| mriedem | belliott: did i hear correctly that rax doesn't use/enable the agent-builds REST API? | 00:01 |
|---|---|---|
| mriedem | it came up during the cells v2 session last week | 00:01 |
| *** shaohe_feng has quit IRC | 00:01 | |
| *** shaohe_feng has joined #openstack-nova | 00:02 | |
| *** doug-fish has quit IRC | 00:03 | |
| belliott | mriedem: checking | 00:03 |
| *** ssurana has joined #openstack-nova | 00:03 | |
| mriedem | doffm: were we moving keypairs to the api db? | 00:04 |
| belliott | mriedem: we don't have the extension enabled, so no | 00:04 |
| mriedem | belliott: ok, thanks | 00:04 |
| mriedem | doffm: nvm, looks like we plan on migrating keypairs to the api db https://review.openstack.org/#/c/294209/ | 00:04 |
| *** salv-orlando has joined #openstack-nova | 00:07 | |
| *** markvoelker has joined #openstack-nova | 00:09 | |
| *** heyongli_ has quit IRC | 00:09 | |
| *** heyongli_ has joined #openstack-nova | 00:10 | |
| *** vilobhmm11 has joined #openstack-nova | 00:10 | |
| *** Sukhdev has quit IRC | 00:11 | |
| *** ijw has joined #openstack-nova | 00:11 | |
| *** shaohe_feng has quit IRC | 00:12 | |
| *** shaohe_feng has joined #openstack-nova | 00:12 | |
| *** salv-orlando has quit IRC | 00:14 | |
| *** salv-orlando has joined #openstack-nova | 00:14 | |
| *** baoli has joined #openstack-nova | 00:14 | |
| *** salv-orlando has quit IRC | 00:15 | |
| *** ijw_ has joined #openstack-nova | 00:15 | |
| *** ijw has quit IRC | 00:15 | |
| *** doug-fish has joined #openstack-nova | 00:16 | |
| *** salv-orlando has joined #openstack-nova | 00:16 | |
| *** thumpba has joined #openstack-nova | 00:16 | |
| *** baoli has quit IRC | 00:16 | |
| *** heyongli_ has quit IRC | 00:20 | |
| *** heyongli_ has joined #openstack-nova | 00:20 | |
| *** thumpba has quit IRC | 00:20 | |
| *** zenoway has joined #openstack-nova | 00:21 | |
| *** shaohe_feng has quit IRC | 00:22 | |
| *** salv-orlando has quit IRC | 00:22 | |
| *** shaohe_feng has joined #openstack-nova | 00:23 | |
| *** salv-orlando has joined #openstack-nova | 00:25 | |
| *** zenoway has quit IRC | 00:25 | |
| *** mylu has quit IRC | 00:28 | |
| *** heyongli_ has quit IRC | 00:30 | |
| *** mrkz has quit IRC | 00:30 | |
| *** Sukhdev has joined #openstack-nova | 00:30 | |
| *** heyongli_ has joined #openstack-nova | 00:30 | |
| *** shaohe_feng has quit IRC | 00:32 | |
| *** shaohe_feng has joined #openstack-nova | 00:33 | |
| *** tjones has left #openstack-nova | 00:37 | |
| *** rbridgeman has quit IRC | 00:38 | |
| *** diogogmt has quit IRC | 00:40 | |
| *** heyongli_ has quit IRC | 00:40 | |
| *** heyongli_ has joined #openstack-nova | 00:41 | |
| *** roeyc has quit IRC | 00:41 | |
| *** diogogmt has joined #openstack-nova | 00:42 | |
| *** shaohe_feng has quit IRC | 00:42 | |
| *** roeyc has joined #openstack-nova | 00:43 | |
| *** shaohe_feng has joined #openstack-nova | 00:45 | |
| *** fawadkhaliq has quit IRC | 00:50 | |
| *** fragatina has quit IRC | 00:50 | |
| *** heyongli_ has quit IRC | 00:50 | |
| *** heyongli_ has joined #openstack-nova | 00:51 | |
| *** shaohe_feng has quit IRC | 00:53 | |
| *** shaohe_feng has joined #openstack-nova | 00:55 | |
| *** cheneydc has joined #openstack-nova | 00:56 | |
| *** smatzek has joined #openstack-nova | 00:56 | |
| *** fishcried has joined #openstack-nova | 00:57 | |
| *** ssurana has quit IRC | 01:00 | |
| *** heyongli_ has quit IRC | 01:01 | |
| *** iceyao has joined #openstack-nova | 01:01 | |
| *** shaohe_feng has quit IRC | 01:03 | |
| *** heyongli_ has joined #openstack-nova | 01:03 | |
| *** shaohe_feng has joined #openstack-nova | 01:03 | |
| *** salv-orlando has quit IRC | 01:04 | |
| *** fishcried1 has joined #openstack-nova | 01:07 | |
| *** baoli has joined #openstack-nova | 01:07 | |
| *** fishcried has quit IRC | 01:08 | |
| *** fishcried1 is now known as fishcried | 01:08 | |
| *** fishcried has quit IRC | 01:08 | |
| *** thorst has quit IRC | 01:08 | |
| *** thorst has joined #openstack-nova | 01:09 | |
| *** david-lyle has joined #openstack-nova | 01:09 | |
| *** heyongli_ has quit IRC | 01:10 | |
| *** fragatina has joined #openstack-nova | 01:12 | |
| *** fragatin_ has joined #openstack-nova | 01:13 | |
| *** shaohe_feng has quit IRC | 01:13 | |
| *** shaohe_feng has joined #openstack-nova | 01:14 | |
| *** ducttape_ has joined #openstack-nova | 01:15 | |
| *** yamahata has quit IRC | 01:15 | |
| *** fragatina has quit IRC | 01:16 | |
| *** nic has quit IRC | 01:17 | |
| *** thorst has quit IRC | 01:17 | |
| *** svenkat has quit IRC | 01:19 | |
| *** EinstCrazy has joined #openstack-nova | 01:22 | |
| mriedem | melwitt: did you have an openstack spec or oslo spec up for review for the transport_url stuff in oslo.messaging? | 01:22 |
| melwitt | mriedem: no, I had assumed a spec wouldn't be needed. shall I write one? | 01:23 |
| *** zz_dimtruck is now known as dimtruck | 01:23 | |
| *** shaohe_feng has quit IRC | 01:23 | |
| mriedem | idk, it's a feature in oslo.messaging right? or needs to be. | 01:24 |
| *** shaohe_feng has joined #openstack-nova | 01:24 | |
| mriedem | i might be confusing this with a spec for policy defaults in code for oslo.policy | 01:24 |
| *** fragatin_ has quit IRC | 01:24 | |
| melwitt | mriedem: yeah it would be the addition of a function that can assemble a transport url from the various config options | 01:25 |
| mriedem | it was listed under the upgrades section in the summit session etherpad https://etherpad.openstack.org/p/newton-nova-cells but i'm not clear on how it's related to upgrades, | 01:26 |
| mriedem | but i think it's because today we have to pass the --transport_url in to the migration command | 01:27 |
| melwitt | like the scheme, user/pass, host/port, virtual host | 01:27 |
| mriedem | and we'd like to be able to automate that | 01:27 |
| mriedem | gdi mortimer will not stop wheezing | 01:27 |
| melwitt | yeah, it would be a lot better to gather it from config rather than make it be passed on the command line | 01:27 |
| *** nelsnelson has joined #openstack-nova | 01:27 | |
| melwitt | lol mortimer | 01:28 |
| *** Jeffrey4l has joined #openstack-nova | 01:28 | |
| *** RuiChen has joined #openstack-nova | 01:29 | |
| *** whenry has joined #openstack-nova | 01:31 | |
| *** fragatina has joined #openstack-nova | 01:31 | |
| *** EinstCra_ has joined #openstack-nova | 01:31 | |
| *** shaohe_feng has quit IRC | 01:34 | |
| *** shaohe_feng has joined #openstack-nova | 01:34 | |
| *** EinstCrazy has quit IRC | 01:35 | |
| *** baoli has quit IRC | 01:37 | |
| *** roeyc has quit IRC | 01:39 | |
| *** roeyc has joined #openstack-nova | 01:40 | |
| *** edmondsw has quit IRC | 01:42 | |
| *** shaohe_feng has quit IRC | 01:44 | |
| *** shaohe_feng has joined #openstack-nova | 01:44 | |
| *** baoli has joined #openstack-nova | 01:45 | |
| *** shaohe_feng has quit IRC | 01:54 | |
| *** ijw_ has quit IRC | 01:54 | |
| *** shaohe_feng has joined #openstack-nova | 01:55 | |
| *** whenry has quit IRC | 01:55 | |
| *** mriedem has quit IRC | 01:57 | |
| *** doug-fish has quit IRC | 01:57 | |
| *** doug-fish has joined #openstack-nova | 01:57 | |
| *** tbachman has quit IRC | 01:57 | |
| *** doug-fis_ has joined #openstack-nova | 02:00 | |
| *** doug-fish has quit IRC | 02:01 | |
| *** fishcried has joined #openstack-nova | 02:03 | |
| *** tbachman has joined #openstack-nova | 02:03 | |
| *** shaohe_feng has quit IRC | 02:04 | |
| *** doug-fis_ has quit IRC | 02:05 | |
| *** shaohe_feng has joined #openstack-nova | 02:05 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova-specs: Re-Propose Expose quiesce/unquiesce API https://review.openstack.org/295595 | 02:06 |
| *** mragupat has joined #openstack-nova | 02:07 | |
| *** mragupat has quit IRC | 02:07 | |
| *** doug-fish has joined #openstack-nova | 02:07 | |
| *** mragupat has joined #openstack-nova | 02:08 | |
| *** mragupat has quit IRC | 02:08 | |
| *** apoorvad has quit IRC | 02:12 | |
| *** ducttape_ has quit IRC | 02:13 | |
| *** zeroDivisible has quit IRC | 02:14 | |
| *** shaohe_feng has quit IRC | 02:15 | |
| *** thorst has joined #openstack-nova | 02:15 | |
| *** shaohe_feng has joined #openstack-nova | 02:15 | |
| *** Sukhdev has quit IRC | 02:18 | |
| *** zenoway has joined #openstack-nova | 02:19 | |
| *** bogdano has quit IRC | 02:20 | |
| *** diogogmt has quit IRC | 02:20 | |
| *** sdake has joined #openstack-nova | 02:20 | |
| *** bogdano has joined #openstack-nova | 02:21 | |
| *** thorst has quit IRC | 02:22 | |
| *** rmart04 has joined #openstack-nova | 02:22 | |
| *** zenoway has quit IRC | 02:24 | |
| *** smatzek has quit IRC | 02:25 | |
| *** shaohe_feng has quit IRC | 02:25 | |
| *** shaohe_feng has joined #openstack-nova | 02:25 | |
| *** rmart04 has quit IRC | 02:25 | |
| *** aarefiev has quit IRC | 02:26 | |
| *** aarefiev has joined #openstack-nova | 02:27 | |
| *** fishcried has quit IRC | 02:29 | |
| *** unicell has quit IRC | 02:30 | |
| *** ducttape_ has joined #openstack-nova | 02:34 | |
| *** shaohe_feng has quit IRC | 02:35 | |
| *** shaohe_feng has joined #openstack-nova | 02:36 | |
| *** ozialien10 has quit IRC | 02:39 | |
| *** shaohe_feng has quit IRC | 02:45 | |
| *** shaohe_feng has joined #openstack-nova | 02:46 | |
| *** ducttape_ has quit IRC | 02:50 | |
| *** vilobhmm11 has quit IRC | 02:55 | |
| *** shaohe_feng has quit IRC | 02:56 | |
| *** shaohe_feng has joined #openstack-nova | 02:56 | |
| *** mspreitz has joined #openstack-nova | 02:59 | |
| *** baoli has quit IRC | 03:01 | |
| *** whenry has joined #openstack-nova | 03:01 | |
| *** shaohe_feng has quit IRC | 03:06 | |
| *** tbachman has quit IRC | 03:06 | |
| *** shaohe_feng has joined #openstack-nova | 03:06 | |
| *** shaohe_feng has quit IRC | 03:16 | |
| *** shaohe_feng has joined #openstack-nova | 03:17 | |
| *** mspreitz has quit IRC | 03:17 | |
| *** fragatina has quit IRC | 03:17 | |
| *** prateek has joined #openstack-nova | 03:19 | |
| *** thorst has joined #openstack-nova | 03:20 | |
| *** pwnall has quit IRC | 03:23 | |
| *** diogogmt has joined #openstack-nova | 03:25 | |
| *** prateek_ has joined #openstack-nova | 03:25 | |
| *** chip_ has quit IRC | 03:26 | |
| *** shaohe_feng has quit IRC | 03:26 | |
| *** shaohe_feng has joined #openstack-nova | 03:27 | |
| *** thorst has quit IRC | 03:28 | |
| *** prateek has quit IRC | 03:28 | |
| *** Jeffrey4l has quit IRC | 03:37 | |
| *** shaohe_feng has quit IRC | 03:37 | |
| *** shaohe_feng has joined #openstack-nova | 03:37 | |
| *** shaohe_feng has quit IRC | 03:47 | |
| *** rfolco has joined #openstack-nova | 03:49 | |
| *** shaohe_feng has joined #openstack-nova | 03:50 | |
| *** shaohe_feng has quit IRC | 03:57 | |
| *** shaohe_feng has joined #openstack-nova | 03:57 | |
| *** artom has quit IRC | 03:57 | |
| *** mdrabe has joined #openstack-nova | 03:58 | |
| *** dave-mccowan has quit IRC | 04:06 | |
| *** shaohe_feng has quit IRC | 04:07 | |
| *** fragatina has joined #openstack-nova | 04:08 | |
| *** shaohe_feng has joined #openstack-nova | 04:08 | |
| *** doug-fish has quit IRC | 04:09 | |
| *** hemna has joined #openstack-nova | 04:13 | |
| *** prateek__ has joined #openstack-nova | 04:14 | |
| *** Jeffrey4l has joined #openstack-nova | 04:14 | |
| *** Jeffrey4l_ has joined #openstack-nova | 04:15 | |
| *** unicell has joined #openstack-nova | 04:15 | |
| *** sacharya has quit IRC | 04:16 | |
| *** Jeffrey4l__ has joined #openstack-nova | 04:16 | |
| *** sacharya has joined #openstack-nova | 04:16 | |
| *** rfolco has quit IRC | 04:16 | |
| *** links has joined #openstack-nova | 04:17 | |
| *** sacharya has quit IRC | 04:17 | |
| *** Jeffrey4l__ has quit IRC | 04:17 | |
| *** Jeffrey4l__ has joined #openstack-nova | 04:17 | |
| *** prateek_ has quit IRC | 04:17 | |
| *** prateek_ has joined #openstack-nova | 04:17 | |
| *** shaohe_feng has quit IRC | 04:18 | |
| *** shaohe_feng has joined #openstack-nova | 04:18 | |
| *** markvoelker has quit IRC | 04:19 | |
| *** Jeffrey4l_ has quit IRC | 04:20 | |
| *** Jeffrey4l has quit IRC | 04:20 | |
| *** prateek__ has quit IRC | 04:21 | |
| *** Jeffrey4l__ has quit IRC | 04:22 | |
| *** unicell1 has joined #openstack-nova | 04:25 | |
| *** thorst has joined #openstack-nova | 04:25 | |
| *** unicell has quit IRC | 04:26 | |
| *** liverpooler has joined #openstack-nova | 04:26 | |
| *** sridharg has joined #openstack-nova | 04:27 | |
| *** shaohe_feng has quit IRC | 04:28 | |
| *** shaohe_feng has joined #openstack-nova | 04:28 | |
| *** vladikr has quit IRC | 04:30 | |
| *** unicell has joined #openstack-nova | 04:31 | |
| *** unicell1 has quit IRC | 04:32 | |
| *** thorst has quit IRC | 04:32 | |
| *** shaohe_feng has quit IRC | 04:38 | |
| *** shaohe_feng has joined #openstack-nova | 04:38 | |
| *** mdrabe has quit IRC | 04:42 | |
| *** unicell has quit IRC | 04:44 | |
| *** vtech has quit IRC | 04:44 | |
| *** shaohe_feng has quit IRC | 04:48 | |
| *** shaohe_feng has joined #openstack-nova | 04:49 | |
| *** prateek_ has quit IRC | 04:50 | |
| *** ducttape_ has joined #openstack-nova | 04:51 | |
| *** josecastroleon has joined #openstack-nova | 04:52 | |
| *** sridharg has quit IRC | 04:53 | |
| *** ducttape_ has quit IRC | 04:55 | |
| *** sree has joined #openstack-nova | 04:56 | |
| *** sridharg has joined #openstack-nova | 04:57 | |
| *** moshele has joined #openstack-nova | 04:58 | |
| *** shaohe_feng has quit IRC | 04:59 | |
| *** shaohe_feng has joined #openstack-nova | 04:59 | |
| *** irenab has joined #openstack-nova | 05:06 | |
| *** shaohe_feng has quit IRC | 05:09 | |
| *** hemna has quit IRC | 05:10 | |
| *** shaohe_feng has joined #openstack-nova | 05:10 | |
| *** whenry has quit IRC | 05:14 | |
| *** sacharya has joined #openstack-nova | 05:17 | |
| *** Jeffrey4l__ has joined #openstack-nova | 05:18 | |
| *** shaohe_feng has quit IRC | 05:19 | |
| *** markvoelker has joined #openstack-nova | 05:19 | |
| *** shaohe_feng has joined #openstack-nova | 05:20 | |
| *** sacharya has quit IRC | 05:22 | |
| *** markvoelker has quit IRC | 05:24 | |
| *** Jeffrey4l__ has quit IRC | 05:24 | |
| *** apoorv has joined #openstack-nova | 05:28 | |
| *** prateek has joined #openstack-nova | 05:28 | |
| *** shaohe_feng has quit IRC | 05:29 | |
| *** thorst has joined #openstack-nova | 05:30 | |
| *** shaohe_feng has joined #openstack-nova | 05:30 | |
| *** sdake has quit IRC | 05:37 | |
| *** thorst has quit IRC | 05:37 | |
| *** iceyao has quit IRC | 05:38 | |
| *** iceyao_ has joined #openstack-nova | 05:38 | |
| *** sdake has joined #openstack-nova | 05:39 | |
| *** iceyao has joined #openstack-nova | 05:39 | |
| *** shaohe_feng has quit IRC | 05:40 | |
| *** shaohe_feng has joined #openstack-nova | 05:40 | |
| *** esracelik has joined #openstack-nova | 05:44 | |
| *** shaohe_feng has quit IRC | 05:50 | |
| *** iceyao has quit IRC | 05:50 | |
| *** shaohe_feng has joined #openstack-nova | 05:53 | |
| *** iceyao has joined #openstack-nova | 05:54 | |
| *** liverpooler has quit IRC | 05:54 | |
| *** liverpooler has joined #openstack-nova | 05:54 | |
| *** shaohe_feng has quit IRC | 06:00 | |
| *** shaohe_feng has joined #openstack-nova | 06:01 | |
| *** sridharg has quit IRC | 06:05 | |
| *** sridharg has joined #openstack-nova | 06:05 | |
| *** shaohe_feng has quit IRC | 06:10 | |
| *** shaohe_feng has joined #openstack-nova | 06:10 | |
| *** salv-orlando has joined #openstack-nova | 06:11 | |
| *** rcernin has joined #openstack-nova | 06:15 | |
| *** EinstCra_ has quit IRC | 06:17 | |
| *** EinstCrazy has joined #openstack-nova | 06:17 | |
| *** sree has quit IRC | 06:19 | |
| *** sree has joined #openstack-nova | 06:19 | |
| *** mjura has joined #openstack-nova | 06:20 | |
| *** markvoelker has joined #openstack-nova | 06:20 | |
| *** shaohe_feng has quit IRC | 06:21 | |
| *** luis5tb has quit IRC | 06:21 | |
| *** shaohe_feng has joined #openstack-nova | 06:21 | |
| *** sree has quit IRC | 06:23 | |
| *** markvoelker has quit IRC | 06:25 | |
| *** mjura has quit IRC | 06:26 | |
| *** mjura has joined #openstack-nova | 06:26 | |
| *** shaohe_feng has quit IRC | 06:31 | |
| *** shaohe_feng has joined #openstack-nova | 06:31 | |
| *** gibi has quit IRC | 06:33 | |
| *** cfriesen has quit IRC | 06:34 | |
| *** gibi has joined #openstack-nova | 06:34 | |
| *** laszloh has joined #openstack-nova | 06:34 | |
| *** thorst has joined #openstack-nova | 06:35 | |
| *** gongysh has quit IRC | 06:36 | |
| *** shaohe_feng has quit IRC | 06:41 | |
| *** thorst has quit IRC | 06:42 | |
| *** tesseract has joined #openstack-nova | 06:44 | |
| *** tesseract is now known as Guest40479 | 06:44 | |
| *** shaohe_feng has joined #openstack-nova | 06:44 | |
| -openstackstatus- NOTICE: Filesystem on logs.openstack.org is broken, we are on the process of repairing it. Please stop checking your jobs until further notice | 06:45 | |
| *** ChanServ changes topic to "Filesystem on logs.openstack.org is broken, we are on the process of repairing it. Please stop checking your jobs until further notice" | 06:45 | |
| *** daemontool_ has quit IRC | 06:45 | |
| *** Guest40479 has quit IRC | 06:47 | |
| *** zenoway has joined #openstack-nova | 06:48 | |
| *** apoorv has quit IRC | 06:48 | |
| *** gibi has quit IRC | 06:48 | |
| *** tesseract- has joined #openstack-nova | 06:49 | |
| *** gibi has joined #openstack-nova | 06:49 | |
| *** mpavone has joined #openstack-nova | 06:51 | |
| *** mpavone has quit IRC | 06:51 | |
| *** shaohe_feng has quit IRC | 06:51 | |
| *** ducttape_ has joined #openstack-nova | 06:52 | |
| *** shaohe_feng has joined #openstack-nova | 06:52 | |
| *** sree has joined #openstack-nova | 06:52 | |
| *** sacharya has joined #openstack-nova | 06:53 | |
| *** gibi has quit IRC | 06:54 | |
| *** gibi has joined #openstack-nova | 06:55 | |
| *** ducttape_ has quit IRC | 06:56 | |
| *** sacharya has quit IRC | 06:58 | |
| *** liverpooler has quit IRC | 06:58 | |
| *** liverpooler has joined #openstack-nova | 06:58 | |
