| *** VW_ has joined #openstack-nova | 00:00 | |
| *** annegentle has joined #openstack-nova | 00:00 | |
| *** _cerberus_ has quit IRC | 00:02 | |
| *** jwcroppe_ has quit IRC | 00:03 | |
| openstackgerrit | Miguel Lavalle proposed openstack/nova-specs: Integrate DNS resolution with Neutron using Nova hostname https://review.openstack.org/90150 | 00:04 |
|---|---|---|
| *** zzzeek has joined #openstack-nova | 00:05 | |
| *** annegentle has quit IRC | 00:05 | |
| *** tjones1 has quit IRC | 00:11 | |
| *** hemna is now known as hemnafk | 00:11 | |
| *** arnaud____ has joined #openstack-nova | 00:12 | |
| *** _diana_ has quit IRC | 00:12 | |
| *** SourabhP has joined #openstack-nova | 00:13 | |
| *** faraujo has quit IRC | 00:15 | |
| *** otter768 has joined #openstack-nova | 00:16 | |
| *** arnaud____ has quit IRC | 00:17 | |
| *** sdake_ has joined #openstack-nova | 00:17 | |
| *** armax has joined #openstack-nova | 00:18 | |
| *** patrickeast has quit IRC | 00:18 | |
| *** _diana_ has joined #openstack-nova | 00:19 | |
| *** sdake has quit IRC | 00:21 | |
| *** markvoelker has joined #openstack-nova | 00:24 | |
| *** zzzeek has quit IRC | 00:26 | |
| *** sdake_ has quit IRC | 00:27 | |
| *** markvoelker has quit IRC | 00:30 | |
| *** armax has quit IRC | 00:30 | |
| *** ijw_ has quit IRC | 00:32 | |
| *** pixelb has quit IRC | 00:32 | |
| *** dsanders has quit IRC | 00:34 | |
| *** tjones1 has joined #openstack-nova | 00:37 | |
| openstackgerrit | Tracy Jones proposed openstack/nova: WIP - convert direct db calls to objects https://review.openstack.org/194851 | 00:38 |
| *** Longgeek has joined #openstack-nova | 00:38 | |
| *** tjones1 has quit IRC | 00:39 | |
| *** tjones1 has joined #openstack-nova | 00:39 | |
| *** IanGovett has quit IRC | 00:42 | |
| *** tjones1 has quit IRC | 00:44 | |
| *** Longgeek has quit IRC | 00:44 | |
| *** Administrator__ has quit IRC | 00:49 | |
| *** Kevin_Zheng has joined #openstack-nova | 00:49 | |
| *** Jeffrey4l has joined #openstack-nova | 00:50 | |
| mriedem | mikal: yes you missed all the fun in the nova meeting this morning talking about specs | 00:54 |
| mikal | mriedem: "fun"? | 00:55 |
| mriedem | 'fun' | 00:56 |
| mriedem | exactly | 00:56 |
| mriedem | s/fun/rage/ | 00:57 |
| claudiub|2 | *inserts nerdy joke here* Everyone deserves a red power ring! :D | 00:57 |
| mikal | Sigh | 00:57 |
| *** tjones1 has joined #openstack-nova | 00:58 | |
| *** otter768 has quit IRC | 00:58 | |
| *** liusheng has joined #openstack-nova | 00:58 | |
| *** tjones1 has quit IRC | 00:59 | |
| *** tjones2 has joined #openstack-nova | 00:59 | |
| *** tjones2 has quit IRC | 00:59 | |
| *** sdake has joined #openstack-nova | 01:00 | |
| *** annegentle has joined #openstack-nova | 01:01 | |
| *** husanu has quit IRC | 01:01 | |
| *** jwcroppe has joined #openstack-nova | 01:02 | |
| *** VW_ has quit IRC | 01:02 | |
| *** husanu has joined #openstack-nova | 01:03 | |
| *** tobe has joined #openstack-nova | 01:05 | |
| *** jwcroppe_ has joined #openstack-nova | 01:05 | |
| *** jwcroppe has quit IRC | 01:06 | |
| *** annegentle has quit IRC | 01:06 | |
| *** david-ly_ has joined #openstack-nova | 01:06 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: glance:check the num_retries option https://review.openstack.org/123318 | 01:08 |
| *** tjones1 has joined #openstack-nova | 01:08 | |
| *** david-lyle has quit IRC | 01:09 | |
| openstackgerrit | Artom Lifshitz proposed openstack/nova-specs: Support --all-tenants in floatingips-list https://review.openstack.org/171389 | 01:13 |
| artom | Could not be more last minute ^^ | 01:15 |
| *** park_heijlong has joined #openstack-nova | 01:20 | |
| *** ctrath has joined #openstack-nova | 01:22 | |
| *** tjones1 has quit IRC | 01:25 | |
| *** SourabhP has quit IRC | 01:25 | |
| *** ijw_ has joined #openstack-nova | 01:32 | |
| *** jasondotstar has joined #openstack-nova | 01:33 | |
| *** ctrath has quit IRC | 01:38 | |
| *** ctrath has joined #openstack-nova | 01:38 | |
| *** ctrath has quit IRC | 01:39 | |
| *** boris-42 has quit IRC | 01:42 | |
| *** radek_ has joined #openstack-nova | 01:52 | |
| *** radek__ has quit IRC | 01:55 | |
| *** husanu has quit IRC | 01:58 | |
| *** jasondotstar has quit IRC | 01:59 | |
| *** husanu has joined #openstack-nova | 02:00 | |
| *** arnaud____ has joined #openstack-nova | 02:01 | |
| *** aluria has quit IRC | 02:03 | |
| *** dosaboy has quit IRC | 02:04 | |
| *** wolsen has quit IRC | 02:05 | |
| *** arnaud____ has quit IRC | 02:05 | |
| *** wolsen has joined #openstack-nova | 02:05 | |
| *** dosaboy has joined #openstack-nova | 02:06 | |
| *** aluria has joined #openstack-nova | 02:06 | |
| *** _diana_ has quit IRC | 02:09 | |
| *** jasondotstar has joined #openstack-nova | 02:10 | |
| *** markvoelker has joined #openstack-nova | 02:14 | |
| *** dims has joined #openstack-nova | 02:14 | |
| *** artom has quit IRC | 02:14 | |
| *** dims_ has quit IRC | 02:15 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: glance:check the num_retries option https://review.openstack.org/123318 | 02:17 |
| *** dsanders has joined #openstack-nova | 02:18 | |
| *** markvoelker has quit IRC | 02:18 | |
| *** dims has quit IRC | 02:20 | |
| *** lan has quit IRC | 02:22 | |
| *** dims has joined #openstack-nova | 02:22 | |
| *** annegentle has joined #openstack-nova | 02:24 | |
| *** lan has joined #openstack-nova | 02:25 | |
| *** RuiChen has quit IRC | 02:30 | |
| *** RuiChen has joined #openstack-nova | 02:30 | |
| *** _diana_ has joined #openstack-nova | 02:31 | |
| *** annashen has joined #openstack-nova | 02:32 | |
| *** jwcroppe_ has quit IRC | 02:33 | |
| *** yamahata has quit IRC | 02:33 | |
| *** dsanders1 has joined #openstack-nova | 02:34 | |
| *** dsanders has quit IRC | 02:34 | |
| *** dsanders1 has quit IRC | 02:35 | |
| *** alexpilotti has quit IRC | 02:36 | |
| *** r-daneel has quit IRC | 02:39 | |
| *** annegentle has quit IRC | 02:39 | |
| *** Longgeek has joined #openstack-nova | 02:40 | |
| *** jasondotstar has quit IRC | 02:42 | |
| *** unicell has quit IRC | 02:42 | |
| *** xyhuang has joined #openstack-nova | 02:43 | |
| *** cfriesen has joined #openstack-nova | 02:44 | |
| *** READ10 has quit IRC | 02:47 | |
| *** Longgeek has quit IRC | 02:52 | |
| *** gyee has quit IRC | 02:58 | |
| *** otter768 has joined #openstack-nova | 02:59 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: add support for muti_statuses_query from migration https://review.openstack.org/169631 | 03:00 |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: DB:support for migration objects multi-statuses https://review.openstack.org/169604 | 03:00 |
| *** dims has quit IRC | 03:01 | |
| *** arnaud____ has joined #openstack-nova | 03:01 | |
| *** mriedem has quit IRC | 03:02 | |
| *** otter768 has quit IRC | 03:04 | |
| *** arnaud____ has quit IRC | 03:06 | |
| *** xyang1 has quit IRC | 03:09 | |
| *** radek_ has quit IRC | 03:13 | |
| *** mdrabe has joined #openstack-nova | 03:13 | |
| *** xyang1 has joined #openstack-nova | 03:14 | |
| *** armax has joined #openstack-nova | 03:17 | |
| *** mdrabe has quit IRC | 03:21 | |
| *** claudiub|2 has quit IRC | 03:22 | |
| *** boris-42 has joined #openstack-nova | 03:22 | |
| *** mestery has joined #openstack-nova | 03:23 | |
| openstackgerrit | junxu proposed openstack/nova: Add hard stop server support in nova api https://review.openstack.org/192972 | 03:24 |
| *** liusheng has quit IRC | 03:26 | |
| *** ijw has quit IRC | 03:29 | |
| *** ijw has joined #openstack-nova | 03:29 | |
| *** ijw__ has joined #openstack-nova | 03:31 | |
| *** ijw has quit IRC | 03:35 | |
| *** VW_ has joined #openstack-nova | 03:35 | |
| *** VW_ has quit IRC | 03:35 | |
| *** norman has joined #openstack-nova | 03:38 | |
| norman | Hi all , sorry, what's the channel of novaclient? I did not find in the https://launchpad.net/python-novaclient | 03:40 |
| *** annegentle has joined #openstack-nova | 03:40 | |
| norman | sdague, ping | 03:41 |
| *** Longgeek has joined #openstack-nova | 03:44 | |
| *** annegentle has quit IRC | 03:45 | |
| *** _diana_ has quit IRC | 03:46 | |
| *** Longgeek has quit IRC | 03:46 | |
| *** Longgeek has joined #openstack-nova | 03:47 | |
| *** jwcroppe has joined #openstack-nova | 03:47 | |
| norman | Hi all, I want to understand the flow of the nova client ,such as when I execute the 'nova floating-ip-associate xxxxx', I want to know which method in class@module will be called eventually, are there some existing document described that? or someone can drop me a note? thanks | 03:48 |
| *** david-ly_ is now known as david-lyle | 03:51 | |
| *** lpetrut has joined #openstack-nova | 03:52 | |
| *** asettle has joined #openstack-nova | 03:53 | |
| *** VW_ has joined #openstack-nova | 03:54 | |
| *** annashen has quit IRC | 03:57 | |
| openstackgerrit | Jesse J. Cook proposed openstack/nova: WIP: cells: weighted cell list randomization https://review.openstack.org/195757 | 03:57 |
| *** rajesht has joined #openstack-nova | 04:00 | |
| *** arnaud____ has joined #openstack-nova | 04:02 | |
| *** markvoelker has joined #openstack-nova | 04:02 | |
| *** mnestratov has quit IRC | 04:03 | |
| *** vilobhmm has joined #openstack-nova | 04:04 | |
| *** isd has joined #openstack-nova | 04:04 | |
| *** markvoelker has quit IRC | 04:07 | |
| *** arnaud____ has quit IRC | 04:07 | |
| *** unicell has joined #openstack-nova | 04:08 | |
| *** liusheng has joined #openstack-nova | 04:10 | |
| *** annashen has joined #openstack-nova | 04:11 | |
| *** VW_ has quit IRC | 04:13 | |
| *** deepthi has joined #openstack-nova | 04:13 | |
| *** rushiagr_away is now known as rushiagr | 04:13 | |
| *** annashen has quit IRC | 04:14 | |
| *** rotbeard has quit IRC | 04:15 | |
| *** xyang1 has quit IRC | 04:15 | |
| *** ildikov has quit IRC | 04:17 | |
| *** unicell has quit IRC | 04:18 | |
| *** annashen has joined #openstack-nova | 04:19 | |
| *** unicell has joined #openstack-nova | 04:20 | |
| *** Marga_ has joined #openstack-nova | 04:20 | |
| *** achanda has joined #openstack-nova | 04:26 | |
| openstackgerrit | ijw-ubuntu proposed openstack/nova-specs: Add a binding_type negotiation when attempting to bind a Neutron port https://review.openstack.org/190917 | 04:29 |
| *** Marga_ has quit IRC | 04:30 | |
| openstackgerrit | ijw-ubuntu proposed openstack/nova-specs: Add a binding type negotiation between Nova and Neutron https://review.openstack.org/190917 | 04:30 |
| *** Marga_ has joined #openstack-nova | 04:31 | |
| openstackgerrit | ijw-ubuntu proposed openstack/nova-specs: Add a binding type negotiation between Nova and Neutron https://review.openstack.org/190917 | 04:34 |
| *** lpetrut has quit IRC | 04:34 | |
| ijw__ | Third time lucky | 04:34 |
| *** ijw__ is now known as ijw | 04:34 | |
| *** baoli has quit IRC | 04:35 | |
| *** vilobhmm has quit IRC | 04:36 | |
| *** tobe has quit IRC | 04:40 | |
| *** arnaud____ has joined #openstack-nova | 04:40 | |
| *** annegentle has joined #openstack-nova | 04:41 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Server list filter with tenant_id implies all_tenants https://review.openstack.org/195412 | 04:42 |
| *** ijw__ has joined #openstack-nova | 04:44 | |
| *** isd has quit IRC | 04:45 | |
| *** annegentle has quit IRC | 04:46 | |
| *** ijw has quit IRC | 04:48 | |
| *** ijw has joined #openstack-nova | 04:48 | |
| *** ijw__ has quit IRC | 04:52 | |
| *** mdbooth has quit IRC | 04:54 | |
| *** rotbeard has joined #openstack-nova | 04:55 | |
| *** _gryf|afk is now known as _gryf | 04:58 | |
| *** otter768 has joined #openstack-nova | 05:00 | |
| *** achanda has quit IRC | 05:00 | |
| *** garyk has joined #openstack-nova | 05:01 | |
| *** mmedvede has quit IRC | 05:03 | |
| *** otter768 has quit IRC | 05:04 | |
| *** dave-mccowan has quit IRC | 05:05 | |
| *** Nikolay_St has quit IRC | 05:09 | |
| *** tojuvone has joined #openstack-nova | 05:14 | |
| *** tobe has joined #openstack-nova | 05:19 | |
| *** sudipto has joined #openstack-nova | 05:19 | |
| *** Sukhdev has joined #openstack-nova | 05:21 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova: Server list filter with tenant_id implies all_tenants https://review.openstack.org/195412 | 05:25 |
| *** Nikolay_St has joined #openstack-nova | 05:26 | |
| *** jmccrory has quit IRC | 05:27 | |
| *** jmccrory has joined #openstack-nova | 05:28 | |
| *** annashen has quit IRC | 05:29 | |
| *** ajayaa has joined #openstack-nova | 05:31 | |
| openstackgerrit | garyk proposed openstack/nova: VMware: improve instance names on VC https://review.openstack.org/166608 | 05:31 |
| *** Nikolay_St has quit IRC | 05:34 | |
| *** sdake has quit IRC | 05:35 | |
| *** baoli has joined #openstack-nova | 05:36 | |
| *** baoli has quit IRC | 05:41 | |
| *** annegentle has joined #openstack-nova | 05:42 | |
| *** rushiagr is now known as rushiagr_away | 05:43 | |
| *** mtanino has joined #openstack-nova | 05:43 | |
| *** achanda has joined #openstack-nova | 05:43 | |
| *** annegentle has quit IRC | 05:47 | |
| *** ijw has quit IRC | 05:48 | |
| *** husanu has quit IRC | 05:49 | |
| *** oro has joined #openstack-nova | 05:50 | |
| *** SourabhP has joined #openstack-nova | 05:51 | |
| *** markvoelker has joined #openstack-nova | 05:51 | |
| *** husanu has joined #openstack-nova | 05:51 | |
| *** kaisers has quit IRC | 05:51 | |
| *** boris-42 has quit IRC | 05:52 | |
| *** kaisers has joined #openstack-nova | 05:53 | |
| *** asettle has quit IRC | 05:53 | |
| *** links has joined #openstack-nova | 05:55 | |
| *** markvoelker has quit IRC | 05:56 | |
| *** arnaud____ has quit IRC | 05:56 | |
| *** Murali has joined #openstack-nova | 05:58 | |
| *** Sukhdev has quit IRC | 05:58 | |
| *** sdake has joined #openstack-nova | 06:00 | |
| *** mtanino is now known as mtanino_zzz | 06:00 | |
| *** arnaud____ has joined #openstack-nova | 06:00 | |
| *** ijw has joined #openstack-nova | 06:01 | |
| *** abhishekk has joined #openstack-nova | 06:01 | |
| *** ajayaa has quit IRC | 06:04 | |
| *** tobe has quit IRC | 06:05 | |
| *** park_heijlong has quit IRC | 06:06 | |
| *** mestery has quit IRC | 06:08 | |
| openstackgerrit | Hiroyuki Eguchi proposed openstack/nova: allow live migration in case of a booted from volume instance https://review.openstack.org/195885 | 06:09 |
| *** radek_ has joined #openstack-nova | 06:12 | |
| *** mtanino_zzz has quit IRC | 06:14 | |
| *** achanda_ has joined #openstack-nova | 06:14 | |
| *** rushiagr_away is now known as rushiagr | 06:15 | |
| *** sdake has quit IRC | 06:16 | |
| *** achanda has quit IRC | 06:16 | |
| *** radek_ has quit IRC | 06:18 | |
| *** sdake has joined #openstack-nova | 06:19 | |
| *** park_heijlong has joined #openstack-nova | 06:21 | |
| *** armax has quit IRC | 06:21 | |
| *** ajayaa has joined #openstack-nova | 06:21 | |
| *** armax has joined #openstack-nova | 06:22 | |
| *** achanda_ has quit IRC | 06:24 | |
| *** boris-42 has joined #openstack-nova | 06:26 | |
| *** ajayaa has quit IRC | 06:27 | |
| *** eliqiao has joined #openstack-nova | 06:28 | |
| *** annashen has joined #openstack-nova | 06:30 | |
| *** ijw__ has joined #openstack-nova | 06:30 | |
| *** radek_ has joined #openstack-nova | 06:32 | |
| *** ijw has quit IRC | 06:33 | |
| openstackgerrit | Ankit Agrawal proposed openstack/nova: Test patch libvirt race condition (do not merge) https://review.openstack.org/191732 | 06:34 |
| *** lpetrut has joined #openstack-nova | 06:36 | |
| *** browne has quit IRC | 06:37 | |
| *** eliqiao has quit IRC | 06:38 | |
| *** tobe has joined #openstack-nova | 06:40 | |
| *** sdake has quit IRC | 06:41 | |
| *** armax has quit IRC | 06:41 | |
| *** sdake has joined #openstack-nova | 06:41 | |
| *** sdake has quit IRC | 06:41 | |
| *** armax has joined #openstack-nova | 06:41 | |
| *** ajayaa has joined #openstack-nova | 06:42 | |
| *** belmoreira has joined #openstack-nova | 06:43 | |
| *** annegentle has joined #openstack-nova | 06:43 | |
| *** coolsvap|away is now known as coolsvap | 06:44 | |
| *** armax has quit IRC | 06:46 | |
| *** eliqiao has joined #openstack-nova | 06:46 | |
| *** annegentle has quit IRC | 06:48 | |
| *** eliqiao1 has joined #openstack-nova | 06:50 | |
| *** eliqiao1 has left #openstack-nova | 06:50 | |
| *** oomichi has joined #openstack-nova | 06:52 | |
| *** eliqiao has quit IRC | 06:53 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: add support for muti_statuses_query from migration https://review.openstack.org/169631 | 06:55 |
| *** scheuran has joined #openstack-nova | 06:56 | |
| *** coolsvap is now known as coolsvap|away | 06:56 | |
| *** arnaud____ has quit IRC | 06:56 | |
| *** lpetrut has quit IRC | 07:00 | |
| *** oomichi has quit IRC | 07:01 | |
| *** otter768 has joined #openstack-nova | 07:01 | |
| *** ijw_ has quit IRC | 07:02 | |
| *** oomichi has joined #openstack-nova | 07:03 | |
| *** ijw has joined #openstack-nova | 07:03 | |
| *** otter768 has quit IRC | 07:06 | |
| *** husanu has quit IRC | 07:06 | |
| *** husanu has joined #openstack-nova | 07:08 | |
| *** oro has quit IRC | 07:10 | |
| *** lpetrut has joined #openstack-nova | 07:10 | |
| openstackgerrit | Michal Rostecki proposed openstack/nova: [WIP] Dedicate aggregates for specific tenants https://review.openstack.org/195783 | 07:12 |
| *** eglynn has quit IRC | 07:12 | |
| *** lpetrut has quit IRC | 07:14 | |
| *** pask81 has joined #openstack-nova | 07:15 | |
| *** claudiub has joined #openstack-nova | 07:19 | |
| *** SourabhP has quit IRC | 07:19 | |
| *** lsmola has joined #openstack-nova | 07:21 | |
| openstackgerrit | javeme proposed openstack/nova: Fix typos detected by toolkit misspellings. https://review.openstack.org/195898 | 07:21 |
| *** annashen has quit IRC | 07:23 | |
| *** bkopilov has quit IRC | 07:23 | |
| *** jcoufal has joined #openstack-nova | 07:25 | |
| *** liusheng has quit IRC | 07:26 | |
| *** liusheng has joined #openstack-nova | 07:27 | |
| *** Nikolay_St has joined #openstack-nova | 07:30 | |
| *** ijw__ has quit IRC | 07:31 | |
| bauzas | aloha nova | 07:32 |
| *** bauzas is now known as bauwser | 07:32 | |
| *** achanda has joined #openstack-nova | 07:32 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: glance:check the num_retries option https://review.openstack.org/123318 | 07:33 |
| *** yamahata has joined #openstack-nova | 07:33 | |
| *** nihilifer has quit IRC | 07:33 | |
| *** Kevin_Zheng has quit IRC | 07:35 | |
| *** rushiagr is now known as rushiagr_away | 07:36 | |
| *** cfriesen has quit IRC | 07:40 | |
| *** nihilifer has joined #openstack-nova | 07:47 | |
| *** oomichi has quit IRC | 07:48 | |
| *** markus_z has joined #openstack-nova | 07:55 | |
| *** Jeffrey4l has quit IRC | 08:00 | |
| *** jlanoux has joined #openstack-nova | 08:03 | |
| *** Marga_ has quit IRC | 08:04 | |
| *** lucasagomes has joined #openstack-nova | 08:04 | |
| *** derekh has joined #openstack-nova | 08:06 | |
| *** oro has joined #openstack-nova | 08:09 | |
| openstackgerrit | Stefan Amann proposed openstack/nova-specs: Support host type specific block volume attachment https://review.openstack.org/194163 | 08:10 |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: hypervisor support matrix: add CLI commands to features https://review.openstack.org/195647 | 08:11 |
| *** ildikov has joined #openstack-nova | 08:11 | |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: hypervisor support matrix: add feature "serial console" https://review.openstack.org/180912 | 08:12 |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: hypervisor support matrix: add feature "evacuate" https://review.openstack.org/195648 | 08:12 |
| *** jwcroppe has quit IRC | 08:14 | |
| *** mdbooth has joined #openstack-nova | 08:15 | |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Added server tags support in nova-api https://review.openstack.org/128940 | 08:16 |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Added method exists to the Tag object https://review.openstack.org/178103 | 08:16 |
| *** mnestratov has joined #openstack-nova | 08:16 | |
| *** jwcroppe has joined #openstack-nova | 08:18 | |
| *** eglynn has joined #openstack-nova | 08:19 | |
| *** jwcroppe has quit IRC | 08:19 | |
| *** jmccrory has quit IRC | 08:19 | |
| *** jwcroppe has joined #openstack-nova | 08:19 | |
| *** claudiub has quit IRC | 08:21 | |
| *** jwcroppe has quit IRC | 08:21 | |
| *** jwcroppe has joined #openstack-nova | 08:22 | |
| *** jwcroppe has quit IRC | 08:22 | |
| *** jwcroppe has joined #openstack-nova | 08:23 | |
| *** jwcroppe has quit IRC | 08:24 | |
| *** jwcroppe has joined #openstack-nova | 08:28 | |
| *** moshele has joined #openstack-nova | 08:28 | |
| *** yassine_ has joined #openstack-nova | 08:29 | |
| *** danpb has joined #openstack-nova | 08:29 | |
| openstackgerrit | Li Yingjun proposed openstack/nova: Validate maximum limit for quota https://review.openstack.org/165069 | 08:29 |
| *** jwcroppe has quit IRC | 08:30 | |
| *** jwcroppe has joined #openstack-nova | 08:30 | |
| *** kylek3h has joined #openstack-nova | 08:31 | |
| *** jmccrory has joined #openstack-nova | 08:34 | |
| openstackgerrit | Li Yingjun proposed openstack/nova: Validate maximum limit for quota https://review.openstack.org/165069 | 08:34 |
| *** park_heijlong has quit IRC | 08:38 | |
| *** yamahata has quit IRC | 08:40 | |
| *** Murali_ has joined #openstack-nova | 08:40 | |
| *** neiljerr` has joined #openstack-nova | 08:42 | |
| *** Marga_ has joined #openstack-nova | 08:43 | |
| *** Murali has quit IRC | 08:43 | |
| *** Murali_ is now known as Murali | 08:43 | |
| *** Guest54958 is now known as BobBall | 08:43 | |
| *** annegentle has joined #openstack-nova | 08:44 | |
| *** annegentle has quit IRC | 08:49 | |
| *** pask81 has quit IRC | 08:50 | |
| *** park_heijlong has joined #openstack-nova | 08:50 | |
| kashyap | markus_z: So, your serial console change worked when I applied it on Compute node (in my 2-node DevStack) env. I'll re-test by applying it on Control node to (which also runs the `nova-compute` process). | 08:50 |
| tdurakov | dhellmann: Hi Doug! got question about https://review.openstack.org/#/c/173834/ Could you explain why this patch haven't merged yet? | 08:53 |
| *** nihilifer has quit IRC | 08:53 | |
| *** lsmola has quit IRC | 08:53 | |
| markus_z | kashyap: yeah, good to hear. Did you need to change anything in the image? "os_command_line" image prop or anything? | 08:54 |
| kashyap | markus_z: Didn't do any change, just pulled in current `git` of all the involved projects. Now I tore down the setup again and re-building with the patch applied on both nodes. | 08:55 |
| *** jwcroppe has quit IRC | 08:56 | |
| *** moshele has quit IRC | 08:58 | |
| *** jwcroppe has joined #openstack-nova | 08:58 | |
| *** moshele has joined #openstack-nova | 08:58 | |
| markus_z | kashyap: cool, thanks for double-checking! I noticed recently an issue with "virsh console <domain> --devname <pty-device>" which concerns me. | 08:59 |
| kashyap | markus_z: What is the issue? | 08:59 |
| markus_z | kashyap: virsh doesn't create the connection to the guest. I have to debug that. | 09:00 |
| kashyap | Okay | 09:00 |
| kashyap | markus_z: Meanwhile, with your change, I still see the "There is no script for 294 version" on the Control node DevStack (running Fedora 22) - http://paste.openstack.org/show/321305/ | 09:01 |
| kashyap | _Without_ your change, DevStack builds just fine. | 09:02 |
| *** otter768 has joined #openstack-nova | 09:02 | |
| markus_z | kashyap: Did you try another change? Maybe there is an issue with the update mechanism. | 09:02 |
| kashyap | markus_z: What do you mean "another change"? Anything other patch apart from yours? If so, no. | 09:02 |
| *** tobe has quit IRC | 09:02 | |
| kashyap | s/Anything/Any/ | 09:02 |
| openstackgerrit | Sergey Nikitin proposed openstack/nova-specs: Implement server instance tagging https://review.openstack.org/177112 | 09:03 |
| markus_z | kashyap: Jepp, plain devstack patched with another patch set than mine. | 09:03 |
| kashyap | markus_z: Just yours, nothing else. | 09:03 |
| *** Marga_ has quit IRC | 09:04 | |
| markus_z | kashyap: I'll test it too. Keeping you in the loop. | 09:04 |
| *** veena has joined #openstack-nova | 09:05 | |
| kashyap | No rush. I'm debugging a unit test too (haven't done much of this before, so learning a few things :-) ) | 09:05 |
| *** otter768 has quit IRC | 09:07 | |
