |
*** mpaolino has quit IRC | 11:36 | |
*** varya has joined #openstack-nova | 11:37 | |
*** thorst has joined #openstack-nova | 11:44 | |
*** jcoufal has quit IRC | 11:44 | |
*** rdo has quit IRC | 11:45 | |
*** rdo has joined #openstack-nova | 11:47 | |
*** tobe has quit IRC | 11:47 | |
*** fawadkhaliq has quit IRC | 11:49 | |
*** heyongli- has joined #openstack-nova | 11:49 | |
*** marcusvrn has joined #openstack-nova | 11:51 | |
*** baoli has joined #openstack-nova | 11:51 | |
*** markvoelker has joined #openstack-nova | 11:54 | |
*** dprince has joined #openstack-nova | 11:59 | |
*** romainh has joined #openstack-nova | 12:00 | |
*** jcoufal has joined #openstack-nova | 12:01 | |
*** baoli_ has joined #openstack-nova | 12:01 | |
*** pradipta has quit IRC | 12:02 | |
*** claudiub has joined #openstack-nova | 12:03 | |
openstackgerrit | Andre Aranha proposed openstack/nova-specs: Pass Down the Instance Name to Ironic Driver https://review.openstack.org/189850 | 12:04 |
*** prashup has joined #openstack-nova | 12:04 | |
*** baoli has quit IRC | 12:04 | |
*** subscope has quit IRC | 12:05 | |
*** rushiagr_away is now known as rushiagr | 12:06 | |
*** varya has quit IRC | 12:08 | |
*** fawadkhaliq has joined #openstack-nova | 12:08 | |
openstackgerrit | Matt Thompson proposed openstack/python-novaclient: Pass full path to pkgutil.iter_modules() https://review.openstack.org/190141 | 12:11 |
openstackgerrit | Daniel Berrange proposed openstack/nova: ironic: convert driver to use nova.objects.ImageMeta https://review.openstack.org/189362 | 12:13 |
openstackgerrit | Daniel Berrange proposed openstack/nova: objects: convert config drive to use ImageMeta object https://review.openstack.org/189323 | 12:13 |
openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: convert driver to use nova.objects.ImageMeta https://review.openstack.org/188119 | 12:13 |
openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: make instance compulsory in blockinfo APIs https://review.openstack.org/188118 | 12:13 |
openstackgerrit | Daniel Berrange proposed openstack/nova: hyperv: convert driver to use nova.objects.ImageMeta https://review.openstack.org/189302 | 12:14 |
jogo | danpb: success https://review.openstack.org/#/c/190110/ | 12:14 |
openstackgerrit | Daniel Berrange proposed openstack/nova: virt: convert hardware module to use nova.objects.ImageMeta https://review.openstack.org/188789 | 12:14 |
openstackgerrit | Daniel Berrange proposed openstack/nova: compute: convert manager to use nova.objects.ImageMeta https://review.openstack.org/189685 | 12:14 |
openstackgerrit | Daniel Berrange proposed openstack/nova: xenapi: convert driver to use nova.objects.ImageMeta https://review.openstack.org/188790 | 12:14 |
openstackgerrit | Daniel Berrange proposed openstack/nova: test: add MatchType helper class as equivalent of mox.IsA https://review.openstack.org/190144 | 12:14 |
openstackgerrit | Daniel Berrange proposed openstack/nova: objects: add virtual 'image_meta' property to Instance object https://review.openstack.org/190145 | 12:14 |
openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: convert to use instance.image_meta property https://review.openstack.org/190146 | 12:14 |
jogo | johnthetubaguy sdague: want to review this and fix live migration https://review.openstack.org/#/c/190110/ | 12:14 |
*** aix has quit IRC | 12:14 | |
* danpb wonders if the live migration test is reliable enough yet that we can make it voting | 12:14 | |
openstackgerrit | Vineet Menon proposed openstack/nova-specs: Add flavor tables to API database https://review.openstack.org/190147 | 12:14 |
*** ivasev has joined #openstack-nova | 12:15 | |
openstackgerrit | Romil Gupta proposed openstack/nova: Revert patch-set "VMware: remove unused parameter for VMOPS spawn" https://review.openstack.org/190148 | 12:15 |
jogo | danpb: sadly no | 12:15 |
jogo | danpb: bug https://bugs.launchpad.net/nova/+bug/1445569 and https://bugs.launchpad.net/nova/+bug/1462305 | 12:15 |
openstack | Launchpad bug 1445569 in OpenStack Compute (nova) "No dhcp lease after shelve unshelve" [High,Confirmed] | 12:15 |
openstack | Launchpad bug 1462305 in OpenStack Compute (nova) "multi-node test causes nova-compute to lockup" [Undecided,Incomplete] - Assigned to Joe Gordon (jogo) | 12:15 |
jogo | danpb: using GMR to figure out the second one | 12:15 |
sdague | jogo: +A | 12:16 |
sdague | danpb: thanks for that fix | 12:16 |
danpb | np | 12:16 |
jogo | danpb: if you have any ideas for either of those bugs, help is appreciated I don't have any good leads for either one | 12:17 |
*** sudipto has quit IRC | 12:17 | |
danpb | sdague: do you have review privs for openstack-infra ? i'm looking for someone to have a look at https://review.openstack.org/#/c/186412/ | 12:18 |
danpb | jogo: ok, i'll have a quick look | 12:18 |
*** oro has joined #openstack-nova | 12:19 | |
danpb | jogo: what is the service deployment setup for the multi-node test case ? | 12:19 |
*** aysyanne has joined #openstack-nova | 12:19 | |
danpb | jogo: both nodes running n-cpu and one node running everything else ? | 12:20 |
sdague | danpb: only +1 on that repo, but it looks fine to me | 12:20 |
*** subscope has joined #openstack-nova | 12:20 | |
jogo | danpb: correct | 12:20 |
*** dulek has joined #openstack-nova | 12:20 | |
jogo | danpb: and second node runs n-net too | 12:20 |
*** mwagner_afk has joined #openstack-nova | 12:21 | |
danpb | ok | 12:21 |
jogo | danpb: see http://logs.openstack.org/10/190110/1/check/check-tempest-dsvm-multinode-full/e0b5720/logs/subnode-2/ | 12:21 |
*** ildikov_ has quit IRC | 12:25 | |
*** raildo has joined #openstack-nova | 12:25 | |
*** bradjones has quit IRC | 12:26 | |
*** garyk has quit IRC | 12:26 | |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: [WIP] cleanup openstack-common.conf and sync updated files https://review.openstack.org/189118 | 12:27 |
*** bradjones has joined #openstack-nova | 12:27 | |
*** bradjones has quit IRC | 12:27 | |
*** bradjones has joined #openstack-nova | 12:27 | |
sdague | jogo: and the metadata service | 12:27 |
kashyap | Can any other Nova core take a look at this? - https://review.openstack.org/#/c/181781/ | 12:28 |
kashyap | (Fixes a crazy race bug, frequently tops the Elastic Recheck charts.) | 12:29 |
*** amotoki has joined #openstack-nova | 12:30 | |
*** mpaolino has joined #openstack-nova | 12:30 | |
*** pilgrimstack has joined #openstack-nova | 12:31 | |
*** claudiub has quit IRC | 12:32 | |
* jogo looks | 12:32 | |
*** marun has joined #openstack-nova | 12:32 | |
jogo | kashyap: done | 12:33 |
kashyap | jogo: Excellent, thanks. | 12:34 |
openstackgerrit | Vineet Menon proposed openstack/nova-specs: Add flavor tables to API database https://review.openstack.org/190147 | 12:36 |
*** VW has joined #openstack-nova | 12:37 | |
*** eharney has joined #openstack-nova | 12:37 | |
*** arnaud____ has joined #openstack-nova | 12:38 | |
*** VW has quit IRC | 12:39 | |
*** rfolco has joined #openstack-nova | 12:40 | |
openstackgerrit | Vineet Menon proposed openstack/nova-specs: Add flavor tables to API database https://review.openstack.org/190147 | 12:40 |
*** VW has joined #openstack-nova | 12:40 | |
jogo | kashyap: next big step is to fix the two known multinode bugs so we can gate on it | 12:40 |
jogo | kashyap: if you have any ideas for either of them help is appreciated | 12:40 |
kashyap | jogo: Yeah, I have the two bugs open in tabs. | 12:41 |
jogo | kashyap: excellent, I am at a loss for root causes for both of them at the moment | 12:41 |
kashyap | jogo: You're testing with nova-net, you also have any jobs with Neutron-net? | 12:41 |
jogo | kashyap: there is a neutron job somewhere but I am ignoring it for now | 12:41 |
jogo | check-tempest-dsvm-neutron-multinode-full | 12:42 |
gilliard | jogo: where are the repos which define the devstack setup for those? | 12:42 |
jogo | gilliard: http://git.openstack.org/cgit/openstack-infra/devstack-gate | 12:43 |
jogo | and git.openstack.org/cgit/openstack-infra/project-config/ | 12:43 |
gilliard | thanks | 12:43 |
*** arnaud____ has quit IRC | 12:43 | |
kashyap | jogo: About that lockup bug, I'll see if I can make a reproducer on my DevStack+Neutron two node setup here. Trying to understand if there's any kind of predictable trigger. | 12:43 |
jogo | gilliard: n-net has nova multihost enabled, and no shared storage | 12:43 |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/python-novaclient: [WIP] cleanup openstack-common.conf and sync updated files https://review.openstack.org/189119 | 12:44 |
jogo | kashyap: that would be helpful, it doesn't happen to often so may take some time | 12:44 |
jogo | kashyap: build_name:"check-tempest-dsvm-neutron-multinode-full" AND (message:"Finished: FAILURE" OR message:"Finished: SUCCESS") | 12:45 |
jogo | only one successful run in 7 days for the neutron job though ... | 12:45 |
*** salv-orlando has quit IRC | 12:45 | |
jogo | so that may be chewing too much off in one bite | 12:46 |
kashyap | jogo: To begin, I'll try by making a test that repeatedly does live block migration. Wonder if that sounds sensible | 12:46 |
*** ildikov_ has joined #openstack-nova | 12:46 | |
jogo | kashyap: yes, but not sure if neutron is stable enough even for that | 12:46 |
jogo | my hunch is the two bugs we are seeing are nova-net related | 12:47 |
jogo | so if you test neutron you will hit a separate class of bugs (ones that need fixing as well) | 12:47 |
kashyap | Yep, noted. The thing is, I hardly ever ran nova-net, just so that I can get used to debugging OVS w/ Neutron. | 12:47 |
jogo | kashyap: well if you want to work on getting the neutron version of the job working, by all means go for it | 12:48 |
jogo | it uses DVR in neutron AFAIK | 12:48 |
jogo | and I don't remember of the top of my head who is working on it over there | 12:48 |
kashyap | jogo: Prefer to begin with small steps. Not sure if this DVR thing _necessary_. I'd begin with a simple OVS+GRE/VXLAN env - that's complex enough. | 12:49 |
jogo | kashyap: you should talk to the neutron team and see who is working on it over there | 12:50 |
jogo | kashyap: yeah ++ to start simple, I think they bit off too much at first myself | 12:50 |
kashyap | jogo: Yep, once I have done some tests and seen some real failures. | 12:50 |
*** aix has joined #openstack-nova | 12:50 | |
jogo | there are plenty of real failures for you to play with | 12:51 |
*** jcoufal has quit IRC | 12:51 | |
*** pilgrimstack has quit IRC | 12:52 | |
jogo | kashyap: another useful thing is what gilliard is working on https://review.openstack.org/#/c/189733/ | 12:52 |
*** jcoufal has joined #openstack-nova | 12:52 | |
jogo | kashyap: which is to make the connectivity tests pass (I think the tempest code may have bit rotted along with real bugs) | 12:52 |
*** jcoufal is now known as Guest62178 | 12:52 | |
*** ociuhandu has joined #openstack-nova | 12:53 | |
gilliard | I'm working with a couple of the -qa team on that... | 12:53 |
*** lucas-hungry is now known as lucasagomes | 12:53 | |
kashyap | jogo: Thanks for the pointer, will look. Just torn between multiple bugs right now. Who isn't? :-) | 12:53 |
jogo | gilliard: awesome! | 12:54 |
*** ociuhandu has quit IRC | 12:54 | |
jogo | gilliard: looking at http://logs.openstack.org/33/189733/1/check/check-tempest-dsvm-full/301a679/logs/testr_results.html.gz when things fail we don't colelct console logs for most of those jobs | 12:55 |
jogo | gilliard: getting that info would make debugging a lot easier | 12:55 |
*** Guest62178 has quit IRC | 12:55 | |
gilliard | jogo kashyap - https://blueprints.launchpad.net/tempest/+spec/ssh-auth-strategy << includes pinging vms | 12:57 |
kashyap | Yep - will look. | 12:59 |
*** cbader has joined #openstack-nova | 13:03 | |
jogo | gilliard: nice! | 13:05 |
*** nihilifer_ has quit IRC | 13:06 | |
lpetrut | Hi, could anyone please take a look over this easy fix targeting the Libvirt SMB volume driver? https://review.openstack.org/#/c/187619/ We have some issues w/ our Cinder CI because of this and would like to get this in as soon as possible | 13:06 |
lpetrut | it's merely 40 LOC | 13:06 |
*** ferest has joined #openstack-nova | 13:06 | |
jogo | gilliard: anything I can do to help with that work? | 13:08 |
*** garyk has joined #openstack-nova | 13:08 | |
BobBall | garyk: Thanks to sdague the xenserver CI is back working now. I've requeued all failures that were affected by this bug. | 13:10 |
garyk | BobBall: sdague: gracias. "you will never walk alone!" | 13:11 |
*** mriedem has joined #openstack-nova | 13:13 | |
jlanoux | jogo: you can always harass qa cores to review and approve the changes :) | 13:13 |
mriedem | danpb: i figured out why we weren't getting volume encryption metadata back from cinder for the ceph case - it's pretty simple actually, | 13:14 |
*** otter768 has joined #openstack-nova | 13:15 | |
mriedem | danpb: nagyz: the problem is the rbd volume driver in cinder doesn't set the 'encrypted' key in the connection_info sent back during initialize_connection like the isci volume driver does | 13:15 |
*** ajayaa has quit IRC | 13:15 | |
mriedem | that's the flag that nova is checking before calling the cinder API | 13:15 |
*** mdrabe has joined #openstack-nova | 13:15 | |
mriedem | so while i could 'fix' that in cinder, there are many volume drivers that don't set it and i'm not sure if that's on purpose or not - the cinder API shows if a volume is encrypted or not | 13:15 |
mriedem | my guess is this is some leftover tight coupling from when the volume service was in nova, but that was before my time | 13:16 |
mriedem | jdurgin: fyi since you commented on the bug ^ | 13:18 |
*** otter768 has quit IRC | 13:19 | |
*** jcoufal_ has joined #openstack-nova | 13:19 | |
*** rushiagr is now known as rushiagr_away | 13:20 | |
danpb | mriedem: yeah, i commented to this effect on the logging patch too | 13:20 |
danpb | mriedem: merely setting the encrypted flag isn't going to do anything useful though | 13:20 |
danpb | because nova will not have any way of supporting encryption with rbd | 13:21 |
*** tojuvone has joined #openstack-nova | 13:21 | |
danpb | at least not without some significant dev work | 13:21 |
mriedem | danpb: yeah, that's what nagyz is planning on | 13:21 |
mriedem | and confirmed that it blows up https://review.openstack.org/#/c/190053/ | 13:21 |
danpb | i think it is really serious bug that you can request encryption via an api call and have that request totally ignored | 13:21 |
mriedem | in cinder? | 13:22 |
danpb | we really must return an error to the user if they request something that cannot be honoured | 13:22 |
danpb | yeah | 13:22 |
mriedem | yeah, that was my justification for the bug | 13:22 |
*** claudiub has joined #openstack-nova | 13:22 | |
sdague | garyk: https://review.openstack.org/#/c/170054/2/nova/virt/vmwareapi/vmops.py,cm - I'm confused on that change, because it seems to be equivalent | 13:22 |
danpb | mriedem: which bug # | 13:22 |
mriedem | danpb: https://bugs.launchpad.net/cinder/+bug/1463525 | 13:22 |
openstack | Launchpad bug 1463525 in Cinder "There is no volume encryption metadata for rbd-backed volumes" [Undecided,Triaged] - Assigned to Matt Riedemann (mriedem) | 13:22 |
*** yingjun has quit IRC | 13:23 | |
mriedem | danpb: it's almost like the cinder volume drivers need capabilities bubbled up to the volume manager to tell it if they support encryption, like what we do with the virt drivers | 13:23 |
mriedem | then if a request is made to create a volume from an encrypted volume type, if the driver doesn't support it it should blow up | 13:23 |