| *** paul-carlton2 has joined #openstack-nova | 06:59 | |
| *** palexster has joined #openstack-nova | 07:00 | |
| *** shaohe_feng has quit IRC | 07:02 | |
| *** shaohe_feng has joined #openstack-nova | 07:03 | |
| *** salv-orlando has quit IRC | 07:04 | |
| *** rcernin_ has joined #openstack-nova | 07:04 | |
| *** dimtruck is now known as zz_dimtruck | 07:10 | |
| *** jed56 has joined #openstack-nova | 07:11 | |
| *** shaohe_feng has quit IRC | 07:12 | |
| *** shaohe_feng has joined #openstack-nova | 07:12 | |
| *** haukebruno has joined #openstack-nova | 07:12 | |
| *** kaisers1 has joined #openstack-nova | 07:17 | |
| *** paul-carlton2 has quit IRC | 07:20 | |
| *** markvoelker has joined #openstack-nova | 07:21 | |
| *** shaohe_feng has quit IRC | 07:22 | |
| *** EinstCrazy has quit IRC | 07:22 | |
| *** EinstCrazy has joined #openstack-nova | 07:23 | |
| *** shaohe_feng has joined #openstack-nova | 07:23 | |
| *** markvoelker has quit IRC | 07:26 | |
| *** zz_dimtruck is now known as dimtruck | 07:26 | |
| *** shaohe_feng has quit IRC | 07:32 | |
| *** rcernin_ has quit IRC | 07:33 | |
| *** shaohe_feng has joined #openstack-nova | 07:33 | |
| * bauzas yawns with a bit of jetlag | 07:33 | |
| bauzas | p!spy | 07:33 |
| *** cristicalin has joined #openstack-nova | 07:34 | |
| *** dimtruck is now known as zz_dimtruck | 07:35 | |
| *** rcernin has quit IRC | 07:36 | |
| *** Oku_OS-away is now known as Oku_OS | 07:39 | |
| *** thorst has joined #openstack-nova | 07:40 | |
| *** shaohe_feng has quit IRC | 07:43 | |
| *** shaohe_feng has joined #openstack-nova | 07:44 | |
| *** thorst has quit IRC | 07:47 | |
| *** paul-carlton2 has joined #openstack-nova | 07:50 | |
| *** shaohe_feng has quit IRC | 07:53 | |
| *** shaohe_feng has joined #openstack-nova | 07:54 | |
| *** itamarl has joined #openstack-nova | 07:54 | |
| *** sdake has quit IRC | 07:56 | |
| *** fragatin_ has joined #openstack-nova | 07:56 | |
| *** xek__ has joined #openstack-nova | 07:56 | |
| *** rhagarty_ has joined #openstack-nova | 07:57 | |
| *** zhiyan_ has joined #openstack-nova | 07:58 | |
| *** pumaranikar_ has quit IRC | 07:58 | |
| *** carl_baldwin has quit IRC | 07:58 | |
| *** scottda has quit IRC | 07:58 | |
| *** yonglihe has quit IRC | 07:58 | |
| *** DuncanT has quit IRC | 07:58 | |
| *** boris-42 has quit IRC | 07:58 | |
| *** zhiyan has quit IRC | 07:58 | |
| *** jed56 has quit IRC | 07:58 | |
| *** fyxim__ has quit IRC | 07:58 | |
| *** GheRivero has quit IRC | 07:58 | |
| *** vdrok has quit IRC | 07:58 | |
| *** woodster_ has quit IRC | 07:58 | |
| *** zzzeek has quit IRC | 07:58 | |
| *** gzhai1 has quit IRC | 07:58 | |
| *** syfy323_ has quit IRC | 07:58 | |
| *** scottda has joined #openstack-nova | 07:58 | |
| *** vdrok_ has joined #openstack-nova | 07:58 | |
| *** sulo has quit IRC | 07:58 | |
| *** vdrok_ has quit IRC | 07:58 | |
| *** vdrok_ has joined #openstack-nova | 07:58 | |
| *** scottda has quit IRC | 07:58 | |
| *** scottda has joined #openstack-nova | 07:58 | |
| *** rook has quit IRC | 07:58 | |
| *** Guest78717 has quit IRC | 07:58 | |
| *** xek_ has quit IRC | 07:58 | |
| *** mmedvede has quit IRC | 07:58 | |
| *** lbeliveau has quit IRC | 07:58 | |
| *** mrhillsman has quit IRC | 07:58 | |
| *** ayoung has quit IRC | 07:58 | |
| *** anthonyper has quit IRC | 07:58 | |
| *** tanee has quit IRC | 07:58 | |
| *** gus has quit IRC | 07:58 | |
| *** rhagarty has quit IRC | 07:58 | |
| *** auggy has quit IRC | 07:58 | |
| *** lbragstad has quit IRC | 07:58 | |
| *** mrodden has quit IRC | 07:58 | |
| *** sulo has joined #openstack-nova | 07:58 | |
| *** gzhai1 has joined #openstack-nova | 07:58 | |
| *** syfy323 has joined #openstack-nova | 07:59 | |
| *** DuncanT_ has joined #openstack-nova | 07:59 | |
| *** BobBall has joined #openstack-nova | 07:59 | |
| *** jistr has joined #openstack-nova | 07:59 | |
| *** fragatina has quit IRC | 07:59 | |
| openstackgerrit | Zhengguang Ou proposed openstack/nova: libvirt: fix detach volume race condition https://review.openstack.org/290510 | 07:59 |
| *** auggy has joined #openstack-nova | 07:59 | |
| *** BobBall is now known as Guest82571 | 07:59 | |
| *** tanee has joined #openstack-nova | 07:59 | |
| *** lbragstad has joined #openstack-nova | 07:59 | |
| *** carl_baldwin has joined #openstack-nova | 07:59 | |
| *** ayoung has joined #openstack-nova | 07:59 | |
| *** anthonyper has joined #openstack-nova | 07:59 | |
| *** mrhillsman has joined #openstack-nova | 08:00 | |
| *** sfinucan has joined #openstack-nova | 08:00 | |
| *** zhiyan_ is now known as zhiyan | 08:00 | |
| *** fyxim__ has joined #openstack-nova | 08:00 | |
| *** mrodden has joined #openstack-nova | 08:00 | |
| *** kashyap` is now known as kashyap | 08:00 | |
| *** lbeliveau has joined #openstack-nova | 08:00 | |
| *** yonglihe has joined #openstack-nova | 08:00 | |
| *** zzzeek has joined #openstack-nova | 08:00 | |
| *** DuncanT_ is now known as DuncanT | 08:02 | |
| *** gus has joined #openstack-nova | 08:02 | |
| *** itamarl_ has joined #openstack-nova | 08:02 | |
| *** matrohon has joined #openstack-nova | 08:02 | |
| *** shaohe_feng has quit IRC | 08:03 | |
| *** pumaranikar has joined #openstack-nova | 08:03 | |
| *** shaohe_feng has joined #openstack-nova | 08:04 | |
| *** woodster_ has joined #openstack-nova | 08:05 | |
| *** itamarl_ has quit IRC | 08:05 | |
| *** jlanoux has joined #openstack-nova | 08:05 | |
| *** itamarl has quit IRC | 08:06 | |
| *** sfinucan has quit IRC | 08:07 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: [WIP] Testing latest u-c https://review.openstack.org/303345 | 08:10 |
| *** boris-42 has joined #openstack-nova | 08:10 | |
| *** GheRivero has joined #openstack-nova | 08:10 | |
| *** mmedvede has joined #openstack-nova | 08:11 | |
| *** jed56 has joined #openstack-nova | 08:11 | |
| *** shaohe_feng has quit IRC | 08:13 | |
| *** shaohe_feng has joined #openstack-nova | 08:14 | |
| *** itamarl has joined #openstack-nova | 08:15 | |
| *** jogo has quit IRC | 08:16 | |
| *** mvk has joined #openstack-nova | 08:18 | |
| *** markvoelker has joined #openstack-nova | 08:22 | |
| *** paul-carlton2 has quit IRC | 08:23 | |
| *** shaohe_feng has quit IRC | 08:24 | |
| *** shaohe_feng has joined #openstack-nova | 08:24 | |
| *** fragatin_ has quit IRC | 08:24 | |
| *** mdavidson has joined #openstack-nova | 08:25 | |
| *** jogo has joined #openstack-nova | 08:25 | |
| *** ChanServ changes topic to "This channel is for Nova development. For support of Nova deployments, please use #openstack. Please see: https://wiki.openstack.org/wiki/Nova/Newton_Release_Schedule" | 08:25 | |
| -openstackstatus- NOTICE: Logs filesystem has been successfully restored, please recheck your jobs | 08:25 | |
| *** markvoelker has quit IRC | 08:27 | |
| *** josecastroleon has quit IRC | 08:28 | |
| *** shaohe_feng has quit IRC | 08:34 | |
| *** shaohe_feng has joined #openstack-nova | 08:35 | |
| -openstackstatus- NOTICE: Filesystem on docs-draft.openstack.org is broken, we are on the process of repairing it. Please stop checking jobs using this filesystem until further notice | 08:35 | |
| *** dmk0202 has joined #openstack-nova | 08:35 | |
| *** paul-carlton2 has joined #openstack-nova | 08:37 | |
| *** fragatina has joined #openstack-nova | 08:37 | |
| *** derekh has joined #openstack-nova | 08:39 | |
| *** yassine has joined #openstack-nova | 08:39 | |
| *** yassine is now known as Guest99400 | 08:39 | |
| *** shaohe_feng has quit IRC | 08:44 | |
| *** thorst has joined #openstack-nova | 08:45 | |
| *** shaohe_feng has joined #openstack-nova | 08:45 | |
| *** zenoway has quit IRC | 08:45 | |
| *** cristicalin has quit IRC | 08:47 | |
| *** zenoway has joined #openstack-nova | 08:48 | |
| *** hkominos has joined #openstack-nova | 08:49 | |
| hkominos | hello Openstackers!. Quick question. Is there any way to actually control the way that a hypervisor schedules the workload in Openstack ? | 08:49 |
| *** apoorv has joined #openstack-nova | 08:49 | |
| *** thorst has quit IRC | 08:52 | |
| *** e0ne has joined #openstack-nova | 08:52 | |
| *** ducttape_ has joined #openstack-nova | 08:53 | |
| *** zenoway has quit IRC | 08:53 | |
| *** sacharya has joined #openstack-nova | 08:54 | |
| *** shaohe_feng has quit IRC | 08:54 | |
| *** danpb has joined #openstack-nova | 08:55 | |
| *** iceyao has quit IRC | 08:55 | |
| *** shaohe_feng has joined #openstack-nova | 08:55 | |
| bauzas | hkominos: what do you want to get ? | 08:56 |
| *** wanghao has joined #openstack-nova | 08:57 | |
| *** ducttape_ has quit IRC | 08:57 | |
| *** sacharya has quit IRC | 08:59 | |
| hkominos | bauzas:I am trying to get my system under load to take some measurements. (CPU-overcommit is now at 1). I have a 2 CPU quad core server. So I am spawning a 8 Vcpu VM and running a CPU intensive job there but the actual physical machine is not using all the cores | 08:59 |
| *** sudipto has joined #openstack-nova | 08:59 | |
| bauzas | hkominos: that's because the scheduler defaults packing the instances instead of spreading them | 09:00 |
| hkominos | its using like 2. maybe less. So i want to force it to use 8 in order to check the extra cost which virtualization induces | 09:00 |
| *** josecastroleon has joined #openstack-nova | 09:00 | |
| hkominos | you mean it packs the load in the CPUS instead of spreading them ? | 09:01 |
| johnthetubaguy | hmm, so you have 8 vCPUs but you can't use all the physical CPUs | 09:01 |
| hkominos | yeap | 09:01 |
| *** sileht has quit IRC | 09:01 | |
| bauzas | oh, I missed your point | 09:02 |
| hkominos | does it make sense ? | 09:02 |
| *** sileht has joined #openstack-nova | 09:02 | |
| bauzas | you mean that you have 8 pCPUs (2 CPUs quad-core) and you spin a 8 vCPU instance flavor, right? | 09:02 |
| hkominos | yeap | 09:03 |
| *** fragatina has quit IRC | 09:03 | |
| bauzas | how many physical hosts ? | 09:03 |
| bauzas | only one ? | 09:03 |
| hkominos | right now yes | 09:03 |
| hkominos | 1 Compute host which runs 1 vm | 09:03 |
| bauzas | ok, so you agree that you could only have one instance using that 8vCPU flavor for your cloud ? | 09:03 |
| hkominos | yes (its just for testing now) | 09:04 |
| bauzas | given your cpu allocation ratio set to 1 | 09:04 |
| bauzas | okay, and you got NoValidHost ? | 09:04 |
| bauzas | are you sure it's because of the CoreFilter ? | 09:04 |
| hkominos | no. the job is running. | 09:04 |
| *** shaohe_feng has quit IRC | 09:05 | |
| bauzas | so I'm misunderstanding your question | 09:05 |
| hkominos | let me specify again | 09:05 |
| johnthetubaguy | bauzas: means on the hypervisor / KVM scheduling I think? | 09:05 |
| johnthetubaguy | I kinda thought it would work how you expect hkominos, so thats curious | 09:05 |
| *** shaohe_feng has joined #openstack-nova | 09:05 | |
| bauzas | yeah | 09:05 |
| johnthetubaguy | I just wonder if its a NUMA related issue | 09:05 |
| bauzas | johnthetubaguy: that wouldn't really change | 09:05 |
| bauzas | johnthetubaguy: ++ if NUMA | 09:06 |
| johnthetubaguy | well, I wonder if its trying to keep all the vCPUs on the same NUMA core | 09:06 |
| johnthetubaguy | hkominos: are you maxing out any of the CPUs on your host? | 09:06 |
| bauzas | I'm curious about the libvirt xml descriptor for that instance | 09:06 |
| johnthetubaguy | hkominos: inside the guest, and the host reported level? | 09:06 |
| hkominos | If i login to my compute host and run one specific job which puts the system on maximum load it runs in 4 seconds. If i provision an 8vCPU machine (same size as my actual physical host) it takes 15 minutes | 09:06 |
| johnthetubaguy | bauzas: thats a good question | 09:06 |
| hkominos | u mean the virsch xml log ? | 09:07 |
| bauzas | hkominos: not the log, the xml file | 09:08 |
| bauzas | hkominos: virsh describe | 09:08 |
| bauzas | (or virsh show, I don't recall it) | 09:08 |
| hkominos | let me get it one sec | 09:08 |
| bauzas | use paste.openstack.org please for the output :) | 09:08 |
| *** palexster1 has joined #openstack-nova | 09:09 | |
| *** palexster has quit IRC | 09:12 | |
| *** itamarl_ has joined #openstack-nova | 09:12 | |
| *** zz_dimtruck is now known as dimtruck | 09:12 | |
| hkominos | i think you meant dumpxml right? | 09:13 |
| hkominos | could not find another command | 09:13 |
| bauzas | my bad, yeah | 09:13 |
| * bauzas still suffers from jet lag | 09:14 | |
| bauzas | I need caffeine again | 09:14 |
| *** wanghao has quit IRC | 09:14 | |
| *** shaohe_feng has quit IRC | 09:15 | |
| *** itamarl has quit IRC | 09:15 | |
| *** itamarl_ is now known as itamarl | 09:15 | |
| hkominos | http://paste.openstack.org/show/495936/ | 09:15 |
| *** wanghao has joined #openstack-nova | 09:15 | |
| *** shaohe_feng has joined #openstack-nova | 09:16 | |
| *** fragatina has joined #openstack-nova | 09:16 | |
| johnthetubaguy | bauzas: yeah, turns out doing a concert the evening after getting on the plane helps make you so sleep it resets your body clock, although I feel like I missed a day of sleep somewhere, which I kinda did! | 09:16 |
| hkominos | So basically if a start the job in my VM and login in to my compute host i can see with top that the physical host in not loaded. it used like 1-2 cpus | 09:17 |
| johnthetubaguy | hkominos: what does the VM show for CPU usage? | 09:17 |
| hkominos | maximum | 09:17 |
| johnthetubaguy | gotcha | 09:17 |
| hkominos | show me the light | 09:18 |
| hkominos | spread the knowledge | 09:18 |
| johnthetubaguy | <topology sockets='8' cores='1' threads='1'/> | 09:18 |
| bauzas | hkominos: what says the guest OS ? I can only see one core | 09:18 |
| bauzas | arrrrh, jinxed | 09:18 |
| hkominos | ! | 09:18 |
| hkominos | guest os though cat proc/cpuinfo reports 8 cpus | 09:19 |
| johnthetubaguy | thats 8 sockets 1 core each though, I would have assumed | 09:19 |
| johnthetubaguy | yeah | 09:19 |
| bauzas | k | 09:19 |
| bauzas | that's qemu thoi | 09:19 |
| bauzas | tho | 09:19 |
| johnthetubaguy | true | 09:19 |
| *** gongysh has joined #openstack-nova | 09:19 | |
| hkominos | So does that mean that it schedules only 1 thread on the physical host ? | 09:19 |
| johnthetubaguy | <vcpu placement='static'>8</vcpu> | 09:19 |
| johnthetubaguy | bauzas: what is that about? | 09:20 |
| hkominos | I assume that describes the Vcpu as set by openstack | 09:20 |
| johnthetubaguy | hkominos: it shouldn't, I think that is 1 thread per core, 1 core per socket, etc | 09:20 |
| hkominos | xm. Btw i have disabled hyperthreading on my physical machine and restarted the it | 09:20 |
| bauzas | johnthetubaguy: that looks like a cpu pinning feature | 09:20 |
| openstackgerrit | Hahyun proposed openstack/nova-specs: Add direct download feature for rbd glance backend https://review.openstack.org/308481 | 09:21 |
| *** e0ne has quit IRC | 09:21 | |
| *** sfinucan has joined #openstack-nova | 09:21 | |
| bauzas | but I wouldn't be worried | 09:21 |
| johnthetubaguy | https://libvirt.org/formatdomain.html#elementsCPUAllocation | 09:21 |
| openstackgerrit | Sudipta Biswas proposed openstack/nova-specs: Report host memory bandwidth as a metric in Nova https://review.openstack.org/312001 | 09:21 |
| bauzas | johnthetubaguy: yeah that's what I recall | 09:22 |
| *** dimtruck is now known as zz_dimtruck | 09:22 | |
| -openstackstatus- NOTICE: Docs-draft filesystem has been restored. Please check your affected jobs again | 09:23 | |
| *** belmoreira has joined #openstack-nova | 09:23 | |
| *** markvoelker has joined #openstack-nova | 09:23 | |
| bauzas | johnthetubaguy: that's because saying "okay, pin those 8 virtual cores to these 8 physical ones", but since we don't care about a specific placement, that looks okay to me | 09:23 |
| hkominos | it looks correct to me | 09:23 |
| hkominos | i am no expert though | 09:23 |
| *** sahid has joined #openstack-nova | 09:23 | |
| hkominos | u made a comment about numa before though | 09:24 |
| johnthetubaguy | bauzas: yeah, it seems correct | 09:24 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: introduces module to handle domain xml migration https://review.openstack.org/299490 | 09:24 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: add "get_job_info" to Guest's object https://review.openstack.org/250740 | 09:24 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: add method to configure max downtime when migrating https://review.openstack.org/250739 | 09:24 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: add "abort_job" to Guest's object https://review.openstack.org/250738 | 09:24 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: add method "migrate" to Guest's object https://review.openstack.org/250365 | 09:24 |
| johnthetubaguy | hkominos: yeah, I was worried it wasn't pinning the 8 virtual cores to the 8 physical ones | 09:24 |
| bauzas | hkominos: so, my point is, I wonder if all of that isn't just due to the fact you still consume the user space by not using KVM | 09:24 |
| bauzas | hkominos: I would make sure to enable kvm, set nova to use it by default and create another instance | 09:25 |
| *** shaohe_feng has quit IRC | 09:25 | |
| *** paul-carlton2 has quit IRC | 09:25 | |
| hkominos | btw i thought qemu/kvm go hand in hand. | 09:25 |
| *** shaohe_feng has joined #openstack-nova | 09:26 | |
| johnthetubaguy | you can run QEMU without KVM, like we do in our testing actually | 09:26 |
| hkominos | so u are saying that i might be misconfigured | 09:27 |
| hkominos | and not using kvm at all ? | 09:27 |
| *** markvoelker has quit IRC | 09:28 | |
| *** kolesovdv has joined #openstack-nova | 09:29 | |
| *** sridharg has quit IRC | 09:29 | |
| danpb | hkominos: look in your /etc/nova/nova.conf to see if it says qemu or kvm for the libvirt driver | 09:29 |
| bauzas | I also wonder which guest OS you're running re: any posisble max_sockets | 09:30 |
| ildikov | johnthetubaguy: hi | 09:30 |