| johnthetubaguy | markus_z: just adding some notes to the evacate change | 09:09 |
| johnthetubaguy | markus_z: loving the updates, just a bit worried about evacuate. | 09:09 |
| *** husanu1 has joined #openstack-nova | 09:09 | |
| markus_z | johnthetubaguy: Yes, please, feel free to do so. | 09:09 |
| markus_z | johnthetubaguy: What are the concerns with "evacuate"? | 09:10 |
| johnthetubaguy | markus_z: I am thinking we should only cover nova evacate, then talk about the meta operations separatly | 09:10 |
| johnthetubaguy | markus_z: the only reason I say this, is because almost everyone is confused about what evacuate does, its a very confusing name! | 09:10 |
| markus_z | johnthetubaguy: agreed | 09:11 |
| abhishekk | johnthetubaguy: hi boss, could you please review the nova-specs, https://review.openstack.org/135387 | 09:11 |
| markus_z | johnthetubaguy: I tried to express an expectation from the end user side. But I'm lacking experience from an operators view. | 09:12 |
| johnthetubaguy | abhishekk: sorry, I am having a day off spec reviews to try get some other things sorted for Nova | 09:13 |
| johnthetubaguy | markus_z: no problems, I am loving the updates here, I think the CLI call is a great way to talk about the operations | 09:13 |
| abhishekk | johnthetubaguy: actually, I heard Liberty-1 milestone release was last day for approval, is it True? | 09:14 |
| markus_z | johnthetubaguy: I'll try to re-phrase the statement for "evacuate". Just put a -1 and maybe if there's time a few hints what you would expect. | 09:14 |
| openstackgerrit | Sergey Nikitin proposed openstack/nova-specs: Implement server instance tagging https://review.openstack.org/177112 | 09:14 |
| johnthetubaguy | markus_z: awesome, I just added a few ideas on that patch, let me know if they make sense | 09:15 |
| markus_z | johnthetubaguy: thanks! | 09:15 |
| *** bradjones has quit IRC | 09:15 | |
| *** lyarwood has quit IRC | 09:15 | |
| *** lyarwood has joined #openstack-nova | 09:15 | |
| *** nihilifer has joined #openstack-nova | 09:16 | |
| *** bradjones has joined #openstack-nova | 09:16 | |
| *** bradjones has quit IRC | 09:16 | |
| *** bradjones has joined #openstack-nova | 09:16 | |
| *** husanu1 has quit IRC | 09:18 | |
| *** neiljerr` has quit IRC | 09:19 | |
| *** ijw_ has joined #openstack-nova | 09:19 | |
| openstackgerrit | Maxim Nestratov proposed openstack/nova: libvirt: rename parallels driver to virtuozzo https://review.openstack.org/184311 | 09:21 |
| *** boris-42 has quit IRC | 09:22 | |
| *** ndipanov has joined #openstack-nova | 09:22 | |
| *** husanu4 has joined #openstack-nova | 09:23 | |
| *** markus_z has quit IRC | 09:23 | |
| *** park_heijlong has quit IRC | 09:23 | |
| *** Murali_ has joined #openstack-nova | 09:25 | |
| *** rotbeard has quit IRC | 09:26 | |
| *** Murali has quit IRC | 09:28 | |
| *** Murali_ is now known as Murali | 09:28 | |
| *** markus_z has joined #openstack-nova | 09:28 | |
| *** husanu4 has quit IRC | 09:28 | |
| *** markvoelker has joined #openstack-nova | 09:29 | |
| *** karimb has joined #openstack-nova | 09:32 | |
| *** exploreshaifali has joined #openstack-nova | 09:33 | |
| *** markvoelker has quit IRC | 09:33 | |
| *** husanu has quit IRC | 09:39 | |
| *** annegentle has joined #openstack-nova | 09:45 | |
| rajesht | dansmith: you around ? | 09:45 |
| *** moshele has quit IRC | 09:46 | |
| *** Marga_ has joined #openstack-nova | 09:48 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: resize-revert can break anti-affinity policy https://review.openstack.org/160110 | 09:48 |
| *** annegentle has quit IRC | 09:50 | |
| *** nihilifer has quit IRC | 09:50 | |
| danpb | johnthetubaguy: as expected, people are already asking me what the process is for requesting spec exceptions... | 09:53 |
| johnthetubaguy | rajesht: he is on the west coast, and not up yet | 09:53 |
| danpb | johnthetubaguy: presumably you will send the usual email telling people the deadline for specs is passed and saying how to request exceptions ? | 09:54 |
| johnthetubaguy | danpb: it was in the nova-meeting logs, I am going to send an email | 09:54 |
| johnthetubaguy | danpb: yeah | 09:54 |
| rajesht | johnthetubaguy: thanks :) | 09:54 |
| danpb | ok i'll tell them to be patient then | 09:54 |
| johnthetubaguy | rajesht: he is online quite early for west cost, so probably in about 2-3 hours or so | 09:54 |
| johnthetubaguy | danpb: just point them at me, if that helps | 09:54 |
| *** norman has quit IRC | 09:55 | |
| johnthetubaguy | danpb: I should get the email out this morning, need to clear out the agenda and actions from the last nova-meeting before then | 09:56 |
| * johnthetubaguy wonders where to get some antidepressants to get through the spec freeze | 09:57 | |
| danpb | johnthetubaguy: i'm sure i've got a few vendors offering pills in my spam folder | 09:58 |
| johnthetubaguy | danpb: heh | 09:58 |
| *** dims has joined #openstack-nova | 09:59 | |
| *** gszasz has joined #openstack-nova | 10:00 | |
| *** Marga_ has quit IRC | 10:04 | |
| openstackgerrit | Peter Savage proposed openstack/python-novaclient: Added marker functionality to flavours and images https://review.openstack.org/193212 | 10:05 |
| *** mpaolino has joined #openstack-nova | 10:07 | |
| *** ijw has quit IRC | 10:07 | |
| openstackgerrit | Bob Ball proposed openstack/nova: XenAPI: Refactor rotate_xen_guest_logs to avoid races https://review.openstack.org/193647 | 10:09 |
| openstackgerrit | Vipin Balachandran proposed openstack/nova: Detach and terminate conn if Cinder attach fails https://review.openstack.org/186742 | 10:09 |
| *** cfriesen has joined #openstack-nova | 10:09 | |
| *** ankit_ag has joined #openstack-nova | 10:11 | |
| BobBall | johnthetubaguy: ^^ Definitely need your input on that one at some point :) (not necessarily today) | 10:14 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: [WIP] Kicking the oslo.service tires https://review.openstack.org/192900 | 10:17 |
| *** dims has quit IRC | 10:18 | |
| *** liusheng has quit IRC | 10:18 | |
| *** Marga_ has joined #openstack-nova | 10:19 | |
| *** Marga_ has quit IRC | 10:22 | |
| *** Marga_ has joined #openstack-nova | 10:23 | |
| *** annashen has joined #openstack-nova | 10:24 | |
| johnthetubaguy | BobBall: yes bug me next week about that, I should defo take a look | 10:25 |
| openstackgerrit | Alvaro Lopez Garcia proposed openstack/nova: scheduler: weight nodes based on their cached images https://review.openstack.org/173304 | 10:26 |
| openstackgerrit | Alvaro Lopez Garcia proposed openstack/nova: Add new image cache monitor in compute nodes https://review.openstack.org/173303 | 10:26 |
| openstackgerrit | Alvaro Lopez Garcia proposed openstack/nova: vmware: add image_cache_manager property https://review.openstack.org/187602 | 10:26 |
| openstackgerrit | Alvaro Lopez Garcia proposed openstack/nova: compute monitors: set update delta from base class https://review.openstack.org/173987 | 10:26 |
| *** annashen has quit IRC | 10:28 | |
| bauwser | BobBall: do you have any updates about the merge issue for Xen Server CI ? | 10:32 |
| bauwser | BobBall: I mean, I really like doing "xen: recheck" but you know, I'm a bit lazy :) | 10:33 |
| openstackgerrit | Matthew Booth proposed openstack/nova: Add the de-pbr utility https://review.openstack.org/195983 | 10:36 |
| *** links has quit IRC | 10:36 | |
| *** mpaolino has quit IRC | 10:37 | |
| BobBall | bauwser: Should be deploying it again next week | 10:38 |
| bauwser | BobBall: sweet ! | 10:38 |
| *** mpaolino has joined #openstack-nova | 10:39 | |
| BobBall | I think the CI must love you | 10:39 |
| bauwser | BobBall: probably, she wants to see me again | 10:39 |
| BobBall | Must miss your attention or something - which is why it wants to make it talk to you | 10:39 |
| bauwser | BobBall: I could be tempted to give her my phone number, but I'm married with children | 10:40 |
| bauwser | BobBall: so I'm unfortunately only able to discuss with her on a professional manner | 10:40 |
| BobBall | I'm sure she'd be content with that! | 10:41 |
| *** achanda_ has joined #openstack-nova | 10:44 | |
| *** ijw has joined #openstack-nova | 10:44 | |
| *** achanda has quit IRC | 10:47 | |
| johnthetubaguy | garyk: I just added your spec to the freature freeze exception list, as an example: https://etherpad.openstack.org/p/liberty-spec-freeze-exceptions | 10:48 |
| johnthetubaguy | garyk: let me know if that looks reasonable | 10:48 |
| johnthetubaguy | kfox1111: have you got a link to your spec, so I can add it to the freeze list? | 10:49 |
| *** ijw has quit IRC | 10:50 | |
| bauwser | johnthetubaguy: can I blame also my spec for it ? :) | 10:51 |
| abhishekk | johnthetubaguy: can I add my spec in the above list? | 10:51 |
| *** stefan_amann has joined #openstack-nova | 10:51 | |
| johnthetubaguy | abhishekk: you can, I am drafting an email to the ML right now | 10:51 |
| bauwser | johnthetubaguy: oh, seen L9-10... | 10:52 |
| abhishekk | johnthetubaguy: thank you | 10:52 |
| johnthetubaguy | bauwser: yeah, I mean we could add them in there, but it seemed silly to list them in two places! | 10:52 |
| bauwser | johnthetubaguy: makes sense | 10:52 |
| bauwser | johnthetubaguy: btw. I made -1 because I think that one doesn't require a spec https://review.openstack.org/#/c/187739/ | 10:53 |
| bauwser | johnthetubaguy: I know you're not reviewing specs today, but just take some time on Monday to review it if you want | 10:53 |
| bauwser | johnthetubaguy: my point is that's just a trivial change for logging, not really needing a spec | 10:53 |
| bauwser | johnthetubaguy: and we know that we have like low-hanging-fruits for adding logs in Nova | 10:54 |
| stefan_amann | johnthetubaguy: is there any chance to get the support-hosttype-on-attach blueprint approved? I updated the spec a couple of days ago. So it should be ready to go. It is actually a very small change. I have done a similar change to Cinder. This change has been merged | 10:54 |
| johnthetubaguy | stefan_amann: it has to go on the exception list: https://etherpad.openstack.org/p/liberty-spec-freeze-exceptions | 10:54 |
| stefan_amann | ok... thanks. I will put it there | 10:55 |
| *** IanGovett has joined #openstack-nova | 10:56 | |
| *** alexschm has joined #openstack-nova | 10:56 | |
| johnthetubaguy | abhishekk: did you just add the unshelved? | 10:57 |
| scheuran | danpb: thanks for your comment on my macvtap spec. baoli made me aware of, that there is a potential api change in this spec, as I have to enrich the vif dictionary with the "mode" and "devicename" information on the neutron side. Is there anything special to consider in this case? | 10:57 |
| abhishekk | johnthetubaguy: yes | 10:57 |
| abhishekk | or should I hadd it once you sent the mail? | 10:58 |
| abhishekk | * add | 10:58 |
| johnthetubaguy | abhishekk: can you please add a reason why on there, as the other examples have | 10:58 |
| *** dims has joined #openstack-nova | 10:59 | |
| abhishekk | johnthetubaguy: ok, I will | 10:59 |
| johnthetubaguy | abhishekk: thank you | 10:59 |
| abhishekk | johnthetubaguy: thank you as well | 10:59 |
| *** mariusv has joined #openstack-nova | 11:00 | |
| *** mariusv has quit IRC | 11:00 | |
| *** mariusv has joined #openstack-nova | 11:00 | |
| *** mariusv has quit IRC | 11:01 | |
| *** otter768 has joined #openstack-nova | 11:03 | |
| *** ijw has joined #openstack-nova | 11:04 | |
| *** baoli has joined #openstack-nova | 11:07 | |
| *** otter768 has quit IRC | 11:07 | |
| *** ijw has quit IRC | 11:08 | |
| *** nihilifer has joined #openstack-nova | 11:08 | |
| *** claudiub has joined #openstack-nova | 11:11 | |
| *** baoli has quit IRC | 11:12 | |
| *** VW_ has joined #openstack-nova | 11:13 | |
| *** slagle_ is now known as slagle | 11:14 | |
| *** ZZelle has quit IRC | 11:14 | |
| *** ZZelle has joined #openstack-nova | 11:15 | |
| *** tojuvone has quit IRC | 11:15 | |
| *** kbyrne has quit IRC | 11:16 | |
| *** Marga_ has quit IRC | 11:17 | |
| *** links has joined #openstack-nova | 11:17 | |
| *** markvoelker has joined #openstack-nova | 11:17 | |
| *** pixelb has joined #openstack-nova | 11:18 | |
| *** links has quit IRC | 11:19 | |
| openstackgerrit | Ritesh proposed openstack/python-novaclient: Improve hypervisor-show print list https://review.openstack.org/193068 | 11:19 |
| *** kbyrne has joined #openstack-nova | 11:20 | |
| *** kbyrne has quit IRC | 11:21 | |
| openstackgerrit | Bob Ball proposed openstack/nova: XenAPI: Refactor rotate_xen_guest_logs to avoid races https://review.openstack.org/193647 | 11:21 |
| *** kbyrne has joined #openstack-nova | 11:21 | |
| *** markvoelker has quit IRC | 11:22 | |
| *** lpetrut has joined #openstack-nova | 11:25 | |
| *** lucasagomes is now known as lucas-afk | 11:28 | |
| *** ajayaa has quit IRC | 11:32 | |
| *** aix has quit IRC | 11:41 | |
| *** bradjones has quit IRC | 11:42 | |
| *** ijw has joined #openstack-nova | 11:44 | |
| *** karimb has quit IRC | 11:44 | |
| *** bradjones has joined #openstack-nova | 11:45 | |
| *** bradjones has quit IRC | 11:45 | |
| *** bradjones has joined #openstack-nova | 11:45 | |
| *** aysyanne has joined #openstack-nova | 11:45 | |
| *** cfriesen has quit IRC | 11:46 | |
| *** annegentle has joined #openstack-nova | 11:47 | |
| *** _diana_ has joined #openstack-nova | 11:49 | |
| *** ijw has quit IRC | 11:50 | |
| tdurakov | danpb, hi Dan! | 11:50 |
| *** _diana_ has quit IRC | 11:50 | |
| tdurakov | danpb: Are you going to merge this spect in current release: https://review.openstack.org/#/c/168982/ | 11:51 |
| *** ajayaa has joined #openstack-nova | 11:51 | |
| *** annegentle has quit IRC | 11:52 | |
| danpb | it will need an excepton request at this point | 11:53 |
| tdurakov | danpb: if it'll not be approved in current release there is no guarantee that my(https://review.openstack.org/#/c/183057/) would be approved in next window. Do we really want to live with such ugly responses so much time? | 11:54 |
| tdurakov | danpb: could you recommend me, how to make my spec approved faster? | 11:55 |
| *** VW_ has quit IRC | 11:55 | |
| *** _diana_ has joined #openstack-nova | 11:56 | |
| *** achanda_ has quit IRC | 11:58 | |
| *** markvoelker has joined #openstack-nova | 12:00 | |
| tdurakov | danpb: I've added both to exception request. Anything else? | 12:03 |
| *** baoli has joined #openstack-nova | 12:08 | |
| *** aix has joined #openstack-nova | 12:08 | |
| *** mpaolino has quit IRC | 12:08 | |
| *** dims is now known as dimsum__ | 12:09 | |
| *** baoli_ has joined #openstack-nova | 12:10 | |
| *** stefan_amann has quit IRC | 12:10 | |
| openstackgerrit | Przemyslaw Czesnowicz proposed openstack/nova: Add PciNumaInfoWeigher https://review.openstack.org/152915 | 12:11 |
| *** baoli has quit IRC | 12:12 | |
| *** dims_ has joined #openstack-nova | 12:13 | |
| *** baoli_ has quit IRC | 12:14 | |
| *** baoli has joined #openstack-nova | 12:14 | |
| *** deepthi has quit IRC | 12:15 | |
| *** dimsum__ has quit IRC | 12:16 | |
| *** Longgeek_ has joined #openstack-nova | 12:17 | |
| *** exploreshaifali has quit IRC | 12:20 | |
| *** rmstar has joined #openstack-nova | 12:20 | |
| *** packet has joined #openstack-nova | 12:20 | |
| *** Longgeek has quit IRC | 12:21 | |
| openstackgerrit | Radoslav Gerganov proposed openstack/nova: VMware: map one nova-compute to one VC cluster https://review.openstack.org/103916 | 12:23 |
| *** _gryf is now known as _gryf|afk | 12:23 | |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: hypervisor support matrix: add feature "evacuate" https://review.openstack.org/195648 | 12:23 |
| *** _diana_ has quit IRC | 12:26 | |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Implements 'microversions' api type - Part 2 https://review.openstack.org/167408 | 12:27 |
| *** _diana_ has joined #openstack-nova | 12:27 | |
| *** aix has quit IRC | 12:28 | |
| openstackgerrit | Radoslav Gerganov proposed openstack/nova: VMware: Use datastore copy when the image is already in vSphere https://review.openstack.org/186716 | 12:31 |
| *** claudiub|2 has joined #openstack-nova | 12:31 | |
| *** claudiub has quit IRC | 12:32 | |
| *** edmondsw has joined #openstack-nova | 12:32 | |
| *** busterswt has joined #openstack-nova | 12:34 | |
| *** vladikr has joined #openstack-nova | 12:34 | |
| *** dave-mccowan has joined #openstack-nova | 12:35 | |
| *** raildo has joined #openstack-nova | 12:37 | |
| *** cfriesen has joined #openstack-nova | 12:39 | |
| *** rotbeard has joined #openstack-nova | 12:40 | |
| *** jaypipes has joined #openstack-nova | 12:41 | |
| *** Longgeek has joined #openstack-nova | 12:41 | |
| *** eglynn-officeafk has joined #openstack-nova | 12:44 | |
| *** ijw has joined #openstack-nova | 12:44 | |
| *** ijuwang has joined #openstack-nova | 12:44 | |
| *** VW_ has joined #openstack-nova | 12:45 | |
| *** tellesnobrega has joined #openstack-nova | 12:45 | |
| *** Longgeek_ has quit IRC | 12:45 | |
| *** rfolco has joined #openstack-nova | 12:45 | |
| *** VW_ has quit IRC | 12:45 | |
| *** VW_ has joined #openstack-nova | 12:46 | |
| openstackgerrit | Claudiu Belu proposed openstack/nova-specs: Adds Hyper-V vTPM devices spec https://review.openstack.org/195068 | 12:48 |
| *** ijw has quit IRC | 12:50 | |
| *** mestery has joined #openstack-nova | 12:52 | |
| *** packet has quit IRC | 12:53 | |
| *** Sukhdev has joined #openstack-nova | 12:54 | |
| *** mestery has quit IRC | 12:57 | |
| *** veena has quit IRC | 13:00 | |
| *** otter768 has joined #openstack-nova | 13:04 | |
| *** tellesnobrega has quit IRC | 13:04 | |
| *** mwagner_afk has quit IRC | 13:05 | |
| *** busterswt has quit IRC | 13:06 | |
| *** dimsum__ has joined #openstack-nova | 13:07 | |
| *** tellesnobrega has joined #openstack-nova | 13:08 | |
| *** otter768 has quit IRC | 13:08 | |
| *** Marga_ has joined #openstack-nova | 13:09 | |
| *** dboik_ has quit IRC | 13:09 | |
| *** dims_ has quit IRC | 13:10 | |
| *** kaisers has quit IRC | 13:11 | |
| *** kaisers has joined #openstack-nova | 13:13 | |
| *** ankit_ag has quit IRC | 13:13 | |
| *** Murali has quit IRC | 13:15 | |
| *** READ10 has joined #openstack-nova | 13:17 | |
| *** ijuwang_ has joined #openstack-nova | 13:18 | |
| *** ijuwang has quit IRC | 13:18 | |
| *** ijuwang_ is now known as ijuwang | 13:18 | |
| openstackgerrit | Eugeniya Kudryashova proposed openstack/python-novaclient: Raise exception when API does't support microversions https://review.openstack.org/188816 | 13:18 |
| *** fawadkhaliq has joined #openstack-nova | 13:18 | |
| openstackgerrit | Eugeniya Kudryashova proposed openstack/python-novaclient: Raise exception when API does't support microversions https://review.openstack.org/188816 | 13:19 |
| *** rlrossit has joined #openstack-nova | 13:24 | |
| *** mriedem has joined #openstack-nova | 13:25 | |
| *** rajesht has quit IRC | 13:26 | |
| openstackgerrit | Michal Rostecki proposed openstack/nova: [WIP] Dedicate aggregates for specific tenants https://review.openstack.org/195783 | 13:26 |
| *** fawadkhaliq has quit IRC | 13:28 | |
| *** ajayaa has quit IRC | 13:29 | |
| *** ildikov has quit IRC | 13:30 | |
| *** dboik has joined #openstack-nova | 13:31 | |
| *** baoli has quit IRC | 13:31 | |
| *** ivasev has joined #openstack-nova | 13:33 | |
| *** edleafe is now known as figleaf | 13:35 | |
| *** _gryf|afk is now known as _gryf | 13:36 | |
| *** Marga_ has quit IRC | 13:39 | |
| *** thangp has joined #openstack-nova | 13:40 | |
| *** browne has joined #openstack-nova | 13:41 | |
| *** _diana_ has quit IRC | 13:41 | |
| *** ijw has joined #openstack-nova | 13:44 | |
| *** achanda has joined #openstack-nova | 13:45 | |
| *** raildo has quit IRC | 13:45 | |
| *** raildo has joined #openstack-nova | 13:47 | |
| *** boris-42 has joined #openstack-nova | 13:47 | |
| snikitin | mriedem: Hi Matt! Your comments in instance tags spec were addressed. https://review.openstack.org/#/c/177112/ Could you look, please? | 13:47 |
| mriedem | snikitin: sure | 13:47 |
| *** annegentle has joined #openstack-nova | 13:48 | |
| garyk | johnthetubaguy: not sure about the exception freeze. do i need to do anything? | 13:49 |
| *** htruta has joined #openstack-nova | 13:49 | |
| johnthetubaguy | garyk: the ML thread should describe whats required, does that help? | 13:49 |
| *** dansmith is now known as superdan | 13:49 | |
| johnthetubaguy | garyk: http://lists.openstack.org/pipermail/openstack-dev/2015-June/068079.html | 13:49 |
| *** ijw has quit IRC | 13:50 | |
| *** nihilifer has quit IRC | 13:50 | |
| garyk | johnthetubaguy: i am still not sure what to do. the spec has 2 +2's. do i need to add it to he list and hop ehtta someone will it +w? | 13:51 |
| johnthetubaguy | garyk: so the plan is added to the etherpad, merges after the next nova-meeting | 13:52 |
| garyk | ok, thanks. | 13:52 |
| *** ctrath has joined #openstack-nova | 13:52 | |
| *** ctrath has left #openstack-nova | 13:52 | |
| *** annegentle has quit IRC | 13:53 | |
| garyk | so just to be 100% sure I just need to wait another week? | 13:53 |
| *** abhishekk has quit IRC | 13:54 | |
| garyk | have a good weekend | 13:54 |
| mriedem | snikitin: a few more questions about deletes in the data model that i didn't think of before | 13:54 |
| johnthetubaguy | garyk: yes, you should get an answer by the end of next week | 13:54 |
| snikitin | mriedem: Let me look | 13:55 |
| mriedem | kfox1111: you might want to tag your spec exception request in the ML with [nova] so it's filtered properly | 13:55 |
| *** annegentle has joined #openstack-nova | 13:55 | |
| *** alexschm has quit IRC | 13:56 | |
| *** ctrath1 has joined #openstack-nova | 13:57 | |
| *** ijw has joined #openstack-nova | 13:58 | |
| *** VW_ has quit IRC | 13:58 | |
| mriedem | scheuran: i added your macvtap vif type spec here https://etherpad.openstack.org/p/liberty-spec-freeze-exceptions | 13:58 |
| mriedem | scheuran: since i think it was agreed in the nova meeting yesterday to no block on it and danpb agreed in the spec | 13:59 |
| *** VW_ has joined #openstack-nova | 13:59 | |
| *** eharney has joined #openstack-nova | 14:00 | |
| *** salv-orl_ has quit IRC | 14:01 | |
| *** artom has joined #openstack-nova | 14:01 | |
| *** achanda has quit IRC | 14:01 | |
| *** ildikov has joined #openstack-nova | 14:02 | |
| *** jecarey has joined #openstack-nova | 14:03 | |
| *** fawadkhaliq has joined #openstack-nova | 14:04 | |
| *** mmedvede has joined #openstack-nova | 14:05 | |
| *** VW_ has quit IRC | 14:05 | |
| *** VW_ has joined #openstack-nova | 14:05 | |
| *** annegentle has quit IRC | 14:07 | |
| *** fawadk has joined #openstack-nova | 14:08 | |
| *** fawadkhaliq has quit IRC | 14:09 | |
| *** krtaylor has quit IRC | 14:10 | |
| *** browne has quit IRC | 14:10 | |
| *** annegentle has joined #openstack-nova | 14:11 | |
| *** annegentle has quit IRC | 14:11 | |
| garyk | johnthetubaguy: i hope that it will be a positive answer ... | 14:12 |
| johnthetubaguy | garyk: looks that way right now | 14:13 |
| garyk | johnthetubaguy: what about the patches that do not have BP's approved? | 14:13 |
| *** VW_ has quit IRC | 14:13 | |
| garyk | those without specs, for example | 14:13 |
| garyk | https://review.openstack.org/147126 | 14:13 |
| *** VW_ has joined #openstack-nova | 14:13 | |
| garyk | that is the biggest pain point that we have at the moment | 14:13 |
| *** annegentle has joined #openstack-nova | 14:13 | |
| *** annegentle has quit IRC | 14:13 | |
| johnthetubaguy | garyk: I am going through those this afternoon, it was approved yesterday at the meeting, I just need to do the paperwork | 14:14 |
| garyk | ok, thanks | 14:14 |
| johnthetubaguy | garyk: just without getting on the meeting agenda, its hard to spot which blueprints are being proposed, hence the process is to put them forward in the nova-meeting | 14:15 |
| *** mmedvede_ has joined #openstack-nova | 14:15 | |
| *** unicell1 has joined #openstack-nova | 14:15 | |
| *** browne has joined #openstack-nova | 14:16 | |
| garyk | ok, thanks. i will be aware of doing that in the future. | 14:16 |
| *** lyarwood has quit IRC | 14:16 | |
| *** lyarwood has joined #openstack-nova | 14:16 | |
| *** unicell has quit IRC | 14:18 | |
| johnthetubaguy | garyk: if something isn't approved, or just unsure, just ping me, than I can get it fixed quicker | 14:18 |
| *** ajayaa has joined #openstack-nova | 14:19 | |