mriedem | i also poked around in the connection_info bits in some of the various cinder drivers, and oh those are not fun if we want to try and model some of that with objects in nova | 13:24 |
*** ociuhandu has joined #openstack-nova | 13:25 | |
mriedem | plus those are not considered any part of the cinder API, it's just a black box and you get what the vendor driver gives you, and you'd better hope it doesn't change | 13:25 |
danpb | mriedem: yeah, thta's exactly why we must model them as objects - this free-for-all cannot continue | 13:25 |
asselin_ | is this wiki page spec implemented? I couldn't find any docs on it. https://wiki.openstack.org/wiki/Spec-provider-firewall | 13:26 |
danpb | we'll hopefully have the VIF stuff modelled as objects soon too | 13:26 |
mriedem | i think i'm going to start by just trying to document the connection_info that comes back from the cinder drivers to get an idea for what's common | 13:26 |
danpb | mriedem: yeah, that's a good idea | 13:26 |
*** mhorban has quit IRC | 13:27 | |
*** changbl_ has quit IRC | 13:27 | |
*** jgrimm is now known as zz_jgrimm | 13:28 | |
dansmith | jogo: thanks! | 13:28 |
nagyz | hey danpb mriedem | 13:28 |
*** mhorban has joined #openstack-nova | 13:29 | |
nagyz | right I just made the RBD driver to pass it through - I still need to write the nova part to handle it well | 13:29 |
nagyz | I'm a bit unsure where that piece should go to inside nova - certainly not into the luks/cryptsetup handlers? | 13:29 |
*** annashen has joined #openstack-nova | 13:30 | |
mriedem | nagyz: i have no idea, apparently we simply just don't have shared storage volume encryption support at all | 13:30 |
mriedem | so this would be new | 13:30 |
*** VW_ has joined #openstack-nova | 13:30 | |
*** VW has quit IRC | 13:31 | |
*** exploreshaifali has joined #openstack-nova | 13:31 | |
garyk | sdague: the issues is when the disk_bus is an empty string. if that is the case then the empty string will be returned and the spawn will be failed. hecne we need the or | 13:31 |
*** ociuhandu has quit IRC | 13:31 | |
BobBall | Can I beg for core reviews on https://review.openstack.org/#/c/189397/ ? It seems to be causing a race which is hurting the XenServer CI at around a 10-20% rate. | 13:31 |
*** ijuwang has joined #openstack-nova | 13:31 | |
jroll | johnthetubaguy: are there notes from the ironic/nova-compute session? and how do you feel about a spec on that being in ironic/nova/both, where should that live (I assume nova)? | 13:32 |
*** jcoufal_ has quit IRC | 13:32 | |
*** teju has joined #openstack-nova | 13:32 | |
sdague | garyk: gotcha | 13:32 |
*** jcoufal_ has joined #openstack-nova | 13:33 | |
mriedem | BobBall: +2 | 13:34 |
BobBall | Yay - thanks :) | 13:34 |
*** annashen has quit IRC | 13:35 | |
nagyz | mriedem, I was thinking to reuse the iSCSI device mapping code that I'm sure happens for iSCSI volumes at the moment? | 13:35 |
nagyz | that's already in tree | 13:35 |
mriedem | BobBall: gross, looks like this also needs to be backported to at least kilo and maybe juno? | 13:35 |
danpb | if you wanted todo that, you'd really need to make cinder return a connection_info with the iscsi driver type, instead of rbd type | 13:35 |
danpb | but that seems like the wrong way to go | 13:35 |
danpb | instead I think you'd want to use the rbd kernel module so that the compute host makes a direct connection to the rbd server | 13:36 |
*** rotbeard has quit IRC | 13:36 | |
danpb | you can then layer the dm crypt on top of that | 13:36 |
*** ferest has quit IRC | 13:36 | |
*** ankit_ag has quit IRC | 13:36 | |
danpb | so you keep iscsi out of the loop still, avoiding that performance hit in the i/o path | 13:36 |
mriedem | BobBall: any idea what changed to tickle that bug? | 13:37 |
*** ociuhandu has joined #openstack-nova | 13:37 | |
danpb | nagyz: ^^ | 13:37 |
nagyz | danieru, that's exactly what I want to do! | 13:37 |
nagyz | danpb | 13:37 |
*** vladikr has joined #openstack-nova | 13:37 | |
BobBall | Good question... | 13:37 |
*** vishwanathj has joined #openstack-nova | 13:37 | |
BobBall | No - I assumed once alaski figured out what was going on that it was recently introduced | 13:37 |
nagyz | danpb, based on the data returned from cinder we should be able to distinguish between an iSCSI and an rbd connection, right? so we can handle it per-case | 13:37 |
mriedem | BobBall: maybe doesn't matter, the code in the bw usage object is wrong, and it goes back to juno so probably needs to be backported either way | 13:37 |
danpb | the cryptsetup part of nova is all in nova/volume/encryptors | 13:38 |
*** oro has quit IRC | 13:38 | |
alaski | BobBall: mriedem it was introduced when the bw usage task moved to objects | 13:38 |
BobBall | Perhaps, but it does raise the question of if actually this is just a symptom rather than the real issue... Which might mean it might not fix the race we're hitting | 13:38 |
alaski | unfortunately that's not at all tested outside of xen CI | 13:38 |
*** vishwana_ has joined #openstack-nova | 13:39 | |
danpb | nagyz: yes connection_info.get('driver_volume_type') | 13:39 |
BobBall | Oh? Why does this CI test this and not others? | 13:39 |
*** tojuvone has quit IRC | 13:40 | |
alaski | most/all(?) other virt drivers bail at http://git.openstack.org/cgit/openstack/nova/tree/nova/compute/manager.py#n5487 | 13:40 |
BobBall | Ah - I see. Makes sense. | 13:41 |
*** vishwanathj has quit IRC | 13:42 | |
*** fawadkhaliq has quit IRC | 13:42 | |
nagyz | danpb, so do you think that in the cryptsetup part let's check the driver_volume_type and if it's rbd then do the kernel mapping? I haven't looked at the code there but there must be a similar code path already for iSCSI in place, right? | 13:43 |
*** pbandzi has quit IRC | 13:43 | |
*** apuimedo has quit IRC | 13:47 | |
danpb | nagyz: i don't think you can do it entirely within the encryptors code - the libvirt volume.py needs to understand this too | 13:47 |
*** dboik_ has quit IRC | 13:48 | |
nagyz | danpb, let me dig around and see what I come up with. I'll shout if I hit a block (which could be soon! :P) | 13:49 |
*** jcoufal_ has quit IRC | 13:49 | |
openstackgerrit | Yusuke Ide proposed openstack/python-novaclient: Add help message for secgroup-add/del-default-rule https://review.openstack.org/189263 | 13:51 |
nagyz | danpb, how come there is both a luks and a cryptsetup one? | 13:51 |
abhishekk | claudiub: hi, you around? | 13:51 |
claudiub | abhishekk: hello | 13:51 |
*** Murali has quit IRC | 13:51 | |
*** zzzeek has joined #openstack-nova | 13:51 | |
danpb | nagyz: i dunno really | 13:52 |
abhishekk | claudib: this is related to your comment on https://review.openstack.org/#/c/183404/2/nova/compute/manager.py | 13:52 |
claudiub | danpb: hello. You have a +2 on this commit and it already has 2x +2. Can you give it a +A? https://review.openstack.org/#/c/189798/1 | 13:52 |
claudiub | abhishekk: ok, thanks! taking a look | 13:53 |
abhishekk | claudiub: yes please | 13:53 |
*** matrohon has quit IRC | 13:53 | |
mriedem | nagyz: i've seen joel coffman's name in the volume encryption stuff in nova and in cinder, he might have some more background, but i don't know his irc nick | 13:56 |
*** igordcard_ has quit IRC | 13:56 | |
*** igordcard_ has joined #openstack-nova | 13:57 | |
*** doude has quit IRC | 13:57 | |
dansmith | it's like jcoffman | 13:57 |
dansmith | and yeah, he'd be the one to ask I think | 13:57 |
*** doude has joined #openstack-nova | 13:58 | |
*** VW_ has quit IRC | 13:58 | |
mriedem | danpb: i sent a thing to the ML asking the cinder devs for help on documenting this connection_info mess | 13:59 |
mriedem | hopefully i get someone to help | 13:59 |
*** cfriesen has joined #openstack-nova | 13:59 | |
mriedem | i didn't really want our object to be made up of 2 fields: driver_volume_type (string) and data (dict of strings) | 13:59 |
mriedem | honestly cinder should really have an object modeling that on their side that is versioned and the volume drivers should be using it, and if someone adds/drops/renames a field it would bump the version | 14:00 |
mriedem | and the caller (nova) would have the version for compat handling | 14:00 |
mriedem | like actually store the version of the connection_info object in cinder and return that in os-initialize_connection | 14:01 |
openstackgerrit | Andreas Scheuring proposed openstack/nova-specs: Spec to add 'libvirt_net' as vif type to novas libvirt driver. https://review.openstack.org/190198 | 14:02 |
*** pilgrimstack has joined #openstack-nova | 14:02 | |
mriedem | maybe that's something that should go in brick so both nova and cinder would be using the same thing | 14:02 |
*** pilgrimstack has quit IRC | 14:02 | |
mriedem | since it's common and could live in a library | 14:02 |
*** pilgrimstack has joined #openstack-nova | 14:02 | |
*** subscope has quit IRC | 14:04 | |
*** VW has joined #openstack-nova | 14:05 | |
*** tojuvone has joined #openstack-nova | 14:05 | |
*** mpaolino has quit IRC | 14:07 | |
danpb | mriedem: indeed - this is the case where we'd define a ConnectionInfo base object, and then subclass it for RBDConnectionInfo, ISCSIConnectionInfo, etc and have everything formally specified | 14:08 |
danpb | some static method on ConnectInfo would look at driver_volume_type to decide which subclass to instantiate perhaps | 14:08 |
*** kiran-r has quit IRC | 14:08 | |
*** dboik has joined #openstack-nova | 14:09 | |
johnthetubaguy | jroll: we have notes in the etherpad, linked from the wiki, but I think we need a nova spec for the nova work, we could probably merge something in the docs around the longer term vision, but I don't have the structure in place for that yet I am afraid | 14:09 |
alex_xu | dansmith: hi, good morning. I read you talk with cfriesen about cpu model yesterday. As my understand you prefer an API can be used to define what cpu model and what cpu features in that model. Then scheduler choice one host can support that model and those features. Then compute-node build up an instance with that model and those features. | 14:10 |
alex_xu | However our propose is reversed with your propose. We propose nova-compute report supported cpu models and features in those models. Then user choice one of it to boot up that instance. Did I understand right? | 14:11 |
*** abhishekk has quit IRC | 14:11 | |
BobBall | danpb: Thanks for the second review (https://review.openstack.org/#/c/189397/); Just checking - are you happy to +A or do you think it needs more eyes? (I suspect you and mriedem reviewed in parallel) | 14:11 |
mriedem | danpb: i'm writing up some ideas on that to the ML now for cinder devs - and i'm going to be bugging jungleboyj about this when he gets into the office since he can't physically hide from me | 14:11 |
danpb | BobBall: ah yes | 14:11 |
BobBall | Joy of joys! | 14:12 |
BobBall | Thanks. | 14:12 |
dansmith | alex_xu: yes, but the problem is not that the models are sourced from the compute, the problem is that the models reported from the compute are not consistent because hypervisor versions, technologies, etc will report different things under the same name | 14:12 |
*** aloga has quit IRC | 14:13 | |
alex_xu | dansmith: ok, I see now, your propose force all hypervisor create instance with the specific model and features. Then resolve all the problem created by inconsistent between hyperivsors, even different version hyepervisors | 14:14 |
*** dguitarbite has joined #openstack-nova | 14:14 | |
dansmith | alex_xu: yes | 14:14 |
*** aloga has joined #openstack-nova | 14:15 | |
dansmith | alex_xu: you can report flags up so the scheduler can place what was requested, but you have to do that by flags and not by the arbitrary model name | 14:15 |
alex_xu | dansmith: ah, thanks, I see now:) and the API you expected is jaypipes proposed resource template? | 14:15 |
dansmith | alex_xu: no, I don't think jaypipes has proposed the flavors/options spec yet, AFAIK | 14:15 |
danpb | dansmith: using flags exclusively only works on x86 | 14:15 |
*** pilgrimstack has quit IRC | 14:16 | |
danpb | dansmith: non-x86 architectures don't really have the same fine grained CPUID flags concept | 14:16 |
danpb | dansmith: so for those the only option is using cpu model names | 14:16 |
*** VW has quit IRC | 14:16 | |
danpb | if we want to expose the ability to influence placement based on cpu, I'm not sure i see any solution which is portable across hypervisors | 14:17 |
dansmith | danpb: we could report flags that correspond to architectures from the compute when we're on such a platform, right? like ppc64native or whatever | 14:17 |
*** amotoki has quit IRC | 14:17 | |
*** amotoki has joined #openstack-nova | 14:17 | |
*** amotoki has quit IRC | 14:18 | |
danpb | dansmith: how does that let the user chose between different ppc64 cpu models with differing capabilities | 14:18 |
alex_xu | dansmith: yea, jaypipes didn't propose that yet. But the concept is resource template, right? | 14:18 |
*** jecarey has joined #openstack-nova | 14:18 | |
danpb | i guess you could introduce an abstraction so that only the cloud admin needs to care about hypervisor specific cpu differences | 14:19 |
dansmith | danpb: you meant that on PPC we only have the machine type, right? | 14:19 |
*** thangp has joined #openstack-nova | 14:19 | |
*** amotoki has joined #openstack-nova | 14:19 | |
jroll | johnthetubaguy: found it, ty | 14:19 |
dansmith | the -M flag to qemu I mean | 14:19 |
danpb | dansmith: no, i meant that on PPC you have a huge choice of cpus http://paste.fedoraproject.org/230740/14339459 | 14:19 |
dansmith | danpb: right, so expose that machine type as a flag | 14:20 |
danpb | and you can't express preferences based on CPU feature names | 14:20 |
*** mpaolino has joined #openstack-nova | 14:20 | |
dansmith | danpb: on intel you may ask for sse4 | 14:20 |
danpb | that's not a machine type - that's cpu model names | 14:20 |
dansmith | danpb: on ppc you may ask for 970mp | 14:20 |
*** amotoki has quit IRC | 14:20 | |
dansmith | fine, whatever, they're just symbols that the compute node exposes as capabilities | 14:20 |
*** amotoki has joined #openstack-nova | 14:20 | |
danpb | those aren't standardized across hyprevisors in any way | 14:20 |
danpb | so we still have the same problem | 14:21 |
*** mtanino has joined #openstack-nova | 14:21 | |
dansmith | we have many fewer hypervisors that can do non-intel architectures, | 14:21 |
danpb | i could see one way to solve this is to not expose any concept of cpu models and features to the tenant user at all | 14:22 |
dansmith | and those that do are massively limited in the CPUs they could expose I think, so just standardizing those model names seems like a reasonable thing for us to do internally to nova | 14:22 |
danpb | instead create a "CPU" object abstraction that the users choose | 14:22 |
danpb | and let the cloud admin populate those abstractions with the hypervisor specific cpu model names / features | 14:22 |
danpb | so the user would see & choose "Xeon with SSE3" | 14:22 |
dansmith | danpb: we still need a way for them to control how to find a cpu like that, but yes that's the heavier approach I'd really like us to take | 14:23 |
danpb | and the cloud admin would have setup a mapping so that uses CPUBlah on libvirt and CPUFoo on vmware | 14:23 |
dansmith | danpb: doing it just based on flags is a medium approach between that and just cheaping out and using the model names the hypervisors provide | 14:23 |