| bauzas | danpb: what if my image is not tagged with max_sockets and I'm running an instance with more sockets than the guest OS would support ? I suppose I'd only see the CPUs in the guest that it can supports ? | 09:31 |
| ildikov | johnthetubaguy: it was a bit late yesterday when we got to a conclusion on when to have a multiattach Hangout call this week, I added you to the meeting invitation, I hope you can make it to the call | 09:32 |
| hkominos | compute_driver=libvirt.LibvirtDriver | 09:32 |
| *** vtech has joined #openstack-nova | 09:32 | |
| *** sambetts|afk is now known as sambetts | 09:32 | |
| danpb | bauzas: qemu always shows the guest the full set of sockets - whether the guest uses them or not is a different matter | 09:33 |
| johnthetubaguy | ildikov: should be OK | 09:33 |
| danpb | hkominos: no you want 'virt_type' - eg grep virt_type /etc/nova/nova.conf | 09:33 |
| bauzas | danpb: okay, that makes sense to me :) | 09:33 |
| ildikov | johnthetubaguy: great, thanks much | 09:33 |
| hkominos | virt type=qemu | 09:34 |
| danpb | well there's the problem then | 09:34 |
| hkominos | should switch to Kvm ? | 09:34 |
| danpb | if you want your guests to have fast performance then yes :) | 09:34 |
| johnthetubaguy | +1 | 09:35 |
| danpb | if you want them to feel like a pentium from 2001, then stick with qemu :-) | 09:35 |
| hkominos | hahhaha. let me try ! | 09:35 |
| bauzas | danpb: yeah the VM was using QEMU hence my point above http://paste.openstack.org/show/495936/ | 09:35 |
| *** shaohe_feng has quit IRC | 09:35 | |
| openstackgerrit | Hahyun proposed openstack/nova-specs: Add direct download feature for rbd glance backend https://review.openstack.org/308481 | 09:35 |
| hkominos | btw the compute_driver field remains the same right ? i dont see another configuration in the opentack docks | 09:36 |
| *** shaohe_feng has joined #openstack-nova | 09:36 | |
| danpb | yes, only change the virt_type | 09:37 |
| hkominos | xm. i wil try booting a vm again and test. Btw should i be seeing KVM in the horizon in the hypervisors tab? | 09:39 |
| kashyap | hkominos: The above assumes you have /dev/kvm exposed (Run `virt-host-validate | grep kvm` should result in 'PASS') on the Compute node. | 09:40 |
| hkominos | Pass it is | 09:40 |
| *** iceyao has joined #openstack-nova | 09:40 | |
| *** fragatina has quit IRC | 09:41 | |
| *** paul-carlton2 has joined #openstack-nova | 09:42 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Add a RequestSpec generation migration script https://review.openstack.org/302828 | 09:42 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Cold migrate using the RequestSpec object https://review.openstack.org/284974 | 09:43 |
| *** apoorv has quit IRC | 09:45 | |
| *** shaohe_feng has quit IRC | 09:46 | |
| hkominos | all right my new friends | 09:46 |
| *** shaohe_feng has joined #openstack-nova | 09:46 | |
| hkominos | that was it. I am out of the pentium 1 era of qemu | 09:47 |
| hkominos | it now works much better. I assume now i can also increase my CPU overcommit ratio right? | 09:47 |
| *** josecastroleon has quit IRC | 09:47 | |
| hkominos | thank you for help | 09:48 |
| *** links has quit IRC | 09:48 | |
| *** thorst has joined #openstack-nova | 09:50 | |
| *** andrearo1 is now known as andrearosa | 09:54 | |
| *** sfinucan has quit IRC | 09:54 | |
| openstackgerrit | Brian Duffy proposed openstack/nova: running through sandbox tutorial here https://review.openstack.org/312005 | 09:54 |
| *** sfinucan has joined #openstack-nova | 09:55 | |
| *** Guest99400 has quit IRC | 09:56 | |
| *** shaohe_feng has quit IRC | 09:56 | |
| *** shaohe_feng has joined #openstack-nova | 09:56 | |
| *** thorst has quit IRC | 09:57 | |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: config: remove deprecated ironic.client_log_level https://review.openstack.org/312006 | 10:00 |
| *** jlanoux has quit IRC | 10:00 | |
| *** jlanoux has joined #openstack-nova | 10:00 | |
| *** cheneydc has quit IRC | 10:02 | |
| *** fishcried has joined #openstack-nova | 10:02 | |
| openstackgerrit | Brian Duffy proposed openstack/nova: running through sandbox tutorial here https://review.openstack.org/312005 | 10:02 |
| *** kengo_sakai has joined #openstack-nova | 10:02 | |
| *** e0ne has joined #openstack-nova | 10:02 | |
| *** iceyao has quit IRC | 10:03 | |
| *** cdent has joined #openstack-nova | 10:03 | |
| johnthetubaguy | hkominos: yeah, increasing the ratio tells the nova scheduler to put more vCPUs than you have physical CPUs on that hypervisor | 10:03 |
| *** links has joined #openstack-nova | 10:04 | |
| hkominos | yeah but i mean that if i only have one VM again then the time it takes to execute a job will not change. Cpu overcommit is only a metric in order to specify another limit in the number of VMs u can spin up per host | 10:05 |
| hkominos | (i think) | 10:05 |
| *** shaohe_feng has quit IRC | 10:06 | |
| *** shaohe_feng has joined #openstack-nova | 10:06 | |
| *** apoorv has joined #openstack-nova | 10:07 | |
| *** links has quit IRC | 10:07 | |
| danpb | cpu overcommit is useful if you have lots of VMs whose CPUs are mostly idle | 10:07 |
| danpb | since it'll give you better utilization of the host overall | 10:07 |
| bauzas | hkominos: using ratios is for workloads not using 100% CPU usage | 10:08 |
| bauzas | arrrh | 10:08 |
| danpb | the downside is that if all VMs decided they need to use their CPUs at the same time, they'll compete for time | 10:08 |
| johnthetubaguy | yeah, +1 what danpb said, it depends what you are optimising for | 10:08 |
| bauzas | hearing the word "job" makes me thinking it's predictable for hkominos :) | 10:08 |
| danpb | generally allowing cpu overcommit is a good thing | 10:08 |
| danpb | as its pretty rare to have guest workloads which need 100% cpu all the time | 10:09 |
| hkominos | hahahaha | 10:09 |
| hkominos | to be honest this job that i am running now is using 100% | 10:10 |
| hkominos | thats its point anyway | 10:10 |
| sfinucan | There are two bugs for the CPU policy stuff that have been waiting for review for a few months. Seeing as ndipanov is gone, could someone else take a look at them, please? https://review.openstack.org/#/c/285321/ https://review.openstack.org/#/c/285232/ | 10:13 |
| *** links has joined #openstack-nova | 10:14 | |
| johnthetubaguy | sfinucan: did we find someone to start taking over the performance vm/NFV/NUMA subteam, I can't 100% remember now? | 10:14 |
| sfinucan | johnthetubaguy: I heard nothing concrete, no. Figured it would be danpb, tbh :) | 10:16 |
| *** shaohe_feng has quit IRC | 10:16 | |
| *** lpetrut has joined #openstack-nova | 10:17 | |
| *** shaohe_feng has joined #openstack-nova | 10:17 | |
| bauzas | johnthetubaguy: sfinucan: vladikr and sahid are probably the best options | 10:19 |
| bauzas | and moshele will be running the new subteam as he claimed in the ML | 10:20 |
| danpb | sfinucan: johnthetubaguy sorry, i don't have enough time to actually organize a subteam myself | 10:20 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Resize API operation passing down original RequestSpec https://review.openstack.org/312009 | 10:22 |
| *** dave-mccowan has joined #openstack-nova | 10:23 | |
| *** Guest82571 is now known as BobBall | 10:23 | |
| *** markvoelker has joined #openstack-nova | 10:24 | |
| *** tbachman has joined #openstack-nova | 10:24 | |
| johnthetubaguy | bauzas: ah, it was moshele, I remember now | 10:24 |
| *** sdague has joined #openstack-nova | 10:25 | |
| *** gszasz has joined #openstack-nova | 10:26 | |
| *** shaohe_feng has quit IRC | 10:27 | |
| moshele | sfinucan: you can put them on the agenda https://etherpad.openstack.org/p/sriov_meeting_agenda, although I am less familiar with NUMA code | 10:27 |
| *** sridharg has joined #openstack-nova | 10:27 | |
| *** hieulq has joined #openstack-nova | 10:27 | |
| *** ljxiash has joined #openstack-nova | 10:27 | |
| *** shaohe_feng has joined #openstack-nova | 10:27 | |
| openstackgerrit | Alexis Lee proposed openstack/nova: Notifications: simple schema https://review.openstack.org/312010 | 10:27 |
| *** markvoelker has quit IRC | 10:28 | |
| *** EinstCrazy has quit IRC | 10:28 | |
| *** hieulq has quit IRC | 10:29 | |
| *** hieulq has joined #openstack-nova | 10:29 | |
| *** jlanoux_ has joined #openstack-nova | 10:30 | |
| *** jlanoux has quit IRC | 10:30 | |
| *** ljxiash has quit IRC | 10:31 | |
| *** jlanoux has joined #openstack-nova | 10:32 | |
| *** jlanoux_ has quit IRC | 10:33 | |
| *** Kennan has quit IRC | 10:35 | |
| *** hieulq has quit IRC | 10:35 | |
| *** Kennan has joined #openstack-nova | 10:35 | |
| *** hieulq has joined #openstack-nova | 10:35 | |
| *** yuywz has quit IRC | 10:36 | |
| *** shaohe_feng has quit IRC | 10:37 | |
| *** shaohe_feng has joined #openstack-nova | 10:37 | |
| *** yuywz has joined #openstack-nova | 10:39 | |
| *** hieulq has quit IRC | 10:40 | |
| *** jlanoux_ has joined #openstack-nova | 10:43 | |
| *** jlanoux has quit IRC | 10:46 | |
| *** shaohe_feng has quit IRC | 10:47 | |
| *** shaohe_feng has joined #openstack-nova | 10:48 | |
| *** prashup has joined #openstack-nova | 10:49 | |
| prashup | Hi guys, I was looking into nova neutrno notification communication. And during some experimentation i stuck with a problem. Instance creation always stuck in spawning stage. | 10:50 |
| prashup | as libvirt emits stop event ; | 10:50 |
| *** gongysh has quit IRC | 10:50 | |
| prashup | can someone tell me, why it is happening, or even better some source which i can go through to better understand this stuff ?? | 10:51 |
| prashup | any help is appreciated :) | 10:51 |
| *** markus_z has joined #openstack-nova | 10:51 | |
| *** ducttape_ has joined #openstack-nova | 10:53 | |
| *** smatzek has joined #openstack-nova | 10:54 | |
| *** thorst has joined #openstack-nova | 10:55 | |
| *** sacharya has joined #openstack-nova | 10:55 | |
| openstackgerrit | sahid proposed openstack/nova: network: handle unauthorized exception from neutron https://review.openstack.org/312014 | 10:57 |
| *** shaohe_feng has quit IRC | 10:57 | |
| *** shaohe_feng has joined #openstack-nova | 10:58 | |
| *** Kennan has quit IRC | 10:58 | |
| *** ducttape_ has quit IRC | 10:58 | |
| *** yuywz has quit IRC | 10:59 | |
| *** lucasagomes is now known as lucas-hungry | 10:59 | |
| *** sacharya has quit IRC | 10:59 | |
| *** rtheis has joined #openstack-nova | 11:01 | |
| *** thorst has quit IRC | 11:02 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: introduces module to handle domain xml migration https://review.openstack.org/299490 | 11:02 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: add "get_job_info" to Guest's object https://review.openstack.org/250740 | 11:02 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: add method to configure max downtime when migrating https://review.openstack.org/250739 | 11:02 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: add "abort_job" to Guest's object https://review.openstack.org/250738 | 11:02 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: add method "migrate" to Guest's object https://review.openstack.org/250365 | 11:02 |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: config: remove deprecated ironic.client_log_level https://review.openstack.org/312006 | 11:03 |
| *** shaohe_feng has quit IRC | 11:08 | |
| *** shaohe_feng has joined #openstack-nova | 11:09 | |
| *** kengo_sakai has quit IRC | 11:09 | |
| *** yuywz has joined #openstack-nova | 11:12 | |
| *** zz_dimtruck is now known as dimtruck | 11:13 | |
| *** Kennan has joined #openstack-nova | 11:14 | |
| *** shaohe_feng has quit IRC | 11:18 | |
| *** shaohe_feng has joined #openstack-nova | 11:19 | |
| *** yassine has joined #openstack-nova | 11:21 | |
| *** yassine is now known as Guest93371 | 11:21 | |
| *** dimtruck is now known as zz_dimtruck | 11:23 | |
| *** links has quit IRC | 11:24 | |
| *** markvoelker has joined #openstack-nova | 11:24 | |
| *** josecastroleon has joined #openstack-nova | 11:27 | |
| *** shaohe_feng has quit IRC | 11:28 | |
| *** markvoelker has quit IRC | 11:29 | |
| *** shaohe_feng has joined #openstack-nova | 11:29 | |
| *** prashup has quit IRC | 11:32 | |
| *** mjura has quit IRC | 11:32 | |
| sdague | BobBall: are you fixing the ci for xenserver? | 11:33 |
| *** sudipto has quit IRC | 11:34 | |
| *** shaohe_feng has quit IRC | 11:38 | |
| *** shaohe_feng has joined #openstack-nova | 11:39 | |
| moshele | jaypipes: can you review some pci cleanup https://review.openstack.org/#/c/310950/ and pci: Deprecate is_new from pci request ? | 11:41 |
| moshele | jaypipes: I mean https://review.openstack.org/#/c/310950/ and https://review.openstack.org/#/c/310990/ | 11:41 |
| jaypipes | moshele: morning! | 11:43 |
| jaypipes | moshele: yup, will do that right now. | 11:43 |
| jaypipes | moshele: I think dansmith did a review on those yesterday. will hop on it now. | 11:43 |
| *** thorst has joined #openstack-nova | 11:43 | |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP resource-pools/placement api https://review.openstack.org/293104 | 11:44 |
| *** mjura has joined #openstack-nova | 11:45 | |
| openstackgerrit | jichenjc proposed openstack/nova: Complete method Verification of os-hypervisors https://review.openstack.org/312024 | 11:46 |
| *** shaohe_feng has quit IRC | 11:49 | |
| *** shaohe_feng has joined #openstack-nova | 11:50 | |
| *** kengo_sakai has joined #openstack-nova | 11:52 | |
| *** lucas-hungry is now known as lucasagomes | 11:54 | |
| *** egarbade- has quit IRC | 11:55 | |
| *** maurosr has quit IRC | 11:55 | |
| *** ekhugen has quit IRC | 11:55 | |
| *** iceyao has joined #openstack-nova | 11:57 | |
| *** baoli has joined #openstack-nova | 11:57 | |
| *** shaohe_feng has quit IRC | 11:59 | |
| openstackgerrit | Moshe Levi proposed openstack/nova: pci: Allow to update pci_requests in instance_extra https://review.openstack.org/312028 | 11:59 |
| *** shaohe_feng has joined #openstack-nova | 12:00 | |
| *** rfolco has joined #openstack-nova | 12:00 | |
| *** kengo_sakai has quit IRC | 12:01 | |
| openstackgerrit | jichenjc proposed openstack/nova: Complete method Verification of consoles https://review.openstack.org/312029 | 12:01 |
| BobBall | sdague: Yes - should be fixed - we're running some more tests before asking to be re-enabled | 12:01 |
| *** links has joined #openstack-nova | 12:03 | |
| *** maurosr has joined #openstack-nova | 12:04 | |
| *** egarbade- has joined #openstack-nova | 12:04 | |
| BobBall | sdague: I can confirm that we're passing the most recent few tests | 12:05 |
| *** ekhugen has joined #openstack-nova | 12:05 | |
| BobBall | sdague: if you have the powah to re-add us to the Nova CI group, that would be great. I've disabled voting for failing jobs (so only passing jobs will vote) and that might help catch some of the rechecks. | 12:06 |
| *** ludovic__ has joined #openstack-nova | 12:06 | |
| *** svenkat has joined #openstack-nova | 12:07 | |
| *** ducttape_ has joined #openstack-nova | 12:09 | |
| *** shaohe_feng has quit IRC | 12:09 | |
| *** shaohe_feng has joined #openstack-nova | 12:10 | |
| *** abalutoiu has joined #openstack-nova | 12:11 | |
| *** egarbade- has quit IRC | 12:11 | |
| johnthetubaguy | BobBall: I think I added you back in now | 12:12 |
| *** ekhugen has quit IRC | 12:12 | |
| BobBall | thanks johnthetubaguy | 12:13 |
| *** maurosr has quit IRC | 12:13 | |
| *** aysyd has joined #openstack-nova | 12:15 | |
| BobBall | FYI I've rechecked all of the devstack failures in the last 72 hours so hopefully it should start churning out passes | 12:15 |
| BobBall | I'll re-enable failed votes later today if I can confirm the pass rate is high enough | 12:15 |
| *** artom has joined #openstack-nova | 12:16 | |
| *** laszloh has quit IRC | 12:17 | |
| *** krtaylor has quit IRC | 12:18 | |
| *** shaohe_feng has quit IRC | 12:19 | |
| openstackgerrit | jichenjc proposed openstack/python-novaclient: Enhance descriptions for get and clear password https://review.openstack.org/312035 | 12:20 |
| *** shaohe_feng has joined #openstack-nova | 12:20 | |
| *** krtaylor has joined #openstack-nova | 12:20 | |
| *** fragatina has joined #openstack-nova | 12:22 | |
| *** markvoelker has joined #openstack-nova | 12:22 | |
| *** burgerk has joined #openstack-nova | 12:23 | |
| *** egarbade- has joined #openstack-nova | 12:23 | |
| *** maurosr has joined #openstack-nova | 12:23 | |
| *** kylek3h has quit IRC | 12:23 | |
| *** ekhugen has joined #openstack-nova | 12:23 | |
| *** apoorv has quit IRC | 12:24 | |
| *** Oku_OS is now known as Oku_OS-away | 12:25 | |
| *** amotoki has joined #openstack-nova | 12:27 | |
| *** ivase has joined #openstack-nova | 12:27 | |
| *** amotoki has quit IRC | 12:27 | |
| openstackgerrit | jichenjc proposed openstack/nova: not use mox for conductor/test_conductor.py https://review.openstack.org/306385 | 12:27 |
| *** ivase has quit IRC | 12:28 | |
| *** ivase has joined #openstack-nova | 12:28 | |
| moshele | dansmith: hi, I have a question regarding you comment https://review.openstack.org/#/c/307124/ | 12:28 |
| *** shaohe_feng has quit IRC | 12:30 | |
| *** shaohe_feng has joined #openstack-nova | 12:30 | |
| *** laszloh has joined #openstack-nova | 12:32 | |
| jaypipes | alaski: around? | 12:33 |
| *** mjura has quit IRC | 12:33 | |
| *** _prad_ has joined #openstack-nova | 12:33 | |
| openstackgerrit | jichenjc proposed openstack/nova: Replace fake_utils by using Fixture https://review.openstack.org/230954 | 12:33 |
| *** _prad_ is now known as pradk_ | 12:33 | |
| *** tbachman has quit IRC | 12:34 | |
| *** hkominos has quit IRC | 12:36 | |
| *** ducttape_ has quit IRC | 12:37 | |
| *** BlackDex has quit IRC | 12:38 | |
| *** shaohe_feng has quit IRC | 12:40 | |
| *** luis5tb has joined #openstack-nova | 12:41 | |
| *** shaohe_feng has joined #openstack-nova | 12:41 | |
| alaski | jaypipes: yes | 12:45 |
| *** toanster is now known as toan | 12:45 | |
| *** mjura has joined #openstack-nova | 12:46 | |