| *** fawadk has quit IRC | 14:19 | |
| openstackgerrit | Jesse J. Cook proposed openstack/nova: Cells: limit slots by io_ops https://review.openstack.org/195756 | 14:20 |
| openstackgerrit | Jesse J. Cook proposed openstack/nova: WIP: cells: weighted cell list randomization https://review.openstack.org/195757 | 14:20 |
| openstackgerrit | Jesse J. Cook proposed openstack/nova: Cells: Add cells filter RamByInstanceTypeFilter https://review.openstack.org/195755 | 14:20 |
| scheuran | mriedem: thanks. dan already gave his +2, just waiting for another | 14:20 |
| garyk | johnthetubaguy: ok, thanks | 14:21 |
| *** lyarwood has quit IRC | 14:22 | |
| *** lyarwood has joined #openstack-nova | 14:23 | |
| *** r-daneel has joined #openstack-nova | 14:26 | |
| *** Sukhdev has quit IRC | 14:29 | |
| *** alexpilotti has joined #openstack-nova | 14:30 | |
| openstackgerrit | Zhenguo Niu proposed openstack/nova: Remove unused decorator on attach/detach volume https://review.openstack.org/165839 | 14:30 |
| thangp | superdan: ping | 14:31 |
| superdan | thangp: I see the pings, I'm on a call right now | 14:32 |
| thangp | superdan: ok, np | 14:32 |
| *** alexpilotti_ has joined #openstack-nova | 14:35 | |
| *** alexpilotti has quit IRC | 14:36 | |
| *** alexpilotti_ is now known as alexpilotti | 14:36 | |
| mriedem | oh man, i'm going to have a hard time remembering to add things to this when new config options are added in a new module https://review.openstack.org/#/c/180013/ | 14:36 |
| openstackgerrit | garyk proposed openstack/nova: virt: update doctrings https://review.openstack.org/196080 | 14:38 |
| garyk | anyone needing a no braininer review ^ | 14:39 |
| mriedem | dimsum__: with https://review.openstack.org/#/c/180013/ - i don't see https://github.com/openstack/nova/blob/master/nova/utils.py#L142 in there - is that a mistake or are the workarounds options magically pulled into the generated config? | 14:39 |
| *** tonytan4ever has joined #openstack-nova | 14:40 | |
| dimsum__ | mriedem: one sec looking | 14:41 |
| mriedem | garyk: i can't help but -1ing :) | 14:41 |
| dimsum__ | mriedem: see opts.py - " ('workarounds', nova.utils.workarounds_opts), " | 14:42 |
| openstackgerrit | Andrey Pavlov proposed openstack/python-novaclient: Fix resolving image.id in servers.boot https://review.openstack.org/196082 | 14:42 |
| mriedem | dimsum__: oh | 14:42 |
| mriedem | dimsum__: so opts.py is like the 'here is where everything else goes' spot? | 14:42 |
| figleaf | garyk: that's the only kind of review I'm qualified for | 14:42 |
| openstackgerrit | garyk proposed openstack/nova: virt: update doctrings https://review.openstack.org/196080 | 14:43 |
| garyk | mriedem: it is final i am an idiot | 14:43 |
| *** yamahata has joined #openstack-nova | 14:43 | |
| *** xyang1 has joined #openstack-nova | 14:43 | |
| garyk | i guess i can blame it on my keyboard | 14:43 |
| dimsum__ | mriedem: for now. i want to break it into list_opts in different modules, then add entries in setup.cfg - once we do that we can generate different kinds of config files with different contents | 14:43 |
| markus_z | hehe | 14:44 |
| mriedem | dimsum__: different config files? | 14:44 |
| dimsum__ | mriedem: also see the huge DEFAULT list, we need to add groups for them | 14:44 |
| mriedem | like nova-compute.conf, nova-cells.conf? | 14:44 |
| dimsum__ | mriedem: right nova-conductor.conf | 14:44 |
| *** ijw has quit IRC | 14:44 | |
| mriedem | why do we need multiple config files? | 14:44 |
| dimsum__ | (if we choose to) | 14:44 |
| markus_z | garyk: Is there some kind of docstring update initiative ongoing? | 14:44 |
| mriedem | markus_z: no | 14:45 |
| garyk | markus_z: not that i am aware of. we are just trying to improve things... | 14:45 |
| dimsum__ | mriedem: you can run nova-conductor on a box by itself and the virt driver settings is of no use in that. (just as an example) | 14:45 |
| markus_z | mriedem: garyk: :,( | 14:45 |
| garyk | the better the code and code base the better the product. yalla. i need to run. have a good weekend | 14:45 |
| mriedem | dimsum__: well, i get that | 14:46 |
| *** krtaylor has joined #openstack-nova | 14:46 | |
| dimsum__ | mriedem: don't have a firm plan, just showing possibilities | 14:46 |
| mriedem | dimsum__: but, imo it would be more confusing for the ops person to know where all of the options need to go | 14:46 |
| *** _cerberus_ has joined #openstack-nova | 14:46 | |
| mriedem | dimsum__: yeah, i know. | 14:46 |
| *** yamahata has quit IRC | 14:46 | |
| mriedem | i assume that in most cases, when someone is told to set an option in nova.conf, they just do it everywhere to be safe | 14:46 |
| mriedem | since they don't know exactly how the code is using it | 14:46 |
| *** yamahata has joined #openstack-nova | 14:46 | |
| dimsum__ | mriedem: right | 14:46 |
| mriedem | this came up yesterday actually when reviewing a blog post from superdan on rolling upgrades, and where the upgrade_levels option should be set - i assume control nodes but we agreed that most people would just set it everywhere | 14:47 |
| markus_z | mriedem: Right know you have to parse the code to see which services uses which config option, or I'm missing something. | 14:48 |
| kfox1111 | johnthetubaguy: is there a way to get a spec marked priority this late in the game? | 14:48 |
| dimsum__ | mriedem: good feedback | 14:48 |
| mriedem | markus_z: yes, and parsing the code is not something an operator wants to do | 14:48 |
| dimsum__ | mriedem: though you agree on the groups right? DEFAULT has too many things | 14:48 |
| mriedem | markus_z: most are using things like puppet/chef/anisble to just set options and deploy | 14:48 |
| mriedem | dimsum__: sure, although i don't see the need for a group of one | 14:48 |
| markus_z | mriedem: me neither... The configuration reference should list which services use an option IMHO | 14:48 |
| *** armax has joined #openstack-nova | 14:48 | |
| mriedem | which is a lot of stuff in DEFAULT | 14:48 |
| mriedem | dimsum__: are there plans to get a pep8 style job that makes sure all options defined in the code would show up in the generated config in case people forget to duplicate the option registration in these other modules? | 14:50 |
| mriedem | b/c i know i will forget that w/o a job | 14:50 |
| *** vilobhmm has joined #openstack-nova | 14:50 | |
| superdan | mriedem: you *must* set it everywhere | 14:51 |
| superdan | mriedem: because compute nodes talk to compute nodes | 14:52 |
| mriedem | ok, was an example | 14:52 |
| mriedem | but proves my point | 14:52 |
| dimsum__ | mriedem: y, we definitely need something, looking at options - one idea was to salvage something from the old generator.py but its so fragile | 14:53 |
| mriedem | wtf do i set this option to make nova work? on the control or compute nodes? oh, and now, in which of the 10 conf files for nova in the given node | 14:53 |
| *** ZZelle has quit IRC | 14:54 | |
| *** nihilifer has joined #openstack-nova | 14:54 | |
| *** jaypipes is now known as leakypipes | 14:56 | |
| leakypipes | mriedem: so, tags should just be hard-deleted when an instance is hard-deleted. that's all. | 14:56 |
| mriedem | leakypipes: instances aren't hard-deleted though, right? | 14:57 |
| *** belmoreira has quit IRC | 14:57 | |
| *** VW_ has quit IRC | 14:57 | |
| leakypipes | mriedem: sure they are, during nova-manage archive (or purge, or whatever it's called..) | 14:58 |
| *** mmedvede_ has quit IRC | 14:58 | |
| mriedem | leakypipes: i thought that moved stuff to shadow tables | 14:58 |
| *** VW_ has joined #openstack-nova | 14:58 | |
| mriedem | regardless | 14:58 |
| *** htruta has quit IRC | 14:58 | |
| leakypipes | mriedem: have I mentioned I hate soft-delete? :) | 14:58 |
| *** agireud has joined #openstack-nova | 14:59 | |
| mriedem | leakypipes: if you have an instance with tags that you soft delete, and the tags are hard deleted in the process, and then you decide to restore that instance so it's no longer soft deleted, then the tags are gone, right? | 14:59 |
| mriedem | leakypipes: that doesn't mean it doesn't exist | 14:59 |
| *** VW_ has quit IRC | 14:59 | |
| leakypipes | mriedem: yes, the tags are gone. | 14:59 |
| mriedem | so unless the tags spec depends on the 'no soft delete' spec from lxsli, which isn't going to make liberty, and i don't think tags should depend on that, there should be some answer | 14:59 |
| *** VW_ has joined #openstack-nova | 15:00 | |
| mriedem | i guess if the answer is, sorry, f you, it's a feature, that could be fine :) | 15:00 |
| leakypipes | mriedem: or we go back and add soft-delete to the tags implementation. | 15:00 |
| lxsli | :< | 15:00 |
| leakypipes | mriedem: you mean the archival-framework blueprint? | 15:00 |
| mriedem | leakypipes: that's the only solution i see right now to be consistent | 15:00 |
| mriedem | whatever it's called | 15:00 |
| leakypipes | snikitin: ^^ please add that. | 15:00 |
| mriedem | no soft delete was the title | 15:00 |
| lxsli | no-more-soft-delete | 15:01 |
| leakypipes | lxsli: https://review.openstack.org/#/c/137669/ | 15:01 |
| mriedem | snikitin: make sure you add the 'f you it's a feature' comment :P | 15:01 |
| mriedem | leakypipes: was referring to this https://review.openstack.org/#/c/184645/ | 15:01 |
| leakypipes | mriedem: no, I mean add in that tags will need to be "backported" to support soft-delete | 15:01 |
| *** schwicke has joined #openstack-nova | 15:02 | |
| *** karimb has joined #openstack-nova | 15:02 | |
| *** Nikolay_St has quit IRC | 15:02 | |
| lxsli | seems fair, since they're linked to instances - much as I'd rather not soft-delete any more things | 15:02 |
| leakypipes | lxsli: yeah. | 15:03 |
| leakypipes | snikitin: ok, you understand what is being asked of you? can you do that real quick? | 15:03 |
| *** ijuwang has quit IRC | 15:03 | |
| *** VW_ has quit IRC | 15:03 | |
| mriedem | leakypipes: did any part of the tags stuff get into a previous release? like the data model changes? | 15:03 |
| * leakypipes struggling to just concentrate today... back is all sorts of fucked up :( | 15:03 | |
| mriedem | shaq says icey hot is the cure | 15:03 |
| *** VW_ has joined #openstack-nova | 15:03 | |
| kfox1111 | leakypipes: Sorry dude. I have a bad back too, so I understand. :? | 15:04 |
| leakypipes | mriedem: yes, the tags data model, DB API, and nova objects has been in Juno. | 15:04 |
| leakypipes | mriedem: sorry, Kilo... | 15:04 |
| leakypipes | mriedem: this blueprint has been around since Icehouse, though :( | 15:04 |
| *** ijuwang has joined #openstack-nova | 15:04 | |
| mriedem | heh, ok, yeah i'd think we'd need to fix the delete issue before the api is available | 15:04 |
| mriedem | leakypipes: yeah i know | 15:04 |
| mriedem | sorry to find this issue | 15:04 |
| *** ajayaa has quit IRC | 15:05 | |
| *** otter768 has joined #openstack-nova | 15:05 | |
| leakypipes | mriedem: no worries | 15:05 |
| *** salv-orlando has joined #openstack-nova | 15:05 | |
| leakypipes | kfox1111: yeah, it started spasming middle of my back this morning. I was down for about an hour, cou;dn't move... | 15:05 |
| *** mestery has joined #openstack-nova | 15:06 | |
| leakypipes | lxsli: were you aware of the archival-framework spec? | 15:06 |
| kfox1111 | that really sucks. :/ | 15:06 |
| *** Marga_ has joined #openstack-nova | 15:06 | |
| lxsli | leakypipes: yeeess but I haven't really read it | 15:07 |
| mriedem | dimsum__: damn, i guess i need oslo.config installed in site-packages to use tox -e genconfig huh | 15:07 |
| lxsli | leakypipes: from how PaulMurray described it, it can stand in complement to my specs | 15:07 |
| lxsli | leakypipes: some people may want purge, some may want a notification | 15:07 |
| leakypipes | lxsli: a notification? | 15:09 |
| * lxsli reads leakypipes' spec real fast | 15:09 | |
| *** jogo is now known as flashgordon | 15:09 | |
| lxsli | leakypipes: so your archival framework has an archive(ListOfObjects) method | 15:09 |
| *** otter768 has quit IRC | 15:10 | |
| lxsli | leakypipes: the call to that is what I meant | 15:10 |
| *** schwicke has quit IRC | 15:10 | |
| openstackgerrit | sulochan-acharya proposed openstack/nova: xapi: ensure pv driver info is present prior to live-migration https://review.openstack.org/189731 | 15:10 |
| *** VW_ has quit IRC | 15:10 | |
| leakypipes | lxsli: ah, gotcha. | 15:11 |
| *** VW_ has joined #openstack-nova | 15:11 | |
| lxsli | leakypipes: you've described the archival framework as being inside Nova; it'd be better as a 3rd party thing imo to allow multiple simple implementations, rather than a single flexible (complex) one in-tree | 15:12 |
| lxsli | leakypipes: but that means we'd need an o.vo client I guess | 15:12 |
| lxsli | leakypipes: or we serialise to JSON and save freeform JSON records | 15:12 |
| *** dimsum__ has quit IRC | 15:13 | |
| *** VW_ has quit IRC | 15:13 | |
| leakypipes | lxsli: yup, agreed. there are now 4 projects using o.vo, though, so that's good at least. | 15:14 |
| *** VW_ has joined #openstack-nova | 15:14 | |
| *** salv-orlando has quit IRC | 15:14 | |
| openstackgerrit | sulochan-acharya proposed openstack/nova: xapi: ensure pv driver info is present prior to live-migration https://review.openstack.org/189731 | 15:15 |
| *** mestery has quit IRC | 15:17 | |
| lxsli | leakypipes: is there a plan to build a lib containing our nova.object definitions? So a non-Nova process can read them? | 15:17 |
| superdan | hell no | 15:18 |
| lxsli | glad we got that cleared up ^^; | 15:18 |
| superdan | :D | 15:18 |
| danpb | lxsli: nova objects are private impl details for nova | 15:18 |
| lxsli | no argument from me | 15:19 |
| danpb | external processes should only be talking to nova via the public rest apis, or other official extension points | 15:19 |
| *** VW_ has quit IRC | 15:20 | |
| *** scheuran has quit IRC | 15:20 | |
| superdan | danpb: he means things that are "friends of nova" that might be examining archive records with object models in them, or the like | 15:20 |
| superdan | but still | 15:20 |
| *** VW_ has joined #openstack-nova | 15:20 | |
| superdan | they can "from nova import objects" if they want to do that, IMHO | 15:20 |
| bauwser | ooooooh | 15:21 |
| * superdan uses friend in the C++ sense here | 15:21 | |
| * bauwser remembers all the pleasure he had to import nova in a separate repo | 15:21 | |
| *** Piet has joined #openstack-nova | 15:21 | |
| bauwser | (for the records, was mainly the fault of oslo.cfg (well, the incubator one) | 15:22 |
| *** VW_ has quit IRC | 15:23 | |
| * danpb mutters something about oslo.cfg, global variables and evilness | 15:23 | |
| leakypipes | lxsli: yeah, what danpb and superdan said. :) | 15:24 |
| *** VW_ has joined #openstack-nova | 15:24 | |
| leakypipes | danpb: yuuuup. If you noticed, I removed all the global CONF crap from os-vif :) | 15:24 |
| lxsli | leakypipes: real sorry to hear about your back btw, they don't kid around :( | 15:24 |
| mriedem | danpb: i'm already raging on oslo.config this morning | 15:25 |
| mriedem | tox -e genconfig is hosed | 15:25 |
| *** VW_ has quit IRC | 15:25 | |
| * bauwser is glad to open a can of worms by a Friday | 15:25 | |
| *** beagles is now known as beagles_brb | 15:25 | |
| bauwser | or troll, whatever | 15:25 |
| *** VW_ has joined #openstack-nova | 15:26 | |
| leakypipes | lxsli: yeah, I'm sitting here in my back brace looking like a compete tard. but it's the only thing that keeps me able to type without pain... | 15:27 |
| *** VW_ has quit IRC | 15:27 | |
| lxsli | leakypipes: teach your pugs that Japanese walking-on-backs thing | 15:27 |
| *** nihilifer has quit IRC | 15:27 | |
| mriedem | oh i see the problem | 15:27 |
| mriedem | eureka | 15:27 |
| leakypipes | lxsli: heh :) | 15:27 |
| *** VW_ has joined #openstack-nova | 15:28 | |
| openstackgerrit | Sergey Nikitin proposed openstack/nova-specs: Implement server instance tagging https://review.openstack.org/177112 | 15:28 |
| *** zul has quit IRC | 15:28 | |
| *** VW_ has quit IRC | 15:31 | |
| *** achanda has joined #openstack-nova | 15:31 | |
| *** VW_ has joined #openstack-nova | 15:31 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Unfudge tox -e genconfig wrt missing versionutils module https://review.openstack.org/196108 | 15:32 |
| mriedem | leakypipes: sdague: ^ let's fix tox -e genconfig | 15:32 |
| *** VW_ has quit IRC | 15:32 | |
| *** lucas___ has joined #openstack-nova | 15:32 | |
| *** zul has joined #openstack-nova | 15:33 | |
| *** VW_ has joined #openstack-nova | 15:33 | |
| *** tellesnobrega has quit IRC | 15:34 | |
| snikitin | mriedem: Matt, now tags is soft-deleted :) https://review.openstack.org/#/c/177112/14 | 15:36 |
| *** VW_ has quit IRC | 15:36 | |
| *** VW_ has joined #openstack-nova | 15:36 | |
| mriedem | looking | 15:36 |
| mriedem | snikitin: you need to list a db migration as a work item right? | 15:38 |
| openstackgerrit | Ryan Rossiter proposed openstack/nova: Add hacking check for greenthread.spawn() https://review.openstack.org/195713 | 15:38 |
| mriedem | snikitin: since the schema needs to change | 15:38 |
| bauwser | mriedem: can I just hit you for a question about https://review.openstack.org/#/c/191251/ ? | 15:38 |
| *** tellesnobrega has joined #openstack-nova | 15:38 | |
| snikitin | mriedem: Yes I need to add one more db migration. | 15:39 |
| *** lyarwood has quit IRC | 15:39 | |
| mriedem | bauwser: so, i'm not sure what happened there or if superdan/jogo got to the bottom of the multi-node job failure there | 15:39 |
| *** lyarwood has joined #openstack-nova | 15:39 | |
| snikitin | mriedem: Do I need to add note about that in spec? | 15:39 |
| superdan | mriedem: we did | 15:39 |
| mriedem | bauwser: when i was checking logstash for the thing that was failing, it was only showing up on your change | 15:39 |
| superdan | mriedem: and when I say we, I mean jogo | 15:39 |
| mriedem | snikitin: please | 15:40 |
| mriedem | snikitin: db migrations should geneerally be called out in specs imo | 15:40 |
| superdan | mriedem: both nodes are getting the same hostname and so we're confused about which node to be talking to | 15:40 |
| mriedem | since we don't like db migrations | 15:40 |
| superdan | mriedem: best part is that it's all mordred's fault | 15:40 |
| mriedem | superdan: is this a latent bug? | 15:40 |
| mriedem | superdan: ok | 15:40 |
| bauwser | mriedem: strange then | 15:40 |
| superdan | nope, not nova's deal | 15:40 |
| mriedem | ok | 15:40 |
| bauwser | mriedem: anyway, I'll need to fire another iteration for that | 15:41 |
| *** VW_ has quit IRC | 15:41 | |
| mriedem | bauwser: removed my -1 | 15:41 |
| bauwser | mriedem: ack | 15:41 |
| mriedem | btw, today is farewall party for flashgordon | 15:41 |
| bauwser | ooooh | 15:41 |
| mriedem | i trust everyone brought their party hats and/or anti-depressents | 15:42 |
| *** zzzeek has joined #openstack-nova | 15:42 | |
| flashgordon | mriedem: o/ | 15:42 |
| superdan | :~( | 15:42 |
| superdan | really tempted to land the spec to see if it helps | 15:42 |
| rlrossit | mriedem: I brought beer, which is an anti-anti-depressant | 15:43 |
| *** unicell1 has quit IRC | 15:43 | |
| *** hemna has joined #openstack-nova | 15:43 | |
| *** unicell has joined #openstack-nova | 15:44 | |
| mriedem | not if it's bud light lime | 15:44 |
| openstackgerrit | Sergey Nikitin proposed openstack/nova-specs: Implement server instance tagging https://review.openstack.org/177112 | 15:45 |
| snikitin | mriedem: Done | 15:46 |
| *** SourabhP has joined #openstack-nova | 15:47 | |
| mriedem | snikitin: +1 | 15:47 |
| mriedem | for whatever that's worth | 15:47 |
| *** SourabhP_ has joined #openstack-nova | 15:47 | |
| flashgordon | mriedem: what a but light & clamato | 15:47 |
| *** ajayaa has joined #openstack-nova | 15:48 | |
| leakypipes | mriedem: approved... | 15:48 |
| mriedem | thanks | 15:48 |
| mriedem | flashgordon: idk. i'm pretty sure i saw a thing once where they made a bud + some energy drink thing | 15:49 |
| snikitin | mriedem: tank you | 15:49 |
| snikitin | mriedem: thank you :) | 15:49 |
| mriedem | b/c budweiser saw all the kids getting red bull vodkas at the bar | 15:49 |
| mriedem | and they wanted to cash in | 15:49 |
| *** SourabhP has quit IRC | 15:52 | |
| *** SourabhP_ is now known as SourabhP | 15:52 | |
| *** sdake has joined #openstack-nova | 15:52 | |
| *** lyarwood has quit IRC | 15:52 | |
| *** lyarwood has joined #openstack-nova | 15:53 | |
| *** mriedem is now known as mriedem_lunch | 15:54 | |
| *** moshele has joined #openstack-nova | 15:55 | |
| *** htruta has joined #openstack-nova | 15:56 | |
| bauwser | mriedem_lunch: here is our antidepressive solution https://en.wikipedia.org/wiki/Chartreuse_%28liqueur%29 | 15:57 |
| *** schwicke has joined #openstack-nova | 15:57 | |
| snikitin | johnthetubaguy: Hi John! Today you +2ed instance tags spec. thank you for that. But after discusion with Matt and Jay we decided to make tags soft-deletable. So if you have a free time please look my spec again. It's only 5 line of changes. https://review.openstack.org/#/c/177112/ | 15:57 |
| flashgordon | mriedem_lunch: that sounds aweful | 15:57 |
| flashgordon | bauwser: chartreuse is fun | 15:58 |
| markus_z | bauwser: so many herbs in that thing, it's almost a salad... | 15:58 |
| *** vilobhmm has quit IRC | 15:58 | |
| kfox1111 | johnthetubaguy: is there a way to get a spec marked priority this late in the game? | 15:58 |
| *** ctrath has joined #openstack-nova | 15:58 | |
| *** ctrath has left #openstack-nova | 15:58 | |
| bauwser | flashgordon: eh | 15:58 |
| johnthetubaguy | kfox1111: it depends if it is one of the current priorities or not | 15:59 |
| *** rfolco has quit IRC | 15:59 | |
| johnthetubaguy | snikitin: OK, I should go read that again soon | 15:59 |
| bauwser | markus_z: I'm pretty sure salads can't get you that way... | 15:59 |
| *** jlanoux has quit IRC | 15:59 | |
| markus_z | :) | 16:00 |
| *** ijuwang has quit IRC | 16:00 | |
| *** ijuwang has joined #openstack-nova | 16:00 | |
| bauwser | too bad I'm not at the midcycle, I could have shipped one :) | 16:00 |
| snikitin | johnthetubaguy: thank you | 16:01 |
| kfox1111 | johnthetubaguy: the instance user thing. still trying to figure out how to land it in liberty since it affects so many other non nova projects. | 16:01 |
| kfox1111 | johnthetubaguy: The merge window is so short after a spec exception, I think we'd have to request a RFE as well. So looking for alternatives. | 16:02 |
| *** achanda has quit IRC | 16:02 | |
| *** achanda has joined #openstack-nova | 16:02 | |
| figleaf | snikitin: found a problem with your spec, but it's a small fix | 16:03 |
| *** fawadkhaliq has joined #openstack-nova | 16:03 | |
| *** xyhuang has left #openstack-nova | 16:04 | |
| *** tjones1 has joined #openstack-nova | 16:05 | |
| snikitin | figleaf: looking | 16:05 |
| johnthetubaguy | figleaf: I should read the spec, why did we add soft_delete for tags? | 16:05 |
| figleaf | johnthetubaguy: if an instance was soft-deleted and then recovered, the tags would be gone | 16:06 |
| johnthetubaguy | figleaf: I figured the tags would only be deleted after the final delete, not the soft delete, then we would be fine? | 16:06 |
| johnthetubaguy | figleaf: I should go read the re-read the spec again | 16:07 |
| figleaf | johnthetubaguy: yeah, that would make sense :) | 16:07 |
| figleaf | johnthetubaguy: it was leakypipes who noticed that a little earlier | 16:08 |
| *** arnaud____ has joined #openstack-nova | 16:08 | |
| leakypipes | figleaf: it was mriedem_lunch :) | 16:08 |
| johnthetubaguy | cools | 16:08 |
| flashgordon | kfox1111: so the spec is really light on the problem statement | 16:08 |
| figleaf | leakypipes: oops - didn't mean to give you undeserved credit :) | 16:08 |
| leakypipes | :) | 16:08 |
| *** rushiagr_away is now known as rushiagr | 16:09 | |
| kfox1111 | flashgordon: ok. I'll work on filling it out a bit more. It has a few different use cases all wrapped up into one, and I was trying to keep it light not to be confusing. | 16:09 |
| *** romainh has joined #openstack-nova | 16:10 | |
| flashgordon | kfox1111: also I disagree that adding a new tiny service is more complex then doing it in nova | 16:10 |
| flashgordon | kfox1111: you ended up making it confusing by not being clear about the thing you want to solve | 16:10 |