danpb | so the cloud admin still gets the pain of dealing with hypervisor specific setup | 14:23 |
danpb | but the tenant user is isolated from it | 14:23 |
*** READ10 has quit IRC | 14:23 | |
dansmith | yes, this is what I originally asked for in the summit conversation | 14:23 |
dansmith | although I think it needs to be modeled in a generic way, like what jaypipes was suggesting as options that sit on top of a flavor | 14:24 |
dansmith | but yes | 14:24 |
danpb | a much bigger bit of work, but it feels like the right approach if we want to islate the user from low level details | 14:24 |
dansmith | yep | 14:24 |
*** arnaud____ has joined #openstack-nova | 14:27 | |
*** nihilifer has quit IRC | 14:28 | |
*** nkrinner has quit IRC | 14:29 | |
*** READ10 has joined #openstack-nova | 14:29 | |
*** varya has joined #openstack-nova | 14:30 | |
alex_xu | dansmith: danpb , so you are both like flavor options? | 14:30 |
dansmith | alex_xu: note that we're talking about a new model here, not options == extra_specs | 14:31 |
alex_xu | dansmith: yea, got that | 14:31 |
*** yingjun has joined #openstack-nova | 14:32 | |
*** arnaud____ has quit IRC | 14:32 | |
alex_xu | dansmith: so user can use that API specific an fake model name like "Xeon with SSE3" and that includes a set of features. | 14:33 |
dansmith | alex_xu: yes, but the model has to expose useful information I think, | 14:33 |
dansmith | so it still needs to expose CPUID flags in the intel case, IMHO | 14:33 |
dansmith | so that you can actually know what is included in that model | 14:33 |
dansmith | "Xeon with SSE3" might be missing some other flag like vmx which is masked out | 14:34 |
alex_xu | dansmith: yea, but what about the real cpu model used for guest? Like libvirt don't know "Xeon with SSE3", how hypervisor decide which real model used for guest? | 14:34 |
dansmith | alex_xu: sure, you still have to turn that into something useful | 14:37 |
openstackgerrit | Andrew Laski proposed openstack/nova: DNM Cells: Provide example boot request flow for reference https://review.openstack.org/190209 | 14:38 |
alex_xu | dansmith: ok so that also an important of live-migratable. If a host finally choice Haswell, but on old host didn't know that model name, then we can't live migration also. | 14:39 |
openstackgerrit | sahid proposed openstack/nova: libvirt: introduce method to get domain XML https://review.openstack.org/189307 | 14:39 |
openstackgerrit | sahid proposed openstack/nova: libvirt: introduce method detach_device to Guest object https://review.openstack.org/188765 | 14:39 |
openstackgerrit | sahid proposed openstack/nova: libvirt: introduce GuestBlock to wrap around Block API https://review.openstack.org/190210 | 14:39 |
openstackgerrit | sahid proposed openstack/nova: libvirt: update swap volume to use Guest https://review.openstack.org/190211 | 14:39 |
openstackgerrit | sahid proposed openstack/nova: libvirt: update live swnapshot to use Guest https://review.openstack.org/190212 | 14:39 |
openstackgerrit | sahid proposed openstack/nova: libvirt: update volume delete swnapshot to use Guest https://review.openstack.org/190213 | 14:39 |
openstackgerrit | sahid proposed openstack/nova: libvirt: introduce method to wait block device job https://review.openstack.org/190214 | 14:39 |
alex_xu | so not only need care about flags | 14:39 |
dansmith | alex_xu: I don't think this will affect the existing libvirt mechanisms and restrictions | 14:40 |
dansmith | this is all about knowing what to ask libvirt for, but in the end it will still be the same thing, AFAIK | 14:40 |
*** openstackgerrit has quit IRC | 14:41 | |
danpb | yeah, this is only about what the user sees | 14:41 |
*** openstackgerrit has joined #openstack-nova | 14:41 | |
danpb | we'd convert this into the libvirt specific format the moment we use it | 14:41 |
danpb | so everything in libvirt will continue to just work | 14:42 |
dansmith | yeah | 14:42 |
alex_xu | emm...ok, I may misunderstand something, let me think about and re-read the talk between you | 14:43 |
alex_xu | dansmith: danpb anyway, thanks for the talk :) | 14:44 |
dansmith | alex_xu: np, thanks for working on this | 14:44 |
alex_xu | :) | 14:44 |
*** electrocucaracha has joined #openstack-nova | 14:45 | |
cfriesen | dansmith: danpb: so are we saying that this feature would have to wait on some generic modelling framework to replace the free-form flavor extra-specs and image metadata? | 14:45 |
danpb | no, its something that should be modelled separately from flavours imho | 14:47 |
danpb | a standard alone object in its own right | 14:47 |
danpb | the flavour extra-specs would merely refer to it via some unique id key | 14:47 |
jaypipes | guh, so much I need to read in scrollback... late start today, sorry guys :( | 14:47 |
*** tsekiyama has joined #openstack-nova | 14:47 | |
danpb | or likewise the image meta could refer to it in the same way | 14:47 |
lxsli | jaypipes: morning! | 14:47 |
jaypipes | is it? :) | 14:48 |
lxsli | jaypipes: I'm rewriting your resource objects spec in terms of inventory objects, this should be a much smaller change than my previous patch (conceptually at least) | 14:48 |
lxsli | jaypipes: somewhere! | 14:48 |
cfriesen | danpb: right, but from what I see the suggestion is that the cpu model/flavor feature is blocked until such an entity exists? | 14:48 |
jaypipes | lxsli: cool, thank you sir. | 14:48 |
danpb | cfriesen: yep, looks that way | 14:49 |
*** lan_ has joined #openstack-nova | 14:49 | |
cfriesen | danpb: dansmith: and based on dansmith's comments last night nobody is actively working on such an entity? | 14:50 |
dansmith | danpb: I'm not sure that the model gets selected via extra_specs, actually, but *something* | 14:50 |
*** diogogmt has joined #openstack-nova | 14:50 | |
dansmith | danpb: you missed the discussion at summit about how this might look in a general way for adding options to flavors that are orthogonal to normal cpu/mem/disk things | 14:50 |
*** VW has joined #openstack-nova | 14:50 | |
dansmith | cfriesen: *you* can work on it to unblock yourself :) | 14:51 |
danpb | dansmith: ok, sure, how we attach them to flavours doesn't matter - just that there's some association there | 14:51 |
dansmith | danpb: yep | 14:51 |
cfriesen | dansmith: yeah, I get that. but realistically I don't think I have that much time to put on this. | 14:52 |
cfriesen | dansmith: alex might be able to help | 14:53 |
dansmith | yep, and he has api chops | 14:53 |
alex_xu | yea, I'm interesting, but I need figure out more~ | 14:53 |
dansmith | alex_xu: you are definitely interesting :D | 14:53 |
alex_xu | dansmith: yes...that's trure | 14:54 |
alex_xu | s/trure/true | 14:54 |
alex_xu | anyway bad news is we block by something, good news is there is something have agreement :) | 14:55 |
nagyz | danpb, mriedem, what you guys discussed would be a longer plan right? :-) if I add an RBDVolumeDriver to libvirt's volume.py and reuse that from the encryptor to actually attach/detach in the encrypted case, is that ok? currently it's using the base variant to get the xml back | 14:55 |
openstackgerrit | Chris Friesen proposed openstack/nova: fix "down" nova-compute service spuriously marked as "up" https://review.openstack.org/163060 | 14:56 |
mriedem | nagyz: today for rbd the LibvirtNetVolumeDriver is used | 14:56 |
*** danpb has quit IRC | 14:56 | |
*** mmedvede has joined #openstack-nova | 14:56 | |
dansmith | alex_xu: well, we need a spec for agreement, but yeah | 14:56 |
openstackgerrit | Vineet Menon proposed openstack/nova-specs: Add flavor tables to API database https://review.openstack.org/190147 | 14:56 |
*** belmoreira has quit IRC | 14:57 | |
bauzas | that flavor stuff wasn't something which was defined as "need an owner/plan" ? :) | 14:57 |
nagyz | mriedem, how come there are iscsi bits in the LibvirtNetVolumeDriver yet there is a separate ISCSIVolumeDriver as well? | 14:57 |
* alex_xu still search dictionary about what is 'chops' mean... | 14:57 | |
*** danpb has joined #openstack-nova | 14:57 | |
gilliard | chops == strength, in this context :) | 14:57 |
mriedem | nagyz: i have no idea, i've been looking at the volume drivers in nova and wondering the same about the lack of hierarchy | 14:58 |
* alex_xu is going to search "strength" | 14:58 | |
dansmith | alex_xu: yes, I meant "alex_xu is skilled at API stuff, and this will be easier for him than some other people" | 14:58 |
*** dguitarbite has quit IRC | 14:58 | |
* alex_xu just joke | 14:58 | |
dansmith | heh | 14:58 |
mriedem | nagyz: because i'm not exactly sure the best way to model these volume connections, e.g. local vs shared vs fibre channel | 14:59 |
*** dguitarbite has joined #openstack-nova | 14:59 | |
mriedem | i don't want to model everything per vendor, since there are dozens of vendor drivers in cindder | 14:59 |
mriedem | *cinder | 14:59 |
*** afazekas has quit IRC | 14:59 | |
nagyz | mriedem, I think the current grouping is fine it's just confusing to have a generic net driver that has rbd and iscsi bits while there is a separate iscsi implementation there | 14:59 |
nagyz | are you against introducing a separate RBDVolumeDriver that can be told if it's an encrypted volume or not and if it's non-encrypted it just calls the base method and if it is then it does the local attach/detach? | 15:00 |
*** lan__ has joined #openstack-nova | 15:00 | |
*** nelsnelson has quit IRC | 15:01 | |
*** exploreshaifali has quit IRC | 15:01 | |
mriedem | nagyz: i'm not against that no | 15:02 |
mriedem | the volume drivers are mapped in the libvirt driver.py | 15:02 |
mriedem | today rbd driver_volume_type is mapped to LibvirtNetVolumeDriver - that could change to RbdVolumeDriver which extends LibvirtNetVolumeDriver | 15:02 |
*** dims_ has quit IRC | 15:03 | |
*** lan_ has quit IRC | 15:03 | |
*** lan_ has joined #openstack-nova | 15:04 | |
openstackgerrit | Dan Smith proposed openstack/nova: API support for migration_type https://review.openstack.org/183200 | 15:05 |
openstackgerrit | Dan Smith proposed openstack/nova: Make live migration create a migration object record https://review.openstack.org/183331 | 15:05 |
openstackgerrit | Dan Smith proposed openstack/nova: Make evacuate leave a record for the source compute host to process https://review.openstack.org/183354 | 15:05 |
*** kmartin has joined #openstack-nova | 15:05 | |
*** lan__ has quit IRC | 15:06 | |
*** lan__ has joined #openstack-nova | 15:06 | |
*** tjones1 has joined #openstack-nova | 15:06 | |
mriedem | ndipanov: danpb: why is it that the connection_info for NFS backends in cinder have a 'mount_point_base' but in the nova libvirt volume module for NFS volumes we also use our own config options, e.g. nfs_mount_point_base, glusterfs_mount_point_base, etc - couldn't we just get that info from cinder after initializing the connection? | 15:07 |
mriedem | i'm totally confused why you have to configure nova for this stuff when it appears that cinder is already passing it back to nova | 15:07 |
*** lan__ has quit IRC | 15:08 | |
*** dims has joined #openstack-nova | 15:08 | |
*** dims has quit IRC | 15:08 | |
*** dims has joined #openstack-nova | 15:08 | |
*** lan_ has quit IRC | 15:09 | |
ndipanov | mriedem, no idea - probably | 15:10 |
ndipanov | some of the cinder stuff might be internal to cinder | 15:10 |
*** patrickeast has joined #openstack-nova | 15:10 | |
ndipanov | i.e. where the volume service will mount it for some of it's operations | 15:10 |
ndipanov | so maybe that's what you are seeing | 15:10 |
ndipanov | (that should not leak back to nova tho | 15:10 |
ndipanov | ) | 15:10 |
*** pedroalvarez_ has joined #openstack-nova | 15:11 | |
mriedem | ndipanov: ok, i'll admit i don't know much about the flow here | 15:11 |
danpb | mriedem: not sure really | 15:12 |
*** mpaolino has quit IRC | 15:12 | |
mriedem | i do see that the LibvirtGlusterfsVolumeDriver is 95% the same as the LibvirtNFSVolumeDriver in nova, i think i'm going to refactor that to clean those up | 15:12 |
mriedem | make glusterfs driver extend nfs driver for the common bits | 15:12 |
mriedem | really the only differences are the options used | 15:12 |
ndipanov | mriedem, or both extend a localmount driver | 15:12 |
*** agireud has joined #openstack-nova | 15:13 | |
*** pedroalvarez has quit IRC | 15:14 | |
*** pedroalvarez_ is now known as pedroalvarez | 15:15 | |
*** otter768 has joined #openstack-nova | 15:15 | |
*** lpetrut has quit IRC | 15:15 | |
*** mikedillion has joined #openstack-nova | 15:16 | |
jogo | sdague: how do I find out what the difference is between each API version? | 15:17 |
jogo | microversion* | 15:17 |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/python-novaclient: [WIP] cleanup openstack-common.conf and sync updated files https://review.openstack.org/189119 | 15:17 |
mriedem | jogo: there is an rst file in tree | 15:17 |
dansmith | and comments in the api_version_request file | 15:17 |
mriedem | jogo: https://github.com/openstack/nova/blob/master/nova/api/openstack/rest_api_version_history.rst | 15:17 |
jogo | ahh nice | 15:18 |
jogo | should we wire that up to http://docs.openstack.org/developer/nova/ | 15:18 |
*** mikedillion has quit IRC | 15:18 | |
mriedem | jogo: yeah | 15:18 |
mriedem | i don't know how though | 15:19 |
* jogo knows the secrets of the sphinx | 15:19 | |
mriedem | there is a joke there somewhere | 15:19 |
*** otter768 has quit IRC | 15:20 | |
lxsli | jaypipes: so in your patches I see IntegerAmount and IntegerUsage. That means that from one of those, there's no way to know what resource it applies to. We could use a dict with key ResourceType and value *Amount; but ugh. Do you think we need something (other than the computenode) to relate these? Specifically I'm looking at the _update_usage example which needs to take a collection of amounts and | 15:20 |
lxsli | know which resources they apply to | 15:21 |
openstackgerrit | Nisha Agarwal proposed openstack/nova-specs: Nova and Ironic-virt-driver to support json list https://review.openstack.org/182572 | 15:21 |
lxsli | We could pass dual comprehension lists but again ugh | 15:22 |
kashyap | mriedem: Heya ill you also backport to stable/kilo once merged in master? I could do that, but wanted to check if you planned to do, so I don't step on toes - https://review.openstack.org/#/c/181781/ | 15:22 |
kashyap | s/ill/will/ | 15:22 |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: [WIP] cleanup openstack-common.conf and sync updated files https://review.openstack.org/189118 | 15:23 |
jaypipes | lxsli: no, I don't think we need anything other than a dict, keyed by the nova.objects.resources.ResourceType classes, of *Amount objects for the user-requested stuff and *Usage (or Inventory) objects for the capacity stuff. | 15:23 |
*** ajayaa has joined #openstack-nova | 15:23 | |
lxsli | jaypipes: OK, thanks | 15:23 |
jaypipes | lxsli: of course, I'm happy to hear your thoughts on alternatives, though... | 15:23 |
jogo | mriedem: lucky guess it was a movie http://www.imdb.com/title/tt0123262/ | 15:24 |
lxsli | jaypipes: no that's fine for now, thanks - if it looks like a problem I'll raise again | 15:25 |
jaypipes | coolio. | 15:25 |
jaypipes | BTW, edleafe, lxsli: did we come to agreement on a new shceduler IRC meeting time? | 15:25 |
*** leopoldj has quit IRC | 15:26 | |
lxsli | poll should have closed by now | 15:26 |
openstackgerrit | Michael Bayer proposed openstack/nova: Add an index to virtual_interfaces.uuid https://review.openstack.org/154243 | 15:26 |