| jaypipes | alaski: sorry mate, never mind :) it was some random transient error about Unauthorized in the cells dsvm gate job that seems to have gone away... | 12:47 |
| danpb | jaypipes: could you take a look at this os-vif change https://review.openstack.org/#/c/308407/ - i want to add reno support for our first release | 12:47 |
| cdent | jaypipes, alaski: I reckon we need a manhattan project to spend 110% of time on "random transient errors" otherwise we are just fooling ourselves | 12:48 |
| alaski | yeah, random errors are concerning | 12:50 |
| *** shaohe_feng has quit IRC | 12:50 | |
| cdent | At the api level when using an environ['nova.context'] does that context already include an active transaction, such that the entire api request is in a transaction, or do I need to do that explicitly? | 12:50 |
| *** shaohe_feng has joined #openstack-nova | 12:51 | |
| *** EinstCrazy has joined #openstack-nova | 12:51 | |
| johnthetubaguy | what do you mean by "transaction"? the DB one? | 12:51 |
| cdent | johnthetubaguy: yeah | 12:51 |
| jaypipes | danpb: done! | 12:51 |
| danpb | cheers | 12:52 |
| johnthetubaguy | cdent: I thought each DB API call is a transaction, and the engine facade decorators deal with that | 12:52 |
| cdent | yes, that appears to be true, but I was wondering if there's anything already happening at the HTTP API level | 12:52 |
| johnthetubaguy | nothing structured that I know of | 12:53 |
| cdent | k, thanks | 12:53 |
| * cdent considers his options | 12:53 | |
| openstackgerrit | Daniel Berrange proposed openstack/os-vif: Start using reno for release notes https://review.openstack.org/308407 | 12:53 |
| *** xyang1 has joined #openstack-nova | 12:53 | |
| johnthetubaguy | the sync stuff is normally a single write db api call, and the async stuff is usually madness | 12:53 |
| *** kylek3h has joined #openstack-nova | 12:53 | |
| johnthetubaguy | not sure that helps your thinking there though | 12:53 |
| jaypipes | cdent: you should not be doing anything directly with the database in the Nova API layer. Use the associated nova.objects interfaces, which should be decorated appropriately with @nova.db.sqlalchemy.api.api_context_manager.writer | 12:54 |
| cdent | jaypipes: that's what I'm doing most of the time but what about the case where I want to update multiple Inventory objects? | 12:55 |
| johnthetubaguy | jaypipes: right, API -> objects (->) db api | 12:55 |
| jaypipes | cdent: grep for api_context_writer in nova/objects.host_mapping.py for an example. | 12:55 |
| johnthetubaguy | cdent: a bulk update on the list object? | 12:55 |
| *** sacharya has joined #openstack-nova | 12:55 | |
| cdent | hmm, how to explain this | 12:56 |
| cdent | a PUT to an existing resource pool contains multiple inventories | 12:56 |
| cdent | some of those inventories may already exist | 12:56 |
| cdent | some may not | 12:56 |
| cdent | for each one of the inventory needs to either be create()d or save()d | 12:57 |
| *** kengo_sakai has joined #openstack-nova | 12:57 | |
| *** rlrossit has joined #openstack-nova | 12:57 | |
| johnthetubaguy | could you not add them to the list, and call save_all() on a list object? | 12:57 |
| *** efried has joined #openstack-nova | 12:57 | |
| *** raildo-afk is now known as raildo | 12:58 | |
| johnthetubaguy | actually, the deletes are kinda dumb in there... | 12:58 |
| cdent | I don't know. | 12:58 |
| cdent | This is new ground for me and I'm struggling to find reasonable examples to steal from | 12:58 |
| tpatzig | mikal: alaski: johnthetubaguy: as discussed during unconference session. could you have a look at https://review.openstack.org/#/c/267673 | 12:58 |
| jaypipes | cdent: see this patch for inspiration: | 12:58 |
| * cdent enjoys inspiration | 12:59 | |
| jaypipes | https://review.openstack.org/#/c/211288/ | 12:59 |
| *** vtech has quit IRC | 12:59 | |
| jaypipes | cdent: we discussed how to do this with the PCIDeviceList batch save of the same kind.. | 12:59 |
| jaypipes | cdent: dansmith has ideas in there. | 12:59 |
| johnthetubaguy | right, thats the sort of thing I was meaning, nice | 12:59 |
| * cdent will read | 13:00 | |
| *** vtech has joined #openstack-nova | 13:00 | |
| jaypipes | cdent: note that it is a remoteable class method on the PciDeviceList objetc. | 13:00 |
| johnthetubaguy | not sure if thats a single transaction though... | 13:00 |
| *** kylek3h_ has joined #openstack-nova | 13:00 | |
| *** shaohe_feng has quit IRC | 13:00 | |
| *** sacharya has quit IRC | 13:00 | |
| cdent | It often feels like all this complexity exists because of a solution to a problem that should have been there in the first place | 13:01 |
| * cdent marches on | 13:01 | |
| *** kengo_sakai has quit IRC | 13:01 | |
| *** shaohe_feng has joined #openstack-nova | 13:01 | |
| cdent | shouldn't! | 13:01 |
| * cdent sighs | 13:01 | |
| *** Oku_OS-away is now known as Oku_OS | 13:02 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: introduces module to handle domain xml migration https://review.openstack.org/299490 | 13:02 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: add "get_job_info" to Guest's object https://review.openstack.org/250740 | 13:02 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: add method to configure max downtime when migrating https://review.openstack.org/250739 | 13:02 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: add "abort_job" to Guest's object https://review.openstack.org/250738 | 13:02 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: add method "migrate" to Guest's object https://review.openstack.org/250365 | 13:02 |
| johnthetubaguy | cdent: it might be easier to model it as an object field, and you just save the object, depending on how that looks? | 13:02 |
| *** kylek3h has quit IRC | 13:03 | |
| *** efried has quit IRC | 13:03 | |
| cdent | is it kosher to save_batch on a batch that mixes saves and created? | 13:04 |
| cdent | creates? | 13:04 |
| *** edmondsw has joined #openstack-nova | 13:04 | |
| johnthetubaguy | save_or_create_batch doesn't have the same ring to it | 13:05 |
| openstackgerrit | sahid proposed openstack/nova: network: handle unauthorized exception from neutron https://review.openstack.org/312014 | 13:05 |
| *** itamarl has quit IRC | 13:05 | |
| cdent | johnthetubaguy: :) | 13:05 |
| *** itamarl has joined #openstack-nova | 13:05 | |
| *** lpetrut has quit IRC | 13:06 | |
| alaski | tpatzig: yep, adding that review to my list | 13:06 |
| *** efried has joined #openstack-nova | 13:06 | |
| bauzas | alaski: jaypipes: dear friends, you should be hitting https://review.openstack.org/#/q/topic:bp/check-destination-on-migrations-newton :D | 13:06 |
| tpatzig | alaski: great, thanks! | 13:07 |
| * alaski opens more tabs for bauzas | 13:07 | |
| bauzas | :p | 13:07 |
| bauzas | Firefox, help me ! | 13:07 |
| openstackgerrit | Peng Li proposed openstack/nova: fix wrong key name in test code https://review.openstack.org/312051 | 13:07 |
| openstackgerrit | sahid proposed openstack/nova: network: handle unauthorized exception from neutron https://review.openstack.org/312014 | 13:07 |
| * cdent takes a think walk | 13:08 | |
| *** ljxiash has joined #openstack-nova | 13:09 | |
| bauzas | jaypipes: cdent: I'm about to review the allocations patches, etc. Which one should I review first ? https://review.openstack.org/#/c/282442/ is -2 | 13:09 |
| bauzas | jaypipes: cdent: btw. I should have time by N-2 to help you guys for implementing the resource-* if the above check-dests patches are merged :) | 13:09 |
| *** shaohe_feng has quit IRC | 13:11 | |
| *** shaohe_feng has joined #openstack-nova | 13:11 | |
| *** mdrabe has joined #openstack-nova | 13:11 | |
| *** itamarl has quit IRC | 13:11 | |
| *** mylu has joined #openstack-nova | 13:11 | |
| *** EinstCrazy has quit IRC | 13:12 | |
| *** mylu has quit IRC | 13:12 | |
| *** ljxiash has quit IRC | 13:13 | |
| openstackgerrit | Sharat Sharma proposed openstack/nova: Added the optional filters to the request parameters. https://review.openstack.org/310420 | 13:13 |
| sahid | hello jaypipes, when you have a moment if you can make this one merged that would be nice https://review.openstack.org/#/c/301247/ | 13:13 |
| *** efried has quit IRC | 13:13 | |
| *** esracelik has quit IRC | 13:13 | |
| *** busterswt has joined #openstack-nova | 13:14 | |
| *** esracelik has joined #openstack-nova | 13:14 | |
| *** busterswt has quit IRC | 13:14 | |
| *** zz_dimtruck is now known as dimtruck | 13:14 | |
| *** jamesdenton has joined #openstack-nova | 13:15 | |
| openstackgerrit | Matthew Booth proposed openstack/nova: Add a lock() context manager to image backend https://review.openstack.org/279625 | 13:15 |
| openstackgerrit | Matthew Booth proposed openstack/nova: Rename Raw backend to NoBacking https://review.openstack.org/279626 | 13:15 |
| openstackgerrit | Matthew Booth proposed openstack/nova: Introduce ImageCacheLocalPool https://review.openstack.org/279669 | 13:15 |
| openstackgerrit | Matthew Booth proposed openstack/nova: Remove fake_imagebackend.Raw and cleanup dependent tests https://review.openstack.org/267661 | 13:15 |
| openstackgerrit | Matthew Booth proposed openstack/nova: Rename Image.check_image_exists to Image.exists() https://review.openstack.org/270998 | 13:15 |
| openstackgerrit | Matthew Booth proposed openstack/nova: Remove deprecated option libvirt.remove_unused_kernels https://review.openstack.org/265886 | 13:15 |
| openstackgerrit | Matthew Booth proposed openstack/nova: Only attempt to inject files if the injection disk exists https://review.openstack.org/250872 | 13:15 |
| *** josecastroleon has quit IRC | 13:15 | |
| *** efried has joined #openstack-nova | 13:16 | |
| *** akshai has joined #openstack-nova | 13:18 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Resize API operation passing down original RequestSpec https://review.openstack.org/312009 | 13:19 |
| *** shaohe_feng has quit IRC | 13:21 | |
| *** shaohe_feng has joined #openstack-nova | 13:22 | |
| *** READ10 has joined #openstack-nova | 13:22 | |
| *** sc68cal has quit IRC | 13:22 | |
| *** ducttape_ has joined #openstack-nova | 13:22 | |
| *** efried has quit IRC | 13:23 | |
| *** dimtruck is now known as zz_dimtruck | 13:24 | |
| *** salv-orlando has joined #openstack-nova | 13:25 | |
| *** smatzek has quit IRC | 13:25 | |
| *** sc68cal has joined #openstack-nova | 13:25 | |
| *** sc68cal has quit IRC | 13:27 | |
| *** ducttape_ has quit IRC | 13:28 | |
| *** eharney has joined #openstack-nova | 13:28 | |
| *** sc68cal has joined #openstack-nova | 13:28 | |
| dansmith | moshele: here now | 13:28 |
| *** vtech has quit IRC | 13:28 | |
| *** vtech has joined #openstack-nova | 13:29 | |
| dansmith | cdent: jaypipes: https://review.openstack.org/#/c/211288/6 | 13:30 |
| dansmith | I never really finished it.. put it up for feedback, but I don't know what shape it's in | 13:31 |
| *** haukebruno has quit IRC | 13:31 | |
| *** shaohe_feng has quit IRC | 13:31 | |
| *** shaohe_feng has joined #openstack-nova | 13:32 | |
| *** salv-orlando has quit IRC | 13:32 | |
| *** pwnall has joined #openstack-nova | 13:34 | |
| *** ccarmack has joined #openstack-nova | 13:35 | |
| *** efried has joined #openstack-nova | 13:35 | |
| sfinucan | johnthetubaguy: Do you know if mriedem had an Etherpad RE: work for EPA features (NUMA, SRIOV, PCI-passthrough etc.)? | 13:35 |
| johnthetubaguy | sfinucan: there is this from the summit session: https://etherpad.openstack.org/p/newton-nova-performance-vms | 13:36 |
| *** b3nt_pin is now known as beagles | 13:36 | |
| openstackgerrit | Sharat Sharma proposed openstack/nova: [api-ref] Added host_ip and hypervisor_type parameters to Show Hypervisor Details API. https://review.openstack.org/310096 | 13:36 |
| *** jlanoux_ has quit IRC | 13:36 | |
| *** sc68cal has quit IRC | 13:36 | |
| openstackgerrit | sahid proposed openstack/nova: pci: make sure device relationships are kept in memory https://review.openstack.org/301859 | 13:36 |
| openstackgerrit | sahid proposed openstack/nova: pci: Make sure PF is 'available' when last VF is freed https://review.openstack.org/303012 | 13:36 |
| openstackgerrit | sahid proposed openstack/nova: pci: related updates are done without DB lookups https://review.openstack.org/301860 | 13:36 |
| johnthetubaguy | sfinucan: helping me out the feature classification matrix might be a good way to formalise the issues | 13:37 |
| bauzas | ++ | 13:37 |
| openstackgerrit | Alexis Lee proposed openstack/nova: Notifications: simple schema https://review.openstack.org/312010 | 13:38 |
| *** sree has quit IRC | 13:39 | |
| *** sree has joined #openstack-nova | 13:40 | |
| sdague | alaski: https://bugs.launchpad.net/nova/+bug/1577311 is somewhat indicative of the confusion on the db configs now. It might be worth figuring out where to document that more. | 13:41 |
| openstack | Launchpad bug 1577311 in OpenStack Compute (nova) "OperationalError: (pymysql.err.OperationalError)" [Undecided,Invalid] | 13:41 |
| sdague | does anyone have a reference for the bug where we were over deleting metadata because of the case issue? | 13:41 |
| *** shaohe_feng has quit IRC | 13:41 | |
| *** palexster1 is now known as palexster | 13:42 | |
| *** rook has joined #openstack-nova | 13:42 | |
| *** sheel has joined #openstack-nova | 13:42 | |
| markus_z | sdague: This one? https://bugs.launchpad.net/nova/+bug/1538011 | 13:42 |
| openstack | Launchpad bug 1538011 in OpenStack Compute (nova) "Metadata keys are not case sensitive" [High,In progress] - Assigned to Augustina Ragwitz (auggy) | 13:42 |
| *** shaohe_feng has joined #openstack-nova | 13:42 | |
| sdague | markus_z: no, I've got that one | 13:43 |
| alaski | sdague: yeah. That also seems like something that should cause some noise on service startup | 13:43 |
| *** baoli has quit IRC | 13:44 | |
| sdague | ah, there is an embedded link - https://bugs.launchpad.net/nova/+bug/1535224 | 13:44 |
| openstack | Launchpad bug 1535224 in OpenStack Compute (nova) "meta data of nova instance with case sensitive has problem when deleting" [Undecided,In progress] - Assigned to Sheel Rana (ranasheel2000) | 13:44 |
| sdague | that's the one I wanted | 13:44 |
| *** baoli has joined #openstack-nova | 13:45 | |
| *** sree has quit IRC | 13:45 | |
| markus_z | right, there was a second one, which is similar | 13:46 |
| *** mriedem has joined #openstack-nova | 13:46 | |
| sheel | https://bugs.launchpad.net/nova/+bug/1538011 | 13:46 |
| openstack | Launchpad bug 1538011 in OpenStack Compute (nova) "Metadata keys are not case sensitive" [High,In progress] - Assigned to Augustina Ragwitz (auggy) | 13:46 |
| sheel | markus_z: ^^ | 13:46 |
| alaski | ccarmack: not sure if you got an answer on https://github.com/openstack-dev/devstack/blob/master/lib/nova#L647 yesterday. Those should not be necessary for cellsv2, and anything that is necessary for v2 would be necessary for Nova and not a special case | 13:47 |
| openstackgerrit | Andrea Rosa proposed openstack/nova: Adding device_present field to bdm table https://review.openstack.org/268053 | 13:47 |
| *** burt has joined #openstack-nova | 13:48 | |
| *** sc68cal has joined #openstack-nova | 13:48 | |
| mriedem | andrearosa: you've got a bad copyright in that migration ^ | 13:50 |
| *** smatzek has joined #openstack-nova | 13:50 | |
| andrearosa | mriedem: thanks I'll fix it | 13:50 |
| *** tbachman has joined #openstack-nova | 13:51 | |
| *** mrkz has joined #openstack-nova | 13:51 | |
| ccarmack | alaski: thanks, that's what I expected. Can we remove setting those configs for newton? I assume we would still need to run nova-manage to set up the db. | 13:51 |
| *** shaohe_feng has quit IRC | 13:52 | |
| *** ametts has joined #openstack-nova | 13:52 | |
| sfinucan | johnthetubaguy: Cheers. Forwarded to sriov meeting, as you saw | 13:52 |
| *** PaulMurray has joined #openstack-nova | 13:52 | |
| *** shaohe_feng has joined #openstack-nova | 13:52 | |
| alaski | ccarmack: those are there for cellsv1 so they can't be removed until that goes away. if anything that could be renamed to indicate it's v1 specific to avoid confusion. | 13:53 |
| ccarmack | alaski: but for newton, wouldn't they go away? | 13:54 |
| *** diana_clarke has quit IRC | 13:54 | |
| *** _diana_ is now known as diana_clarke | 13:54 | |
| alaski | ccarmack: removing them would break the gate-tempest-dsvm-cells job | 13:55 |
| ccarmack | alaski: ok, so we would keep that job for mitaka and earlier? or would it also stay for newton? | 13:56 |
| alaski | v2 isn't replacing v1 yet so nothing really changes there in Newton | 13:56 |
| PaulMurray | live migration meeting in 3 mins on #openstack-meeting-3 PaulMurray paul-carlton2 eliqiao shaohe_feng pkoniszewski jlanoux yuntongjin mdbooth kashyap kmartin claudiub tobasco tdurakov scheuran abhishekk | 13:57 |
| alaski | ccarmack: once we have a migration path tested for v1->v2 we can look at removing those configs and that test job | 13:57 |
| ccarmack | alaski: maybe I'm assuming too much about newton | 13:57 |
| alaski | A lot is going to happen this cycle, but I'm not sure we'll be able to deprecate v1 yet | 13:57 |
| *** matrohon has quit IRC | 13:57 | |
| ccarmack | ok, got it | 13:58 |
| *** esracelik has quit IRC | 13:58 | |
| *** burgerk has quit IRC | 13:58 | |
| ccarmack | …. thanks | 13:58 |
| *** diogogmt has quit IRC | 13:58 | |
| ccarmack | so change devstack to be more clear its a v1 config | 13:58 |
| openstackgerrit | Sean Dague proposed openstack/nova-specs: Add spec for lower case metadata keys only https://review.openstack.org/311529 | 13:59 |
| *** vladikr has joined #openstack-nova | 13:59 | |
| tojuvone_ | sdague,dansmith: I'd still wonder if there is any way to filter in DB level for host_state. I only came to conclusion to filter result after host_state is known. | 13:59 |
| alaski | ccarmack: yeah, that should help avoid future confusion all around | 14:00 |