| bauwser | alaski: johnthetubaguy: just seen https://review.openstack.org/#/c/173243/17 | 16:10 |
| flashgordon | kfox1111: also the I have been working on this for over a year statement, how come the spec went up a month ago | 16:11 |
| bauwser | alaski: I saw your email, I'm going to reply to it, but it deserves some time for thinking about that | 16:11 |
| *** oro has quit IRC | 16:12 | |
| kfox1111 | ~. | 16:12 |
| kfox1111 | ~> | 16:12 |
| bauwser | alaski: the main problem I have is that we have a policy explaining the out-of-tree problem like johnthetubaguy referred | 16:12 |
| kfox1111 | flashgordon: read the spec. it references another. | 16:13 |
| kfox1111 | and there was further work before that spec. | 16:13 |
| kfox1111 | I tried to solve it in barbican and they wanted it resolved more generically. | 16:13 |
| bauwser | alaski: so while I think we can be helping those people, we don't need to change our behaviour because of that | 16:13 |
| kfox1111 | which I don't disagree with, but it makes it harder. since the thing to solve is harder to describe. | 16:13 |
| kfox1111 | flashgordon: Right now, the solution involves barbican, nova, neutron (optionally) and keystone. | 16:14 |
| bauwser | alaski: so, given https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/schemas/v3/scheduler_hints.py | 16:15 |
| kfox1111 | adding a microservice adds yet another service to that list. therefore, it can't be less complex. | 16:15 |
| *** gabriel-bezerra has quit IRC | 16:15 | |
| bauwser | alaski: I'm thinking we should bump a version for changing additionalProperties to true, but not change johnthetubaguy's spec | 16:15 |
| johnthetubaguy | bauwser: you mean relax the validation in v2.1? | 16:16 |
| kfox1111 | my origonal solution only involved nova/keystone, but the keystone folks wanted barbican in the mix. | 16:16 |
| flashgordon | kfox1111: and doing it in nova doesn't involve any complexity? | 16:16 |
| bauwser | johnthetubaguy: alaski has some good concerns about how the hints are validated | 16:16 |
| flashgordon | kfox1111: this sounds like something to hash out on my ML a bit first anyway since it touches so many projects | 16:17 |
| bauwser | johnthetubaguy: and given the discussion you guys have with that, I think we should be clear about how to help out-of-tree filters without really changing our way | 16:17 |
| kfox1111 | I didn't say that. but it doesn't add any more additional complexity to the mix then what's minimally required. having another service adds more complexity then that. | 16:17 |
| flashgordon | and I don't think this is worth a FFE | 16:17 |
| johnthetubaguy | bauwser: ah, yeah, not seen that thread yet | 16:17 |
| *** nihilifer has joined #openstack-nova | 16:17 | |
| bauwser | johnthetubaguy: the main problem is that the current v2.1 API is unhappy with out-of-tree hints because of https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/schemas/v3/scheduler_hints.py#L55 | 16:17 |
| kfox1111 | how unfortunate. :/ | 16:18 |
| johnthetubaguy | bauwser: yep | 16:18 |
| *** derekh has quit IRC | 16:18 | |
| bauwser | johnthetubaguy: so, I assume it's currently not really seen because of the nova client using v2 | 16:18 |
| kfox1111 | so rather then address the problem, we're just continuing to make projects like Octavia work around the problem themselves, | 16:18 |
| kfox1111 | adding further complexity. :/ | 16:18 |
| *** beagles_brb is now known as beagles | 16:18 | |
| johnthetubaguy | bauwser: right | 16:18 |
| flashgordon | my recommendation would be take the problem statement to the ML | 16:18 |
| bauwser | johnthetubaguy: my point is that, I'm OK with bumping a version and changing that field to True | 16:18 |
| snikitin | figleaf: About response status: it should return 404 according API-WG tags guidelines http://specs.openstack.org/openstack/api-wg/guidelines/tags.html | 16:18 |
| kfox1111 | how much discussion does one problem need? Its been discussed on the mailing list a bit, at the summit, in irc, and the spec. | 16:19 |
| *** yamahata has quit IRC | 16:19 | |
| bauwser | johnthetubaguy: meaning that before v2.5, no way to use an out-of-tree filter | 16:19 |
| flashgordon | kfox1111: I may have missed the ML thread on this sorry | 16:19 |
| kfox1111 | We can post for more. but at some point, you get way too many cooks in the kitchen to make progress. :/ | 16:19 |
| *** lucas___ has quit IRC | 16:19 | |
| bauwser | johnthetubaguy: which is not really a problem since we still have the v2 API | 16:19 |
| flashgordon | kfox1111: and getting something wrong has a huge implications here too | 16:19 |
| *** kylek3h has quit IRC | 16:19 | |
| kfox1111 | yeah, but by lack of it, every project is doing it themselves, and then there is way more chance of them doing it wrong, and have to live with it. :/ | 16:20 |
| kfox1111 | I do want to let the spec process work. thats why I've been talking so much with the barbican/keystone folks and getting the ptl support. | 16:21 |
| bauwser | johnthetubaguy: so, related to your spec, I'm thinking on a good way to mimic v2 | 16:21 |
| figleaf | snikitin: that's the current status. It should probably be updated once the http.rst update merges next week, as that talks of resources in general | 16:21 |
| flashgordon | kfox1111: which is why you should discuss this in a place where all relevant parties are involved | 16:21 |
| figleaf | snikitin: tags shouldn't be special snowflakes | 16:21 |
| kfox1111 | but There's a need not to block the process for months because of a freeze in nova too. :/ | 16:21 |
| kfox1111 | flashgordon: is that not the spec itself? | 16:21 |
| bauwser | johnthetubaguy: so, probably, since v2 was just having *no validation at all*, we could consider those additionalProperties as accepted ? | 16:22 |
| johnthetubaguy | kfox1111: this would have been a great Nova session at the summit, its a shame that didn't happen, agreed its an important thing to look at | 16:22 |
| flashgordon | kfox1111: well I don't know because the problem statement is super vague | 16:22 |
| johnthetubaguy | kfox1111: wish I had seen it on the ML before to raise it with Nova folks, but I didn't sorry for that | 16:22 |
| *** achanda has quit IRC | 16:22 | |
| flashgordon | kfox1111: isn't there some openstack security team floating around too | 16:22 |
| kfox1111 | sokay. past is the past. just gota figure out how to make progress. | 16:23 |
| kfox1111 | probably, but I think they are more conserned with security fixes then this sort of thing? | 16:23 |
| johnthetubaguy | kfox1111: agreed | 16:23 |
| flashgordon | kfox1111: IMHO next steps clarify the problem statement and work on a backlog spec | 16:23 |
| kfox1111 | Right now, peoeple are putting certs and passwords in userdata. thats bad. we need to figure out how to make it better. :/ | 16:24 |
| flashgordon | kfox1111: see if we can at least agree where a solution belongs | 16:24 |
| kfox1111 | does putting it in the backlog doom it to M? | 16:24 |
| flashgordon | kfox1111: and be thorough about why this won't work outside of nova etc. | 16:24 |
| kfox1111 | sure. | 16:24 |
| flashgordon | kfox1111: it is realistically not going into Liberty at this point | 16:24 |
| *** haypo has left #openstack-nova | 16:24 | |
| flashgordon | kfox1111: spec went up 1 month ago and not clear consensus yet | 16:25 |
| *** moshele has quit IRC | 16:26 | |
| flashgordon | kfox1111: on a side not I am curious how this is done in !openstack clouds | 16:26 |
| kfox1111 | flashgordon: amazon gives each instance a user. | 16:27 |
| kfox1111 | or so I've been told. | 16:27 |
| flashgordon | kfox1111: that type of detail would be useful in the backlog spec | 16:27 |
| kfox1111 | k. | 16:28 |
| *** tellesnobrega has quit IRC | 16:28 | |
| flashgordon | when reading the spec for the backlog it should make it very clear to the user that yes nova should solve this | 16:28 |
| flashgordon | and doing it outside of nova doesn't make sense etc. | 16:28 |
| kfox1111 | ok. | 16:29 |
| superdan | are we talking about the locking instances by external services? | 16:29 |
| flashgordon | superdan: no, should we? | 16:30 |
| snikitin | figleaf: Sorry, but where you found information which says that deleting a non-existent tag should return a 204 in https://review.openstack.org/#/c/179386/7/guidelines/http.rst? | 16:30 |
| flashgordon | superdan: https://review.openstack.org/186617 | 16:30 |
| *** unicell has quit IRC | 16:30 | |
| flashgordon | kfox1111: FWIW I feel your pain, the way openstack teams are set up it is very hard to do things that touch several projects | 16:30 |
| figleaf | snikitin: https://review.openstack.org/#/c/179386/7/guidelines/http.rst,cmu L40-45 | 16:30 |
| superdan | flashgordon: ah, hadn't seen this one | 16:30 |
| kfox1111 | yeah. it really sucks. :/ | 16:30 |
| superdan | flashgordon: on the face of it it seems like a terrible idea | 16:31 |
| * superdan ducks | 16:31 | |
| kfox1111 | especially when your a project that depends on all the other projects like the ap-catalog. :/ | 16:31 |
| kfox1111 | btw: found this http://docs.aws.amazon.com/IAM/latest/UserGuide/roles-usingrole-ec2instance.html | 16:31 |
| flashgordon | kfox1111: well there are plenty of other fish to fry in app-catalog land | 16:31 |
| kfox1111 | I'll add it to the spec as well. | 16:31 |
| kfox1111 | flashgordon: yeah, but my interest is in creating and providing cloud scale app templates for others to use. | 16:31 |
| kfox1111 | if we can't get any nontrivial coud apps contributed, why bother? | 16:32 |
| flashgordon | kfox1111: so when superdan skims the top of the spec, there should be enough there for him to change his mind about it | 16:32 |
| flashgordon | kfox1111: because we can get nontrivial cloud apps working -- maybe not as secure as we all would like etc. | 16:32 |
| kfox1111 | sure. I'll reword it to explain the set of problems openstack has now, and how this single solution addresses them. | 16:32 |
| flashgordon | plus not all clouds would have barbican | 16:33 |
| *** VW_ has joined #openstack-nova | 16:33 | |
| kfox1111 | yeah. I was hoping for just nova + keystone, but keystone shot that down. :/ | 16:33 |
| flashgordon | kfox1111: does the app catalog have versioning yet? a review system? a better name that doesn't use the word catalog? | 16:33 |
| bauwser | kfox1111: I saw that spec since last week, but can I just ask you a question ? | 16:33 |
| kfox1111 | but +barbican is still b etter then +barbican + some other microservice. | 16:33 |
| *** rfolco has joined #openstack-nova | 16:33 | |
| kfox1111 | flashgordon: Thats all being worked on. but it also looks like glance may be providing alot of that, so its a lot of wait and see... | 16:34 |
| kfox1111 | bauwser: Sure. | 16:34 |
| bauwser | kfox1111: why should Nova be responsible for that, and not a CMS like Puppet/Chef/Ansible/whatever ? | 16:34 |
| bauwser | kfox1111: I mean, fetching creds and so on. | 16:34 |
| flashgordon | kfox1111: the review system? | 16:34 |
| kfox1111 | the secret chicken and the egg probem is not solved with any of the config management systems. :/ | 16:35 |
| * flashgordon wants secret chicken for lunch | 16:35 | |
| kfox1111 | try and get heat autoscaling working with any of them. it cant be done today securely. :/ | 16:35 |
| *** tellesnobrega has joined #openstack-nova | 16:35 | |
| kfox1111 | you end up having to put a chef admin key in userdata. :/ | 16:35 |
| kfox1111 | for example. | 16:35 |
| bauwser | kfox1111: sure, but that can be done on a safe way by granting only a limited set of access | 16:36 |
| bauwser | kfox1111: I mean, Amazon has some specific things for that | 16:36 |
| kfox1111 | cant limit an admin cert currently. | 16:36 |
| kfox1111 | amazon has a lot of featues, no one else has. like IAM. | 16:36 |
| bauwser | kfox1111: and none of them are EC2 for sure, but rather something like in Keystone | 16:36 |
| bauwser | kfox1111: that, IAM | 16:36 |
| bauwser | kfox1111: so, I totally understand passing user creds in userdata is evil | 16:37 |
| snikitin | leakypipes, figleaf: Jay, could you help us? Is this patch https://review.openstack.org/#/c/180094/5/guidelines/http.rst means that if we trying to delete non-existend resource we should get 2xx status? not 404? | 16:37 |
| *** jcoufal has quit IRC | 16:37 | |
| *** gabriel-bezerra has joined #openstack-nova | 16:38 | |
| kfox1111 | have a look at the aws url I just posted. its quite interesting. | 16:38 |
| kfox1111 | its a more advanced version of the nova instance user spec. | 16:38 |
| leakypipes | snikitin: 404. | 16:38 |
| *** VW_ has quit IRC | 16:38 | |
| bauwser | kfox1111: for the records, I played with that | 16:38 |
| kfox1111 | it does role assignment too. :) | 16:38 |
| leakypipes | snikitin: hold up, sorry... | 16:38 |
| kfox1111 | bauwser: cool. I've never used it. what did you think of it? | 16:39 |
| bauwser | kfox1111: heard about Keystone V3 trusts ? | 16:39 |
| *** tonytan4ever has quit IRC | 16:39 | |
| kfox1111 | bauwser: yeah. know a lot about them. | 16:40 |
| *** ctrath1 has quit IRC | 16:40 | |
| leakypipes | sdague: for snikitin's question above... | 16:40 |
| bauwser | kfox1111: okay, my thought is that I can see a lot of things being done by Keystone or Barbican or Heat | 16:40 |
| figleaf | leakypipes: snikitin: deleting a non-existent tag leaves the state the same as if was there. No-ops should be 204. | 16:40 |
| bauwser | kfox1111: what I can't really see is the added value of Nova in that | 16:40 |
| leakypipes | sdague: doing a DELETE /instance/{id}/tags/{tag} for a non-existing resource should result in a 204 or a 404? | 16:40 |
| *** ctrath has joined #openstack-nova | 16:41 | |
| kfox1111 | it plays a part in the "how do you get a secret to the vm so that it can use that secret to get other secrets" | 16:41 |
| leakypipes | figleaf: but the resource does not exist, which is not the same as attempting to operate on an existing resource to put it in a state that it already is in. | 16:41 |
| kfox1111 | nova has an authenticated channel between nova -> the vm. | 16:41 |
| kfox1111 | either via the metadata server, or by config drive. | 16:41 |
| superdan | NO | 16:41 |
| superdan | it does not | 16:41 |
| superdan | that's crazypands | 16:41 |
| superdan | and also crazypants | 16:42 |
| leakypipes | superdan: ? | 16:42 |
| bauwser | kfox1111: nope, it doesn't | 16:42 |
| kfox1111 | I know some people don't trust the metadat server authentication, and config drive is for them. | 16:42 |
| leakypipes | superdan: who are you referring to? | 16:42 |
| superdan | kfox1111: the path from novaclient to nova-api to nova-compute is really far from trustworthy | 16:42 |
| superdan | leakypipes: kfox1111 sorry | 16:42 |
| leakypipes | ah, k | 16:42 |
| figleaf | leakypipes: you're not asking if it exists; you're stating you don't want it to exist | 16:42 |
| *** tjones1 has quit IRC | 16:42 | |
| kfox1111 | superdan, that I agree with. | 16:42 |
| *** exploreshaifali has joined #openstack-nova | 16:43 | |
| kfox1111 | I wasn't talking about the novaclient to metadata side. | 16:43 |
| superdan | kfox1111: and "authenticated" is completely false | 16:43 |
| *** romainh has left #openstack-nova | 16:43 | |
| kfox1111 | just the vm -> metadata server or vm -> config drive. | 16:43 |
| leakypipes | figleaf: heh, not really :) the /instances/.../tags/... URI *is* the resource you wish to operate on. And 404 Not Found should be returned if that resource does not exist. | 16:43 |
| bauwser | kfox1111: I don't see the benefit of Nova's configdrives and metadatas | 16:43 |
| superdan | kfox1111: it's really not | 16:43 |
| bauwser | kfox1111: like SpamapS said, all of that can be done programatically | 16:44 |
| kfox1111 | ? | 16:44 |
| kfox1111 | how is it not authenticated? | 16:44 |
| kfox1111 | can one vm get another's metadata? | 16:44 |
| superdan | kfox1111: anything that goes over RPC is really not trustworthy | 16:44 |
| flashgordon | kfox1111: if we are talking about making things more secure you need to explicitly state your assumptions in what you trust and don't trust | 16:44 |
| flashgordon | superdan: yup (it goes plaintext over the network with no auth) | 16:45 |
| superdan | kfox1111: if one hypervisor has been compromised (a la venom) then it's completely over | 16:45 |
| kfox1111 | sure. So I believe the openstack deployer desides whether to trust config drive or the metadata server for initial deployment. | 16:45 |
| superdan | flashgordon: yeah, plaintext, no auth, etc | 16:45 |
| figleaf | leakypipes: ok, I'll drop the objection. That's not the way I read the api-wg spec, though. | 16:45 |
| kfox1111 | superdan: thats true anyway. no matter what security solution you come up with. | 16:45 |
| markus_z | mdbooth: are you around? | 16:45 |
| leakypipes | snikitin: don't hold up the spec for this, thouhgh... just say something like "DELETE blah will return the HTTP result code that the API working group determines is appropriate. | 16:45 |
| leakypipes | figleaf: it's a weird one, I know... and likely should have clarity. | 16:46 |
| kfox1111 | I don't think that can be solved at all. If you gain physical access to a vm, all of its secrets are compromized. | 16:46 |
| superdan | kfox1111: it's true today, it doesn't have to be true going forward, but my point is that trusting any part of the nova infrastructure for this secret dissemination is really a faulty view | 16:46 |
| openstackgerrit | Joe Gordon proposed openstack/nova: Replace ascii art architecture diagram with svg image https://review.openstack.org/194840 | 16:46 |
| superdan | kfox1111: of course we can solve that | 16:46 |
| *** achanda has joined #openstack-nova | 16:47 | |
| *** tjones1 has joined #openstack-nova | 16:47 | |
| flashgordon | sudo rm -rf / would do it i think | 16:47 |
| kfox1111 | do you have a suggestion? I'm happy to solve it if its solvable. :/ | 16:47 |
| superdan | kfox1111: right now we have no auth between services over RPC, so we can't tell if a compromised compute node is replying to us as if it was conductor | 16:47 |
| kfox1111 | so make every compute host require a private key, and register them with the nova controllers, | 16:48 |
| superdan | kfox1111: if we issue certs, manage revocation, and don't accept messages from compute nodes claiming to be conductors without proper signatures... | 16:48 |
| *** arnaud____ has quit IRC | 16:48 | |
| kfox1111 | and then authenticating metadata requests with that? | 16:48 |
| kfox1111 | ok. so, thats a metadata sever/client implementation detail, and is erelevent to the instance user spec? | 16:48 |
| superdan | if you're running metadata server on the host, then your metadata server is getting the data over RPC anyway, so you'd be much better | 16:48 |
| superdan | kfox1111: well, my point is, I think that adding something like this as a feature of nova before we have even an inkling of infrastructure with which to support it actually being secure is kinda silly | 16:49 |
| *** danpb has quit IRC | 16:49 | |
| superdan | especially when you have much better ways to do this sort of thing, since you already have the ability to communicate authed/encrypted with the guest itself | 16:49 |
| kfox1111 | today its being relied apon for that security. people put crap in userdata that doesnt' belong there., | 16:50 |
| snikitin | leakypipes: Okay, thank you | 16:50 |
| kfox1111 | Lets make that better, and in a way that futher security like what your talking about can be added without breaking things. | 16:50 |
| *** snikitin is now known as snikitin_ | 16:50 | |
| superdan | FWIW, I was mostly jumping on your assertion that these paths are in any way trustworthy | 16:50 |
| superdan | I'm not really sure why this instance user thing belongs in nova even if the bits for doing so are trustworthy | 16:51 |
| kfox1111 | they are aready being relied on as being trustworthy. How trustworthy is open for debate. :/ | 16:51 |
| superdan | so here's the difference | 16:51 |
| superdan | if I give you a cart and you put all your money in the cart and walk down a busy road, | 16:51 |
| *** VW_ has joined #openstack-nova | 16:51 | |
| superdan | you're doing a stupid thing | 16:51 |
| superdan | if I give you an armored vehicle with which to transport your money, and I cheaped out and made it out of impressive-looking tin foil | 16:52 |
| *** shihanzhang has quit IRC | 16:52 | |
| superdan | then I did a stupid thing by making you think it was secure | 16:52 |
| *** dboik_ has joined #openstack-nova | 16:52 | |
| *** shihanzhang has joined #openstack-nova | 16:52 | |
| kfox1111 | there's an api vs a implementation thing. If we can provide an api that works today, that might not be the most secure implementation, but the implementation can be grown to be more secure in the future without breaking the api, | 16:52 |
| kfox1111 | averyone benifits. | 16:52 |
| *** SourabhP has quit IRC | 16:53 | |
| * bauwser google translates | 16:53 | |
| *** dimsum__ has joined #openstack-nova | 16:53 | |
| superdan | kfox1111: well, it's a cart before the horse thing | 16:53 |
| kfox1111 | so, in your analogy, the armored car is still better because it gets some extra seciruty, and then they can, when the can affort to, beaf up the walls. | 16:53 |
| johnthetubaguy | kfox1111: I think we would feel better if we new what the end looks like, at least roughly | 16:53 |
| superdan | kfox1111: it's not because you worry less about the armored vehicle | 16:53 |
| kfox1111 | api's are hard to get adopted. it takes time. | 16:53 |
| *** shihanzhang has quit IRC | 16:53 | |
| kfox1111 | so getting that piece in progress is important. | 16:53 |
| kfox1111 | what backs it is an implementation detail, and if the api's right, can be easily swithced without the user caring. | 16:54 |
| superdan | why does this need to be in nova at all though? | 16:54 |
| kfox1111 | johnthetubaguy: ok. | 16:54 |
| *** kiwik-chenrui has joined #openstack-nova | 16:54 | |
| flashgordon | kfox1111: so this goes back to you wanting this for the app catalog right? realistically best case you won't see adoption of this for a year or two | 16:54 |
| kfox1111 | because nova has in theory, the rains of the vm it manages. by being close to it, it has the most chance of having a trusted channel between them. | 16:54 |
| flashgordon | kfox1111: so you need a way to move forward with whatever your trying to do without this | 16:55 |
| *** shihanzhang has joined #openstack-nova | 16:55 | |
| kfox1111 | flashgordon: And it will be +6 months if we wait another release cycle. | 16:55 |
| *** kiwik has joined #openstack-nova | 16:55 | |
| superdan | kfox1111: but if the giant holes in the armor aren't there, then that statement isn't right either | 16:55 |
| *** yassine_ has quit IRC | 16:55 | |
| *** dboik has quit IRC | 16:55 | |
| *** unicell has joined #openstack-nova | 16:55 | |
| kfox1111 | all I want is for users to start understanding they need to drive instaead of walk. what they drive can be beafed up at a later point. | 16:55 |
| kfox1111 | learning to drive is important, and that takes time. | 16:56 |
| flashgordon | kfox1111: that doesn't change the problem I am referring to | 16:56 |
| superdan | I don't think that has anything to do with where it lives | 16:56 |
| kfox1111 | we provide the vehical they drive. we provide better ones in M vs L. | 16:56 |
| kfox1111 | and moving forward. what they need to learn and do doesn't have to change. | 16:57 |
| *** sharkJing has quit IRC | 16:57 | |
| *** RuiChen has quit IRC | 16:57 | |
| kfox1111 | johnthetubaguy: I'll rewrite the spec to try and address the end game better. | 16:57 |
| *** sharkJing has joined #openstack-nova | 16:58 | |
| *** browne has quit IRC | 16:59 | |
| kfox1111 | really really though, its applications running in vm's need a way to talk to other openstack services. they need keystone tokens to do that. I can't simplify it further then that. there's tons of ways of solving that. as the revisions show. | 16:59 |
| *** kiwik-chenrui has quit IRC | 16:59 | |
| johnthetubaguy | kfox1111: so the context here is that Nova is too big, so we are very worried about expanding the scope, to the bar to prove something should be in Nova is crazy high right now, I am sorry that is not helping you right now, but long term, I think thats good for the whole OpenStack ecosystem | 16:59 |