bauzas | jaypipes: I was not there at the previous meeting, but it seems we have a conflict with timings, by reading the logs | 15:26 |
*** tjones1 has quit IRC | 15:27 | |
*** yamahata has quit IRC | 15:27 | |
*** yamahata has joined #openstack-nova | 15:28 | |
*** arnaud____ has joined #openstack-nova | 15:28 | |
lxsli | oh dear, I missed the last one, must have been quiet! | 15:28 |
*** dsanders has joined #openstack-nova | 15:29 | |
edleafe | jaypipes: there wasn't a slot that *someone* didn't have a conflict | 15:29 |
edleafe | jaypipes: we almost did, but bauzas | 15:29 |
edleafe | ugh | 15:29 |
bauzas | edleafe: which slot ? | 15:30 |
edleafe | but bauzas's vote messed that up | 15:30 |
bauzas | edleafe: I saw the logs but you didn't tell which one | 15:30 |
edleafe | I think it was 1700 UTC Tue | 15:30 |
bauzas | edleafe: which is 7pm by me, really hard to follow due to work/life balance | 15:30 |
bauzas | (at least during summer time) | 15:31 |
mriedem | dansmith: can you view this? https://docs.google.com/drawings/d/1geSKQXz4SqfXllq1Pk5o2YVCycZVf_i6ThY88r9YF4A/edit?usp=sharing | 15:31 |
mriedem | dansmith: you're my guinnea pig | 15:31 |
dansmith | mriedem: yep | 15:31 |
mriedem | cool | 15:31 |
mriedem | it's not ascii art, i know | 15:32 |
bauzas | that's gdocs, evil ! | 15:32 |
mriedem | meh | 15:32 |
mriedem | i apparently uninstalled dia at some point | 15:32 |
mriedem | so this was next | 15:32 |
*** arnaud____ has quit IRC | 15:33 | |
*** sileht has quit IRC | 15:33 | |
*** pradipta has joined #openstack-nova | 15:33 | |
lxsli | jaypipes: oh I hit it. The method that used to be update_compute_node_resource_usages. We want to pass a Dict<Resource, Inventory> to that I think | 15:33 |
*** mpaolino has joined #openstack-nova | 15:33 | |
*** sdake has joined #openstack-nova | 15:33 | |
*** burt has joined #openstack-nova | 15:34 | |
lxsli | jaypipes: would adding a DictOfObjectsField help? hitting my objects knowledge rather hard here sorry | 15:34 |
lxsli | I think we'd still need an object to put that field on | 15:35 |
dansmith | danpb: on this https://review.openstack.org/#/c/162253 | 15:35 |
dansmith | danpb: do we ever give up and cancel after we run out of steps? | 15:35 |
nagyz | mriedem, nice pic :) | 15:35 |
dansmith | danpb: it doesn't look like we do to me, but maybe I'm missing some piece of existing functionality? | 15:35 |
jaypipes | lxsli: no, we'll be passing the compute node object itself, which is versioned and will have already had its resource information set using the (local, not remoteable) update_usages() method. | 15:35 |
bauzas | jaypipes: any change I could put my dirty hands on ? | 15:36 |
lxsli | jaypipes: aha. so really we'll just call .save() ? | 15:36 |
jaypipes | lxsli: yuppers. | 15:36 |
lxsli | Gotcha, thanks | 15:36 |
*** sileht has joined #openstack-nova | 15:36 | |
jaypipes | bauzas: just reviewing my two resource-objects patches please, if you haven't already? | 15:36 |
bauzas | jaypipes: sure, my question was about those 2 | 15:36 |
bauzas | jaypipes: given the convo you had with lxsli | 15:37 |
bauzas | jaypipes: I'm just missing the URLs | 15:37 |
jaypipes | bauzas: I will have third patch in the series submitted by EOD today. will of course add you and lxsli as reviewer | 15:37 |
bauzas | jaypipes: (I was off these 2 last days, so sorry about my lagging) | 15:37 |
jaypipes | bauzas: no worries at all, my friend. | 15:37 |
*** lpetrut has joined #openstack-nova | 15:38 | |
*** lucasagomes has quit IRC | 15:39 | |
*** sileht has quit IRC | 15:39 | |
*** tjones1 has joined #openstack-nova | 15:40 | |
*** sileht has joined #openstack-nova | 15:40 | |
*** unicell has joined #openstack-nova | 15:40 | |
bauzas | jaypipes: hold the trigger for the RequestSpec series, I'm just now struggling with those lovely numa and pci hacks | 15:40 |
nagyz | mriedem, so I see that the volume drivers are passingin a is_block_dev value to their superclass - now in the RBD driver we would need to handle this based on weather it's used for encryption or not | 15:40 |
*** kiran-r has joined #openstack-nova | 15:40 | |
*** rajesht has quit IRC | 15:40 | |
*** dulek has quit IRC | 15:42 | |
*** unicell1 has quit IRC | 15:42 | |
nagyz | mriedem, we could introduce a config flag that sets if it's if we want to use krdb or qemu for rbd mounting, but the nicer thing to do would be to use qemu's rbd for native and krdb for encrypted... however, introducing an LibvirtEncryptedRBDVolumeDriver sounds horrible | 15:43 |
mriedem | nagyz: if you look at where that's used, it's in nova.virt.libvirt.utils.pick_disk_driver_name and it's only in the case of xen | 15:43 |
*** p_deore has quit IRC | 15:43 | |
nagyz | well based on it's name it sounds generic | 15:43 |
*** Daviey has quit IRC | 15:44 | |
danpb | dansmith: this patch doesn't do anything to terminate live migration - it only cares about increasing the downtime value | 15:44 |
dansmith | danpb: okay, so this makes it much better but we could still end up at the end of the steps with a migration that never completes, right? | 15:44 |
danpb | dansmith: the second patch in the series introduces the concept of a overall timeout, along with detection & abort of stuck migrations (ie data is being dirtied faster than it is transferred) | 15:44 |
dansmith | ah, okay | 15:44 |
* dansmith hadn't gotten there yet | 15:44 | |
danpb | the code is kind of hard to understand so i split the work in 2 pieces | 15:45 |
dansmith | danpb: I have a couple minor comments mostly about clarity, but otherwise looks cool | 15:45 |
mriedem | nagyz: is_block_dev looks like it's only ever True for local volume drivers | 15:45 |
mriedem | nagyz: so for rbd that would always be False | 15:45 |
nagyz | iSCSI sets it to true :-) | 15:45 |
*** dsanders has quit IRC | 15:46 | |
*** Sukhdev has joined #openstack-nova | 15:46 | |
openstackgerrit | jichenjc proposed openstack/nova-specs: New API for get service info from id https://review.openstack.org/172412 | 15:46 |
mriedem | gdi | 15:46 |
mriedem | yeah | 15:46 |
mriedem | idk, it seems like a bad name | 15:46 |
mriedem | nagyz: anyway, i don't see what it has to do with encryption | 15:46 |
*** oro has joined #openstack-nova | 15:47 | |
edleafe | bauzas: sorry, on a call. I totally understand the conflict; I just like blaming you for stuff :-P | 15:47 |
nagyz | mriedem, just for the encrypted rbd it's a local block device while for the non-encrypted it's not since that's handled by qemu directly | 15:47 |
*** yingjun has quit IRC | 15:48 | |
*** kiran-r has quit IRC | 15:48 | |
mriedem | nagyz: ok | 15:48 |
*** armax has joined #openstack-nova | 15:49 | |
lxsli | edleafe: who doesn't like blaming the French? | 15:49 |
nagyz | mriedem, so besides introducing a config option, splitting them into two classes sounds horrid, no? | 15:49 |
edleafe | lxsli: :) | 15:49 |
*** Daviey has joined #openstack-nova | 15:50 | |
mriedem | nagyz: splitting seems unnecessary | 15:50 |
mriedem | nagyz: couldn't a single rbd volume driver just have a condional path based on whether or not cinder tells us the volume is encrypted? | 15:50 |
mriedem | via connection_info['data'].get('encrypted')? | 15:51 |
bauzas | edleafe: eh, you're becoming French, if you begin arguing and blaming people for your own duty | 15:51 |
bauzas | edleafe: just try to be a little more grumpy and you'll get your passport | 15:51 |
*** ildikov_ has quit IRC | 15:52 | |
edleafe | bauzas: True, but I don't know if the English in me would allow turning French | 15:52 |
bauzas | edleafe: which state are you in ? | 15:52 |
bauzas | edleafe: I mean, where are you based ? | 15:52 |
*** zz_jgrimm is now known as jgrimm | 15:52 | |
edleafe | bauzas: San Antonio, Texas | 15:52 |
bauzas | edleafe: because nearly 50% of the US was French, a while ago :) | 15:52 |
lxsli | bauzas: England, too... | 15:53 |
bauzas | lxsli: we gave you queens | 15:53 |
lxsli | a bit longer ago though ;) | 15:53 |
edleafe | bauzas: It was Spain here | 15:53 |
*** lucasagomes has joined #openstack-nova | 15:53 | |
edleafe | bauzas: ok, enough geography. Got another call to hop onto. | 15:53 |
bauzas | edleafe: ;) | 15:54 |
openstackgerrit | Joe Gordon proposed openstack/nova: Link to microversion history in docs https://review.openstack.org/190240 | 15:54 |
openstackgerrit | Joe Gordon proposed openstack/nova: Improve formatting of rest_api_version_history https://review.openstack.org/190241 | 15:54 |
dansmith | cburgess: I can't find you in gerrit, so resorting to this... Can you look these over? https://review.openstack.org/#/c/162253 | 15:54 |
jogo | mriedem: there ^ | 15:54 |
*** varya has quit IRC | 15:55 | |
dansmith | rmk: you too ^^ | 15:55 |
dansmith | rmk: cburgess: libvirt robustification of live migration parameters | 15:55 |
*** browne has joined #openstack-nova | 15:56 | |
*** dboik_ has joined #openstack-nova | 15:58 | |
nagyz | mriedem, if I need to pass the block device nature back in the super() then I can't - if as you said that's largely irrelevant as it only pertains to Xen, then right, we can have that easily | 15:58 |
*** lpetrut has quit IRC | 15:59 | |
*** mpaolino has quit IRC | 16:00 | |
*** jlanoux has quit IRC | 16:00 | |
*** arnaud____ has joined #openstack-nova | 16:00 | |
*** Nikolay_St has quit IRC | 16:01 | |
*** dboik has quit IRC | 16:02 | |
openstackgerrit | Mikhail Feoktistov proposed openstack/nova: libvirt: virtuozzo instance resize support https://review.openstack.org/182257 | 16:03 |
dansmith | alaski: so, I think you were +2 on this idea, assuming my robustify spec would make it require-able for an evacuation | 16:04 |
dansmith | alaski: so if you want to review/merge this I can add that to my amendment: https://review.openstack.org/#/c/169836 | 16:04 |
*** mdrabe has quit IRC | 16:04 | |
*** rushiagr_away is now known as rushiagr | 16:06 | |
*** nic has joined #openstack-nova | 16:06 | |
alaski | dansmith: I was, I'll go through it again and rubber stamp if nothing jumps out at me | 16:07 |
mriedem | nagyz: true, since the volume drivers are initialized when the libvirt driver is initialized | 16:07 |
mriedem | and you don't have the connection_info at that point | 16:07 |
dansmith | alaski: cool, thanks | 16:07 |
*** arnaud____ has quit IRC | 16:08 | |
*** kylek3h has joined #openstack-nova | 16:08 | |
*** unicell has quit IRC | 16:09 | |
mriedem | dansmith: with this https://docs.google.com/drawings/d/1geSKQXz4SqfXllq1Pk5o2YVCycZVf_i6ThY88r9YF4A/edit?usp=sharing - the subclassing is a problem for object versioning because of which direction for the version changes? if the parent changes then all of the children have to bump a version or vice-versa? | 16:10 |
*** annegentle has joined #openstack-nova | 16:11 | |
openstackgerrit | Alexis Lee proposed openstack/nova-specs: Amend resource objects proposal https://review.openstack.org/188759 | 16:12 |
openstackgerrit | jichenjc proposed openstack/nova: Add common function for v2.1 API flavor_get. https://review.openstack.org/136568 | 16:12 |
openstackgerrit | jichenjc proposed openstack/nova: Add common function for v2 API flavor_get. https://review.openstack.org/136567 | 16:12 |
dansmith | mriedem: is this supposed to be UML? | 16:13 |
dansmith | mriedem: i.e. the arrows are inheritance? | 16:13 |
bauzas | dansmith: sorry to interrupt you, but I saw you moved most of the fields to o.voo | 16:14 |
mriedem | dansmith: yeah | 16:14 |
mriedem | ConnectionInfo is the base class | 16:14 |
mriedem | the others are sub-classes | 16:14 |
dansmith | mriedem: so why is SharedConnectionInfo not just in the base class? | 16:14 |
bauzas | dansmith: my question is, I guess we would like to avoid any new field being created like ListOfListOfStringsField ? | 16:14 |
*** jistr has quit IRC | 16:14 | |
dansmith | mriedem: or is Shared meaning "NFS like things" ? | 16:14 |
mriedem | dansmith: NFS like things | 16:14 |
bauzas | dansmith: being created in the nova codebase I mean | 16:14 |
dansmith | mriedem: gotcha | 16:14 |
*** varya has joined #openstack-nova | 16:14 | |
mriedem | maybe i should rename that | 16:15 |
dansmith | bauzas: you'll have to create them in both places if you don't want to be blocked, and we'll just have to clean them up as we release o.vo with things you need | 16:15 |
dansmith | bauzas: but also, it'd be great to avoid things like ListOfListOfListOfListOfString | 16:15 |
bauzas | dansmith: okay, my take is that I probably don't need that ugly ListOfListOfStringsField | 16:15 |
dansmith | mriedem: sharedfsconnectioninfo? | 16:15 |
bauzas | dansmith: thanks, I think I got your idea, I'll amend my object to not use that but rather something else | 16:16 |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Enable python34 tests for nova/tests/unit/objects/test*.py https://review.openstack.org/187785 | 16:17 |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Enable python34 tests for nova/tests/unit/scheduler/test*.py https://review.openstack.org/187968 | 16:17 |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Enable python34 tests for nova/tests/unit/virt/vmwareapi/test*.py https://review.openstack.org/188653 | 16:17 |
mriedem | i guess cinder calls the base class a RemoteFS | 16:17 |
mriedem | NFS extends RemoteFS and GlusterFS extends NFS | 16:17 |
dansmith | mriedem: so, I think mount options need not be a dict, but I don't know the details | 16:17 |
dansmith | mriedem: what about the tgt_port_wwns? that has to be a dict? | 16:17 |
*** mdrabe has joined #openstack-nova | 16:17 | |
dansmith | because otherwise, all of that looks like it could easily be in the extra dict | 16:17 |
rmk | dansmith: I shed a tear that you remember my coding days :D | 16:18 |
mriedem | dansmith: i was just looking at how it's modeled in the cinder code today | 16:18 |
lxsli | jaypipes: amounts can be negative. Is it still necessary to have a 'sign' argument to some methods or is that just confusing? | 16:18 |
rmk | I just glanced at this for a second and I already know where this approach came from. | 16:18 |
dansmith | rmk: well, still figure you can evaluate the usefulness or applicability of the config options being added | 16:18 |
rmk | This is exactly what Proxmox does. | 16:18 |
dansmith | rmk: even if you are too big of a deal to concern yourself with code these days :) | 16:19 |
rmk | dansmith: Oh trust me I'd rather do this, they don't let me, not that I take myself too seriously. | 16:19 |
rmk | I'm overhead now. | 16:19 |
dansmith | heh | 16:19 |
mriedem | mount options is probably not a dict, i got confused by the cinder code. tgt_port_wwns is a list | 16:19 |
dansmith | mriedem: okay, lists are easily encoded in a string without much danger I think | 16:19 |
rmk | OK this is going to require a new/current libvirt. | 16:20 |
dansmith | is it? | 16:20 |
dansmith | I guess I thought this had been available in libvirt for some time | 16:21 |
rmk | It was broken. | 16:21 |
dansmith | okay | 16:21 |
rmk | We actually submitted the patch into Libvirt to fix it. Sometime in the 1.2.8+ timeframe. | 16:21 |
mriedem | dansmith: fwiw i think everything could live in a single object with a bunch of optional fields | 16:22 |
mriedem | dansmith: the downside of that is from a developer point of view things aren't typed very well | 16:22 |
dansmith | mriedem: meaning real fields and not just a bag of extra as a dict? | 16:22 |
mriedem | dansmith: yeah | 16:22 |
dansmith | yeah, that gets more confusing I think | 16:23 |
dansmith | because you have options that look formal but don't matter for a given backend | 16:23 |