| dansmith | tojuvone_: I'm sure it's doable, but it's probably worth figuring out how nasty it will be | 14:00 |
| moshele | efried: hi | 14:00 |
| *** pwnall has quit IRC | 14:00 | |
| moshele | efried: I will ping you in 30 min is that cool | 14:00 |
| *** tonytan4ever has joined #openstack-nova | 14:01 | |
| efried | moshele, thanks very much. | 14:01 |
| *** toure is now known as toure_ | 14:01 | |
| tojuvone_ | dansmith: ok, problem is also if filter with multivalue. And that not all is from DB but also SG driver. | 14:02 |
| *** shaohe_feng has quit IRC | 14:02 | |
| dansmith | hmm, yeah | 14:02 |
| *** josecastroleon has joined #openstack-nova | 14:02 | |
| *** shaohe_feng has joined #openstack-nova | 14:02 | |
| tojuvone_ | dansmith: In my mind that made it not that feasible. Taking host_status first apart as service attributes.. it'll be fairly interesting | 14:03 |
| *** sfinucan has quit IRC | 14:04 | |
| *** mageshgv has joined #openstack-nova | 14:06 | |
| *** Jeffrey4l has joined #openstack-nova | 14:06 | |
| *** sfinucan has joined #openstack-nova | 14:06 | |
| *** rama_y has joined #openstack-nova | 14:07 | |
| markus_z | new contributor patch: https://review.openstack.org/#/c/312051/1 | 14:09 |
| *** links has quit IRC | 14:10 | |
| *** lucasagomes is now known as lucas-post-offic | 14:11 | |
| lbeliveau | sfinucan: do we want NFV doc in admin-guide or part of developer-nova ? | 14:12 |
| *** shaohe_feng has quit IRC | 14:12 | |
| lbeliveau | mriedem: ^ opinion ? | 14:13 |
| sfinucan | lbeliveau: Probably the admin-doc, as it's these folks who need the info most | 14:13 |
| *** shaohe_feng has joined #openstack-nova | 14:13 | |
| sfinucan | Do you know if there's documentation for similar feature in the dev docs? | 14:13 |
| lbeliveau | sfinucan: there is one that explains all the filters, and it's not in admin guide | 14:14 |
| sfinucan | lbeliveau: That should probably change so. Link? | 14:15 |
| *** ludovic__ has quit IRC | 14:15 | |
| lbeliveau | sfinucan: http://docs.openstack.org/developer/nova/filter_scheduler.html | 14:16 |
| sfinucan | Ah yes | 14:17 |
| sfinucan | So admin docs for "why you'd want to use this feature", and dev docs for "how this feature works under the hood"? | 14:17 |
| bauzas | lbeliveau: sfinucan: we agreed during the nova session to update the networking guide, right? | 14:17 |
| sfinucan | w/ focus on the former for this cycle | 14:17 |
| *** itamarl has joined #openstack-nova | 14:18 | |
| *** sudipto has joined #openstack-nova | 14:18 | |
| *** itamarl has quit IRC | 14:18 | |
| jaypipes | sahid: done | 14:18 |
| *** cdelatte has joined #openstack-nova | 14:18 | |
| lbeliveau | bauzas: I didn't catch that at the summit, so http://docs.openstack.org/admin-guide/compute-networking-nova.html ? | 14:19 |
| lbeliveau | bauzas: wrong ... this one http://docs.openstack.org/liberty/networking-guide/adv-config-sriov.html | 14:20 |
| bauzas | lbeliveau: yup, this one | 14:20 |
| bauzas | at least that's the minutes of https://etherpad.openstack.org/p/newton-nova-performance-vms | 14:20 |
| lbeliveau | yeap | 14:20 |
| lbeliveau | bauzas: what about numa doc ? | 14:21 |
| bauzas | lbeliveau: that's something missing AFAIK | 14:21 |
| bauzas | probably not related to that SRIOV docs tho | 14:21 |
| bauzas | rather in the compute admin doc | 14:22 |
| lbeliveau | agreed | 14:22 |
| *** shaohe_feng has quit IRC | 14:22 | |
| *** shaohe_feng has joined #openstack-nova | 14:23 | |
| *** jlanoux has joined #openstack-nova | 14:24 | |
| sahid | jaypipes: thank you | 14:25 |
| moshele | efried: free now | 14:26 |
| *** mylu has joined #openstack-nova | 14:27 | |
| jaypipes | bauzas: finishing sahid's serial ports patch series reviews then will do the check-dests stuff | 14:27 |
| *** BjoernT has joined #openstack-nova | 14:27 | |
| openstackgerrit | Balazs Gibizer proposed openstack/nova-specs: Versioned notification transformation https://review.openstack.org/286675 | 14:27 |
| *** laszloh has quit IRC | 14:27 | |
| bauzas | jaypipes: no worries man | 14:28 |
| *** READ10 has quit IRC | 14:28 | |
| bauzas | jaypipes: that's just something I'd see to be implemented and merged soon, so I could help you, guys | 14:28 |
| *** diogogmt has joined #openstack-nova | 14:30 | |
| openstackgerrit | sahid proposed openstack/nova: network: handle unauthorized exception from neutron https://review.openstack.org/312014 | 14:32 |
| *** diogogmt has quit IRC | 14:32 | |
| *** shaohe_feng has quit IRC | 14:33 | |
| *** shaohe_feng has joined #openstack-nova | 14:33 | |
| mriedem | lbeliveau: i'd say admin docs | 14:35 |
| cdent | jaypipes, dansmith: I'm able to figure out how to an InventoryList.save_batch but I'm unclear on how best to integrate that with the conflict handling describe at line 449 in https://review.openstack.org/#/c/300176/2/specs/newton/approved/generic-resource-pools.rst | 14:35 |
| tojuvone_ | dansmith: I could try filter on DB for MAINTENANCE and DOWN. UP and UNKNOWN filter only when SG driver called. Anyhow multivalue filter that has DB and SG driver part could not work as checked in different place (maybe not even make sense) | 14:35 |
| mriedem | lbeliveau: there is some stuff documented here http://docs.openstack.org/admin-guide/compute-flavors.html#extra-specs and here http://docs.openstack.org/liberty/networking-guide/adv-config-sriov.html | 14:36 |
| cdent | jaypipes, dansmith: presumably I'd want to do that within the transaction | 14:36 |
| dansmith | cdent: as an annoying tangent, it sounds like jaypipes and alaski want to move inventory and allocation to the API DB right away, so we probably need to get that change stacked under all of this | 14:36 |
| cdent | hawt | 14:37 |
| jaypipes | cdent: correct. raising an exception during any of those model updates will rollback the transaction. | 14:37 |
| *** irenab_ has joined #openstack-nova | 14:37 | |
| jaypipes | dansmith: hey there, good point. | 14:37 |
| dansmith | tojuvone_: yeah, I think you'd only be able to efficiently do it based on the db state | 14:37 |
| *** pradk_ is now known as pradk | 14:37 | |
| cdent | jaypipes: so on InventoryList we want some kind of save_batch_with_attention_to_allocations() | 14:37 |
| *** burgerk has joined #openstack-nova | 14:37 | |
| alaski | dansmith: cdent yeah it would be easier to not have the migration chasing all of these changes | 14:38 |
| dansmith | yeah, | 14:38 |
| cdent | alaski: I think we're probalby safe on that account, these changes I'm talking about are quite some distance from seeing the light of merge | 14:38 |
| dansmith | and inventory is the easiest to move because it's updated regularly | 14:38 |
| *** cfriesen has joined #openstack-nova | 14:38 | |
| sahid | mriedem: just a pointer since i know that you are involved on the network area https://review.openstack.org/#/c/312014/ | 14:38 |
| alaski | cdent: okay | 14:39 |
| mriedem | sahid: ok, we have a few duplicates of that bug | 14:39 |
| *** irenab has quit IRC | 14:39 | |
| *** irenab_ is now known as irenab | 14:39 | |
| alaski | if doffm is interested I'll leave it to him, if he's not and no one else volunteers I can jump on it | 14:39 |
| alaski | after I knock out the policy stuff | 14:40 |
| cdent | that'll just be a couple minutes work, yeah? | 14:40 |
| * cdent snorts | 14:40 | |
| jaypipes | cdent: InventoryList.save() should be the method name. issue an update statement after checking that the new total - new reserved > sum(allocations) for that resource class on that resource provider. | 14:41 |
| sahid | mriedem: i didn't notice that | 14:41 |
| mriedem | sahid: i think https://bugs.launchpad.net/nova/+bug/1571722 | 14:41 |
| openstack | Launchpad bug 1571722 in OpenStack Compute (nova) "Neutronclient failures throw unhelpful "Unexpected Exception"" [Medium,Confirmed] | 14:41 |
| cdent | jaypipes: dansmith doesn't like save(), see: https://review.openstack.org/#/c/211288/6/nova/objects/pci_device.py | 14:41 |
| dansmith | cdent: man, you make such sweeping generalized statements | 14:42 |
| dansmith | but in this case, yes | 14:42 |
| jaypipes | lol, settle down children. | 14:42 |
| mriedem | sahid: but it's the same idea really, handling a 401 from neutron | 14:42 |
| cdent | dansmith: ? | 14:42 |
| dansmith | I vowed to never have a save() method ona list object | 14:42 |
| *** jdillaman has joined #openstack-nova | 14:43 | |
| jaypipes | cdent: I prefer just save() and leave the fact that it does a batched/atomic save of all listed objects as an implementation detail, but I defer to superdan on the name. | 14:43 |
| *** ducttape_ has joined #openstack-nova | 14:43 | |
| *** shaohe_feng has quit IRC | 14:43 | |
| *** shaohe_feng has joined #openstack-nova | 14:43 | |
| openstackgerrit | Pushkar Umaranikar proposed openstack/nova: Centralize vif,xenpool & vol_utils config options https://review.openstack.org/302446 | 14:44 |
| cdent | I have no idea what's going on. <- that's a sweeping generalized statement that is also true | 14:44 |
| dansmith | jaypipes: the reason I hate save() on a list is because it doesn't make the list in the DB match the list you have (i.e. add,delete,update *everything* so that the two lists are consistent) | 14:45 |
| dansmith | so save_batch() conveys "save all these things please" to me | 14:45 |
| sahid | mriedem: hum yes we can merge all of them probably | 14:45 |
| dansmith | which is vastly different | 14:45 |
| *** baoli has quit IRC | 14:45 | |
| jaypipes | cdent: see dansmith's comment to johnthetubaguy on that file about using main_context_manager (or api_context_manager in this case...) | 14:45 |
| jaypipes | dansmith: sure, I have no prob with save_batch() :) | 14:45 |
| *** mageshgv has quit IRC | 14:46 | |
| jaypipes | dansmith: and I agree with your reasoning there. | 14:46 |
| dansmith | okay | 14:46 |
| * dansmith lowers his nerf gun | 14:46 | |
| *** e0ne has quit IRC | 14:47 | |
| openstackgerrit | Ryan Moore proposed openstack/nova: Add better help to rdp options https://review.openstack.org/253556 | 14:47 |
| *** mtanino has joined #openstack-nova | 14:47 | |
| cdent | I'll wing something and you guys can fight about in the review if you're still in the mood | 14:47 |
| *** READ10 has joined #openstack-nova | 14:47 | |
| *** josecastroleon has quit IRC | 14:47 | |
| *** baoli has joined #openstack-nova | 14:49 | |
| cdent | ah, one more question dansmith: If you have a list of objects, and only some already exist, what is the idiomatic way to choose save or create? | 14:49 |
| *** sacharya has joined #openstack-nova | 14:49 | |
| dansmith | cdent: unfortunately, we just have a few that have a "save_or_create" type method, which I don't like | 14:50 |
| dansmith | BUT | 14:50 |
| dansmith | if you're going to do the batch thing, | 14:50 |
| dansmith | I say you do: try: obj.save() except NotFound: obj.create() | 14:50 |
| dansmith | inside the batch method, which is good I think | 14:50 |
| dansmith | honestly, | 14:50 |
| sdague | mriedem: https://review.openstack.org/#/c/312108/ is the cherry pick of the nova-api death | 14:51 |
| dansmith | we should establish this batching thing as a pattern I think, and do things like that properly | 14:51 |
| cdent | okay, cool, good, that's what I was kinda preferring | 14:51 |
| *** vishwanathj has joined #openstack-nova | 14:51 | |
| *** fawadkhaliq has joined #openstack-nova | 14:51 | |
| cdent | feels like python | 14:51 |
| openstackgerrit | Sean Dague proposed openstack/nova: Test that nova-api ignores paste failures, but continues on https://review.openstack.org/309600 | 14:51 |
| openstackgerrit | Sean Dague proposed openstack/nova: Fail to start nova-api if no APIs were able to be started https://review.openstack.org/309601 | 14:51 |
| *** mdrabe has quit IRC | 14:52 | |
| mriedem | sdague: ok, i'd also want dansmith's chagnes that were stacked on top in master to be backported also | 14:52 |
| *** doug-fish has joined #openstack-nova | 14:52 | |
| mriedem | sdague: see https://review.openstack.org/#/q/topic:1572495 | 14:52 |
| dansmith | mriedem: just those two would be enough though right? | 14:52 |
| sdague | mriedem: ok, those got wedged because of not inheriting parent correctly in gerrit | 14:52 |
| sdague | I just rebased and approved them | 14:52 |
| mriedem | dansmith: which two? the bottom ones? https://review.openstack.org/#/q/Id9a6f5e38b017f4687eb245fd238104379d437d6,n,z ? | 14:53 |
| *** shaohe_feng has quit IRC | 14:53 | |
| *** baoli has quit IRC | 14:53 | |
| sdague | I'll propose the cherry picks once those land | 14:53 |
| *** shaohe_feng has joined #openstack-nova | 14:53 | |
| dansmith | mriedem: actually, you're right, the first opens up a new issue, so we should get them all | 14:53 |
| mriedem | yar | 14:54 |
| dansmith | so, ignore me | 14:54 |
| *** mtanino has quit IRC | 14:57 | |
| openstackgerrit | Pushkar Umaranikar proposed openstack/nova: Centralize vif,xenpool & vol_utils config options https://review.openstack.org/302446 | 14:57 |
| *** cleong has joined #openstack-nova | 14:57 | |
| *** mageshgv has joined #openstack-nova | 14:57 | |
| mriedem | sahid: overall looks good, but we need a functional test for that high level extension.py change in the api | 14:57 |
| *** kengo_sakai has joined #openstack-nova | 14:58 | |
| *** vtech has quit IRC | 14:58 | |
| *** doug-fis_ has joined #openstack-nova | 14:58 | |
| *** vtech has joined #openstack-nova | 14:58 | |
| efried | moshele, sorry, was AFK. Yesterday mriedem and svenkat were discussing CI requirements for SR-IOV. Well, hinting at. I wanted to get some clarification on the expectations there. (<==thorst) | 14:58 |
| *** jdurgin1 has joined #openstack-nova | 14:58 | |
| *** baoli has joined #openstack-nova | 14:59 | |
| moshele | efried: what are you planning to test? is it port direct? | 14:59 |
| *** tbachman has quit IRC | 15:00 | |
| efried | moshele, VFs connected to the instance. | 15:00 |
| *** doug-fi__ has joined #openstack-nova | 15:00 | |
| *** fragatina has quit IRC | 15:00 | |
| efried | moshele, my interest is for nova-powervm btw. | 15:00 |
| moshele | efried: so it VF Passthrough or with port direct ? | 15:01 |
| *** ljxiash has joined #openstack-nova | 15:01 | |
| thorst | moshele: VF Passthrough | 15:01 |
| *** doug-fish has quit IRC | 15:01 | |
| openstackgerrit | Gage Hugo proposed openstack/nova: Fixed multiple issues with server_concepts page https://review.openstack.org/294284 | 15:02 |
| moshele | thorst: I would like that you will be able to run tempest tests | 15:02 |
| *** kengo_sakai has quit IRC | 15:02 | |
| *** doug-fis_ has quit IRC | 15:02 | |
| *** pwnall has joined #openstack-nova | 15:02 | |
| sahid | mriedem: yes i will do that and address your other comments, thanks | 15:02 |
| thorst | efried: I think svenkat was interested in Live Migration for both libvirt driver and powervm. Where as you are interested in base SR-IOV enablement for powervm first. | 15:02 |
| sdague | auggy / mriedem / melwitt - https://review.openstack.org/#/c/311529/ is I think what we came to with final session | 15:03 |
| moshele | thorst: see https://github.com/openstack/tempest/tree/master/tempest/scenario | 15:03 |
| *** PaulMurray has quit IRC | 15:03 | |
| *** shaohe_feng has quit IRC | 15:03 | |
| *** shaohe_feng has joined #openstack-nova | 15:04 | |
| openstackgerrit | Andrea Rosa proposed openstack/nova: Adding device_present field to bdm table https://review.openstack.org/268053 | 15:04 |
| *** tbachman has joined #openstack-nova | 15:04 | |
| *** doug-fi__ has quit IRC | 15:05 | |
| *** mdrabe has joined #openstack-nova | 15:05 | |
| thorst | moshele: We currently run the tempest tests for powervm. I think we would need to reconfigure our CI to use networks with SR-IOV (unless there are tests specific to SR-IOV in tempest). I'm looking through now to find any specific to sr-iov | 15:05 |
| *** ljxiash has quit IRC | 15:05 | |
| *** diogogmt has joined #openstack-nova | 15:06 | |
| andrearosa | sdague: dansmith I don't know if you remeber the missing functional tests you asked for the attach/detach vols for shelved instances, but the change has been around for a while, please have a look when you have time https://review.openstack.org/#/c/277463/ | 15:06 |
| sdague | andrearosa: ok, cool, I'll add it to my review list | 15:06 |
| moshele | thorst: we use say network with SR-IOV you mean port direct right? | 15:06 |
| *** doug-fish has joined #openstack-nova | 15:07 | |
| *** mylu has quit IRC | 15:08 | |
| moshele | thorst: see https://github.com/openstack/tempest/blob/2211ed8c9c2a95849b4475b498ada91bbbe94e74/tempest/config.py#L549 | 15:08 |
| *** BlackDex has joined #openstack-nova | 15:08 | |
| thorst | moshele: Yeah, I think in that we would want the port_vnic_type to be 'direct' | 15:09 |
| *** hemnafk is now known as hemna | 15:09 | |
| *** BlackDex has quit IRC | 15:10 | |
| moshele | thorst: so this is what we test in the Mellanox CI already | 15:11 |
| *** lpetrut has joined #openstack-nova | 15:11 | |
| *** tesseract- has quit IRC | 15:12 | |
| mriedem | thorst: efried: we have a list of what's tested today in the etherpad here https://etherpad.openstack.org/p/newton-nova-performance-vms | 15:12 |
| doffm | ccarmack: The cellsv2 etherpad from the summit is at: https://etherpad.openstack.org/p/newton-nova-cells | 15:12 |
| mriedem | sheel: fyi https://review.openstack.org/#/c/311529/ | 15:13 |
| mriedem | sdague: thanks, will review later | 15:13 |
| thorst | mriedem: Thx, efried svenkat and I will review for the base PVM SR-IOV CI | 15:13 |
| *** shaohe_feng has quit IRC | 15:14 | |
| sheel | mriedem: thanks for pointer | 15:14 |
| mriedem | thorst: also see the performance vm's session recap in the dev list | 15:14 |
| *** shaohe_feng has joined #openstack-nova | 15:14 | |
| doffm | melwitt, alaski: I know its all a blur, but did we agree on how we are going to do the Cell-switching in the API at the summit? I remember a very short discussion at the cells fishbowl, but don't see anything in the etherpad. | 15:15 |