| johnthetubaguy | kfox1111: I agree with your problem, I am not yet sure what Nova needs to do, if anything, if we get it right | 16:59 |
| *** markus_z has quit IRC | 16:59 | |
| *** thorst1 has joined #openstack-nova | 16:59 | |
| superdan | sure, I don't doubt the problem | 16:59 |
| bauwser | kfox1111: reading the convo with SpamapS, can't really see how you explain why it's not needed to be in Novz | 16:59 |
| kfox1111 | I understand, but allowing continued bad security (secrets in userdata) doesn't help though. | 17:00 |
| bauwser | kfox1111: I'm quite on board with him, you only need a one-time key with a particular ACL | 17:00 |
| *** xyang2 has joined #openstack-nova | 17:00 | |
| kfox1111 | bauwser: superdan, flashgordon: if you believe a microservice can securely be made to work with the workflow the barbican/keystone folks agreed to already, | 17:01 |
| bauwser | kfox1111: if the one-time key is compromised, then give another key to your instance | 17:01 |
| *** ijuwang has quit IRC | 17:01 | |
| kfox1111 | please write out a workflow showing how that can work. | 17:01 |
| kfox1111 | I just don't see it. | 17:01 |
| superdan | lol | 17:01 |
| *** Longgeek has quit IRC | 17:02 | |
| *** mriedem_lunch is now known as mriedem | 17:02 | |
| kfox1111 | how do you manage those one-time keys though? | 17:02 |
| kfox1111 | I tried to get additional authentication mechanisms added to barbican. | 17:02 |
| kfox1111 | they pushed back and said authentication belongs to keystone. | 17:02 |
| kfox1111 | I really don't want to require nova + barbican + keystone + some other random service to be required to make this work. | 17:03 |
| *** tellesnobrega has quit IRC | 17:03 | |
| *** annashen has joined #openstack-nova | 17:03 | |
| kfox1111 | openstack already requires a rediculous number of services to be installed by operators. | 17:03 |
| kfox1111 | this is pushing even more work on them. | 17:03 |
| *** lucas__ has joined #openstack-nova | 17:03 | |
| bauwser | kfox1111: that's why projects have mission statements | 17:03 |
| *** gszasz has quit IRC | 17:04 | |
| superdan | kfox1111: you hit the nail on the head | 17:04 |
| bauwser | kfox1111: and from my perspective, managing auths even for the userspace is clearly not scoped by Nova | 17:04 |
| *** patrickeast has joined #openstack-nova | 17:04 | |
| superdan | it's actually squarely outside the scope | 17:04 |
| superdan | given that we don't do anything past the VM barrier | 17:04 |
| kfox1111 | bauwser: I don't like this sort of reasoning much, but... | 17:05 |
| *** eharney has quit IRC | 17:05 | |
| kfox1111 | amazon did it in their nova equivalent. Why would they wast time/money on it if it wasnt the right place? | 17:05 |
| superdan | you're right, that's terrible reasoning :) | 17:05 |
| bauwser | kfox1111: Amazon can do what they want eh :) | 17:05 |
| *** otter768 has joined #openstack-nova | 17:05 | |
| *** zsais has joined #openstack-nova | 17:06 | |
| kfox1111 | openstack has gotten where it has by carefully designing cloud architecture to scale. we've benifited by coping a lot of it. its not a good reason in and of it self, but does have some merit. | 17:06 |
| kfox1111 | amazon I ment. | 17:06 |
| *** zsais has quit IRC | 17:06 | |
| bauwser | okay, that's not a reason for not having our own vision :) | 17:07 |
| kfox1111 | Agreed. but I don't think we do. :/ | 17:07 |
| superdan | amazon also doesn't suffer from being open and doesn't have to worry about review bandwidth and such | 17:07 |
| *** achanda has quit IRC | 17:07 | |
| *** achanda has joined #openstack-nova | 17:07 | |
| bauwser | anyway, I need to run off, I can drop some comments in your spec, but it will basically be restating what Clint said | 17:07 |
| kfox1111 | Yeah. and as such, they have one benifit. they can get solutions in that work without politics. :/ | 17:07 |
| johnthetubaguy | superdan: I was going to say also doesn't support lots of hypervisors, volume drivers, network drivers, etc, etc | 17:08 |
| *** yamahata has joined #openstack-nova | 17:08 | |
| superdan | johnthetubaguy: yep | 17:08 |
| superdan | and of course, | 17:08 |
| johnthetubaguy | kfox1111: so the project split this is almost totally a people issue at this point, agreed | 17:08 |
| superdan | just because it's in the unified EC2 API doesn't mean anything about where it lives in the backend | 17:08 |
| kfox1111 | johnthetubaguy: they may under the hood. we just don't know. ;) | 17:08 |
| *** ajayaa has quit IRC | 17:08 | |
| johnthetubaguy | kfox1111: it may also be in separate git repos too, we just don't know | 17:08 |
| kfox1111 | I would be somewhat surprised if all their storage was backed by the same vendor and plugin. :) | 17:09 |
| kfox1111 | yeah. | 17:09 |
| superdan | we're pretty far off in the weeds now, right? :) | 17:09 |
| johnthetubaguy | so lets get back to makeing progress.. | 17:09 |
| kfox1111 | hehe. yeah. | 17:09 |
| kfox1111 | +1 :) | 17:09 |
| johnthetubaguy | we need a good problem definition, I think we have agreement on the problem | 17:09 |
| johnthetubaguy | we don't have agreement of it being in nova's scope, because that depends on the implementation | 17:10 |
| kfox1111 | true. | 17:10 |
| johnthetubaguy | now, that means we need to see the end game | 17:10 |
| *** otter768 has quit IRC | 17:10 | |
| johnthetubaguy | as I see it, there needs a VM <-> barbican trust relationship right? | 17:10 |
| johnthetubaguy | assuming barbican has all the keys | 17:11 |
| kfox1111 | the ultimate endgame is vm's need a keystone account of some sort, and a way to get a keystone token bound to that account. | 17:11 |
| *** achanda has quit IRC | 17:11 | |
| kfox1111 | all the rest are implemenatation details. Keystone would like barbican to be a CA and allocate user certs. | 17:11 |
| johnthetubaguy | kfox1111: OK, so the instance has access to some keystone account then | 17:11 |
| kfox1111 | then the user cert be used to "login" to keystone by the vm. | 17:12 |
| johnthetubaguy | so lets talk about what happens in the VM first | 17:12 |
| johnthetubaguy | ignoring the fact Nova has anything to do with it | 17:12 |
| kfox1111 | since the barbican CA needs to be registerd with keystone, there should only be one CA ideally. | 17:12 |
| *** dboik_ has quit IRC | 17:12 | |
| kfox1111 | k. | 17:12 |
| johnthetubaguy | lets back right up | 17:13 |
| johnthetubaguy | what does the VM need to get the trust relationship? | 17:13 |
| johnthetubaguy | if it were done manually | 17:13 |
| kfox1111 | the vm needs a certificate. | 17:13 |
| johnthetubaguy | needs some secrete thats registered with keystone I guess? | 17:13 |
| kfox1111 | yes. | 17:13 |
| kfox1111 | well, sort of. | 17:13 |
| johnthetubaguy | right, so the user could manually create that via an API, and figure that inside their VM already? | 17:14 |
| kfox1111 | it needs a certificate signed by a ca that is registered with keystone. | 17:14 |
| *** armax has quit IRC | 17:14 | |
| kfox1111 | correct. | 17:14 |
| johnthetubaguy | kfox1111: do we have all the APIs to do that today? | 17:14 |
| kfox1111 | barbican yes. keystone, yes and no. Yes they have a way, their prefered way is landing in liberty. | 17:15 |
| johnthetubaguy | OK, so thats a problem, but anyways, lets leave that on a back burner for now | 17:15 |
| kfox1111 | which is the CA trusted domain stuff. | 17:15 |
| kfox1111 | k. | 17:15 |
| morganfainberg | johnthetubaguy: it is a liberty target for the only really viable implemntation in keystone | 17:15 |
| morganfainberg | johnthetubaguy: short of creating an account per VM (please don't do this) | 17:16 |
| kfox1111 | (they are willing to tweak the API specifically to solve this problem if we need to) | 17:16 |
| *** ameade is now known as ameade_vacation | 17:16 | |
| kfox1111 | yeah. | 17:16 |
| johnthetubaguy | Now how does the user get the certificate inside the VM? | 17:16 |
| kfox1111 | metadata/configdrive. | 17:16 |
| johnthetubaguy | morganfainberg: OK, I should probably get more context on that API | 17:16 |
| johnthetubaguy | kfox1111: I am saying we are ignoreing Nova exists | 17:16 |
| johnthetubaguy | what is the data that VM needs, some keystone credentials? | 17:17 |
| kfox1111 | how else does a vm that's being provisioned get data? | 17:17 |
| johnthetubaguy | kfox1111: the user puts it there manually | 17:17 |
| kfox1111 | the cert is the keystone credential. | 17:17 |
| kfox1111 | ssh? ok... | 17:17 |
| morganfainberg | johnthetubaguy: sure. let me know when - but it's really just handling an x509 client cert in the same way as a federated user is handled | 17:17 |
| johnthetubaguy | OK, for the same user that created the VM? | 17:17 |
| kfox1111 | no. a different user. | 17:17 |
| kfox1111 | the user doesn't want to give all his/her power to the vm. | 17:18 |
| *** isd has joined #openstack-nova | 17:18 | |
| *** VW_ has quit IRC | 17:18 | |
| johnthetubaguy | kfox1111: OK, so the user has some sub user with less roles, and it wants to give that access to the specific VM | 17:18 |
| kfox1111 | so, that can be acomplished with trusts, | 17:18 |
| *** VW_ has joined #openstack-nova | 17:18 | |
| kfox1111 | but in some cases, even less. | 17:18 |
| kfox1111 | so say the vm has its own user with no access at all. | 17:19 |
| kfox1111 | the user can associate a single secret with that user with an acl. | 17:19 |
| johnthetubaguy | right, thats fine | 17:19 |
| kfox1111 | k. | 17:19 |
| johnthetubaguy | so how do you get the secret in the VM, manually I mean | 17:19 |
| *** rotbeard has quit IRC | 17:20 | |
| kfox1111 | assuming you have the cert already, | 17:20 |
| johnthetubaguy | you should ssh it in there, I suppose, or some trusted channel | 17:20 |
| kfox1111 | you authenticate with keystone using the cert. | 17:20 |
| *** nic has joined #openstack-nova | 17:20 | |
| kfox1111 | you get back a keystone token. | 17:20 |
| johnthetubaguy | kfox1111: I am saying how do you get the cert there, or generate the cert there | 17:20 |
| *** eharney has joined #openstack-nova | 17:21 | |
| kfox1111 | you then present the keystone token to barbican with a request for the secret, | 17:21 |
| kfox1111 | then get it. | 17:21 |
| kfox1111 | ah. ok. | 17:21 |
| johnthetubaguy | OK, so I am being dumb or missing something, but you lost me there | 17:21 |
| kfox1111 | so I only know 3 ways currently. metadata server,config drive or ssh. | 17:21 |
| johnthetubaguy | OK, right | 17:21 |
| kfox1111 | the first two work with heat right now, the last doesn't. | 17:22 |
| johnthetubaguy | now ssh is great, the others are really bad, as the thing hangs around for ages and lives in a database, etc | 17:22 |
| kfox1111 | ssh has other problems. | 17:22 |
| johnthetubaguy | so this is were bauwser I think was trying to suggest some ideas | 17:22 |
| johnthetubaguy | kfox1111: agreed | 17:22 |
| johnthetubaguy | at least I think he was | 17:22 |
| kfox1111 | controller -> vm via ssh has real problems with SDN. | 17:22 |
| johnthetubaguy | kfox1111: agreed | 17:22 |
| kfox1111 | you have to use floating ip's or tunnel. :/ | 17:22 |
| *** VW_ has quit IRC | 17:23 | |
| kfox1111 | so that tends to leave metadata/configdrive. | 17:23 |
| johnthetubaguy | so what if we had some one-time token that allowed you to register a cert for a given uuid? | 17:23 |
| johnthetubaguy | and each vm has its own cert | 17:23 |
| superdan | that's the suggestion in the review comments | 17:23 |
| superdan | a one-time url interface in barbican makes a lot of sense to me | 17:24 |
| *** xgsa has quit IRC | 17:24 | |
| superdan | put that into the metadata, the instance pulls it | 17:24 |
| superdan | url expires in five minutes or once pulled | 17:24 |
| johnthetubaguy | kfox1111: so for context, passing the onetime thing seems OK for metadata service | 17:24 |
| kfox1111 | ok... lets explore that for a bit. | 17:24 |
| johnthetubaguy | superdan: right, maybe both | 17:24 |
| johnthetubaguy | kfox1111: yep | 17:24 |
| kfox1111 | the cert must be signed by the barbican CA in order to be trusted by keystone. | 17:25 |
| johnthetubaguy | kfox1111: unless we tell it do allow self-signed right? | 17:25 |
| superdan | johnthetubaguy: I meant both yeah | 17:25 |
| *** eharney has quit IRC | 17:25 | |
| *** EmilienM is now known as EmilienM|brb | 17:25 | |
| johnthetubaguy | superdan: cools | 17:25 |
| kfox1111 | johnthetubaguy: then every instance's local CA cert has to be added to keystone. :/ | 17:25 |
| johnthetubaguy | kfox1111: so the other thing, is the one time thing generates a cert and gives it to the VM? | 17:25 |
| kfox1111 | it could potentally be made to do that. but, how do you restrict too many certs from being allowcated by a user? | 17:26 |
| johnthetubaguy | kfox1111: although I think I just came up with what morganfainberg told me not to, which sucks | 17:26 |
| kfox1111 | they are intended for instances, not for any other useage. | 17:26 |
| *** eharney has joined #openstack-nova | 17:26 | |
| kfox1111 | johnthetubaguy: yeah. been down that path. :/ | 17:27 |
| johnthetubaguy | kfox1111: well you would have a quota, I suspect | 17:27 |
| morganfainberg | johnthetubaguy: you can't create users for each vm - the reason is that identity management (users/groups) will never be designated nor required for an openstack cloud | 17:27 |
| kfox1111 | johnthetubaguy: why would it be different then nova's vm quota? | 17:27 |
| morganfainberg | johnthetubaguy: so this becomes non-interoperable from the get-go (and not useful in almost any production environment, because creating users is scary levels of trust) | 17:27 |
| johnthetubaguy | kfox1111: its a separate API service, I mean a quota of generated certs on the keystone side | 17:27 |
| *** ajayaa has joined #openstack-nova | 17:28 | |
| *** sudipto has quit IRC | 17:28 | |
| kfox1111 | johnthetubaguy: or worse, with a quota, if a user has a quota of 2 vm's, and only uses an instance user on one, then he's free to use that url for a non instance. :/ | 17:28 |
| morganfainberg | just as a quick background of why you shouldn't make a user per vm like that | 17:28 |
| johnthetubaguy | morganfainberg: but can't all those certs map back to a single user, its just like lots of federation? | 17:28 |
| *** xyang2 has quit IRC | 17:28 | |
| morganfainberg | johnthetubaguy: it could - but i was advocting each cert ends up being an ephemeral user | 17:28 |
| johnthetubaguy | so federation doesn't normally do that I guess, but I just threw that in there | 17:28 |
| kfox1111 | johnthetubaguy: the users have to be seperate though, since seperate vm's get seperate permissions to things. | 17:28 |
| morganfainberg | johnthetubaguy: so you don't delegate full trust to every cert, only the ones you want | 17:29 |
| johnthetubaguy | morganfainberg: oh, user dies with the cert you mean, thats quite nice | 17:29 |
| morganfainberg | johnthetubaguy: yep | 17:29 |
| kfox1111 | you dont' want vm A to download secrets in barbican associated with vm B. | 17:29 |
| kfox1111 | yeah. exactly. :) | 17:29 |
| morganfainberg | johnthetubaguy: federated users don't hold information in keysotne (well they hold a minimal amount, but that is separate) | 17:29 |
| johnthetubaguy | morganfainberg: could the user die with the CA cert, oh dear, that got nasty quickly | 17:29 |
| kfox1111 | so the cert is quite literally the instance user in keystone. | 17:29 |
| morganfainberg | johnthetubaguy: so basically the cert (part of the DN, and the domain they map into) is the user id. | 17:30 |
| *** ijw has joined #openstack-nova | 17:30 | |
| *** tdurakov has quit IRC | 17:30 | |
| kfox1111 | and the user id I think in this case is the nova instance id. | 17:30 |
| openstackgerrit | Claudiu Belu proposed openstack/nova-specs: Adds Hyper-V vNUMA enable spec https://review.openstack.org/196162 | 17:30 |
| morganfainberg | kfox1111: not quite | 17:30 |
| morganfainberg | kfox1111: doesn't work like that | 17:30 |
| *** annegentle has joined #openstack-nova | 17:31 | |
| morganfainberg | kfox1111: it'll be a sha256 of domain_id, and part of the dn | 17:31 |
| * morganfainberg just realized there is a hole in this plan | 17:31 | |
| morganfainberg | crap | 17:31 |
| morganfainberg | you can't grant a trust or an assignment to an ephemeral user atm | 17:31 |
| kfox1111 | morganfainberg: no, the domain's dn is the nova instance id, the keystone user id is the sha256. | 17:31 |
| * flashgordon wants Beer as a Service | 17:31 | |
| morganfainberg | kfox1111: this is going to require a bunch more work on the keystone side i think. | 17:31 |
| morganfainberg | kfox1111: than i initially thought | 17:32 |
| kfox1111 | morganfainberg: I asked about that earlier. crap. | 17:32 |
| johnthetubaguy | flashgordon: I am fairly sure you provided that for me at one point at the summit | 17:32 |
| kfox1111 | morganfainberg: this is wahy I wanted an api that handed out keystone tokens. | 17:32 |
| kfox1111 | because then the authentication to keystone is hidden from the vm. you can start with acutal keystone users, and switch to certs or something under the hood and the api doesn't have to change. | 17:33 |
| flashgordon | johnthetubaguy: true | 17:33 |
| johnthetubaguy | morganfainberg: so I have a feeling we should thing through the per instance user, I mean the API load seems a bit high, but I like the limited scope there | 17:33 |
| morganfainberg | kfox1111: this is going ot be a significant amount of work in either case. | 17:33 |
| *** nikhil_k is now known as nikhil_k-eto | 17:33 | |
| flashgordon | johnthetubaguy: but there as no multitenancy REST API | 17:33 |
| morganfainberg | johnthetubaguy: well i'll say that most environments a user per vm wont ever fly | 17:33 |
| *** nikhil_k-eto is now known as nikhil_k-away | 17:33 | |
| johnthetubaguy | flashgordon: talking of which, its 6.30pm on a friday, and I have two birthday presents still left over in my fridge | 17:33 |
| morganfainberg | johnthetubaguy: even minimally. | 17:34 |
| johnthetubaguy | morganfainberg: because of the audit of so many users, or something like that? | 17:34 |
| morganfainberg | johnthetubaguy: because we don't guarnatee you can create users in keystone | 17:34 |
| kfox1111 | morganfainberg: its less then that though. its a user per vm that nees instance users. which is a subset of "all". | 17:34 |
| morganfainberg | many deployments are moving away from SQL backed | 17:34 |
| morganfainberg | so user data is "read-only" | 17:34 |
| kfox1111 | morganfainberg: heat relies on being able to create users. | 17:34 |
| kfox1111 | it has its own domain. | 17:34 |
| flashgordon | johnthetubaguy: you should go drink them | 17:34 |
| morganfainberg | johnthetubaguy: and we're working to remove/kill the SQL backend | 17:35 |
| morganfainberg | it's not something we want to do | 17:35 |
| kfox1111 | we do it all the time. read only ldap domain for users, sql domain for heat. | 17:35 |
| johnthetubaguy | morganfainberg: ah, so that totally makes sense now, got you | 17:35 |
| kfox1111 | ah. | 17:35 |
| openstackgerrit | Jesse J. Cook proposed openstack/nova: WIP: cells: weighted cell list randomization https://review.openstack.org/195757 | 17:35 |
| morganfainberg | the sql backend is a very very very bad implementation | 17:35 |
| morganfainberg | and lots of identity providers do it better | 17:35 |
| morganfainberg | so we don't want to try and keep up | 17:35 |
| kfox1111 | johnthetubaguy: morgan and I had the conversation earlier. they want to get out of the identity provider game. | 17:35 |
| johnthetubaguy | morganfainberg: totally with you there, thats a good call | 17:36 |
| kfox1111 | so instances need their own identity provider. | 17:36 |
| morganfainberg | the x509 cert thing solves the immidiate issue for service users | 17:36 |
| morganfainberg | because they become ephemeral | 17:36 |
| johnthetubaguy | morganfainberg: its like Nova building a hypervisor, because that would be fun | 17:36 |
| morganfainberg | johnthetubaguy: yep | 17:36 |
| kfox1111 | morganfainberg: I'm ok with trusts not working out of the gate. | 17:36 |
| *** annegentle has quit IRC | 17:36 | |
| morganfainberg | kfox1111: the issue is out of the gate nothing but barbican/swift could work | 17:37 |
| kfox1111 | for my 2 use cases, I only need acl support in barbican/zaqar with no trusts. | 17:37 |
| *** dboik has joined #openstack-nova | 17:37 | |
| morganfainberg | because you'd get no roles | 17:37 |
| morganfainberg | so strict service ACLs on user | 17:37 |
| *** fawadk has joined #openstack-nova | 17:37 | |
| kfox1111 | yeah. but thats purely a keystone thing that can be solved over time. | 17:37 |
| johnthetubaguy | so multiple certs mapping to a single user, then, that sounds like what we want here | 17:37 |
| johnthetubaguy | although thats not possible right now, probably | 17:37 |
| morganfainberg | johnthetubaguy: that should be doable - but yes, not doable today | 17:37 |
| *** browne has joined #openstack-nova | 17:37 | |
| morganfainberg | johnthetubaguy: target is liberty-ish | 17:37 |
| kfox1111 | I think the x509 thing will work for now, and can be made more useful in the future. | 17:38 |
| johnthetubaguy | morganfainberg: thats cool | 17:38 |
| johnthetubaguy | kfox1111: so here is a possible plan.... | 17:38 |
| johnthetubaguy | for now heat could pass the nasty thing into the VM via user data, and have lots of warnings all over the place | 17:38 |
| johnthetubaguy | but we work out a plan for per user certs in the meantime | 17:38 |
| kfox1111 | there's a hole in th workflow still. lets address that first. | 17:39 |
| johnthetubaguy | my thinking is, heat generates a onetime (with expiry) URL to collect a cert thats tied to a specific user for some group of instances? | 17:39 |
| johnthetubaguy | kfox1111: OK, the whats the hole? | 17:40 |
| *** fawadkhaliq has quit IRC | 17:40 | |
| kfox1111 | the quota thing I'm still not sure about. | 17:40 |
| johnthetubaguy | kfox1111: quota for what? | 17:40 |
| kfox1111 | do we allow these instance users to be created outside of a vm workflow? | 17:40 |
| *** ijuwang has joined #openstack-nova | 17:40 | |
| kfox1111 | ie, non instance, instance users? | 17:40 |
| johnthetubaguy | now thats an interesting issue | 17:41 |
| kfox1111 | ie, quota for this other service = nova instance quota. | 17:41 |
| kfox1111 | user launches 2 vm of 3 quota. | 17:41 |
| johnthetubaguy | kfox1111: it should be totally separate quotas | 17:41 |
| kfox1111 | he still has 1 instance user quota free, he could create user himself and use it outside of the vm. | 17:41 |
| johnthetubaguy | kfox1111: at some point someone will want to give a VM two users | 17:41 |
| *** ndipanov has quit IRC | 17:42 | |
| kfox1111 | why? | 17:42 |
| johnthetubaguy | kfox1111: I can invent a use case, but I suspect we could find a good one | 17:42 |
| kfox1111 | keystone should let you scope the instance user to some subset of permissions to give to other services. | 17:42 |
| kfox1111 | maybe we can. just trying to think that one through. | 17:42 |
| johnthetubaguy | kfox1111: I have my web tier shares with mid tier, mid shares with DB, or something like that | 17:43 |
| kfox1111 | so one of the things I was hoping for was a way to keep instance users tied to instances. | 17:43 |
| johnthetubaguy | so the mid tier needs two users | 17:43 |
| kfox1111 | our site policy makes it hard to allow users to be able to create new users for themselves. :/ | 17:43 |
| johnthetubaguy | kfox1111: certs tied to instances, possibly, but only possibly | 17:44 |
| *** lucas-afk has quit IRC | 17:44 | |
| kfox1111 | yeah, but if the instance cert is ever given to a user, they can just launch a vm, and download the cert. | 17:44 |
| *** zsais has joined #openstack-nova | 17:44 | |
| *** ZZelle has joined #openstack-nova | 17:44 | |
| kfox1111 | with the origional proposal, the metadata server held the cert, and only gave the vm keystone tokens, which were time limited. | 17:45 |
| kfox1111 | so if the user downloaded a keystone token it wouldn't last forever. | 17:45 |
| johnthetubaguy | kfox1111: we just said you can only ever download the cert once, I thought? | 17:45 |
| johnthetubaguy | its only stored by the person who downloaded it, I think | 17:45 |