dansmith | mriedem: so does that mean you're leaning towards subclasses? let me answer your question above: | 16:23 |
rmk | Basically libvirt would only allow setting maxdowntime while the domain was migrating. Well, two problems with that: the migrate libvirt call is a blocking operation, and then if you tried to set it during migration via libvirt, it would respond back that qemu monitor was unresponsive. | 16:23 |
mriedem | dansmith: at a min we could have a single object with a few attributes like what connection_info is today, driver_volume_type, serial and 'data' | 16:23 |
mriedem | where 'data' is a big dict of gorp | 16:23 |
dansmith | mriedem: you can tie the versions of all of those objects together so that they all bump when anything bumps, which makes it a little easier | 16:23 |
rmk | The limitation around setting maxdowntime only during migration isn't a reality for qemu itself, libvirt was imposing a limitation that didn't exist. | 16:23 |
mriedem | but i don't want that b/c data has passwords in it and i want to specifically list those separately so we can mask them in the logs | 16:23 |
dansmith | mriedem: yeah | 16:24 |
*** teju has left #openstack-nova | 16:24 | |
mriedem | i want to flatten the nested data dict out as much as possible so that the only thing left in 'data' from today is the vendor-specific thingies | 16:24 |
rmk | I have to read more of this code but I don't see how this works. I'm sure danbp sorted it but I'd think this all needs to run in a thread of its own. | 16:24 |
dansmith | rmk: thanks for justifying me pinging you for review :) | 16:25 |
rmk | Haha np | 16:25 |
mriedem | dansmith: so if the fibrechannel object version changes, we'd also bump the version in the shared object? | 16:25 |
*** varya has quit IRC | 16:25 | |
dansmith | mriedem: yeah | 16:25 |
mriedem | that seems weird | 16:25 |
*** vilobhmm has joined #openstack-nova | 16:25 | |
dansmith | mriedem: doing that means we can automate some of the dependency stuff and make it less manually intensive when we need to tweak things | 16:26 |
dansmith | mriedem: actually, you have this as inheriting from the same class, not a nested object relationship | 16:26 |
*** mpaolino has joined #openstack-nova | 16:26 | |
dansmith | mriedem: so it's a little less of a problem that way I think | 16:27 |
dansmith | mriedem: the painful part comes in when one object embeds another, and the embedded object version changes, the parent has to be bumped as well to signal the change, and so that we can efficiently backport the whole thing and not just the leaves as we hit them | 16:27 |
mriedem | dansmith: right, no nested objects right now | 16:27 |
dansmith | yeah, we've not done it that way as often, usually because we don't have this sort of arrangement | 16:28 |
mriedem | i originally thought we'd have a Data object that was nested within ConnectionInfo to model like how it is serialized today, but then i realized the way it is today sucks | 16:28 |
dansmith | so yeah, I say give it a go | 16:28 |
*** mingy has joined #openstack-nova | 16:29 | |
openstackgerrit | Alexis Lee proposed openstack/nova-specs: Amend resource objects proposal https://review.openstack.org/188759 | 16:29 |
mriedem | cool | 16:29 |
mriedem | this is going to ruin me | 16:29 |
lxsli | jaypipes: ok, in your copious free time :) | 16:30 |
*** exploreshaifali has joined #openstack-nova | 16:30 | |
*** annashen has joined #openstack-nova | 16:32 | |
*** browne has quit IRC | 16:33 | |
openstackgerrit | Kevin Fox proposed openstack/nova-specs: Instance Users for Cloud Interaction https://review.openstack.org/186617 | 16:33 |
*** atuvenie has quit IRC | 16:37 | |
*** annashen has quit IRC | 16:37 | |
mingy | hmm, in the nova scheduler, what's the difference between the tests in nova/tests/scheduler/test_filters[...].py and nova/tests/scheduler/filters/[...].py | 16:37 |
*** unicell has joined #openstack-nova | 16:39 | |
mingy | specifically for example, there are test files in two different directories that test functions in nova/scheduler/filters/utils.py | 16:40 |
mriedem | mingy: one is probably just older than the other | 16:40 |
mriedem | mingy: like nova.tests.unit.compute.test_compute is old, and newer tests for the compute manager are in nova.tests.unit.compute.test_compute_mgr (mock compute.api), and test_compute_api (mock compute.manager) | 16:41 |
openstackgerrit | Philip Marc Schwartz proposed openstack/nova: Implement online schema migrations https://review.openstack.org/154521 | 16:41 |
mingy | mriedem: oh, what's the rationale for organizing it that way? | 16:42 |
mriedem | mingy: for the compute manager tests? | 16:42 |
mriedem | layer isolation | 16:42 |
mriedem | so we can mock out the things that we aren't testing | 16:42 |
mriedem | if i'm testing nova.compute.api, i hit that in test_compute_api and mock out the calls to conductor | 16:42 |
mriedem | test_compute (old way) doesn't mock that stuff out, so requests go through comptue.api then through conductor and into compute.manager | 16:43 |
mriedem | which is more of a functional test | 16:43 |
*** tonytan4ever has joined #openstack-nova | 16:43 | |
mriedem | so generally more complex and racier | 16:44 |
*** sahid has quit IRC | 16:46 | |
mingy | hmm, I see, that makes sense, but in the filter tests both don't mock anything | 16:47 |
*** romainh has left #openstack-nova | 16:48 | |
*** ajayaa has quit IRC | 16:48 | |
*** haypo has quit IRC | 16:48 | |
mingy | they're both functional tests | 16:48 |
mingy | I mean they | 16:49 |
mingy | 're both unit test | 16:49 |
*** Nikolay_St has joined #openstack-nova | 16:52 | |
*** gszasz has quit IRC | 16:52 | |
mriedem | mingy: i guess i don't know, i don't venture in there often. are you trying to figure out where to put a test? | 16:54 |
mriedem | when i was new i figured out where to put new tests based on where the code i'm trying to test is currently tested | 16:54 |
*** esmute has joined #openstack-nova | 16:55 | |
*** annegentle has quit IRC | 16:56 | |
mingy | mriedem: mm somewhat, I was trying to figure out where the tests for the methods were, just to see how they are invoked | 16:56 |
*** asselin has joined #openstack-nova | 16:56 | |
mriedem | mingy: this jaypipes guy once said that the unit tests for the scheduler were bonkers | 16:57 |
*** alex_klimov has quit IRC | 16:57 | |
mriedem | falls in with my beware from yesterday | 16:57 |
*** marun has quit IRC | 16:58 | |
*** moshele has quit IRC | 16:58 | |
mingy | I guess I could put them in the same file, since they look like very similar. I was confused because for a moment I thought half of the methods weren't tested | 16:58 |
mriedem | jogo: your sphinx skillz are maddening | 16:59 |
*** esmute has quit IRC | 17:00 | |
*** claudiub has quit IRC | 17:00 | |
*** dsanders has joined #openstack-nova | 17:00 | |
*** Mike_D_laptop has joined #openstack-nova | 17:01 | |
jaypipes | mriedem: they are slightly better now than they were before. ... slightly. | 17:01 |
jaypipes | mriedem: at least now they are all no-db test cases... | 17:01 |
mriedem | > 0 is improvement | 17:02 |
cfriesen | dansmith: danpb: could you take a look at my latest comment on https://review.openstack.org/#/c/168982/ ? I'm trying to figure out if we can agree on backend stuff while we work on how the end-user specifies what they want. | 17:02 |
edleafe | mingy: test_filters.py actually tests the nova/filters.py code | 17:02 |
edleafe | mingy: not sure why that's not nova/scheduler/filters.py, but it is what it is | 17:03 |
edleafe | mingy: the tests in nova/tests/unit/scheduler/filters test the individual filters | 17:03 |
*** derekh has quit IRC | 17:04 | |
bauzas | edleafe: because filters.py is not only for the scheduler :) | 17:04 |
mingy | edleafe: I was wondering about nova/tests/unit/scheduler/test_filters_utils.py vs nova/tests/unit/scheduler/filters/test_utils.py | 17:05 |
bauzas | edleafe: the scheduler uses it as a base handler for the filters | 17:05 |
*** tonytan4ever has quit IRC | 17:05 | |
ndipanov | jaypipes, a very old one - responded to you so if you can come back to it that would be great | 17:06 |
*** tonytan4ever has joined #openstack-nova | 17:06 | |
ndipanov | https://review.openstack.org/#/c/83516/ | 17:06 |
*** achanda has joined #openstack-nova | 17:07 | |
*** lucasagomes is now known as lucas-dinner | 17:07 | |
*** claudiub has joined #openstack-nova | 17:07 | |
edleafe | mingy: ah. That looks like people adding tests at different times and not consolidating the files | 17:07 |
mingy | also wondering about what the directory nova/tests/unit/functional does | 17:08 |
edleafe | bauzas: I didn't realize that cells uses it too | 17:08 |
bauzas | yep | 17:08 |
jaypipes | ndipanov: +W'd | 17:09 |
patrickeast | hi everyone, cross posting from #openstack in the hopes maybe someone here might know whats going on | 17:10 |
edleafe | mingy: there are also 2 utils files: nova/scheduler/utils and nova/scheduler/filters/utils | 17:10 |
patrickeast | i’m having some trouble with nova-network dhcp in multi-host flat dhcp configuration, i have force_dhcp_release=True and see when instances are terminated that nova-network is calling the dhcp_release script (and its returning 0), but in nova-dhcpbridge.log i only sometimes actually see a ‘del’ | 17:10 |
patrickeast | and then later if another vm is assigned the old ip address, the dnsmasq proccess gives error output in syslog saying the address is already leased to another mac address (the old vm that had been terminated) | 17:10 |
ndipanov | jaypipes, many thanks | 17:10 |
patrickeast | any ideas on what might be causing that? | 17:10 |
*** tonytan4ever has quit IRC | 17:10 | |
jaypipes | patrickeast: sorry, I don't recognize what the issue might be there. have you posted to the openstack[-operators] ML about it? | 17:11 |
patrickeast | jaypipes: not yet, i’ll ask there next | 17:11 |
mingy | edleafe: I think the former is tested in nova/tests/unit/scheduler/test_scheduler_utils.py, while the latter is tested in the two files from earlier | 17:11 |
*** lpetrut has joined #openstack-nova | 17:11 | |
*** ZZelle_ has joined #openstack-nova | 17:12 | |
*** prashant__ has joined #openstack-nova | 17:12 | |
jaypipes | n0ano: any idea what your Intel colleague Michael Dulko's IRC nick might be? | 17:12 |
openstackgerrit | Karim Boumedhel proposed openstack/nova: report an error for hw:numa_nodes=0 https://review.openstack.org/190267 | 17:13 |
vilobhmm | jaypipes : dulek | 17:13 |
jaypipes | vilobhmm: cheers :) | 17:13 |
vilobhmm | :) | 17:13 |
vilobhmm | cheers | 17:13 |
*** prashup has quit IRC | 17:13 | |
*** browne has joined #openstack-nova | 17:13 | |
n0ano | jaypipes, looks like you got it (I would have responded quicker if my keyboard hadn't decided to stop working) | 17:14 |
cfriesen | dansmith: you mentioned a couple times that HP and Rackspace had different cpu features exposed. Do you know anything about how they're implementing that? Are they defining compute nodes with custom models, or turning off features in the BIOS, or... | 17:14 |
*** achanda has quit IRC | 17:15 | |
edleafe | mingy: yeah, just pointing out even more potential confusion | 17:16 |
alaski | cfriesen: I can't say for sure for Rackspace, but I don't think it's anything we do intentionally | 17:16 |
alaski | cfriesen: different versions of hypervisors would be my guess | 17:16 |
edleafe | mingy: this might be a good place to start contributing: sorting this out and making the structure saner. | 17:16 |
alaski | I've never investigated though | 17:16 |
openstackgerrit | Karim Boumedhel proposed openstack/nova: report an error for hw:numa_nodes=0 https://review.openstack.org/190267 | 17:17 |
mingy | edleafe: alright, I will file a bug and deal with it :) | 17:17 |
cfriesen | alaski: thanks. Dan mentioned the hypervisor version thing in a different context..my mind is blown that they would change the model/feature mapping after a model has been released. | 17:17 |
alaski | cfriesen: agreed, unless it was just a matter of support. but I could be totally off on what's actually happening | 17:20 |
*** tonytan4ever has joined #openstack-nova | 17:21 | |
johnthetubaguy | cfriesen: so from the rackspace side we use, Xen (well XenServer) | 17:21 |
*** exploreshaifali has quit IRC | 17:21 | |
johnthetubaguy | cfriesen: the old gear is AMD, the new stuff is Intel | 17:21 |
johnthetubaguy | cfriesen: mostly we don't do any CPU masking, except a few cases where I think we needed some for live-migrate | 17:22 |
johnthetubaguy | cfriesen: Xen CPU masking required a reboot, so its quite minimal | 17:22 |
*** esmute has joined #openstack-nova | 17:22 | |
johnthetubaguy | its possible Xen just hides some stuff when using its crazy QEMU fork, which causes extra confusion, but thats beyond where I have looked | 17:22 |
*** annashen has joined #openstack-nova | 17:23 | |
johnthetubaguy | cfriesen: I think as we get new hardware, some of them get new CPUs too, and to alaski's point, different version of Xen might give us some differences, but I didn't think they were massive, we have a few versions out there | 17:24 |
*** pradipta has quit IRC | 17:24 | |
cfriesen | johnthetubaguy: thanks. I'm trying to get context for the "setting cpu model/features" spec. | 17:25 |
*** markus_z has quit IRC | 17:25 | |
johnthetubaguy | cfriesen: yeah, thats cool | 17:26 |
johnthetubaguy | cfriesen: appreciate the effort to get a good solution thats workable across clouds | 17:26 |
johnthetubaguy | cfriesen: its the CPU masking for live-migrate that means the models anyone expose might not actually be something anyone has ever made | 17:26 |
danpb | cfriesen: fyi, intel have issued microcode updates in the past which removed features | 17:27 |
cfriesen | johnthetubaguy: it's a pain in the butt. :) So it's starting to look like we're going to have to rely on the admin to come up with a hypervisor-specific "base_model+features" that gets exported as some sort of named entity for end-users to use. | 17:28 |
danpb | this caused an unbelievable amount of pain because peoples guests configured to use those featurs no longer booted :-( | 17:28 |
cfriesen | danpb: the haswell tsx stuff? or were there others like that? | 17:28 |
danpb | cfriesen: yeah haswell | 17:28 |
*** ssurana has joined #openstack-nova | 17:28 | |
johnthetubaguy | what happened with haswell? | 17:28 |
johnthetubaguy | people want it I guess? | 17:29 |
danpb | they screwed up the impl of a feature in silicon and had to disble it | 17:29 |
cfriesen | johnthetubaguy: Intel realized the TSX instructions were broken | 17:29 |
johnthetubaguy | oh... | 17:29 |
*** claudiub has quit IRC | 17:29 | |
johnthetubaguy | ooops | 17:29 |
danpb | it also turns out they actually shipped cpus with the same names but without the feature present at all | 17:29 |
johnthetubaguy | like the old floating point bug or something, that happened before | 17:29 |
danpb | ie, you could actually buy haswell cpus which never had tsx in the first place | 17:29 |
johnthetubaguy | danpb: gotcha, yuck | 17:29 |
danpb | yeah | 17:29 |
*** esmute_ has joined #openstack-nova | 17:30 | |
*** pixelbeat_ has quit IRC | 17:30 | |
*** esmute has quit IRC | 17:30 | |
*** esmute_ has quit IRC | 17:30 | |
*** belmoreira has joined #openstack-nova | 17:30 | |
johnthetubaguy | honestly, its tempting to just talk about features | 17:30 |
johnthetubaguy | its just very very long winded | 17:30 |
danpb | johnthetubaguy: nope, that doesn't work on non-x86 | 17:30 |
johnthetubaguy | danpb: really? they are all model based then? | 17:31 |