| doffm | Except about pagination. :) | 15:15 |
| *** tbachman has quit IRC | 15:16 | |
| thorst | mriedem: thx | 15:16 |
| alaski | doffm: for the mq we agreed to do it all at once in get_cell_client | 15:16 |
| mriedem | doffm: see the cells v2 dev list recap | 15:16 |
| mriedem | yeah, basically, turn it on and see what breaks | 15:16 |
| mriedem | was the agreement | 15:16 |
| *** unicell has joined #openstack-nova | 15:17 | |
| *** kolesovdv has quit IRC | 15:18 | |
| alaski | we didn't touch on the db side of it. we could potentially get fancy with something in db/api but the db code in objects will need to be done piecemeal | 15:18 |
| kashyap | mriedem: When you get a moment, for Kilo, to retain the instance metadata for shared storage, should this work? In the 'else' block, instead of image_meta = {}, just do: 'image_meta = self.image_api.get(context, image_ref)' -- https://github.com/openstack/nova/blob/stable/kilo/nova/compute/manager.py#L3029,L3032 | 15:18 |
| mriedem | kashyap: is that a security bug? if not, we're freezing kilo this week for the final release | 15:19 |
| mriedem | and then EOL | 15:19 |
| kashyap | mriedem: Darn, thought you'd ask it, no not a security bug | 15:19 |
| *** links has joined #openstack-nova | 15:19 | |
| mriedem | kashyap: also, as noted on the stable/mitaka backport, i'd like to see us land a tempest patch for rebuild of a volume-backed instance in tempest before doing all of those backports | 15:19 |
| kashyap | mriedem: Trying to fix for downstream | 15:19 |
| *** zz_dimtruck is now known as dimtruck | 15:19 | |
| mriedem | maybe not land, but at least propose and have a dependency on the nova change to show it's working | 15:19 |
| kashyap | mriedem: I was not planning to fix for Kilo upstream, given that it's Security-Supported only | 15:20 |
| kashyap | mriedem: Okay, will investigate that. | 15:20 |
| *** prateek has quit IRC | 15:21 | |
| *** rbradfor has joined #openstack-nova | 15:21 | |
| mriedem | kashyap: for your question, if image_ref is None doing a get on it will fail | 15:21 |
| sdague | tojuvone_: where is the base spec adding reason to the host object? https://review.openstack.org/#/c/310510/ is only about the bits leaking through to server. | 15:21 |
| mriedem | kashyap: you'd have to get the image meta out of the instance object's system_metadata i'd think | 15:21 |
| melwitt | alaski: iiuc, the thing about doing it all at once in get_cell_client is it would persist the connection info in the instance (not in the db sense) where in the beginning we were trying to make sure the connection info only stayed in the request context a minimum time | 15:22 |
| kashyap | mriedem: I see, okay, will read the relevant Kilo code to get a sense of it | 15:22 |
| *** sdake has joined #openstack-nova | 15:23 | |
| *** rama_y has quit IRC | 15:23 | |
| *** shaohe_feng has quit IRC | 15:24 | |
| doffm | alaski: I'm going to update the 'miscellaneous' db migrations spec to include the work that needs to be done passing the keypair down to the cell. | 15:24 |
| *** shaohe_feng has joined #openstack-nova | 15:24 | |
| doffm | alaski: Is it worth moving keypairs out to its own spec and splitting 'miscellaneous' in to keyparis and instance groups specs. | 15:25 |
| *** pushkaru has joined #openstack-nova | 15:25 | |
| doffm | Certificates / agent builds are probably going away anyway. | 15:25 |
| alaski | doffm: it's looking like it | 15:25 |
| alaski | if keypairs is more than just a db migration having a separate spec would be good | 15:26 |
| *** sdake_ has joined #openstack-nova | 15:26 | |
| doffm | Ok, will do that today before there are any real reviews on the miscellaneous spec. | 15:26 |
| doffm | There is probably discussion to be had about HOW to pass down the keypair. instance_extra sounds fine to me but... who knows. | 15:26 |
| *** belmoreira has quit IRC | 15:26 | |
| *** tonytan_brb has joined #openstack-nova | 15:27 | |
| doffm | Separate spec for instance groups makes sense too. As there are things to be sorted out there. | 15:27 |
| alaski | melwitt: I need to refresh myself on that flow to ensure we haven't been talking nonsense here. That's a great point about the lifetime of the transport switch | 15:27 |
| alaski | doffm: agreed | 15:28 |
| *** sdake has quit IRC | 15:28 | |
| *** zenoway has joined #openstack-nova | 15:29 | |
| *** tbachman has joined #openstack-nova | 15:29 | |
| *** ducttape_ has quit IRC | 15:30 | |
| *** kmartin has quit IRC | 15:30 | |
| *** ducttape_ has joined #openstack-nova | 15:31 | |
| *** tonytan4ever has quit IRC | 15:31 | |
| tojuvone_ | sdague: My misunderstanding. I caughtg | 15:31 |
| auggy | sdague: cool i'll add myself as reviewer and take a look | 15:32 |
| tojuvone_ | sdague: I thought just this disabled_reason from service to be exposed in servers | 15:32 |
| *** zenoway has quit IRC | 15:32 | |
| cfriesen | bauzas: jaypipes: I think I hit a race calling LibvirtDriver._get_disk_over_committed_size_total(). https://bugs.launchpad.net/nova/+bug/1577642 Basically it's looping over instances without holding the resource semaphore. | 15:33 |
| openstack | Launchpad bug 1577642 in OpenStack Compute (nova) "race between disk_available_least and instance operations" [Undecided,New] | 15:33 |
| *** sudipto has quit IRC | 15:34 | |
| *** shaohe_feng has quit IRC | 15:34 | |
| *** shaohe_feng has joined #openstack-nova | 15:35 | |
| bauzas | cfriesen: otp but looking at the bug report, sounds a good probability | 15:35 |
| bauzas | cfriesen: because we're using the dict given by the virt driver | 15:36 |
| *** mtanino has joined #openstack-nova | 15:36 | |
| *** salv-orlando has joined #openstack-nova | 15:36 | |
| bauzas | which couldn't be in the synchronous section | 15:36 |
| *** iceyao has quit IRC | 15:39 | |
| *** tbachman has quit IRC | 15:39 | |
| *** vtech has quit IRC | 15:39 | |
| *** mylu has joined #openstack-nova | 15:41 | |
| *** lucas-post-offic is now known as lucasagomes | 15:41 | |
| *** tbachman has joined #openstack-nova | 15:44 | |
| *** shaohe_feng has quit IRC | 15:44 | |
| *** mylu has quit IRC | 15:45 | |
| *** shaohe_feng has joined #openstack-nova | 15:45 | |
| *** kmartin has joined #openstack-nova | 15:45 | |
| openstackgerrit | Andrea Rosa proposed openstack/nova: Mask the token used to allow access to consoles https://review.openstack.org/220622 | 15:46 |
| *** mjura has quit IRC | 15:47 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Skip network allocation if 'none' is requested https://review.openstack.org/309638 | 15:47 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: nova-net: handle 'auto' network request in allocate_for_instance https://review.openstack.org/297906 | 15:47 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Skip network validation if explicitly requesting no networks https://review.openstack.org/309602 | 15:47 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Add helpers to NetworkRequest(List) objects for auto/none cases https://review.openstack.org/297871 | 15:47 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: neutron: validate auto-allocate is available https://review.openstack.org/309574 | 15:48 |
| openstackgerrit | Andrea Rosa proposed openstack/nova: Mask the token used to allow access to consoles https://review.openstack.org/220622 | 15:48 |
| *** pradk has quit IRC | 15:49 | |
| *** tonytan_brb has quit IRC | 15:49 | |
| *** haplo37 has joined #openstack-nova | 15:49 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: Acquire TCP ports for console during live migration https://review.openstack.org/275801 | 15:49 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: release serial console ports when destroying guests https://review.openstack.org/301247 | 15:49 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: Add serial ports to the migration data object https://review.openstack.org/215102 | 15:49 |
| *** paul-carlton2 has quit IRC | 15:50 | |
| mriedem | i could have sworn someone was removing this block of code that isn't used https://github.com/openstack/nova/blob/master/nova/compute/resource_tracker.py#L734-L739 | 15:51 |
| *** doug-fish has quit IRC | 15:52 | |
| *** ssurana has joined #openstack-nova | 15:52 | |
| *** doug-fish has joined #openstack-nova | 15:52 | |
| *** tonytan4ever has joined #openstack-nova | 15:52 | |
| *** tonytan4ever has quit IRC | 15:53 | |
| *** doug-fish has quit IRC | 15:53 | |
| *** shaohe_feng has quit IRC | 15:55 | |
| *** shaohe_feng has joined #openstack-nova | 15:55 | |
| cburgess | mriedem From Yesterday re https://review.openstack.org/#/c/309832/ Yes we saw it. Doesn't make sense to us. Nic -1 with comments | 15:56 |
| *** vilobhmm11 has joined #openstack-nova | 15:56 | |
| *** abalutoiu has quit IRC | 15:57 | |
| *** vilobhmm111 has joined #openstack-nova | 15:59 | |
| *** fragatina has joined #openstack-nova | 16:01 | |
| *** vilobhmm11 has quit IRC | 16:01 | |
| tojuvone_ | sdague: So my understanding was to just use disabled_reason as it is already there, but obviously I was wrong? | 16:02 |
| jaypipes | cfriesen: looks like a nice catch, dude. you proposing a fix? | 16:02 |
| *** dmk0202 has quit IRC | 16:03 | |
| *** salv-orlando has quit IRC | 16:03 | |
| *** rena9067 has joined #openstack-nova | 16:03 | |
| sfinucan | jaypipes: Hey. Have you any specs/BPs that detail the work needed to get "nested" resource trackers (or at least mention the future need for them?) | 16:04 |
| *** BjoernT has quit IRC | 16:05 | |
| *** earlephilhower has quit IRC | 16:05 | |
| *** shaohe_feng has quit IRC | 16:05 | |
| *** doug-fish has joined #openstack-nova | 16:05 | |
| *** doug-fish has quit IRC | 16:05 | |
| *** shaohe_feng has joined #openstack-nova | 16:06 | |
| *** doug-fish has joined #openstack-nova | 16:06 | |
| jaypipes | sfinucan: no, not yet. it's on my todo list right after finishing the dynamic-resource-classes one I'm about 70% done with. | 16:06 |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: config options: improve help text of database (related) options (1/3) https://review.openstack.org/301238 | 16:06 |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: config options: centralize section "database" + "api_database" https://review.openstack.org/263804 | 16:06 |
| cfriesen | jaypipes: not yet...not seeing an obvious fix. Can we somehow move the scan within the code path where we hold COMPUTE_RESOURCE_SEMAPHORE? | 16:07 |
| *** doug-fish has quit IRC | 16:08 | |
| *** doug-fish has joined #openstack-nova | 16:08 | |
| *** ssurana has quit IRC | 16:09 | |
| cfriesen | (and do we have any documentation anywhere on exactly what the rules/expectations are for when we should/shouldn't hold that semaphore?) | 16:09 |
| sfinucan | jaypipes: Lovely. Could you stick me on that review when you do get it out, please :) | 16:09 |
| jaypipes | sfinucan: of course :) | 16:10 |
| jaypipes | cfriesen: not sure yet. will need to look deeper into it | 16:10 |
| *** mylu has joined #openstack-nova | 16:11 | |
| *** e0ne has joined #openstack-nova | 16:11 | |
| *** roeyc has joined #openstack-nova | 16:12 | |
| *** mylu has quit IRC | 16:13 | |
| *** fragatin_ has joined #openstack-nova | 16:13 | |
| *** ssurana has joined #openstack-nova | 16:13 | |
| *** earlephilhower has joined #openstack-nova | 16:13 | |
| cfriesen | jaypipes: me too. also looking at a related issue around disk space tracking not properly taking into account backing file space requirements | 16:14 |
| *** mtecer has quit IRC | 16:14 | |
| *** shaohe_feng has quit IRC | 16:15 | |
| *** shaohe_feng has joined #openstack-nova | 16:16 | |
| *** mylu has joined #openstack-nova | 16:16 | |
| *** fragatina has quit IRC | 16:17 | |
| *** mylu has quit IRC | 16:18 | |
| *** pbrooko has joined #openstack-nova | 16:21 | |
| *** apoorvad has joined #openstack-nova | 16:21 | |
| *** apoorvad has quit IRC | 16:22 | |
| *** fawadkhaliq has quit IRC | 16:22 | |
| kashyap | mriedem: Is that what you meant previously when you said "get the image meta out of the instance object's system_metadata": http://paste.openstack.org/show/495978/ | 16:22 |
| *** apoorvad has joined #openstack-nova | 16:22 | |
| mriedem | kashyap: yeah, assuming the instance is an object | 16:22 |
| mriedem | in kilo | 16:22 |
| kashyap | mriedem: Hmm, clues as to how would I know that? | 16:23 |
| mriedem | kashyap: is anything above in the method calling instance.*? | 16:23 |
| * kashyap looks | 16:23 | |
| kashyap | mriedem: I see a bunch of instance.host, instance.node, instance.save(), | 16:24 |
| mriedem | ok, you're good then | 16:24 |
| *** jdurgin1 has quit IRC | 16:25 | |
| kashyap | mriedem: Thanks! Now to go check the unit tests :-) | 16:25 |
| *** shaohe_feng has quit IRC | 16:25 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: rt: remove unused image_meta parameter https://review.openstack.org/312153 | 16:26 |
| *** shaohe_feng has joined #openstack-nova | 16:26 | |
| sdague | tojuvone_: so we have disabled_reason on the Service, but not on the ComputeNode... it sort of feels like this should also go on the compute node directly | 16:27 |
| *** unicell has quit IRC | 16:28 | |
| johnthetubaguy | sdague: isn't that because the compute node moves to the scheduler, the service record stays in Nova? | 16:28 |
| johnthetubaguy | sdague: although, I totally feel like I have lost track of the split there now | 16:29 |
| mriedem | kashyap: so on your change, maybe a tempest test isn't needed, since we aren't going to get into the case where image_ref is None when doing a rebuild, since the image_ref is required for rebuild | 16:29 |
| mriedem | it's only None for evacuate (or recreate in that code), and we can't test that in Tempest | 16:30 |
| sdague | johnthetubaguy: right, but in this case it's specifically about the desire to mark a physical host as in maintenance | 16:30 |
| mriedem | so a tempest test for build won't even hit that path that you're fixing... | 16:30 |
| kashyap | mriedem: Yeah, was trying to study if it really is the case. Thanks for confirming | 16:30 |
| mriedem | rebuild of a volume-backed instance is still good, but not required here | 16:30 |
| kashyap | mriedem: Do you mind adding a comment to this effect, with this new detail on the Mitaka backport? | 16:30 |
| mriedem | yeah | 16:30 |
| kashyap | Merci | 16:30 |
| kashyap | dansmith: I'm about to print this and paste on the wall next to me -- http://www.danplanet.com/blog/2016/03/03/evacuate-in-nova-one-command-to-confuse-us-all/ | 16:31 |
| *** unicell has joined #openstack-nova | 16:31 | |
| *** ssurana has quit IRC | 16:31 | |
| dansmith | kashyap: I permit you a license to print and display one copy | 16:32 |
| kashyap | :-) Thanks! | 16:32 |
| mriedem | kashyap: you could be hero by getting an evacuate test in tempest that uses the multinode job | 16:32 |
| mriedem | *be a hero | 16:32 |
| johnthetubaguy | sdague: true, thinking about the DB model, its confusing (https://github.com/openstack/nova/blob/master/nova/db/sqlalchemy/models.py#L76) | 16:33 |
| cfriesen | Can we rename evacuate as 'resurrect' ? | 16:33 |
| mriedem | no | 16:33 |
| cburgess | lol | 16:33 |
| sdague | cfriesen: that was my proposal from vancouver :) | 16:33 |
| cburgess | Can we add a facility to the API to alias commands? | 16:33 |
| cfriesen | c'mon, it'd just be a little microversion change. :) | 16:34 |
| mriedem | just alias them in novaclient | 16:34 |
| kashyap | mriedem: :-) Evacuate has got me 'wrapped around the axle' (learnt this phrase recently, trying to see if it works in real) . | 16:34 |
| mriedem | hell, we have nova image-create which is really snapshot | 16:34 |
| sdague | mriedem: as a microversion change, I think it's totally a good idea | 16:34 |
| mriedem | sdague: i feel like we went over this before and decided it wasn't worth the churn | 16:35 |
| kashyap | mriedem: Yeah, that was the first thing I tried to figure when I began playing with Nova | 16:35 |
| sdague | mriedem: I think the fact that we have to explain it every couple of months all over again, it is | 16:35 |
| cfriesen | and why is "live migration" a whole separate command from "migrate" instead of an option flag? | 16:35 |
| johnthetubaguy | sdague: mriedem: I like it as an addition, its the removing the old one that seems not worth the churn, except maybe the CLI | 16:35 |
| markus_z | cfriesen: http://www.danplanet.com/blog/2016/03/03/evacuate-in-nova-one-command-to-confuse-us-all/ | 16:36 |
| *** shaohe_feng has quit IRC | 16:36 | |
| cfriesen | markus_z: yeah, I've got it open | 16:36 |
| *** doug-fish has quit IRC | 16:36 | |
| johnthetubaguy | cfriesen: because it was implemented separately in two different drivers, thats not a good answer, but thats the real reason | 16:36 |
| markus_z | oh, whoops, should have read the backscroll | 16:36 |
| *** pbrooko has quit IRC | 16:37 | |
| *** shaohe_feng has joined #openstack-nova | 16:37 | |
| *** doug-fish has joined #openstack-nova | 16:37 | |
| *** doug-fish has quit IRC | 16:37 | |
| johnthetubaguy | dansmith: I don't think regular users usually get migrate, they do get resize though | 16:37 |
| *** ssurana has joined #openstack-nova | 16:38 | |
| johnthetubaguy | but that might just be me | 16:38 |
| dansmith | johnthetubaguy: I'm missing context I think | 16:38 |
| johnthetubaguy | dansmith: oops, yeah, http://www.danplanet.com/blog/2016/03/03/evacuate-in-nova-one-command-to-confuse-us-all/ | 16:39 |
| *** jlanoux has quit IRC | 16:39 | |
| moshele | mriedem: when you have time can you review the patches in the section of ready for core review L28 in https://etherpad.openstack.org/p/sriov_meeting_agenda | 16:39 |
| dansmith | johnthetubaguy: that whole document is mostly written from the perspective of the admin user, right? | 16:39 |
| mriedem | moshele: sure | 16:39 |