| kfox1111 | yeah. its where it can be used tough. if you can download it to the vm, then ssh to the vm and download it to your workstation, then it can be used there too. | 17:45 |
| *** oro has joined #openstack-nova | 17:46 | |
| *** ctrath has quit IRC | 17:46 | |
| kfox1111 | by having the nova metadata server hold the cert, the vm can request a token from the metadata server, the metadata server can authenticate with the vm cert to keystone and return the token to the vm, | 17:46 |
| kfox1111 | and the vm never has direct access to any credentials. | 17:46 |
| kfox1111 | you could even switch out the cert for username/password or anything else in the nova and the vm wouldn't know...... | 17:47 |
| johnthetubaguy | kfox1111: but thats bad | 17:47 |
| *** zsais has quit IRC | 17:48 | |
| johnthetubaguy | kfox1111: security wise, we wanted it to have a single use (time bounded) url to get/generate that cert tied to the instance uuid and the user who created the onetime URL | 17:48 |
| kfox1111 | k. +1. | 17:49 |
| johnthetubaguy | kfox1111: so its almost 7pm now on a Friday evening, so I am going to sign off | 17:49 |
| johnthetubaguy | kfox1111: if you could add some of these ideas (even if only into the alternatives section) that would be awesome | 17:49 |
| kfox1111 | so how do we get a one time credential to the instance (userdata)? that allows the cert to be created? | 17:49 |
| johnthetubaguy | kfox1111: we totally still haven't answered what Nova needs to do | 17:49 |
| kfox1111 | "secret to the vm to get a secret? " :/ | 17:49 |
| johnthetubaguy | kfox1111: so my default answer is kinda rude, heat could do that | 17:49 |
| kfox1111 | s/nova/heat/ :/ | 17:50 |
| johnthetubaguy | its possible that a dumb idea, but we need to say way | 17:50 |
| kfox1111 | that whole, project passing the buck thing I was talking about. ;) | 17:50 |
| kfox1111 | I understand though. | 17:50 |
| johnthetubaguy | kfox1111: I am saying we have to actively decide why that doesn't work | 17:50 |
| kfox1111 | we can talk about it more monday. | 17:50 |
| kfox1111 | sure. | 17:51 |
| johnthetubaguy | kfox1111: heat could call keystone, using the users context, then inject the URL into the user data, and jobs done, with zero nova code | 17:51 |
| *** zsais has joined #openstack-nova | 17:51 | |
| kfox1111 | but the one time pw stuff needs to live somewhere. | 17:51 |
| kfox1111 | is nova or barbican or keystone the right placd? | 17:51 |
| *** Marga_ has quit IRC | 17:51 | |
| kfox1111 | each project is saying the other project. | 17:51 |
| kfox1111 | or worse, a new project. :/ | 17:52 |
| johnthetubaguy | right, they are both good answers | 17:52 |
| kfox1111 | too many projects. :/ | 17:52 |
| johnthetubaguy | swift already does it, FWIW | 17:52 |
| johnthetubaguy | although thats probably not a good answer | 17:52 |
| johnthetubaguy | ah, so it doesn't do the one time bit | 17:52 |
| kfox1111 | I was afraid osomeone might say that. | 17:52 |
| kfox1111 | I was thinking that. | 17:52 |
| kfox1111 | but then the solution is now nova + swift + barbican + keystone + heat. :/ | 17:52 |
| johnthetubaguy | kfox1111: so the produce management working group would be a great group to try and push this forward, in theory at least | 17:52 |
| *** agireud has quit IRC | 17:53 | |
| superdan | kfox1111: I think you're missing docker in that chain | 17:53 |
| * kfox1111 chuckles | 17:53 | |
| kfox1111 | exactly. | 17:53 |
| johnthetubaguy | kfox1111: I really only said that to be annoying :) | 17:53 |
| kfox1111 | and ironic. ;) | 17:53 |
| johnthetubaguy | kfox1111: the swift bit I mean | 17:53 |
| kfox1111 | johnthetubaguy: heh. but it is a possible solution. :/ | 17:53 |
| kfox1111 | ok. lets not go down that rabbit hole any further. :/ | 17:54 |
| johnthetubaguy | kfox1111: well only if you assume everyone deploys swift, which seems like a bad plan | 17:54 |
| kfox1111 | agreed. | 17:54 |
| kfox1111 | needing barbican seems a bit heavy too, but I udnerstand keystone not wanting to maintain a ca when barbican already soes. | 17:54 |
| kfox1111 | thanks for the great discussion. I really appreciate it. | 17:55 |
| johnthetubaguy | kfox1111: so the one time thing, it fells very like barbican, except for the fact it would be better if it was one time cert generation, so if the URL is not used you don't need to even generate the cert, which in my head means keystone, but I could be wrong | 17:55 |
| johnthetubaguy | ah, right | 17:55 |
| kfox1111 | even better might be the vm itself generates the private key, | 17:55 |
| kfox1111 | and a one time csr goes to the one time url for signing. | 17:56 |
| *** ijuwang has quit IRC | 17:56 | |
| johnthetubaguy | kfox1111: so I actually like that (I will ignore the whole key generation on VMs being bad due to lack of randomness stuff) | 17:56 |
| *** lucas-afk has joined #openstack-nova | 17:56 | |
| kfox1111 | but then how does the single use url verify that the nova instance id in the cert matches the vm that is asking for the signing? | 17:57 |
| dimsum__ | "produce management" :) | 17:57 |
| *** nic has quit IRC | 17:57 | |
| *** tjones2 has joined #openstack-nova | 17:57 | |
| kfox1111 | thats encoded in the one time url I guess? | 17:57 |
| johnthetubaguy | kfox1111: so steps forward, that alternatives section is a great way to make the chosen solution look better | 17:57 |
| *** zsais has left #openstack-nova | 17:57 | |
| johnthetubaguy | kfox1111: I am not sure it has to, yeah, its in the URL | 17:57 |
| kfox1111 | ... so a problem. | 17:58 |
| *** schwicke has quit IRC | 17:58 | |
| kfox1111 | the temp url needs to have the nova intance id encoded into the url. | 17:58 |
| johnthetubaguy | you don' t get the uuid till you create the instance | 17:58 |
| johnthetubaguy | yeah, so thats like a good thing to mention in the spec | 17:58 |
| kfox1111 | but you need to create the vm to get it. | 17:58 |
| *** mriedem has quit IRC | 17:59 | |
| kfox1111 | so you'd need to be able to create the vm paused without user data, then add userdata later. :/ | 17:59 |
| *** achanda has joined #openstack-nova | 17:59 | |
| *** tjones2 has quit IRC | 17:59 | |
| johnthetubaguy | kfox1111: fwiw, you could have a separate uuid, then put the mapping somewhere, but yeah... | 17:59 |
| johnthetubaguy | kfox1111: so happy spec updating, I really need to stop, its hit 7pm now | 17:59 |
| kfox1111 | johnthetubaguy: Thanks. I'll see if I can find more holes. I'm sure there are still a few. Have a good weekend. | 18:00 |
| * johnthetubaguy bravely runs away | 18:00 | |
| johnthetubaguy | kfox1111: good luck, its a fun problem | 18:00 |
| superdan | flashgordon: oomichi had a good point about why you leaving breaks all of nova: https://review.openstack.org/#/c/194774/1 | 18:00 |
| kfox1111 | like I said in the pec, I think the tempurl thing can be made to work, but is more complicated to actually implement :/ | 18:01 |
| *** mriedem has joined #openstack-nova | 18:01 | |
| *** Longgeek has joined #openstack-nova | 18:02 | |
| *** Marga_ has joined #openstack-nova | 18:02 | |
| kfox1111 | I'll dig a bit into aws's implemenation too. maybe they have some implementation details leak through that might be interesting to discuss. | 18:03 |
| *** EmilienM|brb is now known as EmilienM | 18:04 | |
| *** rushiagr is now known as rushiagr_away | 18:04 | |
| *** hogepodge has quit IRC | 18:06 | |
| *** Marga_ has quit IRC | 18:07 | |
| *** Marga_ has joined #openstack-nova | 18:07 | |
| *** annashen has quit IRC | 18:08 | |
| *** Longgeek has quit IRC | 18:08 | |
| *** annashen has joined #openstack-nova | 18:08 | |
| *** ctrath has joined #openstack-nova | 18:12 | |
| *** sdake has quit IRC | 18:13 | |
| *** eglynn has quit IRC | 18:14 | |
| *** Marga_ has quit IRC | 18:14 | |
| rlrossit | mriedem: for the hacking check that I'm adding in https://review.openstack.org/#/c/195713/, is eventlet.spawn() != greenthread.spawn()? | 18:14 |
| *** Marga_ has joined #openstack-nova | 18:14 | |
| *** karimb has quit IRC | 18:14 | |
| *** Marga_ has quit IRC | 18:15 | |
| *** Marga_ has joined #openstack-nova | 18:15 | |
| mriedem | rlrossit: i'd have to look | 18:16 |
| *** achanda has quit IRC | 18:16 | |
| rlrossit | because if both are "evil" with respect to pep8 I can catch both and then I can just ignore nova.utils for that check | 18:17 |
| *** rushiagr_away is now known as rushiagr | 18:17 | |
| openstackgerrit | Giridhar Jayavelu proposed openstack/nova: VMware: add in vnic resource limitations https://review.openstack.org/180202 | 18:18 |
| openstackgerrit | Giridhar Jayavelu proposed openstack/nova: VMware: create common object for limits, reservations and shares https://review.openstack.org/179059 | 18:18 |
| openstackgerrit | Giridhar Jayavelu proposed openstack/nova: VMware: store extra_specs object https://review.openstack.org/179061 | 18:18 |
| openstackgerrit | Giridhar Jayavelu proposed openstack/nova: Derive resource quota from Flavor's extra_spec when attaching interface to an instance https://review.openstack.org/196171 | 18:18 |
| *** oro has quit IRC | 18:26 | |
| *** tonytan4ever has joined #openstack-nova | 18:28 | |
| *** artom has quit IRC | 18:29 | |
| *** arif-ali has quit IRC | 18:31 | |
| *** annegentle has joined #openstack-nova | 18:32 | |
| *** arif-ali has joined #openstack-nova | 18:36 | |
| *** annegentle has quit IRC | 18:37 | |
| *** arif-ali has quit IRC | 18:37 | |
| *** arif-ali has joined #openstack-nova | 18:38 | |
| *** sdake has joined #openstack-nova | 18:38 | |
| *** rushiagr is now known as rushiagr_away | 18:39 | |
| *** hogepodge has joined #openstack-nova | 18:41 | |
| mriedem | rlrossit: i have another thing you could work on which might be 'fun' | 18:42 |
| mriedem | s/fun/educational/ | 18:42 |
| *** ildikov has quit IRC | 18:42 | |
| rlrossit | what isn't fun about openstack?!?! | 18:42 |
| mriedem | specs, apparently | 18:42 |
| rlrossit | as someone on the sidelines, I kinda find that fun | 18:43 |
| mriedem | so it's a follow on to this race fix | 18:44 |
| mriedem | https://review.openstack.org/#/c/194843/ | 18:44 |
| mriedem | rlrossit: there is still a race when associating a fixed ip to an instance | 18:45 |
| mriedem | http://git.openstack.org/cgit/openstack/nova/tree/nova/network/manager.py#n909 | 18:45 |
| mriedem | because the fixed_ip.allocated flag isn't set until here http://git.openstack.org/cgit/openstack/nova/tree/nova/network/manager.py#n925 | 18:45 |
| mriedem | which is after we've associated the fixed IP and instance | 18:45 |
| mriedem | what we want to do is make that all atomic | 18:45 |
| mriedem | in the db api | 18:45 |
| mriedem | so to do that, we need to update the fixed ip associate/associate_pool methods in the fixed IP object to pass in the vif and plumb that down to the db api | 18:46 |
| larsks | What sets the 'injected' metadata on a network (e.g., as in network.get_meta('injected'))? | 18:46 |
| openstackgerrit | Jay Pipes proposed openstack/nova: Adds MonitorMetricTypeField enum field https://review.openstack.org/196179 | 18:46 |
| openstackgerrit | Jay Pipes proposed openstack/nova: Adds MonitorMetric object https://review.openstack.org/196180 | 18:46 |
| openstackgerrit | Jay Pipes proposed openstack/nova: Rework monitor plugin interface and API https://review.openstack.org/196181 | 18:46 |
| openstackgerrit | Jay Pipes proposed openstack/nova: Use stevedore for loading monitor extensions https://review.openstack.org/196182 | 18:46 |
| rlrossit | mriedem: so we need to link the vif to the fixed ip in the DB? or am I off? | 18:47 |
| mriedem | rlrossit: right, we want to set fixed_ip.virtual_interface_id and the allocated boolean in the db api associate/_pool method if there is a vif passed in | 18:48 |
| *** tjones1 has quit IRC | 18:48 | |
| mriedem | rlrossit: so these 2 associate methdos in the fixed IP object would need to change to take an optional vif parameter | 18:49 |
| mriedem | http://git.openstack.org/cgit/openstack/nova/tree/nova/objects/fixed_ip.py#n150 | 18:49 |
| mriedem | which also means a version bump on that object at the top | 18:49 |
| *** tjones1 has joined #openstack-nova | 18:49 | |
| mriedem | http://git.openstack.org/cgit/openstack/nova/tree/nova/objects/fixed_ip.py#n44 | 18:49 |
| rlrossit | wait mriedem the fixed ip object already has a nullable virtual_interface field? | 18:50 |
| mriedem | rlrossit: yes | 18:50 |
| leakypipes | dhellmann: https://review.openstack.org/196182 is stevedore related. would love your review on it (no priority, though) | 18:50 |
| *** Nikolay_St has joined #openstack-nova | 18:51 | |
| *** annegentle has joined #openstack-nova | 18:51 | |
| mriedem | rlrossit: hmm, yeah, maybe don't need the method sig change | 18:51 |
| mriedem | rlrossit: so if you have the vif before calling fip.associate, you could set the vif and allocated flags on the fip before calling associate | 18:52 |
| mriedem | but i think you'd still need to pass the vif from the fip object down into the db api so it doesn't overwrite those values | 18:53 |
| mriedem | or that they aren't set in the db if the associate call fails | 18:53 |
| kfox1111 | arg... it looks like amazon's doing exactly what i was suggesting in the first place. :/ you get an instance user's token from http://169.254.169.254/latest/meta-data/iam/security-credentials | 18:53 |
| rlrossit | mriedem: so.... wouldn't that just mean http://git.openstack.org/cgit/openstack/nova/tree/nova/network/manager.py#n925 would just get moved to be above the associate? | 18:54 |
| kfox1111 | the only difference is the vm may have a few different users assigned to it. | 18:54 |
| kfox1111 | so they did extend their metadata server. :/ | 18:54 |
| rlrossit | mriedem: then once that's set on the fip before the associate happens it would be on the fip object? | 18:54 |
| mriedem | rlrossit: yeah, but would have to see if that's actually persisted | 18:54 |
| rlrossit | is that what fip.save() is doing? | 18:54 |
| mriedem | yes | 18:55 |
| larsks | superdan: around for a nova networking question? | 18:55 |
| mriedem | but i don't think you can call fip.save() before associate | 18:55 |
| superdan | larsks: around yeah | 18:55 |
| mriedem | rlrossit: basically, i want that stuff to all be set in the same db transaction | 18:55 |
| larsks | N/m, saw your comment elsewhere. | 18:55 |
| mriedem | rlrossit: like, i think you still need to update the db api methods to pass a vif in here http://git.openstack.org/cgit/openstack/nova/tree/nova/db/sqlalchemy/api.py#n1112 | 18:56 |
| *** agireud has joined #openstack-nova | 18:56 | |
| superdan | larsks: there is a config element for injected | 18:56 |
| *** Nikolay_St has quit IRC | 18:56 | |
| *** agireud has quit IRC | 18:56 | |
| *** rushiagr_away is now known as rushiagr | 18:56 | |
| larsks | superdan: Which one? There is the injected_template, for one... | 18:57 |
| mriedem | rlrossit: and if there is a vif passed in, you'd update virtual_interface_id and allocated in params here http://git.openstack.org/cgit/openstack/nova/tree/nova/db/sqlalchemy/api.py#n1139 | 18:57 |
| larsks | There is a flat_injected option, which looks like it may be nova-network only, but I'm not sure yet... | 18:57 |
| mriedem | so that when the update is called on the table, it's all set together in the same db session | 18:57 |
| superdan | larsks: I think there is an injected conf variable and a property you can set on a network when you create it too | 18:58 |
| larsks | superdan: The thing I'm trying to figure out is if that is available for neutron networks. There isn't an obvious 'inject this network' flag to net-create... | 18:58 |
| superdan | larsks: well, flat_injected is used in neutronapi | 18:59 |
| mriedem | rlrossit: once the db api is done, the fixed ip object is returned with those fields set | 18:59 |
| mriedem | http://git.openstack.org/cgit/openstack/nova/tree/nova/objects/fixed_ip.py#n150 | 18:59 |
| larsks | superdan: Oh yah? Well, that is interesting. | 18:59 |
| superdan | just a sec | 18:59 |
| mriedem | rlrossit: so actually you can't move http://git.openstack.org/cgit/openstack/nova/tree/nova/network/manager.py#n925 before fip.associate because you don't have the fip yet, associate is a classmethod | 18:59 |
| superdan | larsks: https://github.com/openstack/nova/blob/master/nova/network/neutronv2/api.py#L1571 | 18:59 |
| mriedem | rlrossit: so you have to pass the vif into FixedIP.associate() | 18:59 |
| rlrossit | mriedem: so if the fip object would no longer have the vif field explicitly set on it then right? we would just pass it in to associate and let the db handle everything? | 18:59 |
| larsks | Yeah, just looking at that. | 19:00 |
| mriedem | rlrossit: yeah | 19:00 |
| larsks | Let me give it a shot and see what happens. | 19:00 |
| mriedem | rlrossit: return cls._from_db_object(context, cls(context), db_fixedip) takes care of setting the fields in the fixed IP object | 19:00 |
| *** ijuwang has joined #openstack-nova | 19:00 | |
| mriedem | returned from associate back to the network manager | 19:00 |
| *** claudiub|2 has quit IRC | 19:00 | |
| rlrossit | mriedem: ok that's what I thought | 19:00 |
| mriedem | rlrossit: so it's still a version bump in the object | 19:00 |
| mriedem | b/c the method sig changes | 19:00 |
| rlrossit | even though they're optional it still changes? | 19:01 |
| *** ijw__ has joined #openstack-nova | 19:01 | |
| mriedem | rlrossit: yeah, you'd need a vif=None kwarg | 19:01 |
| rlrossit | is there a bug associated w/ this? or is it the same bug that other change was linked to? | 19:02 |
| mriedem | rlrossit: same bug | 19:02 |
| *** scottda has quit IRC | 19:03 | |
| mriedem | rlrossit: it's at least good exposure to touching many parts of the code base | 19:03 |
| mriedem | and dealing with an objects versoin change | 19:03 |
| rlrossit | mriedem: alright. I think I know enough to at least get a start on this then | 19:03 |
| mriedem | ok | 19:04 |
| mriedem | i'd start in the db api and work backwards from there | 19:04 |
| *** ijw has quit IRC | 19:04 | |
| rlrossit | mriedem: yeah that's for sure.... not too clear on how I can test this change out yet, but we'll wait until I actually have it started before I try that :P | 19:04 |
| mriedem | rlrossit: well, there are unit tests for the db api which run against an actual db connection so you can see if you got the db api stuff correct | 19:06 |
| mriedem | unit tests for the fixed ip object will use mock for the db api calls | 19:06 |
| *** otter768 has joined #openstack-nova | 19:06 | |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: discover supported versions to choose latest one https://review.openstack.org/184281 | 19:06 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Implements 'microversions' api type - Part 2 https://review.openstack.org/167408 | 19:06 |
| mriedem | rlrossit: tempest changes upstream will hit the changes to the network manager that pulls it all together | 19:06 |
| mriedem | *tempest tests | 19:07 |
| rlrossit | mriedem: are there unit tests that I should be adding around this? (whether it be db or manager related) | 19:09 |
| *** harlowja has quit IRC | 19:09 | |
| larsks | superdan: flat_injected does set the 'injected' key on the network. so that's good. | 19:09 |
| superdan | larsks: ....but? | 19:09 |
| larsks | ...but, since I'm running on my laptop, my guest is still booting up :) | 19:09 |
| *** ijw__ has quit IRC | 19:09 | |
| superdan | heh | 19:10 |
| *** shihanzhang has quit IRC | 19:10 | |
| openstackgerrit | Marian Horban proposed openstack/nova: [WIP] Kicking the oslo.service tires https://review.openstack.org/192900 | 19:10 |
| larsks | ...but, still not showing up in the metadata on the config drive, which is what I was trying to have happen. | 19:11 |
| *** otter768 has quit IRC | 19:11 | |
| *** sharkJing has quit IRC | 19:13 | |
| *** sharkJing has joined #openstack-nova | 19:14 | |
| *** kiwik has quit IRC | 19:15 | |
| *** baoli has joined #openstack-nova | 19:15 | |
| *** RuiChen has joined #openstack-nova | 19:16 | |
| *** r-daneel has quit IRC | 19:17 | |
| mriedem | rlrossit: there are unit tests for everything | 19:17 |
| *** RuiChen has quit IRC | 19:17 | |
| mtreinish | mriedem: are there unit tests for the unit tests? | 19:18 |
| *** RuiChen has joined #openstack-nova | 19:18 | |
| *** rushiagr is now known as rushiagr_away | 19:19 | |
| mriedem | rlrossit: see the section on unit tests here https://github.com/openstack/nova/blob/master/HACKING.rst | 19:19 |
| mriedem | mtreinish: i think that's a side project repo of yours | 19:19 |
| mriedem | os-testr^2 | 19:19 |
| mtreinish | mriedem: os-testr has unit tests, but not very good coverage | 19:19 |
| *** markvoelker has quit IRC | 19:20 | |
| mriedem | s/os-testr/nova-network/ :) | 19:20 |
| *** annegentle has quit IRC | 19:20 | |
| *** sharkJing has quit IRC | 19:20 | |
| *** Piet has quit IRC | 19:21 | |
| *** sharkJing has joined #openstack-nova | 19:21 | |
| *** kiwik-chenrui has joined #openstack-nova | 19:21 | |
| *** harlowja has joined #openstack-nova | 19:22 | |
| mriedem | ctrath: something else to think about with this bulk stats thing is the rpc message size limit | 19:22 |
| mriedem | ctrath: i posted something to the ML on that in the same thread | 19:23 |
| mriedem | ctrath: we could start pulling some of this together in a backlog spec | 19:23 |
| mriedem | since it's not going to make liberty | 19:23 |
| *** kiwik-chenrui has quit IRC | 19:23 | |
| *** kiwik-chenrui has joined #openstack-nova | 19:24 | |
| *** RuiChen has quit IRC | 19:24 | |
| larsks | Ah, because you don't get injection if dhcp_server is True in any case. | 19:24 |
| *** baoli has quit IRC | 19:26 | |
| *** alexpilotti has quit IRC | 19:26 | |
| *** markvoelker has joined #openstack-nova | 19:26 | |
| kfox1111 | superdan, bauwser, flashgordon: a couple more things to think about. what about cert expiration? In light of suspending the vm for a while and then needing to resume it later? also, what about snapshotting? A user might not realize that snapshoting a vm would capture their cert. | 19:26 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Implements 'microversions' api type - Part 2 https://review.openstack.org/167408 | 19:28 |
| *** sharkJing has quit IRC | 19:28 | |
| *** kiwik-chenrui has quit IRC | 19:29 | |
| ctrath | mriedem: are you thinking the return message is going to exceed the rpc message limit? | 19:29 |
| *** sharkJing has joined #openstack-nova | 19:29 | |
| mriedem | ctrath: i'm thinking it could | 19:29 |
| *** kiwik-chenrui has joined #openstack-nova | 19:29 | |
| ctrath | mriedem: do you have a link to the ML? | 19:29 |
| mriedem | ctrath: do you have that setup yet? | 19:30 |
| *** kiwik-chenrui has quit IRC | 19:30 | |
| mriedem | ctrath: http://lists.openstack.org/pipermail/openstack-dev/2015-June/068129.html | 19:30 |
| ctrath | thx.. haven't configured it yet on my new machine | 19:31 |
| *** cfriesen has quit IRC | 19:31 | |
| *** ijw has joined #openstack-nova | 19:31 | |
| *** markvoelker has quit IRC | 19:32 | |
| *** markvoelker has joined #openstack-nova | 19:32 | |
| *** sharkJing has quit IRC | 19:33 | |
| *** sharkJing has joined #openstack-nova | 19:34 | |
| *** RuiChen has joined #openstack-nova | 19:34 | |
| mriedem | ctrath: here is a thread in the rabbitmq ML on message size http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/2012-March/018699.html | 19:35 |
| mriedem | sounds like it's more or less bounded by our system resources, but there are obviously fallbacks to sending things too big | 19:35 |
| *** RuiChen has quit IRC | 19:35 | |
| mriedem | i could have sworn we used to see failures in unit tests from time to time when messages got over 64K though | 19:35 |
| mriedem | superdan: ^ do you remember that? | 19:35 |
| *** RuiChen has joined #openstack-nova | 19:35 | |
| mriedem | bnemec: ^ | 19:35 |
| mriedem | i could have sworn that bnemec fixed it like 2 years ago | 19:36 |
| ctrath | mriedem: I wonder if the limits are the same for text messages as they are for binary messages | 19:36 |
| *** ijw has quit IRC | 19:36 | |
| flashgordon | who wants to review https://review.openstack.org/194840 before its too late? | 19:36 |
| flashgordon | mriedem: https://review.openstack.org/194840 | 19:36 |
| superdan | mriedem: with rabbit? no | 19:36 |