*** esmute has joined #openstack-nova | 17:31 | |
*** sdake has quit IRC | 17:31 | |
danpb | primarily yes - fine grained cpuid like features are only used/exposed on x86 currently | 17:31 |
cfriesen | danpb: so based on your comment it seems like we're going to need to let admins specify "base_mode+features" for the "high level model"? (In which case it'd be up to the admins to ensure that what they come up with will actually run?) | 17:31 |
danpb | this is why libvirt tries very hard to discourage you from using features at all and focus on cpu model names | 17:32 |
cfriesen | s/base_mode/base_model | 17:32 |
johnthetubaguy | danpb: that makes sense now... ouch | 17:32 |
danpb | well i'd expect admins would mostly just specify a base libvirt model - i doubt they'd often customize feature set | 17:32 |
*** aix has quit IRC | 17:33 | |
*** earlephilhower has joined #openstack-nova | 17:33 | |
*** achanda has joined #openstack-nova | 17:33 | |
danpb | for vmware i think you need to specify the raw cpuid hex masks | 17:33 |
danpb | and likewise for xenapi iirc | 17:33 |
cfriesen | danpb: dansmith was opinionated about wanting to be able to define something that would be common across hypervisors...that seemed to imply the ability to tweak features | 17:34 |
danpb | we could possibly provide some canned cpu models in nova, so admin doesn't have to figure this out in a plain deployment | 17:34 |
danpb | we could provide a tool which takes the libvirt cpu xml definitions and generates the suitable cpuid masks for xen & vmware to make their life easier | 17:35 |
cfriesen | danpb: otherwise if we're not going to tweak features then it's basically my current proposal where you just use the CPU models actually present on hosts in the cloud. | 17:35 |
*** krtaylor has quit IRC | 17:37 | |
*** zul has quit IRC | 17:38 | |
*** changbl has joined #openstack-nova | 17:40 | |
*** zul has joined #openstack-nova | 17:40 | |
openstackgerrit | Marian Horban proposed openstack/nova: Associating of floating IPs corrected https://review.openstack.org/149358 | 17:40 |
*** danpb has quit IRC | 17:42 | |
*** yamahata has quit IRC | 17:42 | |
*** achanda_ has joined #openstack-nova | 17:43 | |
cfriesen | danpb: so my current thinking is that the virt driver will have to support base_model+features. Do you think we'd let the cloud operator limit the supported base models to a subset? Or would we need to support the full set that the hypervisor is capable of running? | 17:44 |
nagyz | mriedem, even if I do two classes, I need to register them in the driver.py... is it sensible to have a separate encrypted RBD option? | 17:44 |
nagyz | mriedem, I don't like this one bit. :) | 17:44 |
kashyap | cfriesen: Seems like his IRC client got disconnected. | 17:44 |
mriedem | nagyz: i honestly don't have my head fully wrapped around it but i don't see why two rbd drivers are necessary | 17:45 |
cfriesen | kashyap: thx, missed that. | 17:45 |
*** achanda has quit IRC | 17:45 | |
kashyap | cfriesen: I've seen those Haswell related patches fly by on QEMU and related lists, not quite pretty :-) | 17:45 |
*** Sukhdev has quit IRC | 17:46 | |
kashyap | (The issue, not the fixes.) | 17:46 |
*** Sukhdev has joined #openstack-nova | 17:46 | |
*** electrocucaracha has quit IRC | 17:49 | |
sdague | johnthetubaguy: so... I thought there were some quite old instructions masked out on rax, which was part of our fun coming up with a custom model for the gate live migration testing | 17:50 |
sdague | same with HP, but different ones | 17:50 |
sdague | neither looks anything like anything in the libvirt model set | 17:51 |
*** moshele has joined #openstack-nova | 17:56 | |
nagyz | mriedem, is there a document describing the flow in nova when someone wants to attach an encrypted volume? | 17:56 |
*** arnaud____ has joined #openstack-nova | 17:56 | |
*** annegentle has joined #openstack-nova | 17:56 | |
mriedem | nagyz: i very much doubt it | 17:56 |
mriedem | nagyz: it's pretty simple to trace though from the nova.compute.manager.attach_volume method | 17:57 |
nagyz | ok, will do just that | 17:57 |
mriedem | the nova.virt.block_device.attach method does the encryptor lookup stuff | 17:57 |
mriedem | the do_driver_attach flag tells block_device to call the attach_volume method on the virt driver (libvirt) | 17:57 |
mriedem | and that uses the encryptor (if present) to do the stuff in nova.volume.encryptors | 17:58 |
nagyz | ok, so the libvirt volume driver is before the encryptor - so it has time to actually attach the device if needed, cool | 17:58 |
*** harlowja has quit IRC | 17:58 | |
*** krtaylor has joined #openstack-nova | 17:59 | |
mriedem | nagyz: yeah that happens here http://git.openstack.org/cgit/openstack/nova/tree/nova/virt/libvirt/driver.py#n1044 | 18:00 |
*** tojuvone has quit IRC | 18:00 | |
mriedem | nova.virt.block_device is what calls out to cinder's os-initialize_connection API where we get the connection_info dict back | 18:01 |
mriedem | http://git.openstack.org/cgit/openstack/nova/tree/nova/virt/block_device.py#n238 | 18:01 |
*** arnaud____ has quit IRC | 18:01 | |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: [WIP] cleanup openstack-common.conf and sync updated files https://review.openstack.org/189118 | 18:01 |
mriedem | then that gets persisted in the block_device_mapping table in the nova db | 18:01 |
*** harlowja has joined #openstack-nova | 18:02 | |
mriedem | well, saved in the db after calling the virt driver | 18:02 |
*** annegentle has quit IRC | 18:02 | |
mriedem | once we get back from the virt driver's attach_volume, we call the attach API in cinder to mark the volume as in-use | 18:03 |
mriedem | nagyz: maybe you want to throw a nice flow diagram into the devref after all :) | 18:03 |
nagyz | :D | 18:04 |
nagyz | I appreciate your help; I need to take care of some stuff but will dig into this later tonight and tomorrow and then push some patches for review... it can't be that hard :-) | 18:04 |
mriedem | i think the compute api actually tells cinder to mark the volume as attaching before we even get to the compute manager in case of races | 18:05 |
mriedem | and on failure that's rolled back | 18:05 |
*** Mike_D_laptop has quit IRC | 18:06 | |
*** claudiub has joined #openstack-nova | 18:08 | |
*** yassine has quit IRC | 18:09 | |
*** annashen has quit IRC | 18:09 | |
mriedem | jaypipes: are you ready? https://review.openstack.org/#/c/154243/ | 18:10 |
jaypipes | mriedem: done. | 18:10 |
*** exploreshaifali has joined #openstack-nova | 18:11 | |
*** annegentle has joined #openstack-nova | 18:12 | |
mriedem | jaypipes: thanks | 18:14 |
mriedem | melwitt: do you know if there are any things in novaclient today which are checking for microversions to do stuff? | 18:14 |
mriedem | like with v2.4 the os-fixed-ips response will have a 'reserved' field in it | 18:15 |
melwitt | mriedem: not yet. some patches have been up for a long time but Im not sure if they do all the things needed to fully support microversions | 18:16 |
mriedem | melwitt: i arguably don't need to know the version since this isn't a request change, it's just checking a thing in the response | 18:16 |
*** achanda_ has quit IRC | 18:16 | |
mriedem | which i could easily do w/o a version check | 18:16 |
mriedem | if it's not in the response body we just don't show it | 18:16 |
*** agireud has quit IRC | 18:17 | |
*** annashen has joined #openstack-nova | 18:17 | |
melwitt | mriedem: yeah, I remember the spec mentioned that | 18:17 |
*** baoli_ has quit IRC | 18:17 | |
*** yamahata has joined #openstack-nova | 18:20 | |
*** iamjarvo has joined #openstack-nova | 18:21 | |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/python-novaclient: [WIP] cleanup openstack-common.conf and sync updated files https://review.openstack.org/189119 | 18:21 |
*** Sukhdev has quit IRC | 18:22 | |
*** ijw has joined #openstack-nova | 18:22 | |
*** exploreshaifali has quit IRC | 18:22 | |
melwitt | mriedem: curious, did you understand what that fixedip-post-req.json.tpl required by the api sample tests is for? I was looking at the review last night and was puzzled by it | 18:22 |
*** rushiagr is now known as rushiagr_away | 18:23 | |
melwitt | mriedem: nvm, it seems to be for schema validation stuff. maybe | 18:23 |
*** bradjones has quit IRC | 18:23 | |
mriedem | melwitt: i more or less just copied what the existing fixed_ips api samples tests were doing | 18:24 |
mriedem | i try not to think about how those are setup too much | 18:24 |
melwitt | heh | 18:24 |
*** bradjones has joined #openstack-nova | 18:24 | |
*** bradjones has quit IRC | 18:24 | |
*** bradjones has joined #openstack-nova | 18:24 | |
*** phschwartz has joined #openstack-nova | 18:27 | |
*** amotoki has quit IRC | 18:28 | |
mriedem | melwitt: looks like i'd need to implement __contains__ in base.Resource in novaclient if i wanted to do a kind of: if 'reserved' in fixed_ip: | 18:29 |
dims | melwitt: i remember running into it once, see api_samples_test_base._verify_response | 18:29 |
mriedem | otherwise it looks like you'll get an AttributeError | 18:29 |
*** mpaolino has quit IRC | 18:30 | |
*** dprince has quit IRC | 18:31 | |
mriedem | or use to_dict() if i want to be a brute | 18:31 |
*** electrocucaracha has joined #openstack-nova | 18:34 | |
jaypipes | dansmith: https://review.openstack.org/#/c/163060/ looks ready to me. | 18:34 |
openstackgerrit | Jay Pipes proposed openstack/nova: Conditionally expose "last_seen_up" in service list https://review.openstack.org/168418 | 18:34 |
*** achanda has joined #openstack-nova | 18:35 | |
dansmith | jaypipes: okay, it's been a while since I've looked at that | 18:35 |
dansmith | jaypipes: I'll add it to my tab queue | 18:36 |
jaypipes | dansmith: yes, understood. I pinged cfriesen last night for him to rebase that. | 18:36 |
mriedem | oh ew http://git.openstack.org/cgit/openstack/oslo-incubator/tree/openstack/common/apiclient/base.py is o-i | 18:36 |
mriedem | nevermind adding __contains__ | 18:36 |
melwitt | mriedem: looking through it still. I have used getattr before to do similar, __contains__ would be nicer. if it's in o-i we would have to add it there first are you saying? | 18:39 |
*** agireud has joined #openstack-nova | 18:39 | |
*** cbader has quit IRC | 18:39 | |
melwitt | mriedem: I'm also thinking of python api users, without microversion they wouldn't know whether to expect 'reserve' to be there or not | 18:40 |
*** prashant__ has quit IRC | 18:40 | |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/python-novaclient: [WIP] cleanup openstack-common.conf and sync updated files https://review.openstack.org/189119 | 18:40 |
openstackgerrit | jichenjc proposed openstack/nova-specs: New API for get service info from id https://review.openstack.org/172412 | 18:44 |
*** Nikolay_St has quit IRC | 18:45 | |
melwitt | dims: thanks. the thing I mentioned was a request template though. looking at test_fixed_ips.py I see now that it's just a sample request for reserving an ip, to test the reserve ip call | 18:46 |
dims | melwitt: ah cool | 18:47 |
openstackgerrit | Hans Lindgren proposed openstack/nova: Make compute_api.trigger_members_refresh() issue a single db call https://review.openstack.org/72382 | 18:48 |
openstackgerrit | Hans Lindgren proposed openstack/nova: Remove unused db.security_group_rule_get_by_security_group_grantee() https://review.openstack.org/72383 | 18:48 |
openstackgerrit | Hans Lindgren proposed openstack/nova: New db api method instance_get_all_by_grantee_security_groups() https://review.openstack.org/72380 | 18:48 |
openstackgerrit | Hans Lindgren proposed openstack/nova: Add new instance object method get_by_grantee_security_groups() https://review.openstack.org/72381 | 18:48 |
openstackgerrit | Hans Lindgren proposed openstack/nova: Remove and deprecate conductor security_groups_trigger_members_refresh() https://review.openstack.org/182213 | 18:48 |
openstackgerrit | Hans Lindgren proposed openstack/nova: Revert "Make nova-network use conductor for security groups refresh" https://review.openstack.org/182212 | 18:48 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: adjust os_command_empty test to ImageMetaProps object https://review.openstack.org/190017 | 18:49 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: objects: raise a meaningful exception when validating VIF properties https://review.openstack.org/190018 | 18:49 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: virt: introduce libosinfo library to set hardware policy https://review.openstack.org/149625 | 18:49 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: use osinfo when configuring network model https://review.openstack.org/149627 | 18:49 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: adding libosinfo configuration https://review.openstack.org/149626 | 18:49 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: use osinfo when configuring the disk bus https://review.openstack.org/149628 | 18:49 |
*** vladikr has quit IRC | 18:49 | |
*** iamjarvo has quit IRC | 18:53 | |
*** zzzeek has quit IRC | 18:54 | |
openstackgerrit | Jens Rosenboom proposed openstack/nova: Add microversion to allow server search option ip6 for non-admin https://review.openstack.org/179569 | 18:56 |
openstackgerrit | Hans Lindgren proposed openstack/nova: Remove/deprecate unused parts of the compute node object https://review.openstack.org/186757 | 18:57 |
openstackgerrit | Hans Lindgren proposed openstack/nova: Make ComputeNode.service_id nullable to match db schema https://review.openstack.org/169284 | 18:57 |
*** tonytan4ever has quit IRC | 18:58 | |
openstackgerrit | Nicolas Simonds proposed openstack/nova: Use utf8_bin collation on the flavor extra-specs table in MySQL https://review.openstack.org/190300 | 19:00 |
*** arnaud____ has joined #openstack-nova | 19:00 | |
*** vladikr has joined #openstack-nova | 19:01 | |
*** achanda has quit IRC | 19:02 | |
*** achanda has joined #openstack-nova | 19:03 | |
*** pece has joined #openstack-nova | 19:04 | |
*** iamjarvo has joined #openstack-nova | 19:04 | |
openstackgerrit | Hans Lindgren proposed openstack/nova: Avoid querying for Service in resource tracker https://review.openstack.org/169288 | 19:04 |
*** iamjarvo has quit IRC | 19:05 | |
*** Nikolay_St has joined #openstack-nova | 19:05 | |
*** iamjarvo has joined #openstack-nova | 19:05 | |
*** arnaud____ has quit IRC | 19:05 | |
*** iamjarvo has quit IRC | 19:05 | |
*** iamjarvo has joined #openstack-nova | 19:06 | |
*** iamjarvo has quit IRC | 19:06 | |
openstackgerrit | Ming Yang proposed openstack/nova: Two separate test files [nova/tests/unit/scheduler/test_filters_utils.py] and [nova/tests/unit/scheduler/filters/test_utils.py] appear to be testing methods from the same module nova.scheduler.filters.utils in a very similar fashion. This commit consolida https://review.openstack.org/190301 | 19:06 |
*** iamjarvo has joined #openstack-nova | 19:07 | |
*** iamjarvo has quit IRC | 19:07 | |
dansmith | mingy: ^ short first lines in your commits, and wrapped at 79 cols after that :) | 19:07 |
*** iamjarvo has joined #openstack-nova | 19:07 | |
mingy | dansmith: ok | 19:07 |
*** iamjarvo has quit IRC | 19:07 | |
*** nihilifer has joined #openstack-nova | 19:07 | |
*** iamjarvo has joined #openstack-nova | 19:08 | |
dansmith | mingy: https://wiki.openstack.org/wiki/GitCommitMessages | 19:08 |