| dansmith | johnthetubaguy: also, I thought some people expose migrate just as a way to "find me a new neighbor, mine is noisy" | 16:40 |
| johnthetubaguy | dansmith: ah, gotcha | 16:40 |
| johnthetubaguy | dansmith: thats true I guess | 16:41 |
| cfriesen | how about the ability to put hosts in "maintenance_mode", that only works with xen | 16:41 |
| markus_z | cfriesen: worst - API - ever! | 16:41 |
| johnthetubaguy | cfriesen: to be clear, that doesn't actually do anything | 16:42 |
| cfriesen | in xen it does...it triggers what should be "evacuate" | 16:42 |
| cfriesen | ie it pushes everything off | 16:42 |
| mriedem | does this work for people for the mid-cycle rsvp? http://goo.gl/forms/MxrriHsABq | 16:43 |
| mriedem | i've never done a google form for rsvp | 16:43 |
| mriedem | not sure if you have to have an account or not | 16:43 |
| johnthetubaguy | cfriesen: it just changes the icon in the manager screen | 16:43 |
| *** ssurana has quit IRC | 16:43 | |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: Config options: base path configuration: improve help text https://review.openstack.org/299236 | 16:44 |
| cfriesen | johnthetubaguy: I'm looking at virt.xenapi.host.Host.host_maintenance_mode() | 16:45 |
| cfriesen | it sure seems to be doing live migrations | 16:45 |
| *** unicell has quit IRC | 16:45 | |
| mriedem | jaypipes: sdague: alaski: doffm: dansmith: johnthetubaguy: bauzas: test this out please http://goo.gl/forms/MxrriHsABq | 16:45 |
| openstackgerrit | Julian proposed openstack/nova-specs: Creating Nova Servers should emit Nova Notifications that contain scheduler hint information. https://review.openstack.org/312169 | 16:45 |
| * mriedem gets lunch | 16:45 | |
| cfriesen | mriedem: I can see it at least | 16:46 |
| *** shaohe_feng has quit IRC | 16:46 | |
| dansmith | mriedem: wfm | 16:46 |
| *** shaohe_feng has joined #openstack-nova | 16:47 | |
| *** ssurana has joined #openstack-nova | 16:47 | |
| openstackgerrit | Moshe Levi proposed openstack/nova: pci: Move PCI devices and PCI requests into migration context https://review.openstack.org/307124 | 16:50 |
| diana_clarke | mikal: Thanks for reviewing the database poison warning patches. Much appreciated! | 16:51 |
| *** paul-carlton2 has joined #openstack-nova | 16:52 | |
| *** moshele has quit IRC | 16:52 | |
| tojuvone_ | sdague: Ok, make sense. So would it then be to introduce a new param. Have API to set it and expose to server also or still use the disabled_reason. | 16:52 |
| cdent | gibi: the notification meeting moved to where? | 16:52 |
| alaski | mriedem: fancy, and works for me | 16:52 |
| *** rbridgeman has joined #openstack-nova | 16:52 | |
| gibi | cdent: openstack-meeting-4 | 16:53 |
| cdent | thanks! | 16:53 |
| sfinucan | lbeliveau: Damn, that was quick :) https://review.openstack.org/#/c/312168/1 | 16:54 |
| gibi | notification subteam meeting start in 5 minutes on #openstack-meeting-4 | 16:54 |
| lbeliveau | sfinucan: your review was quick as well :) | 16:55 |
| lbeliveau | thanks | 16:55 |
| lbeliveau | sfinucan: working on pci passthrough now, this is not addressed at all atm | 16:55 |
| jaypipes | mriedem: done. I'm a vegan who works for The Bomb, Inc. | 16:56 |
| sfinucan | lbeliveau: Networking guide also? | 16:56 |
| mriedem | jaypipes: gdi | 16:56 |
| *** shaohe_feng has quit IRC | 16:56 | |
| mriedem | now if there ends up being only 1 vegan respondant i'll have to remember to tell intel it was a joke | 16:56 |
| mriedem | you bastard | 16:56 |
| lbeliveau | sfinucan: not sure ... a PCI device can be something else than a networking device, like Coleto Creek, a GPU, etc . | 16:56 |
| jaypipes | mriedem: hey, you asked for a test. :P | 16:57 |
| mriedem | i'm testing in production | 16:57 |
| jaypipes | mriedem: don't complain when you get test data ;) | 16:57 |
| *** doug-fish has joined #openstack-nova | 16:57 | |
| *** akshai has quit IRC | 16:57 | |
| *** shaohe_feng has joined #openstack-nova | 16:57 | |
| jaypipes | mriedem: I can edit it :) | 16:57 |
| mriedem | i will complain regardless | 16:57 |
| jaypipes | k, good. :L) | 16:57 |
| sfinucan | lbeliveau: my thoughts exactly | 16:58 |
| *** Jeffrey4l has quit IRC | 16:58 | |
| *** palexster has quit IRC | 16:58 | |
| sfinucan | I'm mostly copied this [1], but still unsure where to "slot it in" https://01.org/openstack/blogs/stephenfin/2015/smarter-cpu-pinning | 16:58 |
| *** palexster has joined #openstack-nova | 16:58 | |
| *** paul-carlton2 has quit IRC | 16:58 | |
| *** akshai has joined #openstack-nova | 16:59 | |
| *** Oku_OS is now known as Oku_OS-away | 17:00 | |
| *** doug-fish has quit IRC | 17:00 | |
| *** tonytan4ever has joined #openstack-nova | 17:01 | |
| lbeliveau | sfinican: part of it should be in the flavor guide, and perhaps a new guide under compute ? | 17:02 |
| *** jistr has quit IRC | 17:04 | |
| gibi | rlrossit: ping | 17:04 |
| sfinucan | lbeliveau: Yeah, both would be good. The flavor guide to explain how, and a new guide to explain the implications of it | 17:04 |
| *** doug-fish has joined #openstack-nova | 17:06 | |
| *** shaohe_feng has quit IRC | 17:06 | |
| *** ericksonsantos has joined #openstack-nova | 17:06 | |
| *** shaohe_feng has joined #openstack-nova | 17:07 | |
| *** aghadge has joined #openstack-nova | 17:07 | |
| *** mvk has quit IRC | 17:08 | |
| *** sambetts is now known as sambetts|afk | 17:09 | |
| *** mriedem1 has joined #openstack-nova | 17:10 | |
| vilobhmm111 | jaypipes, alaski : please point me to the blueprint/spec which has details about generation-id thing | 17:10 |
| *** mriedem has quit IRC | 17:11 | |
| *** sfinucan has quit IRC | 17:11 | |
| *** paul-carlton2 has joined #openstack-nova | 17:11 | |
| *** doug-fish has quit IRC | 17:12 | |
| *** earlephilhower has quit IRC | 17:13 | |
| jaypipes | vilobhmm111: https://review.openstack.org/#/c/283253/ | 17:14 |
| vilobhmm111 | jaypipes : thank you ! :) | 17:14 |
| jaypipes | any time :) | 17:14 |
| *** earlephilhower has joined #openstack-nova | 17:15 | |
| *** zhenq has joined #openstack-nova | 17:16 | |
| *** roeyc has quit IRC | 17:16 | |
| *** doug-fish has joined #openstack-nova | 17:16 | |
| *** shaohe_feng has quit IRC | 17:17 | |
| *** unicell has joined #openstack-nova | 17:17 | |
| *** pushkaru has quit IRC | 17:17 | |
| *** shaohe_feng has joined #openstack-nova | 17:17 | |
| *** lpetrut has quit IRC | 17:17 | |
| *** ducttape_ has quit IRC | 17:20 | |
| mriedem1 | melwitt: where is that jpenick character? | 17:20 |
| *** mriedem1 is now known as mriedem | 17:20 | |
| melwitt | mriedem1: he's been sick with a cold that started on saturday | 17:20 |
| *** roeyc has joined #openstack-nova | 17:22 | |
| mriedem | melwitt: ok, i watched https://youtu.be/RqpeiVB5xKg last night and wondered if he had seen http://specs.openstack.org/openstack/nova-specs/specs/mitaka/implemented/oslo_db-enginefacade.html | 17:22 |
| mriedem | sounds like yahoo is on juno, and ^ is in mitaka, so would probably be awhile before we could get any feedback on performance improvements with the DB | 17:22 |
| dansmith | melwitt: ah, so he's why I'm sick now I guess :/ | 17:23 |
| alaski | dansmith: melwitt I commented on https://review.openstack.org/#/c/274955 . I think we need a slightly different approach to do an all at once switchover | 17:23 |
| melwitt | mriedem: ah, okay. yeah, my understanding is mitaka upgrade isn't going to be soon. they're still trying to get on liberty | 17:24 |
| melwitt | dansmith: uh oh :( | 17:24 |
| *** roeyc has quit IRC | 17:25 | |
| *** shaohe_feng has quit IRC | 17:27 | |
| *** lucasagomes is now known as lucas-dinner | 17:27 | |
| *** salv-orlando has joined #openstack-nova | 17:27 | |
| *** shaohe_feng has joined #openstack-nova | 17:28 | |
| *** ccarmack has quit IRC | 17:29 | |
| *** e0ne has quit IRC | 17:31 | |
| *** ijw has joined #openstack-nova | 17:31 | |
| doffm | melwitt, alaski: I guess I don't properly understand the difference between the 'all at once' method and the 'one-at-a-time' method of switching. Whether we use 'target_cell' or 'get_cell_client', we still need to create a decorator and add it to lots of methods somewhere in the code. | 17:32 |
| *** ijw has quit IRC | 17:32 | |
| *** ijw has joined #openstack-nova | 17:32 | |
| markus_z | nova-bugs team meeting in 30 min in #openstack-meeting-4. New contributors are welcome as well. https://wiki.openstack.org/wiki/Meetings/Nova/BugsTeam | 17:32 |
| melwitt | doffm, alaski: my understanding is (and I could be wrong) is that one would set the connection info in the context at Instance object creation time so it carries it around everywhere | 17:33 |
| *** vishwanathj has quit IRC | 17:33 | |
| mriedem | sdague: we talked about an api-ref review push next week right? | 17:33 |
| mriedem | s/push/sprint/ | 17:33 |
| alaski | doffm: true. our conversation started with an assumption that melwitts patch was merged, but we could backtrack to doing it all at once if that was modified | 17:33 |
| melwitt | doffm, alaski: and switching can be done by just looking at the instance's context. but I was wondering if that's cool because at first we were talking about limiting the connection info to the minimum time set in the context | 17:34 |
| openstackgerrit | Julian proposed openstack/nova-specs: instance.create notifications should emit with scheduler hints https://review.openstack.org/312169 | 17:34 |
| alaski | melwitt: in my head I was thinking that the context would only carry connection info for the lifetime of a context manager | 17:34 |
| doffm | melwitt: I see. That is a in-fact a fundamental difference in approach. | 17:34 |
| doffm | Which I was totally missing. | 17:34 |
| *** paul-carlton2 has quit IRC | 17:34 | |
| *** sheel has quit IRC | 17:34 | |
| *** serverascode has quit IRC | 17:34 | |
| *** jamesden_ has joined #openstack-nova | 17:34 | |
| *** vishwanathj has joined #openstack-nova | 17:35 | |
| melwitt | alaski: right. I think we were concerned about it getting dumped out somewhere accidentally etc if we left it around | 17:35 |
| *** woodster_ has quit IRC | 17:35 | |
| *** zhiyan has quit IRC | 17:35 | |
| *** dougwig has quit IRC | 17:35 | |
| alaski | yeah, or being used by accident in a place it wasn't intended for | 17:35 |
| openstackgerrit | Julian proposed openstack/nova-specs: instance.create notifications should emit with scheduler hints https://review.openstack.org/312169 | 17:36 |
| *** serverascode has joined #openstack-nova | 17:36 | |
| *** ccarmack has joined #openstack-nova | 17:36 | |
| alaski | so when we discussed passing in instance I wasn't thinking of using the context from instance, just being able to pull the cell_mapping at that point | 17:36 |
| *** zhiyan has joined #openstack-nova | 17:36 | |
| *** sheel has joined #openstack-nova | 17:37 | |
| *** fmccrthy has quit IRC | 17:37 | |
| *** patrickeast has quit IRC | 17:37 | |
| doffm | The 'all-at-once' approach seems like a useful hack. I mean, yeah when we create an instance object we know which cell we should use for it, but setting the context at that point is doing a major end-around. | 17:37 |
| *** shaohe_feng has quit IRC | 17:37 | |
| *** woodster_ has joined #openstack-nova | 17:37 | |
| *** jamesdenton has quit IRC | 17:37 | |
| *** shaohe_feng has joined #openstack-nova | 17:37 | |
| markus_z | sdague: Would you give those a push please? https://review.openstack.org/#/q/topic:bp/centralize-config-options-newton+status:open+label:Workflow%253D%252B1 They got stuck in the gate. | 17:38 |
| *** patrickeast has joined #openstack-nova | 17:38 | |
| melwitt | alaski: oh, gotcha. hm | 17:39 |
| *** akshai has quit IRC | 17:39 | |
| *** dougwig has joined #openstack-nova | 17:39 | |
| *** moshele has joined #openstack-nova | 17:39 | |
| *** tonytan4ever has quit IRC | 17:39 | |
| *** ctrath has joined #openstack-nova | 17:39 | |
| *** danpb has quit IRC | 17:40 | |
| *** ducttape_ has joined #openstack-nova | 17:41 | |
| *** akshai has joined #openstack-nova | 17:41 | |
| *** rbridgeman has quit IRC | 17:41 | |
| *** fmccrthy has joined #openstack-nova | 17:41 | |
| melwitt | doffm, alaski: so I did indeed misunderstand what it meant by "just get the connection info from the instance." the pull of CellMapping in get_cell_client seems like it would be okay | 17:41 |
| *** sileht has quit IRC | 17:42 | |
| *** ctrath has quit IRC | 17:42 | |
| doffm | melwitt: Ok, we are all on the same page. :). The cell mapping will go on the instance and be used by the 'get_cell_client' method / decorator. | 17:42 |
| doffm | We will add the decorator where required. | 17:42 |
| lbeliveau | mriedem: bauzas: do you guys want a bp to track pci doc ? or we don't care | 17:42 |
| doffm | And hopefully not miss anywhere. | 17:43 |
| mriedem | lbeliveau: we don't need a bp for doc cleanups | 17:43 |
| mriedem | open a bug if the docs are wrong i guess | 17:43 |
| lbeliveau | ok | 17:43 |
| markus_z | sdague: nevermind, I see now that the other changes in the chain are not merged yet. I'm taking a look. | 17:43 |
| *** rlrossit has quit IRC | 17:44 | |
| *** roeyc has joined #openstack-nova | 17:44 | |
| *** moshele has quit IRC | 17:45 | |
| *** ctrath has joined #openstack-nova | 17:45 | |
| melwitt | alaski, doffm: so we're thinking put cell_mapping on instance? although we don't want that going over rpc for the sensitive info. maybe cell uuid on instance? cell_name is there already | 17:46 |
| melwitt | to do the lookup | 17:47 |
| *** shaohe_feng has quit IRC | 17:47 | |
| *** sileht has joined #openstack-nova | 17:48 | |
| *** shaohe_feng has joined #openstack-nova | 17:48 | |
| doffm | Hmmm. Vuage recollection of discussion about this at summit. Does it ever have to go over RPC? Or is it just a case of making sure it never goes over RPC. | 17:48 |
| doffm | UUID is fine by me. | 17:48 |
| *** ljxiash has joined #openstack-nova | 17:48 | |
| *** sridharg has quit IRC | 17:49 | |
| melwitt | oh, hm. I was thinking if instance goes over rpc so would cell_mapping | 17:49 |
| *** rlrossit has joined #openstack-nova | 17:49 | |
| melwitt | if we put it on instance | 17:49 |
| rlrossit | gibi: the versioned notifications meeting was 50 minutes ago, wasn't it... | 17:50 |
| doffm | Does it have to? If we move it over rpc it is probably in the conductor or compute node that are already in the cell? | 17:50 |
| gibi | rlrossit: yes, but it is still on :) | 17:50 |
| *** salv-orlando has quit IRC | 17:53 | |
| *** ljxiash has quit IRC | 17:53 | |
| melwitt | doffm: I don't think I understood your question | 17:54 |
| doffm | melwitt: I'm not completely sure I did either. :) Maybe we can discuss again at the meeting tomorrow. Once I have had a think about it. | 17:55 |
| melwitt | sure | 17:55 |
| *** shaohe_feng has quit IRC | 17:58 | |
| *** akshai has quit IRC | 17:58 | |
| *** paul-carlton2 has joined #openstack-nova | 17:59 | |
| *** shaohe_feng has joined #openstack-nova | 17:59 | |
| *** tpepper has joined #openstack-nova | 17:59 | |
| *** pushkaru has joined #openstack-nova | 17:59 | |
| *** baoli has quit IRC | 18:00 | |
| markus_z | nova-bugs team meeting in #openstack-meeting-4. New contributors are welcome as well. https://wiki.openstack.org/wiki/Meetings/Nova/BugsTeam | 18:00 |
| *** rbridgeman has joined #openstack-nova | 18:00 | |
| *** baoli has joined #openstack-nova | 18:01 | |
| *** akshai has joined #openstack-nova | 18:01 | |
| *** paul-carlton2 has quit IRC | 18:04 | |
| *** markvoelker has quit IRC | 18:04 | |
| *** markvoelker has joined #openstack-nova | 18:05 | |
| *** BlackDex has joined #openstack-nova | 18:05 | |
| *** zenoway has joined #openstack-nova | 18:06 | |
| *** markvoelker_ has joined #openstack-nova | 18:07 | |
| *** gszasz has quit IRC | 18:08 | |
| *** markvoelker has quit IRC | 18:08 | |
| *** shaohe_feng has quit IRC | 18:08 | |
| *** roeyc has quit IRC | 18:11 | |
| *** shaohe_feng has joined #openstack-nova | 18:11 | |
| *** tpepper has quit IRC | 18:12 | |
| *** roeyc has joined #openstack-nova | 18:13 | |
| *** tpepper has joined #openstack-nova | 18:13 | |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP resource-pools/placement api https://review.openstack.org/293104 | 18:17 |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP Add update_inventories to ResourceProvider https://review.openstack.org/312193 | 18:17 |
| *** ctrath has quit IRC | 18:18 | |
| *** shaohe_feng has quit IRC | 18:18 | |
| *** tonytan4ever has joined #openstack-nova | 18:19 | |
| *** markus_z has quit IRC | 18:19 | |
| alaski | doffm: melwitt I hadn't considered putting anything actually on the instance, just using the objects.InstanceMapping.get_by_instance_uuid(uuid) method and then pulling the cell_mapping off that object | 18:20 |
| melwitt | alaski: ah, cool. that works | 18:21 |
| *** melwitt has left #openstack-nova | 18:21 | |
| *** melwitt has joined #openstack-nova | 18:21 | |
| *** shaohe_feng has joined #openstack-nova | 18:22 | |
| *** sudipto has joined #openstack-nova | 18:22 | |
| *** vilobhmm111 has quit IRC | 18:23 | |
| mriedem | dansmith: you want to +W these? https://review.openstack.org/#/q/topic:1572495+branch:stable/mitaka | 18:24 |
| *** ctrath has joined #openstack-nova | 18:24 | |
| dansmith | mriedem: yes | 18:25 |
| melwitt | alaski: I had focused on the CellMapping being queried only by uuid and forgot the InstanceMapping having it | 18:26 |
| alaski | melwitt: I'll admit that confused me too for a second, but I knew we had definitely implemented some way to get it | 18:27 |
| melwitt | :) | 18:28 |
| *** shaohe_feng has quit IRC | 18:28 | |
| *** shaohe_feng has joined #openstack-nova | 18:29 | |
| *** zenoway has quit IRC | 18:30 | |
| *** tbachman has quit IRC | 18:31 | |
| *** armax_ has joined #openstack-nova | 18:31 | |
| mriedem | rlrossit: https://review.openstack.org/#/c/312014/ | 18:32 |
| *** bwallis has joined #openstack-nova | 18:32 | |
| *** ircuser-1 has quit IRC | 18:32 | |
| *** edmondsw has quit IRC | 18:32 | |