| mriedem | superdan: yeah, whatever we used in the unit tests | 19:37 |
| mriedem | fake rpc | 19:37 |
| superdan | mriedem: hmm not sure why that would be | 19:37 |
| mriedem | i think things would go crazy with serialization of the instance objects or something | 19:38 |
| mriedem | when an error occurred and that masked the real issue | 19:38 |
| mriedem | like in the integrated api samples tests | 19:38 |
| *** kiwik-chenrui has joined #openstack-nova | 19:38 | |
| mriedem | note that i'm talking about like havana level code | 19:38 |
| *** markvoelker_ has joined #openstack-nova | 19:38 | |
| mriedem | bnemec would probably know what i'm talking about | 19:38 |
| mriedem | flashgordon: so this is your thingy now huh http://docs-draft.openstack.org/40/194840/3/check/gate-nova-docs/4758715//doc/build/html/architecture.html | 19:39 |
| *** READ10 has quit IRC | 19:39 | |
| superdan | hmm, I dunno | 19:39 |
| bnemec | mriedem: I fixed an issue with qpid, but at the time I didn't think rabbit had the same problem. | 19:39 |
| flashgordon | mriedem: yup | 19:39 |
| mriedem | bnemec: ah maybe it was just qpid, i can't remember | 19:40 |
| ctrath | mriedem: where did you put the example payload from virsh? | 19:40 |
| *** kiwik has joined #openstack-nova | 19:40 | |
| *** RuiChen has quit IRC | 19:40 | |
| mriedem | ctrath: it's in a paste in the mailing list | 19:40 |
| mriedem | same thread | 19:40 |
| *** markvoelker has quit IRC | 19:40 | |
| ctrath | got it.. thx. | 19:41 |
| *** achanda has joined #openstack-nova | 19:41 | |
| ctrath | mriedem: so, is there a limit to the number of nodes? | 19:41 |
| mriedem | flashgordon: shouldn't neutron be in your hexagon list of other external services? | 19:42 |
| mriedem | flashgordon: and for the left side 'network' box that would maybe be prefixed as (nova-)network | 19:42 |
| mriedem | since conductor isn't talking to neutron | 19:43 |
| *** kiwik-chenrui has quit IRC | 19:43 | |
| mriedem | comments inline | 19:44 |
| *** markvoelker has joined #openstack-nova | 19:44 | |
| *** sharkJing has quit IRC | 19:44 | |
| mriedem | ctrath: which nodes? | 19:44 |
| mriedem | compute nodes? | 19:44 |
| *** markvoelker_ has quit IRC | 19:45 | |
| bnemec | Wow, nostalgia. https://review.openstack.org/#/c/28711/ was my first incubator patch. | 19:45 |
| mriedem | bnemec: i knew you'd remember :) | 19:45 |
| ctrath | mriedem: sorry.. my terminology isn't correct yet…. guests? | 19:45 |
| bnemec | I will always remember that one because the gate started blowing up on it about 24 hours after that finally merged. :-) | 19:45 |
| *** kiwik has quit IRC | 19:45 | |
| *** sharkJing has joined #openstack-nova | 19:46 | |
| *** kiwik has joined #openstack-nova | 19:46 | |
| *** Marga_ has quit IRC | 19:46 | |
| *** baoli has joined #openstack-nova | 19:47 | |
| *** _diana_ has joined #openstack-nova | 19:48 | |
| *** Marga_ has joined #openstack-nova | 19:48 | |
| mriedem | ctrath: there isn't a limit to the number of instances per compute node, no | 19:48 |
| *** arnaud____ has joined #openstack-nova | 19:48 | |
| mriedem | ctrath: you're only bounded by resource limits, although it'd be dumb to grow vertically in case that one compute node goes down | 19:48 |
| mriedem | ctrath: which is why many people don't like the clustered hypervisor approach since it grows vertically behind the scenes rather than horizontally per compute nodes | 19:49 |
| *** RuiChen has joined #openstack-nova | 19:49 | |
| *** _diana_ has quit IRC | 19:49 | |
| mriedem | ctrath: superdan and leakypipes haven't had to explain that yet this week, so i'm sure they'd be all for it | 19:49 |
| *** kiwik has quit IRC | 19:51 | |
| *** sharkJing has quit IRC | 19:52 | |
| *** sharkJing has joined #openstack-nova | 19:52 | |
| *** RuiChen has quit IRC | 19:52 | |
| *** RuiChen has joined #openstack-nova | 19:52 | |
| *** Marga_ has quit IRC | 19:53 | |
| *** arnaud____ has quit IRC | 19:53 | |
| superdan | heh | 19:53 |
| *** markvoelker has quit IRC | 19:54 | |
| *** mestery has joined #openstack-nova | 19:54 | |
| *** _diana_ has joined #openstack-nova | 19:56 | |
| *** Piet has joined #openstack-nova | 19:56 | |
| *** RuiChen has quit IRC | 19:56 | |
| *** baoli has quit IRC | 19:57 | |
| *** sharkJing has quit IRC | 19:57 | |
| *** markvoelker has joined #openstack-nova | 19:58 | |
| *** achanda has quit IRC | 19:58 | |
| *** sharkJing has joined #openstack-nova | 19:58 | |
| *** ijw has joined #openstack-nova | 19:58 | |
| *** ijuwang has quit IRC | 19:59 | |
| *** RuiChen has joined #openstack-nova | 19:59 | |
| kfox1111 | so, I'm trying to rewite the spec, but it isn't really designed to have lots of different implementations specified out in it. and the generic "what do we want to acomplish" is so generic it doesn't really get us anywhare. What should I do? | 20:01 |
| *** kiwik has joined #openstack-nova | 20:02 | |
| *** zhenguo has quit IRC | 20:03 | |
| *** Longgeek has joined #openstack-nova | 20:04 | |
| *** RuiChen has quit IRC | 20:04 | |
| *** kiwik has quit IRC | 20:05 | |
| *** kiwik has joined #openstack-nova | 20:05 | |
| *** Marga_ has joined #openstack-nova | 20:06 | |
| *** kiwik-chenrui has joined #openstack-nova | 20:07 | |
| *** sharkJing has quit IRC | 20:08 | |
| *** ivasev has quit IRC | 20:09 | |
| *** Longgeek has quit IRC | 20:09 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Support Network objects in set_network_host https://review.openstack.org/192396 | 20:09 |
| *** markvoelker_ has joined #openstack-nova | 20:09 | |
| mriedem | kfox1111: couldn't you just dump them into the alternatives section? | 20:09 |
| *** Marga_ has quit IRC | 20:10 | |
| *** kiwik has quit IRC | 20:10 | |
| *** eglynn has joined #openstack-nova | 20:10 | |
| *** Marga_ has joined #openstack-nova | 20:10 | |
| *** markvoelker has quit IRC | 20:11 | |
| *** markvoelker_ has quit IRC | 20:12 | |
| *** radek_ has quit IRC | 20:14 | |
| *** kiwik-chenrui has quit IRC | 20:15 | |
| *** browne has quit IRC | 20:15 | |
| *** kiwik-chenrui has joined #openstack-nova | 20:15 | |
| *** oro has joined #openstack-nova | 20:16 | |
| kfox1111 | mriedem: I'm worried its going to be huge... there are a lot of permutations. :/ | 20:16 |
| *** kiwik-chenrui has quit IRC | 20:16 | |
| *** kiwik-chenrui has joined #openstack-nova | 20:16 | |
| openstackgerrit | Andrey Pavlov proposed openstack/python-novaclient: Fix resolving image.id in servers.boot https://review.openstack.org/196082 | 20:16 |
| *** kiwik-chenrui has quit IRC | 20:17 | |
| *** kiwik-chenrui has joined #openstack-nova | 20:18 | |
| mriedem | ugh | 20:18 |
| *** ctrath1 has joined #openstack-nova | 20:20 | |
| *** krtaylor has quit IRC | 20:21 | |
| *** Marga_ has quit IRC | 20:22 | |
| *** vladikr has quit IRC | 20:22 | |
| *** ijuwang has joined #openstack-nova | 20:23 | |
| *** Marga_ has joined #openstack-nova | 20:23 | |
| flashgordon | mriedem: thanks, so should I have a nova-network diagram and a neutro none? | 20:23 |
| flashgordon | mriedem: thanks, so should I have a nova-network diagram and a neutron one? | 20:23 |
| mriedem | flashgordon: see my comments in the review? | 20:23 |
| flashgordon | looking | 20:24 |
| *** kiwik has joined #openstack-nova | 20:24 | |
| flashgordon | mriedem: I'll just add a note saying this is nova-networking not neutron | 20:24 |
| *** ctrath1 has quit IRC | 20:24 | |
| mriedem | flashgordon: well, i think you could have both in there with an asterisk | 20:25 |
| mriedem | and in the * note you mention that it's either nova-network or neutron | 20:25 |
| *** hemna has quit IRC | 20:25 | |
| mriedem | but that's how it's modeled for each | 20:25 |
| mriedem | idk | 20:25 |
| mriedem | nova-network* on left | 20:25 |
| mriedem | neutron* on right in the hexagon with glance/cinder | 20:25 |
| *** lucas__ has quit IRC | 20:25 | |
| mriedem | * pick one, don't f it up | 20:26 |
| mriedem | as your note | 20:26 |
| *** nic has joined #openstack-nova | 20:26 | |
| kfox1111 | so... how is config drive implemented today? Iso image? | 20:26 |
| flashgordon | mriedem: I am just trying to make the non-ascii version of http://docs.openstack.org/developer/nova/architecture.html | 20:27 |
| *** kiwik-chenrui has quit IRC | 20:27 | |
| openstackgerrit | Andrey Pavlov proposed openstack/python-novaclient: Fix resolving image.id in servers.boot https://review.openstack.org/196082 | 20:27 |
| flashgordon | mriedem: but I'll update the comments | 20:28 |
| mriedem | flashgordon: yeah, but, you can also make it about 3 years more current | 20:28 |
| *** ijuwang has quit IRC | 20:29 | |
| mriedem | flashgordon: but that's fine if you just want to make the note that it's the nova-networking model | 20:29 |
| mriedem | flashgordon: you didn't want to point out that the compute is talking to libvirt/xen/etc | 20:29 |
| mriedem | ? | 20:29 |
| *** kiwik has quit IRC | 20:29 | |
| *** lucas__ has joined #openstack-nova | 20:29 | |
| flashgordon | mriedem: wasn't sure if the libvirt level of detail in the diagram really helps thoughts? | 20:30 |
| mriedem | flashgordon: i dont think it hurts | 20:31 |
| mriedem | flashgordon: you could ask ctrath, he's been looking at this stuff lately with fresh eyes | 20:31 |
| openstackgerrit | Thang Pham proposed openstack/nova: VMware: Add support for swap disk https://review.openstack.org/142174 | 20:31 |
| mriedem | ctrath: compare old http://docs.openstack.org/developer/nova/architecture.html to new proposed http://docs-draft.openstack.org/40/194840/3/check/gate-nova-docs/4758715//doc/build/html/architecture.html | 20:31 |
| mriedem | ctrath: do you think it's helpful to have the compute part pointing out that it's talking to libvirt/xen/etc? | 20:32 |
| *** kiwik has joined #openstack-nova | 20:32 | |
| *** ijw__ has joined #openstack-nova | 20:32 | |
| ctrath | mriedem: probably | 20:32 |
| ctrath | also, where's the controller in the proposed diagram? | 20:33 |
| mriedem | ctrath: well, that's a deployment topology concept | 20:33 |
| mriedem | ctrath: for the purpose of that diagram, you could consider it single-node | 20:33 |
| ctrath | alright.. fair enough | 20:33 |
| flashgordon | so right now the explanation on the bottom mentions the hypervisor | 20:34 |
| mriedem | flashgordon: oh right | 20:34 |
| ctrath | could the explanation below in include a subnote that specifies specific hypervisors? | 20:35 |
| *** kiwik has quit IRC | 20:35 | |
| mriedem | ctrath: no, probably note | 20:35 |
| mriedem | *not | 20:35 |
| mriedem | ctrath: b/c that will be stale over time | 20:35 |
| *** kiwik has joined #openstack-nova | 20:35 | |
| *** ijw has quit IRC | 20:35 | |
| mriedem | flashgordon: so i guess i'm good if you just make a note that this is the nova-network model | 20:35 |
| ctrath | mriedem: true.. Does this documentation get updated each release? | 20:35 |
| mriedem | flashgordon: if you wanted to go the extra mile and add a separate diagram for neutron, that'd be cool too, but whatevs | 20:35 |
| mriedem | ctrath: it can, but doesn't normally | 20:36 |
| *** annegentle has joined #openstack-nova | 20:37 | |
| ctrath | mriedem: what about bare-metal? Is the firmware in the bare-metal instance considered a hypervisor? I'm wondering if that should be spelled out | 20:38 |
| mriedem | ctrath: ironic is the hypervisor in that case | 20:38 |
| mriedem | it's a black box | 20:38 |
| ctrath | alright.. I didn't realize ironic was considered a hypervisor | 20:39 |
| *** oro has quit IRC | 20:40 | |
| *** rlrossit has left #openstack-nova | 20:40 | |
| *** RuiChen has joined #openstack-nova | 20:40 | |
| *** kiwik has quit IRC | 20:41 | |
| *** isd1 has joined #openstack-nova | 20:41 | |
| *** ijuwang has joined #openstack-nova | 20:41 | |
| flashgordon | ctrath: that is part of the problem ... it is a special child | 20:41 |
| *** isd has quit IRC | 20:41 | |
| *** fawadk has quit IRC | 20:42 | |
| melwitt | flashgordon: thanks for making that diagram, it's a vast improvement | 20:42 |
| *** RuiChen has quit IRC | 20:42 | |
| *** RuiChen has joined #openstack-nova | 20:43 | |
| *** tonytan4ever has quit IRC | 20:43 | |
| *** fawadkhaliq has joined #openstack-nova | 20:43 | |
| *** fawadkhaliq has quit IRC | 20:44 | |
| flashgordon | melwitt: I have been wanting to make it not nearly as bad for a long time | 20:45 |
| flashgordon | new version coming soon | 20:45 |
| mtreinish | melwitt, flashgordon: I'm not sure, I have a really hard time figuring out what it's supposed to look like when I open the svg in vim. It worked fine with the old version though | 20:46 |
| *** RuiChen has quit IRC | 20:46 | |
| *** RuiChen has joined #openstack-nova | 20:47 | |
| melwitt | mtreinish: oh, I cheated and opened the draft docs by clicking on the gate-nova-docs job | 20:47 |
| flashgordon | mriedem: how is this http://imgh.us/architecture.svg | 20:47 |
| *** _diana_ has quit IRC | 20:47 | |
| flashgordon | mtreinish: you can't mentally render the XML? | 20:47 |
| *** tonytan4ever has joined #openstack-nova | 20:48 | |
| flashgordon | ctrath: ^ new draft | 20:48 |
| mtreinish | flashgordon: you're making it so I'm going to have to learn how to | 20:48 |
| * bauwser whispers by thinking how HTML5 could help but there is something called IE | 20:48 | |
| *** kiwik-chenrui has joined #openstack-nova | 20:48 | |
| *** toure has quit IRC | 20:48 | |
| flashgordon | dia isn't that hard to use, and it has shapes for IBM mainframes | 20:49 |
| mriedem | flashgordon: ok, in the neutron diagram i might have still just lumped neutron in with glance/cinder on the right side | 20:49 |
| melwitt | \o/ mainframes shapes | 20:49 |
| mriedem | not sure if i like the compute square eating the hypervisor, but now i'm just nit picking | 20:49 |
| flashgordon | mriedem: I thought this would make it easier to spot the difference versus nova-net | 20:49 |
| *** ajayaa has quit IRC | 20:49 | |
| *** tjones2 has joined #openstack-nova | 20:50 | |
| flashgordon | yeah you are nitpicking | 20:50 |
| *** jecarey has quit IRC | 20:50 | |
| mtreinish | mriedem: but that's how it actually works isn't it? | 20:50 |
| flashgordon | and I couldn't find the pacman symbol | 20:50 |
| mriedem | flashgordon: i think the difference would be easier to spot if you completely just remove hte left side network thing in the neutron case | 20:50 |
| mriedem | and lump neutron in with glance and cinder as external services that nova compute and api talk to | 20:50 |
| flashgordon | keystone has its own thing anyway | 20:50 |
| *** xyang1 has quit IRC | 20:50 | |
| mriedem | yeah, whatevs, it's fine | 20:51 |
| *** kiwik has joined #openstack-nova | 20:51 | |
| mriedem | it's better than what was there earlire | 20:51 |
| mriedem | i'll just tear it all up after you're officially gone :) | 20:51 |
| flashgordon | mriedem: I will still be on IRC so be careful | 20:51 |
| *** markvoelker has joined #openstack-nova | 20:51 | |
| *** Marga_ has quit IRC | 20:51 | |
| mriedem | comstud: hi | 20:51 |
| mriedem | he | 20:51 |
| *** RuiChen has quit IRC | 20:52 | |
| openstackgerrit | Joe Gordon proposed openstack/nova: Replace ascii art architecture diagram with svg image https://review.openstack.org/194840 | 20:52 |
| flashgordon | mriedem: don't forget vishy | 20:52 |
| mriedem | flashgordon: vishy already ignores me on purpose | 20:52 |
| mriedem | i was asking vish-network questions earlier in the week | 20:52 |
| *** lucas__ has quit IRC | 20:52 | |
| flashgordon | haha | 20:52 |
| *** kiwik-chenrui has quit IRC | 20:53 | |
| *** Marga_ has joined #openstack-nova | 20:54 | |
| *** kiwik-chenrui has joined #openstack-nova | 20:54 | |
| *** tjones2 has quit IRC | 20:54 | |
| *** thangp has quit IRC | 20:55 | |
| melwitt | boo to flashgordon leaving | 20:56 |
| flashgordon | 3 hours | 20:56 |
| flashgordon | I will mess y'all | 20:56 |
| *** kiwik has quit IRC | 20:56 | |
| melwitt | flashgordon: you'll be missed. best of luck at your new gig | 20:57 |
| *** mestery has quit IRC | 20:57 | |
| openstackgerrit | Giridhar Jayavelu proposed openstack/nova: VMware: add in vnic resource limitations https://review.openstack.org/180202 | 20:57 |
| openstackgerrit | Giridhar Jayavelu proposed openstack/nova: VMware: create common object for limits, reservations and shares https://review.openstack.org/179059 | 20:57 |
| openstackgerrit | Giridhar Jayavelu proposed openstack/nova: VMware: Resource limits for memory https://review.openstack.org/179060 | 20:57 |
| openstackgerrit | Giridhar Jayavelu proposed openstack/nova: VMware: store extra_specs object https://review.openstack.org/179061 | 20:57 |
| openstackgerrit | Giridhar Jayavelu proposed openstack/nova: VMware: add resource limits for disk https://review.openstack.org/179062 | 20:57 |
| *** jwcroppe_ has joined #openstack-nova | 20:58 | |
| *** ijw__ has quit IRC | 20:58 | |
| *** RuiChen has joined #openstack-nova | 20:58 | |
| *** kiwik-chenrui has quit IRC | 20:58 | |
| flashgordon | melwitt: thanks | 20:59 |
| *** Marga_ has quit IRC | 20:59 | |
| *** nic has quit IRC | 20:59 | |
| *** jwcroppe has quit IRC | 21:00 | |
| mriedem | melwitt: thanks for clearing up the confusion on https://review.openstack.org/#/c/171389/ | 21:01 |
| *** lucas__ has joined #openstack-nova | 21:01 | |
| *** Marga_ has joined #openstack-nova | 21:01 | |
| *** lucas__ has quit IRC | 21:01 | |
| figleaf | flashgordon: so what's the new gig? | 21:01 |
| *** jwcroppe_ has quit IRC | 21:02 | |
| *** thorst1 has quit IRC | 21:02 | |
| *** lucas__ has joined #openstack-nova | 21:02 | |
| *** jwcroppe has joined #openstack-nova | 21:02 | |
| flashgordon | figleaf: pinterest | 21:03 |
| melwitt | mriedem: it took me awhile to understand it myself. the original use case was ceilometer and I guess if they query Nova proxy, they don't have to know whether it's Neutron or nova-network. though I don't know how ceilometer works, whether it's already polling Neutron/Glance/etc anyway | 21:03 |
| figleaf | flashgordon: well, nothing but the best to you. You will be missed. | 21:04 |
| *** lucas___ has joined #openstack-nova | 21:04 | |
| mriedem | melwitt: yeah and them artom said the ceilometer thing was reverted and not coming back | 21:04 |
| *** kiwik-chenrui has joined #openstack-nova | 21:04 | |
| mriedem | melwitt: basically, the spec says 'it was a thing in novaclient that didn't work but people have said they want it to' | 21:04 |
| flashgordon | figleaf: thanks, I will miss all the fun (and yelling) we have in nova | 21:04 |
| mriedem | melwitt: which i think just means, use neutronclient then | 21:04 |
| *** krtaylor has joined #openstack-nova | 21:04 | |
| *** jwcroppe_ has joined #openstack-nova | 21:05 | |
| figleaf | flashgordon: if it gets too bad, pop back here and we'll yell at you | 21:05 |
| *** kiwik-chenrui has quit IRC | 21:05 | |
| flashgordon | figleaf: I may take you up on that :) | 21:05 |
| melwitt | mriedem: yeah, though the other spec was the same person who did the original change that got reverted, which was a ceilometer person I thought. so there's a "hint" that ceilometer wants it again but it's really not clear. and I think the specs should be really clear about "why" of the change | 21:06 |
| mriedem | melwitt: agreed | 21:06 |
| *** lucas__ has quit IRC | 21:06 | |
| *** raildo has quit IRC | 21:06 | |
| *** lucas___ has quit IRC | 21:06 | |
| mriedem | sounds like some synergy 2.0 is required here | 21:06 |
| *** jwcroppe has quit IRC | 21:06 | |
| bauwser | need some help on understanding Ceilo ? | 21:07 |
| *** kiwik-chenrui has joined #openstack-nova | 21:07 | |
| *** otter768 has joined #openstack-nova | 21:07 | |
| *** RuiChen has quit IRC | 21:07 | |
| bauwser | melwitt: ^ ? | 21:07 |
| * bauwser misses the whole context | 21:07 | |
| *** browne has joined #openstack-nova | 21:08 | |
| *** lpetrut has quit IRC | 21:09 | |
| melwitt | bauwser: not really, we were talking about this spec that wants to add "all_tenants" for listing floating ips to Nova API. and the use case "might be" Ceilometer, and I was just saying I didn't know if ceilometer polls all the things (Nova, Neutron, Glance, etc) or if it only queries Nova for information. Neutron already supports "all_tenants" so if Ceilometer can query Neutron, things should be fine | 21:10 |
| bauwser | melwitt: http://docs.openstack.org/developer/ceilometer/architecture.html#gathering-the-data HTH | 21:11 |
| *** romainh has joined #openstack-nova | 21:11 | |
| *** RuiChen has joined #openstack-nova | 21:11 | |
| melwitt | bauwser: cool, thanks | 21:11 |
| *** otter768 has quit IRC | 21:12 | |
| *** htruta has quit IRC | 21:13 | |
| *** kiwik-chenrui has quit IRC | 21:14 | |
| *** mwagner_afk has joined #openstack-nova | 21:15 | |
| *** kiwik-chenrui has joined #openstack-nova | 21:18 | |
| *** markvoelker has quit IRC | 21:19 | |
| *** Longgeek has joined #openstack-nova | 21:22 | |
| *** wxy has joined #openstack-nova | 21:22 | |
| *** RuiChen has quit IRC | 21:22 | |
| *** harlowja has quit IRC | 21:23 | |
| *** lucas__ has joined #openstack-nova | 21:24 | |
| *** openstack has joined #openstack-nova | 21:25 | |
| *** ijw has joined #openstack-nova | 21:26 | |
| *** Longgeek has quit IRC | 21:26 | |
| *** wxy has quit IRC | 21:27 | |
| superdan | dammit leakypipes it's friday | 21:28 |
| superdan | don't make me do work | 21:28 |
| *** lucas__ has quit IRC | 21:28 | |
| *** wxy has joined #openstack-nova | 21:28 | |
| *** aysyanne has quit IRC | 21:28 | |
| *** markvoelker has joined #openstack-nova | 21:28 | |
| *** wxy has quit IRC | 21:28 | |
| *** RuiChen has joined #openstack-nova | 21:29 | |
| *** wxy has joined #openstack-nova | 21:29 | |
| *** _diana_ has joined #openstack-nova | 21:29 | |
| *** tjones2 has joined #openstack-nova | 21:29 | |
| mriedem | 5:30 on a friday is when leakypipes gets shit done | 21:29 |
| mriedem | the phone has stopped rining | 21:29 |
| mriedem | *ringing | 21:30 |
| figleaf | mriedem: it's the rest of the week when he doesn't do jack | 21:30 |
| mriedem | the pugs are asleep in their diapers | 21:30 |
| *** markvoelker has quit IRC | 21:30 | |
| *** kiwik-chenrui has quit IRC | 21:30 | |
| *** Raildo has joined #openstack-nova | 21:31 | |
| superdan | heh | 21:31 |
| melwitt | lol | 21:32 |
| *** zxen has joined #openstack-nova | 21:33 | |
| *** Raildo is now known as raildo | 21:34 | |
| *** earlephilhower has joined #openstack-nova | 21:34 | |
| *** wxy has quit IRC | 21:35 | |
| *** kiwik-chenrui has joined #openstack-nova | 21:35 | |
| *** harlowja has joined #openstack-nova | 21:35 | |
| figleaf | leakypipes: reviewed. LGTM! | 21:36 |
| superdan | wow, that was fast | 21:37 |
| *** wxy has joined #openstack-nova | 21:38 | |
| *** hemna has joined #openstack-nova | 21:38 | |
| *** RuiChen has quit IRC | 21:39 | |
| *** annegentle has quit IRC | 21:39 | |
| *** kiwik-chenrui has quit IRC | 21:40 | |
| *** annegentle has joined #openstack-nova | 21:40 | |
| *** zxen has quit IRC | 21:40 | |
| *** kiwik-chenrui has joined #openstack-nova | 21:40 | |
| figleaf | superdan: only about the 10th time I've reviewed that code :) | 21:41 |
| openstackgerrit | Joe Gordon proposed openstack/nova: Plan to move to one network stack https://review.openstack.org/196258 | 21:42 |
| superdan | leakypipes: it seems kinda silly to bury the compute_node.save() in the scheduler client like that | 21:43 |
| * flashgordon braces for impact | 21:43 | |
| superdan | leakypipes: is this because we're expecting we'll change the scheduler client to do something other than a save() there in the future? | 21:43 |
| *** zxen has joined #openstack-nova | 21:43 | |
| *** tjones2 has left #openstack-nova | 21:43 | |
| *** _diana_ has quit IRC | 21:44 | |
| superdan | flashgordon: I applaud your pragmatism and leadership good sir | 21:44 |
| *** artom has joined #openstack-nova | 21:45 | |
| flashgordon | superdan: thank you :), but I just thought of a better commit message | 21:45 |
| tjones1 | flashgordon: nice working with you. Thanks for the bug help. Hope you enjoy your new gig!! If not - ping me. we are hiring :-) | 21:45 |
| openstackgerrit | Joe Gordon proposed openstack/nova: One network stack to rule them all https://review.openstack.org/196258 | 21:45 |
| *** _diana_ has joined #openstack-nova | 21:46 | |
| flashgordon | tjones1: thanks, its been great working with you too! | 21:46 |
| flashgordon | superdan: ^^ | 21:46 |