*** iamjarvo has quit IRC | 19:08 | |
*** iamjarvo has joined #openstack-nova | 19:09 | |
*** iamjarvo has quit IRC | 19:09 | |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: adjust os_command_empty test to ImageMetaProps object https://review.openstack.org/190017 | 19:09 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: objects: raise a meaningful exception when validating VIF properties https://review.openstack.org/190018 | 19:09 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: virt: introduce libosinfo library to set hardware policy https://review.openstack.org/149625 | 19:09 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: use osinfo when configuring network model https://review.openstack.org/149627 | 19:09 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: adding libosinfo configuration https://review.openstack.org/149626 | 19:09 |
openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: use osinfo when configuring the disk bus https://review.openstack.org/149628 | 19:09 |
*** iamjarvo has joined #openstack-nova | 19:09 | |
*** iamjarvo has quit IRC | 19:09 | |
*** atuvenie has joined #openstack-nova | 19:09 | |
*** iamjarvo has joined #openstack-nova | 19:10 | |
*** iamjarvo has quit IRC | 19:10 | |
openstackgerrit | Earle F. Philhower, III proposed openstack/nova: Add support for HGST Solutions volumes to libvirt https://review.openstack.org/186594 | 19:10 |
*** iamjarvo has joined #openstack-nova | 19:10 | |
*** iamjarvo has quit IRC | 19:11 | |
openstackgerrit | Ming Yang proposed openstack/nova: Consolidates scheduler utils tests into a single file. https://review.openstack.org/190301 | 19:11 |
*** iamjarvo has joined #openstack-nova | 19:11 | |
*** iamjarvo has quit IRC | 19:11 | |
dansmith | mingy: yes, better :) | 19:11 |
sdague | mriedem: so... https://review.openstack.org/#/c/190304/1 - db2 merge conflict spam | 19:12 |
*** iamjarvo has joined #openstack-nova | 19:12 | |
*** iamjarvo has quit IRC | 19:12 | |
sdague | I guess I should just take it to infra and get them banned | 19:12 |
mingy | dansmith: I feel that maybe the changes should be split into different commits. I'm not sure. I'm headed out for lunch now, but will come back to take a look at it. Thanks for guiding! | 19:12 |
*** iamjarvo has joined #openstack-nova | 19:12 | |
dansmith | mingy: yep, when you say things like "additional stuff I did in here:" that usually means they should be separate :) | 19:12 |
*** iamjarvo has quit IRC | 19:13 | |
*** iamjarvo has joined #openstack-nova | 19:13 | |
*** iamjarvo has quit IRC | 19:14 | |
*** iamjarvo has joined #openstack-nova | 19:14 | |
*** iamjarvo has quit IRC | 19:14 | |
*** Sukhdev has joined #openstack-nova | 19:14 | |
*** iamjarvo has joined #openstack-nova | 19:15 | |
*** iamjarvo has quit IRC | 19:15 | |
*** iamjarvo has joined #openstack-nova | 19:15 | |
*** iamjarvo has quit IRC | 19:15 | |
*** iamjarvo has joined #openstack-nova | 19:16 | |
*** iamjarvo has joined #openstack-nova | 19:17 | |
*** iamjarvo has quit IRC | 19:17 | |
*** iamjarvo has joined #openstack-nova | 19:17 | |
*** iamjarvo has quit IRC | 19:17 | |
*** iamjarvo has joined #openstack-nova | 19:18 | |
melwitt | does anyone think we should revert behavior and instead do a deprecation period for something like this in novaclient? creating a Client object with "invalid version" used to silently ignore and use the default/latest version, and recently we started raising an exception on invalid version https://bugs.launchpad.net/python-novaclient/+bug/1463809 | 19:18 |
openstack | Launchpad bug 1463809 in python-novaclient "Invalid message returned when attempting to use version 3" [Undecided,In progress] - Assigned to Matt Thompson (mattt416) | 19:18 |
openstackgerrit | Chris Friesen proposed openstack/nova: fix "down" nova-compute service spuriously marked as "up" https://review.openstack.org/163060 | 19:18 |
*** iamjarvo has quit IRC | 19:18 | |
*** iamjarvo has joined #openstack-nova | 19:18 | |
*** iamjarvo has joined #openstack-nova | 19:19 | |
*** iamjarvo has quit IRC | 19:19 | |
*** iamjarvo has joined #openstack-nova | 19:20 | |
*** iamjarvo has quit IRC | 19:20 | |
*** iamjarvo has joined #openstack-nova | 19:21 | |
*** iamjarvo has quit IRC | 19:21 | |
*** achanda has quit IRC | 19:21 | |
*** iamjarvo has joined #openstack-nova | 19:21 | |
*** iamjarvo has quit IRC | 19:22 | |
*** iamjarvo has joined #openstack-nova | 19:22 | |
*** vladikr has quit IRC | 19:26 | |
openstackgerrit | garyk proposed openstack/nova: VMware: use vCenter instead of VC https://review.openstack.org/189593 | 19:26 |
alaski | melwitt: using version 3 was never really supported right? just based on that bug I'm not sure it's worth a revert | 19:27 |
garyk | melwitt: maybe that is something that should be raised on the list | 19:27 |
melwitt | alaski: it wasn't, it was being silently changed to '2'. concern is people who wrote client = client.Client('3', ...) and client.do_stuff() code used to work and now it doesn't | 19:28 |
melwitt | so I wasn't sure | 19:28 |
*** amotoki has joined #openstack-nova | 19:29 | |
alaski | melwitt: gotcha, was just reading further and seeing that. definitely not revert worthy imo, better to fail clearly than continue incorrect behavior | 19:29 |
alaski | we broke a use case that was completely incorrect and didn't do what users thought it did is what I mean to say | 19:30 |
bauzas | dansmith: around ? was just wondering how to provide something like a NamedTuple as a field ? | 19:30 |
dansmith | bauzas: dict? | 19:30 |
bauzas | dansmith: eg. I have a list of tuples like (host, node) | 19:30 |
bauzas | dansmith: yeah, there is a ListofDicts | 19:31 |
melwitt | alaski: yeah, what you're saying makes sense. | 19:31 |
bauzas | dansmith: but I was thinking it was not cool since any dict could be done | 19:31 |
dansmith | bauzas: yeah, maybe | 19:31 |
dansmith | bauzas: what do you need to store that for? | 19:31 |
bauzas | dansmith: that's for the Retry object used by the ReqSpec | 19:31 |
dansmith | bauzas: which would just be a list of hosts if we didn't have the uuber stupid host,node thing :( | 19:32 |
bauzas | dansmith: there, we have a list of lists like [['host1, 'node1'], ['host2, 'node2']] which is quite ugly | 19:32 |
dansmith | it's ugly regardless | 19:32 |
bauzas | yeah... :( | 19:32 |
dansmith | jaypipes: host-and-node continues to assault infantile felines ^ :( | 19:33 |
bauzas | dansmith: wondering if providing a list of ComputeNode object would be cooler | 19:33 |
bauzas | dansmith: since all of that is actually just a Computenode | 19:34 |
*** amotoki has quit IRC | 19:34 | |
dansmith | bauzas: yeah, it would be a sparse compute node object | 19:34 |
dansmith | with just those two things | 19:34 |
*** achanda has joined #openstack-nova | 19:34 | |
dansmith | that's much heavier in terms of messaging load, but maybe not that bad | 19:34 |
dansmith | how many things will be in the list? | 19:34 |
*** nihilifer has quit IRC | 19:35 | |
bauzas | dansmith: for the moment, we're just checking the list and add the tuple if reschedule need | 19:35 |
dansmith | bauzas: so it would just be $num_retries long/? | 19:36 |
bauzas | dansmith: so I think a not-saved ComputeNode with just 2 fields set (host, hypervisor_hostname) is enough | 19:36 |
*** tonytan4ever has joined #openstack-nova | 19:36 | |
*** baoli has joined #openstack-nova | 19:36 | |
bauzas | dansmith: nope, we need to know which nodes were bad | 19:36 |
*** pixelbeat_ has joined #openstack-nova | 19:36 | |
openstackgerrit | Jorge Niedbalski R. proposed openstack/nova: Detach iSCSI latest path for latest disk https://review.openstack.org/135382 | 19:38 |
*** haypo has joined #openstack-nova | 19:43 | |
*** oro has quit IRC | 19:46 | |
openstackgerrit | Vilobh Meshram proposed openstack/nova-specs: Servicegroup foundational refactoring for Control Plane https://review.openstack.org/190322 | 19:48 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: API: make sure a blank volume with no size is rejected https://review.openstack.org/188913 | 19:48 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: Make Nova better at keeping track of volume sizes in BDM https://review.openstack.org/188914 | 19:48 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: compute: Make swap_volume with resize updates BDM size https://review.openstack.org/188915 | 19:48 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: snapshot: Add device_name to the snapshot bdms https://review.openstack.org/190323 | 19:49 |
openstackgerrit | Nikola Dipanov proposed openstack/nova: Fix overloading of block device on boot by device name https://review.openstack.org/190324 | 19:49 |
dansmith | bauzas: okay | 19:49 |
dansmith | bauzas: it means your version is tied to the compute node version, but otherwise it seems like a win | 19:49 |
bauzas | dansmith: that's exactly my thoughts :/ | 19:50 |
bauzas | dansmith: I don't like to have a tied Computenode object but it seems it would be far cleaner IMHO | 19:50 |
dansmith | yeah | 19:50 |
*** iamjarvo has quit IRC | 19:50 | |
vilobhmm | jaypipes : As per your feedback on https://review.openstack.org/#/c/138607/28 and we have spoken on this before…wrote a new spec…Servicegroup foundational refactoring for Control Plane https://review.openstack.org/190322 | 19:51 |
bauzas | dansmith: ack, let's see how to fix that uglyness | 19:52 |
bauzas | dansmith: thanks for your help | 19:52 |
*** aix has joined #openstack-nova | 19:53 | |
*** achanda has quit IRC | 19:55 | |
*** achanda has joined #openstack-nova | 19:56 | |
*** Mike_D_laptop has joined #openstack-nova | 19:57 | |
*** dboik has joined #openstack-nova | 19:57 | |
*** pixelbeat_ has quit IRC | 19:57 | |
mriedem | sdague: i haven't heard back from the db2 ci maintainer since yesterday | 19:58 |
*** haypo has left #openstack-nova | 19:58 | |
*** annashen has quit IRC | 19:59 | |
*** achanda has quit IRC | 20:00 | |
*** dboik_ has quit IRC | 20:00 | |
mriedem | sdague: i see the ML thread, i'll post that in the internal issue | 20:00 |
*** setuid has quit IRC | 20:01 | |
*** iamjarvo has joined #openstack-nova | 20:01 | |
*** arnaud____ has joined #openstack-nova | 20:01 | |
*** setuid has joined #openstack-nova | 20:01 | |
mriedem | dims: so how deprecated is apiclient in oslo-incubator? | 20:02 |
mriedem | dims: given novaclient is not deprecated and novaclient is using it | 20:02 |
dims | mriedem: doing minimum changes to it | 20:02 |
dims | what do you need mriedem? | 20:04 |
mriedem | dims: was thinking about adding __contains__ to base.Resource so in novaclient i can do: if 'reserved' in fixed_ip: | 20:04 |
mriedem | it looks like hasattr will work though? | 20:05 |
*** arnaud____ has quit IRC | 20:06 | |
mriedem | i'll just use that | 20:06 |
mriedem | nvm | 20:06 |
dims | mriedem: ack | 20:06 |
*** sdake has joined #openstack-nova | 20:06 | |
*** marun has joined #openstack-nova | 20:08 | |
ijw | Can anyone recall why we didn't make it possible to add and enumerate nova VIF plugging drivers? I seem to recall there was a reason you couldn't just add a new VIF plugging type by adding a chunk of non-core code, but I can't remember what it was | 20:09 |
ijw | Cos it seems to be a problem we repeatedly stumble over when doing new nova-neutron stuff | 20:10 |
ijw | (nd I'm thinking specifically libvirt here) | 20:10 |
*** burt has quit IRC | 20:11 | |
openstackgerrit | Vilobh Meshram proposed openstack/nova-specs: Servicegroup foundational refactoring for Control Plane https://review.openstack.org/190322 | 20:13 |
mriedem | ijw: no idea, it's gross though, and the same thing happens with cinder volume drivers | 20:14 |
mriedem | fwiw getting changes into nova to talk to a new vif type is generally not controversial | 20:15 |
mriedem | ijw: you probably want to talk to danpb when he's around | 20:15 |
openstackgerrit | Vilobh Meshram proposed openstack/nova-specs: Servicegroup foundational refactoring for Control Plane https://review.openstack.org/190322 | 20:15 |
ijw | Yeah, I do, and he'll know, but I checked first to see if he was there... | 20:15 |
ijw | I want to get to the point that we can add trivial drivers in Nova, test them, and *then* commit them upstream. But the other point is that when you add a (now out of tree) new driver in Neutron you actually can't supply 3rd party VIF plugging for Nova. | 20:16 |
mriedem | oh i see what you're saying, when the config option was removed | 20:17 |
*** VW has quit IRC | 20:17 | |
mriedem | to load the driver | 20:17 |
ijw | Yeah - which was, for what it's worth, not a bad thing - the generic driver (which is not really a driver any more) was the right approach | 20:18 |
ijw | But there is no ability to add types to it without changing Nova, because they're all in a single file and there's no discovery | 20:18 |
ijw | I dream of the day where VIF drivers register themselves with a (hypervisor, VIF type) and Nova tells Neutron what types it can offer... | 20:18 |
ijw | I believe the worry was that if that's done by autodiscovery then the behaviour of Openstack is at the whim of what else you might have installed on the system. | 20:19 |
*** VW has joined #openstack-nova | 20:19 | |
dansmith | no, the worry was that people would consider it a stable plugin interface | 20:20 |
dansmith | which is what you're describing | 20:20 |
mriedem | this is the change that deprecates the option in icehouse https://github.com/openstack/nova/commit/fdffaab6171562487a404963dbf6b7f1f9469a65 - https://github.com/openstack/nova/commit/7561c8ded211d53e8745d1420a73b82bd0fc35cf was the commit to remove it | 20:20 |
mriedem | ijw: this was a ML thread on the topic https://www.mail-archive.com/openstack-dev@lists.openstack.org/msg30174.html | 20:21 |
*** marun has quit IRC | 20:22 | |
mriedem | ^ was all pre-neutron moving drivers/plugins out of tree | 20:24 |
mriedem | so that muddles things a bit | 20:24 |
*** VW has quit IRC | 20:24 | |
*** setuid has quit IRC | 20:26 | |
*** annegentle has quit IRC | 20:26 | |
openstackgerrit | Andrew Laski proposed openstack/nova: DNM Cells: Provide example request flows for v2 https://review.openstack.org/190209 | 20:26 |
*** setuid has joined #openstack-nova | 20:27 | |
cfriesen | jaypipes: lifeless: you guys remember the glance-api file descriptor thing from yesterday? I just reproduced it again and this time it's been almost 10 minutes and the file descriptor is still hanging around. | 20:30 |
*** VW has joined #openstack-nova | 20:30 | |
lifeless | \o/ | 20:30 |
lifeless | cfriesen: strace -p $glance-api-pid ? | 20:31 |
cfriesen | lifeless: one sec, have to disable some monitoring stuff that doesn't like strace. | 20:31 |
*** lpetrut has quit IRC | 20:32 | |
cfriesen | lifeless: http://paste.openstack.org/show/282092/ | 20:32 |
cfriesen | ~10 secs between poll calls | 20:33 |
lifeless | so no obvious cile copying or anything | 20:34 |
cfriesen | nope | 20:34 |
cfriesen | lsof shows: | 20:34 |
cfriesen | python 86732 root 16r REG 147,3 4294967296 262155 /opt/cgcs/glance/images/5e039b35-8fb8-4e08-bdec-7e4d7b922978 (deleted) | 20:34 |
*** baoli_ has joined #openstack-nova | 20:34 | |
lifeless | jaypipes: ^ | 20:34 |
*** ivasev has quit IRC | 20:34 | |
lifeless | cfriesen: ok so thats not the cached copy AIUI ? | 20:35 |
cfriesen | this is with the cache middleware enabled | 20:35 |
cfriesen | but that's the actual glance image file. In this test there wasn't enough room to store the copy in /etc/nova/instances/_base on the compute node | 20:36 |
cfriesen | then I deleted the image before the fd had gone away | 20:37 |