| superdan | flashgordon: heh | 21:46 |
| figleaf | superdan: why is that 'buried'? | 21:47 |
| *** wxy has quit IRC | 21:47 | |
| *** isd1 has quit IRC | 21:47 | |
| superdan | figleaf: because it calls through two layers to do just a compute_node.save() ? | 21:47 |
| *** wxy has joined #openstack-nova | 21:47 | |
| figleaf | superdan: but it's an RPC call, and the caller shouldn't know what's involved. It does the same thing as it did before | 21:48 |
| figleaf | superdan: just a lot cleaner | 21:48 |
| superdan | figleaf: where is an rpc call? | 21:49 |
| openstackgerrit | Giridhar Jayavelu proposed openstack/nova: Derive resource quota from Flavor's extra_spec when attaching interface to an instance https://review.openstack.org/196171 | 21:49 |
| figleaf | superdan: the update_resource_stats | 21:50 |
| superdan | figleaf: not that I see | 21:50 |
| figleaf | superdan: this is preserving the current flow | 21:50 |
| superdan | figleaf: you think that's sending the whole compute node back to scheduler rpc just to save it to the DB? | 21:50 |
| *** zxen has quit IRC | 21:51 | |
| *** kiwik-chenrui has quit IRC | 21:51 | |
| *** zxen has joined #openstack-nova | 21:51 | |
| *** kiwik-chenrui has joined #openstack-nova | 21:51 | |
| *** nihilifer has quit IRC | 21:51 | |
| superdan | figleaf: the query bit of the client is an rpc call | 21:52 |
| superdan | figleaf: the report bit is not, as far as I can tell | 21:52 |
| superdan | figleaf: and the long-standing confusion around this code is that the "client" part of that is just writing to the db from the compute node (through conductor of course) | 21:53 |
| figleaf | superdan: hmmm... I thought they both were | 21:53 |
| *** wxy has quit IRC | 21:53 | |
| superdan | this doesn't do anything in that layer anymore, | 21:53 |
| superdan | which makes it confusing for no reaosn | 21:53 |
| figleaf | superdan: the confusion comes from tracking the same information is a bunch of different places | 21:53 |
| superdan | so unless this is critical for setting something up... | 21:53 |
| superdan | figleaf: only ten times through the patch huh? :) | 21:54 |
| figleaf | superdan: since March? at least | 21:54 |
| *** Marga_ has quit IRC | 21:55 | |
| figleaf | superdan: I get that this isn't ideal, but it is a huge clean-up from how it is now | 21:55 |
| figleaf | superdan: I'd love to ditch the current approach to keeping the scheduler in sync with compute | 21:56 |
| figleaf | superdan: this doesn't fix it, it just makes it suck a little less :) | 21:56 |
| *** rwsu has quit IRC | 21:56 | |
| *** kiwik-chenrui has quit IRC | 21:56 | |
| *** kiwik-chenrui has joined #openstack-nova | 21:57 | |
| *** wxy has joined #openstack-nova | 21:57 | |
| melwitt | figleaf: why is the self._write_ext_resources not needed anymore during the init? that is, what takes care of doing that in the new way? | 21:58 |
| *** baoli has joined #openstack-nova | 21:59 | |
| superdan | yeah, haven't found that either | 21:59 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/194325 | 21:59 |
| *** zxen has quit IRC | 21:59 | |
| *** ijw has quit IRC | 21:59 | |
| figleaf | melwitt: which init are you looking at? | 22:01 |
| *** zxen has joined #openstack-nova | 22:01 | |
| melwitt | figleaf: _init_compute_node | 22:01 |
| *** hemna_ has joined #openstack-nova | 22:01 | |
| superdan | the method is modified even, | 22:02 |
| superdan | but the call to it is gone, AFAICT | 22:02 |
| *** hemna has quit IRC | 22:02 | |
| *** patrickeast has quit IRC | 22:03 | |
| *** kiwik-chenrui has quit IRC | 22:03 | |
| melwitt | yeah, that's what I find too. I followed into self.compute_node.update_from_virt_driver(resources) to see if that does something like it but it doesn't seem to | 22:04 |
| *** wxy has quit IRC | 22:04 | |
| superdan | yeah | 22:04 |
| *** patrickeast has joined #openstack-nova | 22:04 | |
| superdan | and the fact that it's gone and passing tests is kinda disturbing too | 22:04 |
| *** RuiChen has joined #openstack-nova | 22:04 | |
| *** sharkJing has joined #openstack-nova | 22:04 | |
| bauwser | so, there was a discussion and I wasn't seeing it /o\ | 22:05 |
| bauwser | because I was reviewing leakypipes's change | 22:05 |
| bauwser | ... | 22:05 |
| bauwser | so, yeah, I think we can leave _write_ext_resources() but drop it in a later patch | 22:05 |
| *** kiwik-chenrui has joined #openstack-nova | 22:06 | |
| *** wxy has joined #openstack-nova | 22:06 | |
| melwitt | bauwser: so it's no longer needed for some reason? I just wasn't sure if it was a mistake | 22:06 |
| *** wxy has quit IRC | 22:07 | |
| superdan | if it's no longer needed then the method should be removed from RT | 22:07 |
| bauwser | melwitt: it's not really needed anymore because it was for ERT | 22:07 |
| *** kiwik-chenrui has quit IRC | 22:07 | |
| *** wxy has joined #openstack-nova | 22:07 | |
| superdan | but I don't understand why it's no longer needed | 22:07 |
| bauwser | Extensible RT | 22:07 |
| mriedem | if it's needed and busted but not failing tests, add tests | 22:07 |
| superdan | we spent so much time getting that crap in there | 22:07 |
| superdan | I know it's not really useful, but if it's changing what is being reported, | 22:07 |
| superdan | and people have scheduler filters written to read it... | 22:08 |
| *** kiwik-chenrui has joined #openstack-nova | 22:08 | |
| *** edmondsw has quit IRC | 22:08 | |
| bauwser | superdan: the ERT doesn't have the scheduler method for getting them | 22:08 |
| *** edmondsw has joined #openstack-nova | 22:08 | |
| bauwser | superdan: but that's out of the current change | 22:08 |
| melwitt | bauwser: oh, okay. I was reviewing for the conversion to ComputeNode, so when I see things disappearing with no replacement they look like a bug | 22:08 |
| superdan | I don't know what that means | 22:08 |
| *** edmondsw has quit IRC | 22:08 | |
| *** kiwik-chenrui has quit IRC | 22:09 | |
| superdan | melwitt: it is a bug | 22:09 |
| bauwser | okay, little story behind that, it was for providing extensible resources to the scheduler | 22:09 |
| *** RuiChen has quit IRC | 22:09 | |
| *** baoli has quit IRC | 22:09 | |
| *** wxy has quit IRC | 22:09 | |
| *** zxen has quit IRC | 22:10 | |
| *** wxy has joined #openstack-nova | 22:10 | |
| bauwser | so paul added it in the RT, but the scheduler change was removedf | 22:10 |
| *** RuiChen has joined #openstack-nova | 22:10 | |
| superdan | melwitt: new line 488 makes a change to that method, but the method is never called | 22:10 |
| *** sharkJing has quit IRC | 22:10 | |
| *** dave-mcc_ has joined #openstack-nova | 22:10 | |
| superdan | bauwser: well, this patch commit message mentions none of that, and modifies the method in the process | 22:10 |
| superdan | bauwser: so if we're removing this, it should be in a separate patch | 22:10 |
| *** sharkJing has joined #openstack-nova | 22:11 | |
| melwitt | superdan: ah, yeah. if there's no call left at all then that's inconsistent | 22:11 |
| *** zxen has joined #openstack-nova | 22:11 | |
| bauwser | superdan: yup, that's what I said :) | 22:11 |
| bauwser | melwitt: there is one left call, L583 | 22:11 |
| bauwser | in update() | 22:11 |
| bauwser | which is fine | 22:11 |
| bauwser | sorry if I'm unclear, 0:11am here | 22:12 |
| superdan | I see it | 22:12 |
| *** dave-mccowan has quit IRC | 22:12 | |
| superdan | we should remove that before this patch if that's what we're going to do | 22:12 |
| superdan | gerrit search fails for me sometimes | 22:13 |
| melwitt | bauwser: I see. maybe I saw it before too, not sure. but I guess if we take one call that supports ERT out and leave another, I don't understand that either | 22:13 |
| *** RuiChen has quit IRC | 22:13 | |
| *** RuiChen has joined #openstack-nova | 22:14 | |
| *** ijuwang has quit IRC | 22:14 | |
| *** wxy has quit IRC | 22:14 | |
| figleaf | melwitt: good point. At least it should have an explanation | 22:14 |
| melwitt | anyway, my point was it takes a lot longer to review a patch when it's doing extra things that aren't explained, they look like bugs | 22:14 |
| *** exploreshaifali has quit IRC | 22:15 | |
| bauwser | melwitt: so my take on that is to say 'okay, whatever we like ERT or not, it's not by the scope of the change" | 22:15 |
| *** sharkJing has quit IRC | 22:16 | |
| *** earlephilhower has quit IRC | 22:16 | |
| bauwser | melwitt: so, we can leave it since it's still called, and provide another change for removing it | 22:16 |
| *** Marga_ has joined #openstack-nova | 22:16 | |
| superdan | I think it would be much better to put something ahead of this to just remove it | 22:16 |
| bauwser | melwitt: I think Paul would appreciate some discussion on keeping ERT or not, while he wouldn't want to pollute the objectify change | 22:16 |
| superdan | now wait a sec | 22:16 |
| *** sharkJing has joined #openstack-nova | 22:17 | |
| superdan | you just said it's completely unused(able) right? | 22:17 |
| melwitt | I have to step out for an appointment right now, bbl | 22:17 |
| *** fawadkhaliq has joined #openstack-nova | 22:17 | |
| bauwser | superdan: AFAICT, yes | 22:17 |
| bauwser | superdan: I remember the story, the change has landed on the scheduler side, but has been reverted | 22:18 |
| *** Nikolay_St has joined #openstack-nova | 22:18 | |
| superdan | okay, but it doesn't matter | 22:18 |
| superdan | either this patch doesn't remove this call, or it's gone before this patch, right? | 22:18 |
| *** SourabhP has joined #openstack-nova | 22:19 | |
| bauwser | superdan: so, the objectify change doesn't impact the call yup | 22:19 |
| *** wxy has joined #openstack-nova | 22:19 | |
| bauwser | superdan: I mean, the RT.update() call still calls the ERT handler | 22:19 |
| *** fawadk has joined #openstack-nova | 22:19 | |
| superdan | but it's not done from init anymore | 22:19 |
| *** pixelb has quit IRC | 22:19 | |
| bauwser | superdan: what's changed by the objectify patch is just a removed call in the _init() method | 22:20 |
| bauwser | superdan: correct | 22:20 |
| bauwser | superdan: I see your point | 22:20 |
| *** fawadkhaliq has quit IRC | 22:20 | |
| bauwser | superdan: so, yeah, removing that call shouldn't be part of the change, since it's not impacted | 22:21 |
| *** Marga_ has quit IRC | 22:21 | |
| superdan | what's up with the removal of compute_node_create? | 22:21 |
| *** sharkJing has quit IRC | 22:21 | |
| *** RuiChen has quit IRC | 22:21 | |
| flashgordon | mriedem melwitt: tests are in, want to review https://review.openstack.org/#/c/194840/ | 22:21 |
| bauwser | superdan: ah that | 22:21 |
| *** annashen has quit IRC | 22:22 | |
| *** RuiChen has joined #openstack-nova | 22:22 | |
| *** sharkJing has joined #openstack-nova | 22:22 | |
| bauwser | superdan: if you see the code, it just calls the object .create() instead | 22:22 |
| *** lucas-afk has quit IRC | 22:22 | |
| *** annashen has joined #openstack-nova | 22:22 | |
| superdan | oh right, I was looking in the wrong place for it | 22:22 |
| *** zxen has quit IRC | 22:23 | |
| mriedem | flashgordon: +2 | 22:23 |
| *** tjones2 has joined #openstack-nova | 22:23 | |
| superdan | I want to hear what leakypipes has to say about this ERT business | 22:23 |
| bauwser | superdan: but that's fair to say that removing the call to write_ext_res() in init() makes the create() missing the ERT stuff | 22:23 |
| *** ctrath has quit IRC | 22:23 | |
| *** wxy has quit IRC | 22:24 | |
| *** zxen has joined #openstack-nova | 22:24 | |
| *** kiwik-chenrui has joined #openstack-nova | 22:26 | |
| *** zxen has quit IRC | 22:27 | |
| *** kiwik-chenrui has quit IRC | 22:27 | |
| *** zxen has joined #openstack-nova | 22:27 | |
| *** kiwik-chenrui has joined #openstack-nova | 22:27 | |
| flashgordon | not my patch, but an awesome one https://review.openstack.org/#/c/175569 | 22:27 |
| flashgordon | removes 2k lines | 22:27 |
| flashgordon | and lets nova blame cinder for more problems | 22:28 |
| *** RuiChen has quit IRC | 22:29 | |
| *** Marga_ has joined #openstack-nova | 22:30 | |
| *** tonytan4ever has quit IRC | 22:30 | |
| *** kiwik-chenrui has quit IRC | 22:30 | |
| *** tpeoples has quit IRC | 22:31 | |
| *** wxy has joined #openstack-nova | 22:31 | |
| *** kiwik-chenrui has joined #openstack-nova | 22:31 | |
| *** kiwik-chenrui has quit IRC | 22:32 | |
| *** sharkJing has quit IRC | 22:32 | |
| *** kiwik-chenrui has joined #openstack-nova | 22:33 | |
| mriedem | flashgordon: yeah it's in the review queue | 22:34 |
| *** kiwik-chenrui has quit IRC | 22:34 | |
| *** wxy has quit IRC | 22:34 | |
| leakypipes | superdan: yes, the ert stuff will go away in later patches... | 22:34 |
| mriedem | flashgordon: i've been in this libvirt volume driver code a lot lately so i need to review that one | 22:34 |
| leakypipes | superdan: and yes, the scheduler client will eventually be more than compute_node.save() :) | 22:34 |
| *** wxy has joined #openstack-nova | 22:34 | |
| superdan | leakypipes: so we've been sending the ERT stuff over the wire for a couple cycles for no reson? | 22:34 |
| superdan | er, reason | 22:34 |
| bauwser | superdan: exactly :) | 22:35 |
| *** zxen has quit IRC | 22:35 | |
| leakypipes | superdan: yup :( | 22:35 |
| superdan | leakypipes: that's really unfortunate | 22:35 |
| *** kiwik-chenrui has joined #openstack-nova | 22:35 | |
| superdan | leakypipes: so why don't we just yank that stuff in a precursor patch so that this doesn't git-blame to touch any of it? | 22:35 |
| *** zxen has joined #openstack-nova | 22:35 | |
| flashgordon | mriedem: cool, this should clean that code up a lot | 22:35 |
| bauwser | superdan: the story about providing a separate scheduler method just for calling save() is just for having a clear interface | 22:35 |
| bauwser | superdan: it was actually done when we were talking about splitting the sched y'know< | 22:36 |
| superdan | bauwser: yeah, it's just hiding what is going on right now.. is that interface still there for actual reasons? | 22:36 |
| bauwser | superdan: honestly, no real compelling reasons but a good interface | 22:37 |
| *** zxen has quit IRC | 22:37 | |
| *** Nikolay_St has quit IRC | 22:37 | |
| mriedem | future requirements are the best requirements | 22:37 |
| *** zxen has joined #openstack-nova | 22:37 | |
| superdan | it doesn't seem like it's a good interface, since this patch modifies it | 22:38 |
| superdan | and passing the compute_node object doesn't really seem like the right level of detail for an independent scheduler client | 22:38 |
| superdan | but... | 22:39 |
| bauwser | superdan: now that we have a clear object usage, maybe that can be discussed agreed | 22:39 |
| *** wxy has quit IRC | 22:39 | |
| *** annashen has quit IRC | 22:39 | |
| *** wxy has joined #openstack-nova | 22:39 | |
| *** fawadk has quit IRC | 22:40 | |
| superdan | leakypipes: anyway, it would be nice to keep this thing from touching ERT in any way, ideally by removing it first if it's really not needed (should be easy right?) | 22:41 |
| *** annegentle has quit IRC | 22:41 | |
| *** kiwik has joined #openstack-nova | 22:41 | |
| superdan | leakypipes: if not, maybe a comment in the commit message about why it seems to drop some calls because we don't care about it anymore | 22:41 |
| bauwser | okay, France lost at the quarters of the Soccer WC, can say bye to you folks ! | 22:42 |
| superdan | lost to who? | 22:42 |
| bauwser | Germany | 22:42 |
| *** zxen has quit IRC | 22:42 | |
| * superdan reserves a joke | 22:42 | |
| bauwser | :D | 22:42 |
| * figleaf was thinking of the same joke | 22:42 | |
| *** wxy has quit IRC | 22:43 | |
| bauwser | superdan: figleaf: to be clear, France was playing at 11 and Germany at 12 - including the referree | 22:44 |
| *** claudiub has joined #openstack-nova | 22:44 | |
| *** kiwik has quit IRC | 22:44 | |
| *** kiwik-chenrui has quit IRC | 22:44 | |
| *** wxy has joined #openstack-nova | 22:44 | |
| *** kiwik has joined #openstack-nova | 22:44 | |
| *** armax has joined #openstack-nova | 22:44 | |
| * bauwser is a French, y'know so arguing is like a 2nd nature | 22:45 | |
| *** packet has joined #openstack-nova | 22:47 | |
| *** armax has quit IRC | 22:47 | |
| mriedem | the french team obviously needs a tom brady and bill bellichick | 22:48 |
| *** wxy has quit IRC | 22:48 | |
| mriedem | https://www.youtube.com/watch?v=oDpzJk3X020 | 22:48 |
| *** dboik has quit IRC | 22:48 | |
| *** RuiChen has joined #openstack-nova | 22:49 | |
| *** wxy has joined #openstack-nova | 22:49 | |
| *** dboik has joined #openstack-nova | 22:49 | |
| *** tjones2 has quit IRC | 22:50 | |
| *** RuiChen has quit IRC | 22:50 | |
| *** wxy has quit IRC | 22:50 | |
| *** RuiChen has joined #openstack-nova | 22:50 | |
| *** wxy has joined #openstack-nova | 22:50 | |
| *** annashen has joined #openstack-nova | 22:50 | |
| *** kiwik has quit IRC | 22:52 | |
| *** zxen has joined #openstack-nova | 22:52 | |
| *** zxen has quit IRC | 22:53 | |
| *** RuiChen has quit IRC | 22:53 | |
| *** zxen has joined #openstack-nova | 22:53 | |
| *** RuiChen has joined #openstack-nova | 22:53 | |
| *** tjones2 has joined #openstack-nova | 22:53 | |
| *** zxen has quit IRC | 22:54 | |
| *** _diana_ has quit IRC | 22:54 | |
| *** ctrath has joined #openstack-nova | 22:54 | |
| *** zxen has joined #openstack-nova | 22:54 | |
| *** otter768 has joined #openstack-nova | 22:55 | |
| *** wxy has quit IRC | 22:55 | |
| *** kiwik-chenrui has joined #openstack-nova | 22:55 | |
| *** tjones2 has quit IRC | 22:56 | |
| *** _diana_ has joined #openstack-nova | 22:56 | |
| *** zxen has quit IRC | 22:57 | |
| *** bauwser is now known as bauzas | 22:57 | |
| *** RuiChen has quit IRC | 22:58 | |
| *** zxen has joined #openstack-nova | 22:59 | |
| *** wxy has joined #openstack-nova | 23:00 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move the LibvirtQuobyteVolumeDriver into the quobyte module https://review.openstack.org/193236 | 23:00 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move volume/remotefs/quobyte modules under volume subdir https://review.openstack.org/193223 | 23:00 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move the LibvirtScalityVolumeDriver into it's own module https://review.openstack.org/193249 | 23:00 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtGlusterfsVolumeDriver into it's own module https://review.openstack.org/193380 | 23:00 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtFibreChannelVolumeDriver into it's own module https://review.openstack.org/193278 | 23:00 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move lvm/dmcrypt/rbd_utils modules under storage subdir https://review.openstack.org/193195 | 23:00 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move fibre channel specific utilities into fc module https://review.openstack.org/193317 | 23:00 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move the LibvirtGPFSVolumeDriver into it's own module https://review.openstack.org/193240 | 23:00 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: move LibvirtAOEVolumeDriver into it's own module https://review.openstack.org/196276 | 23:00 |
| *** romainh has left #openstack-nova | 23:01 | |
| flashgordon | pow | 23:01 |
| mriedem | yeah, eat that | 23:01 |
| flashgordon | why not do the brick thing first? | 23:01 |
| flashgordon | you will merge conflict with it I think | 23:01 |
| mriedem | flashgordon: it doesn't cover all of it | 23:01 |
| *** annegentle has joined #openstack-nova | 23:01 | |
| mriedem | i know i will on the fc stuff | 23:01 |
| mriedem | so i think i'm going to drop https://review.openstack.org/193317 | 23:02 |
| mriedem | which is why i have a -2 on it, just haven't figured out if i want to drop it yet | 23:02 |
| mriedem | os-brick doesn't completely remove everything | 23:02 |
| flashgordon | right, it leaves a little bit | 23:02 |
| mriedem | and it doesn't touch any of the file system type drivers | 23:02 |
| mriedem | of wihch we have 4 | 23:02 |
| mriedem | that are about 80% the same code | 23:02 |
| flashgordon | :/ | 23:03 |
| mriedem | so at the end of that refactor series i'm going to pull up the common stuff into a fs base class | 23:03 |
| *** zxen has quit IRC | 23:03 | |
| *** kiwik has joined #openstack-nova | 23:04 | |
| flashgordon | man our 3rd party CI stinks | 23:06 |
| *** kiwik has quit IRC | 23:07 | |
| flashgordon | VMware fails more often then it works | 23:07 |
| *** kiwik-chenrui has quit IRC | 23:07 | |
| *** zxen has joined #openstack-nova | 23:07 | |
| *** kiwik has joined #openstack-nova | 23:07 | |
| flashgordon | same for hyperv | 23:07 |
| flashgordon | IBM is actually really good | 23:07 |
| flashgordon | as is xenproject | 23:08 |
| *** wxy has quit IRC | 23:09 | |
| *** zxen has quit IRC | 23:10 | |
| *** zxen has joined #openstack-nova | 23:10 | |
| *** kiwik has quit IRC | 23:11 | |
| *** kiwik has joined #openstack-nova | 23:11 | |
| *** eglynn has quit IRC | 23:12 | |
| *** zxen has quit IRC | 23:12 | |
| *** _diana_ has quit IRC | 23:12 | |
| *** zxen has joined #openstack-nova | 23:12 | |
| *** annashen has quit IRC | 23:14 | |
| *** zzzeek has quit IRC | 23:15 | |
| *** wxy has joined #openstack-nova | 23:16 | |
| *** wxy has quit IRC | 23:17 | |
| *** wxy has joined #openstack-nova | 23:17 | |
| *** zxen has quit IRC | 23:17 | |
| *** Marga_ has quit IRC | 23:17 | |
| *** ekarlso has quit IRC | 23:18 | |
| *** ekarlso has joined #openstack-nova | 23:18 | |
| *** annegentle has quit IRC | 23:19 | |
| *** _diana_ has joined #openstack-nova | 23:19 | |
| *** sdake_ has joined #openstack-nova | 23:21 | |
| *** kiwik has quit IRC | 23:22 | |
| *** zxen has joined #openstack-nova | 23:22 | |
| *** Longgeek has joined #openstack-nova | 23:22 | |
| *** kiwik has joined #openstack-nova | 23:22 | |
| *** zxen has quit IRC | 23:23 | |
| *** sdake has quit IRC | 23:25 | |
| *** zxen has joined #openstack-nova | 23:25 | |
| *** wxy has quit IRC | 23:26 | |
| *** annashen has joined #openstack-nova | 23:26 | |
| *** kiwik has quit IRC | 23:26 | |
| *** kiwik has joined #openstack-nova | 23:26 | |
| *** wxy has joined #openstack-nova | 23:29 | |
| *** RuiChen has joined #openstack-nova | 23:30 | |
| *** annashen has quit IRC | 23:31 | |
| *** kiwik has quit IRC | 23:32 | |
| *** wxy has quit IRC | 23:32 | |
| *** zxen has quit IRC | 23:32 | |
| *** wxy has joined #openstack-nova | 23:32 | |
| mriedem | xen has some intermittent failures | 23:33 |
| mriedem | flashgordon: where are you looking for stats? | 23:33 |
| *** RuiChen has quit IRC | 23:34 | |
| *** RuiChen has joined #openstack-nova | 23:35 | |
| *** ctrath has quit IRC | 23:36 | |
| *** zxen has joined #openstack-nova | 23:37 | |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: discover supported versions to choose latest one https://review.openstack.org/184281 | 23:38 |
| *** ZZelle has quit IRC | 23:38 | |
| *** RuiChen has quit IRC | 23:39 | |
| *** RuiChen has joined #openstack-nova | 23:39 | |
| *** wxy has quit IRC | 23:40 | |
| *** wxy has joined #openstack-nova | 23:40 | |
| *** sdake_ has quit IRC | 23:42 | |
| *** zxen has quit IRC | 23:42 | |
| *** wxy has quit IRC | 23:44 | |
| *** RuiChen has quit IRC | 23:44 | |
| *** wxy has joined #openstack-nova | 23:44 | |
| *** RuiChen has joined #openstack-nova | 23:44 | |
| *** Longgeek has quit IRC | 23:44 | |
| *** wxy has quit IRC | 23:45 | |
| *** kiwik-chenrui has joined #openstack-nova | 23:47 | |
| *** wxy has joined #openstack-nova | 23:48 | |
| flashgordon | mriedem: woops forgot to post the data | 23:49 |
| *** kiwik-chenrui has quit IRC | 23:49 | |
| flashgordon | running last comment | 23:49 |
| *** kiwik-chenrui has joined #openstack-nova | 23:50 | |
| *** RuiChen has quit IRC | 23:50 | |
| flashgordon | mriedem: http://paste.openstack.org/show/321474 | 23:51 |
| *** kiwik-chenrui has quit IRC | 23:51 | |
| *** kiwik-chenrui has joined #openstack-nova | 23:52 | |
| flashgordon | mriedem: documented here https://github.com/jogo/lastcomment | 23:53 |
| * flashgordon looks at the clock and runs away | 23:54 | |
| *** RuiChen has joined #openstack-nova | 23:54 | |
| *** packet has quit IRC | 23:55 | |
| *** kiwik-chenrui has quit IRC | 23:56 | |
| *** zxen has joined #openstack-nova | 23:58 | |
| *** kiwik-chenrui has joined #openstack-nova | 23:58 | |
| *** raildo has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