*** sdake has quit IRC | 20:37 | |
*** hemna is now known as hemnafk | 20:38 | |
*** tjones1 has quit IRC | 20:39 | |
*** baoli has quit IRC | 20:39 | |
lifeless | cfriesen: lsof time on the server | 20:39 |
lifeless | cfriesen: lets see if the http fd is still open | 20:39 |
*** Mike_D_laptop has quit IRC | 20:40 | |
*** marun has joined #openstack-nova | 20:40 | |
cfriesen | lifeless: just the glance-api process has the file open | 20:41 |
lifeless | cfriesen: not the file | 20:41 |
lifeless | cfriesen: the socket | 20:41 |
cfriesen | for that glance-api process? | 20:41 |
lifeless | yeah | 20:41 |
cfriesen | lifeless: http://paste.openstack.org/show/282131/ | 20:42 |
*** oro has joined #openstack-nova | 20:43 | |
lifeless | so python 86732 root 13u IPv4 73359764 0t0 TCP controller:9292->compute-0:52632 (ESTABLISHED) | 20:43 |
lifeless | thats suspect to me | 20:43 |
lifeless | cfriesen: can you see how long its been open ? | 20:43 |
ijw | mriedem: In fact, if you look at the stevedore-driven plugging mechanisms, there's a configuration file (except, when deployed, it usually *isn't* a configuration file, it's a source code file that shouldn't be changed and will be overwritten on upgrade) listing options. My concern is that this makes it impossible to plug 3rd party code into certain plugpoints where it would be valuable. | 20:45 |
ijw | (This may be a Neutron-only thing, but I don't think so.) | 20:46 |
openstackgerrit | Dmitry Guryanov proposed openstack/nova: libvirt: resize ploop disks with prl_disk_tool https://review.openstack.org/182390 | 20:46 |
mriedem | ijw: as pointed out in that mailing list, there was also a pluggable volume driver thing too that was removed | 20:46 |
cfriesen | lifeless: how do I check connection established timestamp? netstat? | 20:46 |
ijw | mriedem: Yeah. Perhaps we should address the underlying problem, then, and have a config item that lists drivers that should register themselves as options. | 20:47 |
ijw | (for each category) | 20:47 |
ijw | I fantasise, given that half the day has passed and I've yet to find the time for the 10 minute journey to the office ;) | 20:48 |
mriedem | ijw: my understanding from the ML thread and history is that there is no stable API for this and that's part of the reason for killing it | 20:48 |
mriedem | the pluggability i mean | 20:48 |
dansmith | mriedem: yep | 20:48 |
mriedem | didn't compute manager and/or conductor used to be pluggable? | 20:48 |
ijw | mriedem: I would phrase that as 'we choose not to define that API as stable', but yes, it's a fair point. | 20:49 |
dansmith | not ever conductor, that I know of, but yeah | 20:49 |
mriedem | so compute manager | 20:49 |
dansmith | ijw: right, that's exactly it | 20:49 |
mriedem | i vaguely remember that causing some powervc shenanigans | 20:49 |
*** alex_klimov has joined #openstack-nova | 20:49 | |
dansmith | mriedem: jaypipes had a nice list of "pluggable interfaces" that were to be nuked in that thread | 20:50 |
ijw | The libvirt plugging drivers are awkward, because they have to return a load of crap that (in the case of libvirt) can be folded into a config file in a way that suits the rest of the config file over which they have no control | 20:50 |
ijw | I shall rephrase and make that tuple (hypervisor, interface-version, plug-type) | 20:50 |
*** dboik_ has joined #openstack-nova | 20:51 | |
ijw | As long as the registration API remained stable, then you could put the new version in place prior to an upgrade and have it picked up when nova changed version | 20:51 |
ijw | (and ignored beforehand) | 20:51 |
lifeless | cfriesen: sorry, my daughter is using me as a climbing frame atm | 20:52 |
cfriesen | lifeless: no worries. I've got my own. | 20:52 |
ijw | Pluggable interfaces serve a purpose when they're not there because they're architecture wank (technical term) | 20:52 |
lifeless | cfriesen: the thing is to figure out if there is a stuck http connection open from nova | 20:52 |
lifeless | cfriesen: and then use that to eliminate glance-api as the source of the bug | 20:52 |
lifeless | cfriesen: another way - bounce the nova-compute process | 20:52 |
*** dboik_ has quit IRC | 20:52 | |
cfriesen | lifeless: that I can do | 20:52 |
lifeless | cfriesen: if I'm right, the file will disappear immediately more or less | 20:52 |
cfriesen | file gone on server | 20:53 |
openstackgerrit | Matt Riedemann proposed openstack/python-novaclient: WIP: Show reserved status for a fixed_ip if available https://review.openstack.org/190346 | 20:53 |
*** mwagner_afk has quit IRC | 20:53 | |
*** thangp has quit IRC | 20:54 | |
cfriesen | lifeless: so somehow we ended up with a socket in nova that didn't get closed | 20:54 |
cfriesen | lifeless: and that was keeping the glance one open | 20:54 |
*** dboik has quit IRC | 20:54 | |
bauzas | oh oh oh, live migrate doesn't check the retries | 20:54 |
*** zzzeek has joined #openstack-nova | 20:54 | |
bauzas | it also doesn't claim | 20:55 |
bauzas | uh uh | 20:55 |
lifeless | cfriesen: yes, exactly | 20:55 |
* bauzas just wonders if we should just say EXPERIMENTAL for live-migrate :) | 20:55 | |
lifeless | cfriesen: so - this is the thing I was proposing when I looked at the iterator stuff in nova's side | 20:55 |
*** pixelbeat_ has joined #openstack-nova | 20:56 | |
lifeless | cfriesen: so what I'd suggest as the next step is to figure out the type of the iterator in nova's code | 20:56 |
lifeless | so we can look at it and see if it has an appropraite close hooked in somewhere | 20:56 |
lifeless | rather than relying on gc | 20:56 |
openstackgerrit | Dmitry Guryanov proposed openstack/nova: libvirt: resize ploop disks with prl_disk_tool https://review.openstack.org/182390 | 20:57 |
*** annashen has joined #openstack-nova | 20:57 | |
*** jdurgin1 has joined #openstack-nova | 21:01 | |
openstackgerrit | Philip Marc Schwartz proposed openstack/nova: Implement online schema migrations https://review.openstack.org/154521 | 21:03 |
*** belmoreira has quit IRC | 21:04 | |
*** ildikov_ has joined #openstack-nova | 21:04 | |
*** belmoreira has joined #openstack-nova | 21:04 | |
*** belmoreira has quit IRC | 21:04 | |
*** jdurgin has quit IRC | 21:04 | |
*** thorst has quit IRC | 21:05 | |
cfriesen | lifeless: I'm stopped with pdb in nova-compute, but image_chunks.__class__ just returns <type 'generator'> | 21:05 |
lifeless | ouch | 21:06 |
lifeless | may need to work back up the code to the point where its acquired or something | 21:06 |
openstackgerrit | Vilobh Meshram proposed openstack/nova-specs: Servicegroup foundational refactoring for Control Plane https://review.openstack.org/190322 | 21:08 |
*** angdraug has joined #openstack-nova | 21:11 | |
*** angdraug has quit IRC | 21:11 | |
*** angdraug has joined #openstack-nova | 21:11 | |
openstackgerrit | Nicolas Simonds proposed openstack/nova: Use utf8_bin collation on the flavor extra-specs table in MySQL https://review.openstack.org/190300 | 21:14 |
*** Sukhdev has quit IRC | 21:14 | |
*** ociuhandu has quit IRC | 21:15 | |
*** agentle has joined #openstack-nova | 21:16 | |
*** VW has quit IRC | 21:16 | |
*** oro has quit IRC | 21:18 | |
cfriesen | lifeless: not totally certain, but I think we're hitting glanceclient.v1.images.ImageManager.data() | 21:18 |
lifeless | cfriesen: yup, and its buggy | 21:20 |
openstackgerrit | Matt Riedemann proposed openstack/nova: Add libvirt-python to test-requirements.txt to enforce min version https://review.openstack.org/183706 | 21:21 |
*** ijw has quit IRC | 21:22 | |
*** sdake has joined #openstack-nova | 21:22 | |
lifeless | cfriesen: try this https://review.openstack.org/190359 | 21:23 |
cfriesen | lifeless: will try. what's the actual underlying thing that we're calling close() on? | 21:24 |
lifeless | a requests body object | 21:24 |
lifeless | pretty sure it has close on it, but hey, we'll find otu | 21:24 |
cfriesen | heh | 21:24 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/190362 | 21:26 |
cfriesen | lifeless: if this works you should reference my bug in your fix | 21:27 |
*** moshele has quit IRC | 21:28 | |
lifeless | cfriesen: tag, you're it. | 21:29 |
lifeless | cfriesen: the client probably needs tests etc. I was just demonstrating through code. | 21:30 |
lifeless | cfriesen: (review -d it, add yourself as co-authored-by, write tests, profit) | 21:30 |
*** zzzeek has quit IRC | 21:30 | |
*** amotoki has joined #openstack-nova | 21:31 | |
cfriesen | lifeless: works. will let you know how it goes | 21:31 |
*** bnemec has quit IRC | 21:31 | |
*** tonytan4ever has quit IRC | 21:32 | |
cfriesen | lifeless: thanks for the help | 21:32 |
*** jdurgin has joined #openstack-nova | 21:32 | |
*** tonytan4ever has joined #openstack-nova | 21:32 | |
*** jdurgin1 has quit IRC | 21:32 | |
lifeless | assuming it works. it may not | 21:33 |
lifeless | now, I have to disappear | 21:33 |
*** amotoki has quit IRC | 21:36 | |
*** tonytan4ever has quit IRC | 21:37 | |
*** baoli_ has quit IRC | 21:39 | |
openstackgerrit | Karim Boumedhel proposed openstack/nova: report an error for hw:numa_nodes=0 https://review.openstack.org/190267 | 21:39 |
*** baoli has joined #openstack-nova | 21:39 | |
*** baoli has quit IRC | 21:46 | |
*** baoli has joined #openstack-nova | 21:46 | |
openstackgerrit | Ming Yang proposed openstack/nova: Consolidates scheduler utils tests into a single file. https://review.openstack.org/190301 | 21:50 |
openstackgerrit | melanie witt proposed openstack/nova: Send Instance object to cells instance_update_at_top https://review.openstack.org/176518 | 21:50 |
*** arnaud____ has joined #openstack-nova | 21:50 | |
*** iamjarvo has quit IRC | 21:51 | |
openstackgerrit | Karim Boumedhel proposed openstack/nova: report an error for hw:numa_nodes=0 https://review.openstack.org/190267 | 21:51 |
*** svalleru has joined #openstack-nova | 21:51 | |
openstackgerrit | Ming Yang proposed openstack/nova: Minor refactor in nova.scheduler.filters.utils https://review.openstack.org/190369 | 21:52 |
*** aysyanne has quit IRC | 21:52 | |
openstackgerrit | Karim Boumedhel proposed openstack/nova: report an error for hw:numa_nodes=0 https://review.openstack.org/190267 | 21:54 |
*** marcusvrn has quit IRC | 21:54 | |
*** mriedem is now known as mriedem_away | 21:54 | |
*** arnaud____ has quit IRC | 21:54 | |
*** iamjarvo has joined #openstack-nova | 21:55 | |
*** agireud has quit IRC | 21:57 | |
*** esmute has quit IRC | 22:00 | |
*** dguitarbite has quit IRC | 22:01 | |
openstackgerrit | Ming Yang proposed openstack/nova: Minor refactor in nova.scheduler.filters.utils https://review.openstack.org/190369 | 22:01 |
*** lsmola has quit IRC | 22:01 | |
*** changbl has quit IRC | 22:05 | |
*** kylek3h has quit IRC | 22:05 | |
*** salv-orlando has joined #openstack-nova | 22:05 | |
marun | does anyone else find it insane that devstack runs an instance of nova-api per core by default? | 22:06 |
*** Sukhdev has joined #openstack-nova | 22:07 | |
marun | I can remember a devstack aio needing 1gb of ram, now it's pushing 4gb. And that's without any instances! | 22:07 |
openstackgerrit | Qiu Yu proposed openstack/nova: Force iSCSI disconnect after timeout https://review.openstack.org/167815 | 22:08 |
openstackgerrit | Earle F. Philhower, III proposed openstack/nova: Add support for HGST Solutions volumes to libvirt https://review.openstack.org/186594 | 22:09 |
openstackgerrit | Ming Yang proposed openstack/nova: Minor refactor in nova.scheduler.filters.utils https://review.openstack.org/190369 | 22:10 |
*** alexpilotti has quit IRC | 22:13 | |
*** alexpilotti has joined #openstack-nova | 22:17 | |
*** salv-orlando has quit IRC | 22:18 | |
openstackgerrit | Earle F. Philhower, III proposed openstack/nova: Add support for HGST Solutions volumes to libvirt https://review.openstack.org/186594 | 22:23 |
*** artom has quit IRC | 22:23 | |
*** artom has joined #openstack-nova | 22:24 | |
*** mdrabe has quit IRC | 22:25 | |
*** harlowja has quit IRC | 22:27 | |
*** agentle has quit IRC | 22:28 | |
*** baoli has quit IRC | 22:28 | |
*** Alexandra_ has joined #openstack-nova | 22:30 | |
*** Alexandra_ has quit IRC | 22:30 | |
*** amotoki has joined #openstack-nova | 22:32 | |
*** harlowja has joined #openstack-nova | 22:32 | |
*** jgrimm is now known as zz_jgrimm | 22:34 | |
*** salv-orlando has joined #openstack-nova | 22:36 | |
*** baoli has joined #openstack-nova | 22:36 | |
*** amotoki has quit IRC | 22:36 | |
*** ssurana has left #openstack-nova | 22:43 | |
*** jecarey has quit IRC | 22:46 | |
*** lucas-dinner has quit IRC | 22:48 | |
*** baoli has quit IRC | 22:50 | |
*** baoli has joined #openstack-nova | 22:51 | |
*** angdraug has quit IRC | 22:51 | |
*** krtaylor has quit IRC | 22:54 | |
*** ZZelle_ has quit IRC | 22:57 | |
*** baoli has quit IRC | 22:58 | |
*** hanzhang has joined #openstack-nova | 23:00 | |
*** shz has quit IRC | 23:02 | |
*** n0ano has quit IRC | 23:04 | |
*** nic has quit IRC | 23:05 | |
*** alex_klimov has quit IRC | 23:06 | |
*** marun has quit IRC | 23:06 | |
openstackgerrit | Ming Yang proposed openstack/nova: Consolidates scheduler utils tests into a single file. https://review.openstack.org/190301 | 23:09 |
*** ijw has joined #openstack-nova | 23:09 | |
*** sdake has quit IRC | 23:10 | |
*** atuvenie has quit IRC | 23:14 | |
*** iamjarvo has quit IRC | 23:20 | |
openstackgerrit | Miguel Lavalle proposed openstack/nova-specs: Integrate DNS resolution with Neutron using Nova hostname https://review.openstack.org/90150 | 23:26 |
*** angdraug has joined #openstack-nova | 23:28 | |
*** asettle has joined #openstack-nova | 23:28 | |
*** svalleru has quit IRC | 23:31 | |
*** browne has quit IRC | 23:34 | |
vilobhmm | dansmith : ping | 23:37 |
dansmith | vilobhmm: ack | 23:38 |
*** krtaylor has joined #openstack-nova | 23:38 | |
*** arnaud____ has joined #openstack-nova | 23:39 | |
vilobhmm | does nova maintain *instance snapshot QUOTA* because when i did nova quota-show —tenant <id> i cannot find *instance snapshot* as one the resource http://paste.openstack.org/show/282434/ | 23:39 |
vilobhmm | dansmith : ^^ | 23:39 |
dansmith | vilobhmm: I think that would count against your glance quota | 23:39 |
*** david-lyle has quit IRC | 23:40 | |
dansmith | or cinder if snapshotting to a volume of course | 23:40 |
*** electrocucaracha has quit IRC | 23:40 | |
vilobhmm | dansmith : but looks like glance does not maintain any quota…atleast no cli to display quota | 23:41 |
vilobhmm | second part I agree | 23:41 |
dansmith | vilobhmm: https://review.openstack.org/#/c/37993/ | 23:42 |
dansmith | what version? | 23:42 |
vilobhmm | dansmith : juno | 23:43 |
vilobhmm | thanks for the link | 23:43 |
dansmith | seems like that should be in there though | 23:43 |
dansmith | but yeah, I dunno | 23:43 |
dansmith | I don't know a whole lot about nova quotas, but if it's not in the quota output, I'd assume we rely on glance | 23:43 |
*** arnaud____ has quit IRC | 23:43 | |
vilobhmm | sure thanks for the pointer as always useful…will jump on the glance channel and check | 23:43 |
openstackgerrit | p-draigbrady proposed openstack/nova: rbd: use rbd snapshots and cloning for ephemeral disk snapshots https://review.openstack.org/125963 | 23:43 |
openstackgerrit | Ming Yang proposed openstack/nova: Minor refactor in nova.scheduler.filters.utils https://review.openstack.org/190369 | 23:44 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/190362 | 23:49 |
*** harlowja has quit IRC | 23:54 | |
*** harlowja has joined #openstack-nova | 23:55 | |
*** gmann has quit IRC | 23:56 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!