| *** antonyfm has joined #openstack-nova | 18:33 | |
| *** armax has quit IRC | 18:33 | |
| *** armax_ is now known as armax | 18:33 | |
| *** aghadge has quit IRC | 18:34 | |
| *** sudipto has quit IRC | 18:35 | |
| *** ircuser-1 has joined #openstack-nova | 18:37 | |
| *** fragatin_ has quit IRC | 18:37 | |
| *** ivase_ has joined #openstack-nova | 18:38 | |
| *** fragatina has joined #openstack-nova | 18:38 | |
| *** shaohe_feng has quit IRC | 18:39 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Skip network allocation if 'none' is requested https://review.openstack.org/309638 | 18:40 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: nova-net: handle 'auto' network request in allocate_for_instance https://review.openstack.org/297906 | 18:40 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Skip network validation if explicitly requesting no networks https://review.openstack.org/309602 | 18:40 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Add helpers to NetworkRequest(List) objects for auto/none cases https://review.openstack.org/297871 | 18:40 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: neutron: validate auto-allocate is available https://review.openstack.org/309574 | 18:40 |
| *** salv-orlando has joined #openstack-nova | 18:40 | |
| *** shaohe_feng has joined #openstack-nova | 18:40 | |
| *** ivase has quit IRC | 18:41 | |
| *** d34dh0r53 is now known as m1r4nt15_b0y | 18:42 | |
| *** jichen has joined #openstack-nova | 18:42 | |
| *** marcusvrn_ has quit IRC | 18:43 | |
| mriedem | mdbooth: what are you using for tracking the libvirt storage pools work? which spec/bp(s)? | 18:45 |
| *** salv-orlando has quit IRC | 18:45 | |
| mriedem | https://review.openstack.org/#/c/310505/ ? | 18:45 |
| *** moshele has joined #openstack-nova | 18:45 | |
| mriedem | https://review.openstack.org/#/c/302117 ? | 18:47 |
| *** salv-orlando has joined #openstack-nova | 18:48 | |
| *** salv-orlando has quit IRC | 18:48 | |
| dansmith | mriedem: those are dependent | 18:48 |
| dansmith | so .. both | 18:49 |
| *** shaohe_feng has quit IRC | 18:49 | |
| mriedem | hot damn! | 18:49 |
| mriedem | and it's already all dep'ed up in https://blueprints.launchpad.net/nova/+spec/libvirt-instance-storage | 18:49 |
| * mriedem is doing the newton priorities thing in nova-specs | 18:49 | |
| *** Guest93371 has quit IRC | 18:49 | |
| mriedem | dansmith: regarding priorities, i'm not sure all 3 blueprints in newton is realistic | 18:50 |
| *** shaohe_feng has joined #openstack-nova | 18:50 | |
| mriedem | the bottom refactor one certainly | 18:50 |
| dansmith | could be, depends on how much attention it gets | 18:50 |
| *** armax has quit IRC | 18:51 | |
| *** salv-orlando has joined #openstack-nova | 18:51 | |
| *** baoli has quit IRC | 18:52 | |
| mriedem | i guess our priorities doc doesn't actually call out specific specs, it's just broad | 18:52 |
| mriedem | so we can wiggle if needed | 18:53 |
| mriedem | i know you guys like to wiggle | 18:53 |
| *** salv-orlando has quit IRC | 18:54 | |
| dansmith | mriedem: well, like anything, we get done what we can get done and punt the rest | 18:55 |
| *** baoli has joined #openstack-nova | 18:56 | |
| dansmith | there's a lot of code up for it already.. not sure how close it is, but it's better than nothing | 18:56 |
| *** links has quit IRC | 18:58 | |
| *** doug-fis_ has joined #openstack-nova | 18:58 | |
| *** baoli has quit IRC | 18:58 | |
| *** ijw has quit IRC | 18:59 | |
| *** shaohe_feng has quit IRC | 18:59 | |
| *** armax has joined #openstack-nova | 18:59 | |
| *** shaohe_feng has joined #openstack-nova | 18:59 | |
| *** doug-fi__ has joined #openstack-nova | 19:00 | |
| doffm | dansmith: Sooo, the problem you have with the '_hosts' and '_metadata' properties in the Aggregate model. (https://review.openstack.org/#/c/294806/5/nova/db/sqlalchemy/api_models.py) | 19:01 |
| dansmith | doffm: on a call atm, fyi | 19:01 |
| *** ijw has joined #openstack-nova | 19:02 | |
| doffm | dansmith: Ok, sorry to bother you. Will post comments on patch instead. | 19:02 |
| *** doug-fish has quit IRC | 19:02 | |
| *** doug-fis_ has quit IRC | 19:03 | |
| *** doug-fi__ has quit IRC | 19:04 | |
| cfriesen | in LibvirtDriver._get_disk_over_committed_size_total() we call self._host.list_instance_domains() to get the list of domains to loop over. This will ignore any non-running domains. Wouldn't we still want to include their disks in the calculation? | 19:05 |
| *** baoli has joined #openstack-nova | 19:05 | |
| *** vtech has joined #openstack-nova | 19:06 | |
| *** ssurana has quit IRC | 19:06 | |
| *** pwnall has quit IRC | 19:08 | |
| *** pwnall has joined #openstack-nova | 19:08 | |
| *** shaohe_feng has quit IRC | 19:09 | |
| *** ducttape_ has quit IRC | 19:10 | |
| *** shaohe_feng has joined #openstack-nova | 19:10 | |
| *** READ10 has quit IRC | 19:18 | |
| openstackgerrit | Sean Dague proposed openstack/nova-specs: Deprecate API Proxies https://review.openstack.org/312209 | 19:19 |
| *** shaohe_feng has quit IRC | 19:20 | |
| *** shaohe_feng has joined #openstack-nova | 19:20 | |
| *** ducttape_ has joined #openstack-nova | 19:21 | |
| *** ducttape_ has quit IRC | 19:22 | |
| *** matrohon has joined #openstack-nova | 19:23 | |
| *** Sukhdev has joined #openstack-nova | 19:23 | |
| openstackgerrit | dane-fichter proposed openstack/nova-specs: Signature verification for cached images https://review.openstack.org/312210 | 19:28 |
| *** ducttape_ has joined #openstack-nova | 19:29 | |
| mriedem | sdague: when you're around, we should talk about details for the api-ref review sprint next week, | 19:29 |
| mriedem | i think it's going to have to be more targeted than the entire week | 19:29 |
| sdague | mriedem: sure | 19:29 |
| sdague | I'm around now | 19:29 |
| mriedem | and then i can advertise it in the dev list and schedule wiki | 19:29 |
| mriedem | sdague: ok, so, 1-2 days for this? | 19:30 |
| mriedem | maybe mon / wed? | 19:30 |
| *** shaohe_feng has quit IRC | 19:30 | |
| mriedem | my thinking is, put a day in between for time to respond to comments, or add new things for the 2nd day | 19:30 |
| *** shaohe_feng has joined #openstack-nova | 19:30 | |
| mriedem | was thinking mon/wed b/c tuesday is full of meetings | 19:31 |
| sdague | I think doing mon / wed like that is worth a try | 19:31 |
| *** vilobhmm11 has joined #openstack-nova | 19:31 | |
| sdague | assuming we knock each of the first 3 phases out, on the 47 docs, we're still looking at > 100 patches | 19:32 |
| sdague | which is why it felt like a week to me | 19:32 |
| sdague | be we can try the more targetted thing | 19:32 |
| cdent | mriedem: you said something overnight about me using fancy new libraries on the placement api. there's really only two new things there: gabbi for the tests, and selector, which is a library for doing what Routes does, but cleaner (and handles 405 correctly) | 19:34 |
| cdent | the other main difference is that is function rather than class based, where possible | 19:35 |
| openstackgerrit | Matt Riedemann proposed openstack/nova-specs: Add in the Newton review priorities https://review.openstack.org/312217 | 19:37 |
| *** pbrooko has joined #openstack-nova | 19:38 | |
| *** ericksonsantos has quit IRC | 19:39 | |
| *** BjoernT has joined #openstack-nova | 19:39 | |
| *** shaohe_feng has quit IRC | 19:40 | |
| *** shaohe_feng has joined #openstack-nova | 19:41 | |
| *** rmk has quit IRC | 19:42 | |
| *** matrohon has quit IRC | 19:42 | |
| *** jichen has quit IRC | 19:42 | |
| *** fragatina has quit IRC | 19:42 | |
| *** antonyfm has quit IRC | 19:42 | |
| *** vishwanathj has quit IRC | 19:42 | |
| *** palexster has quit IRC | 19:42 | |
| *** cleong has quit IRC | 19:42 | |
| *** jdillaman has quit IRC | 19:42 | |
| *** vladikr has quit IRC | 19:42 | |
| *** rook has quit IRC | 19:42 | |
| *** kylek3h_ has quit IRC | 19:42 | |
| *** yuywz has quit IRC | 19:42 | |
| *** rhagarty_ has quit IRC | 19:42 | |
| *** david-lyle has quit IRC | 19:42 | |
| *** rwmjones has quit IRC | 19:42 | |
| *** zul has quit IRC | 19:42 | |
| *** cfarquhar has quit IRC | 19:42 | |
| *** diana_clarke has quit IRC | 19:42 | |
| *** Yingxin has quit IRC | 19:42 | |
| *** purplerbot has quit IRC | 19:42 | |
| *** kfox1111 has quit IRC | 19:42 | |
| *** Prophidys has quit IRC | 19:42 | |
| *** gmann has quit IRC | 19:42 | |
| *** pcm_ has quit IRC | 19:42 | |
| *** ducttape_ has quit IRC | 19:42 | |
| *** akshai has quit IRC | 19:42 | |
| *** mriedem has quit IRC | 19:42 | |
| *** mtanino has quit IRC | 19:42 | |
| *** burgerk has quit IRC | 19:42 | |
| *** mrkz has quit IRC | 19:42 | |
| *** eharney has quit IRC | 19:42 | |
| *** thorst has quit IRC | 19:42 | |
| *** rtheis has quit IRC | 19:42 | |
| *** fishcried has quit IRC | 19:42 | |
| *** jogo has quit IRC | 19:42 | |
| *** xek__ has quit IRC | 19:42 | |
| *** aarefiev has quit IRC | 19:42 | |
| *** ByPasS has quit IRC | 19:42 | |
| *** kaisers has quit IRC | 19:42 | |
| *** yhvh- has quit IRC | 19:42 | |
| *** pav0 has quit IRC | 19:42 | |
| *** mag009_ has quit IRC | 19:42 | |
| *** lennyb has quit IRC | 19:42 | |
| *** FL1SK has quit IRC | 19:42 | |
| *** zhangjn has quit IRC | 19:42 | |
| *** andrearosa has quit IRC | 19:42 | |
| *** mirrorbox has quit IRC | 19:42 | |
| *** crinkle has quit IRC | 19:42 | |
| *** rdo has quit IRC | 19:42 | |
| *** tobasco has quit IRC | 19:42 | |
| *** keekz has quit IRC | 19:42 | |
| *** timello has quit IRC | 19:42 | |
| *** kevinbenton has quit IRC | 19:42 | |
| *** jgriffith has quit IRC | 19:42 | |
| *** adreznec has quit IRC | 19:42 | |
| *** sneti has quit IRC | 19:42 | |
| *** andreaf has quit IRC | 19:42 | |
| *** markmcclain has quit IRC | 19:42 | |
| *** pbrooko has quit IRC | 19:42 | |
| *** rtheis has joined #openstack-nova | 19:46 | |
| *** ericksonsantos has joined #openstack-nova | 19:47 | |
| *** moshele has quit IRC | 19:48 | |
| *** Swami has joined #openstack-nova | 19:48 | |
| *** baoli has quit IRC | 19:49 | |
| *** phschwartz_ is now known as phschwartz | 19:49 | |
| *** baoli has joined #openstack-nova | 19:50 | |
| *** cdent has quit IRC | 19:50 | |
| *** shaohe_feng has quit IRC | 19:50 | |
| *** cdent has joined #openstack-nova | 19:51 | |
| *** rmk has joined #openstack-nova | 19:51 | |
| *** dmk0202 has joined #openstack-nova | 19:51 | |
| *** david-lyle has joined #openstack-nova | 19:51 | |
| *** ducttape_ has joined #openstack-nova | 19:51 | |
| *** akshai has joined #openstack-nova | 19:51 | |
| *** mriedem has joined #openstack-nova | 19:51 | |
| *** mtanino has joined #openstack-nova | 19:51 | |
| *** burgerk has joined #openstack-nova | 19:51 | |
| *** mrkz has joined #openstack-nova | 19:51 | |
| *** eharney has joined #openstack-nova | 19:51 | |
| *** thorst has joined #openstack-nova | 19:51 | |
| *** fishcried has joined #openstack-nova | 19:51 | |
| *** jogo has joined #openstack-nova | 19:51 | |
| *** xek__ has joined #openstack-nova | 19:51 | |
| *** aarefiev has joined #openstack-nova | 19:51 | |
| *** ByPasS has joined #openstack-nova | 19:51 | |
| *** kaisers has joined #openstack-nova | 19:51 | |
| *** yhvh- has joined #openstack-nova | 19:51 | |
| *** pav0 has joined #openstack-nova | 19:51 | |
| *** mag009_ has joined #openstack-nova | 19:51 | |
| *** lennyb has joined #openstack-nova | 19:51 | |
| *** FL1SK has joined #openstack-nova | 19:51 | |
| *** zhangjn has joined #openstack-nova | 19:51 | |
| *** andreaf has joined #openstack-nova | 19:51 | |
| *** andrearosa has joined #openstack-nova | 19:51 | |
| *** mirrorbox has joined #openstack-nova | 19:51 | |
| *** crinkle has joined #openstack-nova | 19:51 | |
| *** rdo has joined #openstack-nova | 19:51 | |
| *** tobasco has joined #openstack-nova | 19:51 | |
| *** keekz has joined #openstack-nova | 19:51 | |
| *** timello has joined #openstack-nova | 19:51 | |
| *** kevinbenton has joined #openstack-nova | 19:51 | |
| *** jgriffith has joined #openstack-nova | 19:51 | |
| *** adreznec has joined #openstack-nova | 19:51 | |
| *** sneti has joined #openstack-nova | 19:51 | |
| *** markmcclain has joined #openstack-nova | 19:51 | |
| *** purplerbot has joined #openstack-nova | 19:51 | |
| *** jichen has joined #openstack-nova | 19:51 | |
| *** fragatina has joined #openstack-nova | 19:51 | |
| *** antonyfm has joined #openstack-nova | 19:51 | |
| *** vishwanathj has joined #openstack-nova | 19:51 | |
| *** palexster has joined #openstack-nova | 19:51 | |
| *** cleong has joined #openstack-nova | 19:51 | |
| *** jdillaman has joined #openstack-nova | 19:51 | |
| *** vladikr has joined #openstack-nova | 19:51 | |
| *** rook has joined #openstack-nova | 19:51 | |
| *** kylek3h_ has joined #openstack-nova | 19:51 | |
| *** yuywz has joined #openstack-nova | 19:51 | |
| *** rhagarty_ has joined #openstack-nova | 19:51 | |
| *** rwmjones has joined #openstack-nova | 19:51 | |
| *** zul has joined #openstack-nova | 19:51 | |
| *** cfarquhar has joined #openstack-nova | 19:51 | |
| *** diana_clarke has joined #openstack-nova | 19:51 | |
| *** Yingxin has joined #openstack-nova | 19:51 | |
| *** kfox1111 has joined #openstack-nova | 19:51 | |
| *** Prophidys has joined #openstack-nova | 19:51 | |
| *** gmann has joined #openstack-nova | 19:51 | |
| *** pcm_ has joined #openstack-nova | 19:51 | |
| *** shaohe_feng has joined #openstack-nova | 19:51 | |
| *** rmk has quit IRC | 19:52 | |
| *** rmk has joined #openstack-nova | 19:52 | |
| *** matrohon has joined #openstack-nova | 19:52 | |
| sdague | mriedem: I love how gerrit decided that the word 'this' in priorities is original, and everything else is changed | 19:53 |
| *** moshele has joined #openstack-nova | 19:53 | |
| *** ssurana has joined #openstack-nova | 19:55 | |
| *** moshele has quit IRC | 19:57 | |
| mriedem | oh yeah weird | 19:57 |
| jaypipes | holy fucksticks this synamic-resource-classes spec is getting long. | 19:58 |
| mriedem | sdague: anyway, thoughts on the api-ref review sprint words above? | 19:58 |
| mriedem | jaypipes: you do get an award at some point | 19:58 |
| sdague | mriedem: which words? | 19:58 |
| jaypipes | mriedem: unwanted award for sure. | 19:58 |
| mriedem | sdague: my proposal for mon/wed? | 19:58 |
| mriedem | did i drop off irc and not get a reply? | 19:58 |
| sdague | maybe? | 19:58 |
| sdague | <mriedem> was thinking mon/wed b/c tuesday is full of meetings | 19:58 |
| sdague | <sdague> I think doing mon / wed like that is worth a try | 19:58 |
| sdague | <sdague> assuming we knock each of the first 3 phases out, on the 47 docs, we're still looking at > 100 patches | 19:58 |
| sdague | <sdague> which is why it felt like a week to me | 19:58 |
| sdague | <sdague> be we can try the more targetted thing | 19:58 |
| mriedem | wow yeah didn't get any of that | 19:59 |
| mriedem | sdague: we could do another sprint day afterward, based on how this goes | 20:00 |
| sdague | sure | 20:00 |
| *** shaohe_feng has quit IRC | 20:01 | |
| mriedem | just, we have a hard enough time getting people to focus on a single day sprint when we're not in person, so a full week is rough | 20:01 |
| *** shaohe_feng has joined #openstack-nova | 20:01 | |
| sdague | yeh, my feeling was it would probably naturally be more of the api subteam folks | 20:01 |
| *** roeyc has quit IRC | 20:01 | |
| sdague | and maybe some new people | 20:01 |
| mriedem | alright, i'll send a thing to the dev list and put it in the release schedule, and thurs meeting agenda | 20:02 |
| sdague | sounds great | 20:02 |
| openstackgerrit | dane-fichter proposed openstack/nova-specs: Trusted flavors https://review.openstack.org/312225 | 20:02 |
| mriedem | cdent: i also missed what you said, but just read it in the irc logs | 20:03 |
| mriedem | i didn't know about selector, so 2 new libs for nova qualifies as 'a bunch' in my book | 20:03 |
| cdent | selector isn't really required, it just makes my life a ton easier | 20:03 |
| *** baoli has quit IRC | 20:04 | |
| *** baoli has joined #openstack-nova | 20:04 | |
| cdent | I'd fight for it, but not hard | 20:05 |
| bauzas | mriedem: any chance to see Intel providing us some hotels for the midcycle ? | 20:05 |
| bauzas | or should I find those by myself ? | 20:05 |
| *** tbachman has joined #openstack-nova | 20:05 | |
| mriedem | bauzas: when we did it in hillsboro before, there were some group rates | 20:05 |
| bauzas | k | 20:05 |
| bauzas | btw. sorry I'm a bit off those evenings because of the jetlag :( | 20:06 |
| *** moshele has joined #openstack-nova | 20:06 | |
| mriedem | rlrossit: ctrath: jimmy legs https://www.youtube.com/watch?v=EnG8jrpLJDE | 20:06 |
| *** Sukhdev has quit IRC | 20:07 | |
| *** doug-fish has joined #openstack-nova | 20:08 | |
| *** doug-fish has quit IRC | 20:08 | |
| *** rmk has quit IRC | 20:08 | |
| *** cdelatte has quit IRC | 20:10 | |
| *** shaohe_feng has quit IRC | 20:11 | |
| *** shaohe_feng has joined #openstack-nova | 20:11 | |
| sdague | the group rate as quite good last time, so I'd wait for them to tell us what the story is | 20:14 |
| *** rmk has joined #openstack-nova | 20:15 | |
| openstackgerrit | dane-fichter proposed openstack/nova-specs: Trusted flavors https://review.openstack.org/312225 | 20:17 |
| *** shaohe_feng has quit IRC | 20:21 | |
| *** shaohe_feng has joined #openstack-nova | 20:22 | |
| mikal | Morning | 20:23 |
| cfriesen | afternoon | 20:23 |
| *** matrohon has quit IRC | 20:24 | |
| *** smatzek has quit IRC | 20:25 | |
| *** ljxiash has joined #openstack-nova | 20:26 | |
| *** matrohon has joined #openstack-nova | 20:27 | |
| -openstackstatus- NOTICE: restarting apache on review.openstack.org to pick up security patches. Gerrit web ui may disappear for a short time. | 20:27 | |
| alaski | mriedem: are we good to merge https://review.openstack.org/#/c/312217/1 or do you want more +2s? | 20:28 |
| mriedem | do it | 20:28 |
| mriedem | i'm writing up the changes to the newton schedule in the releases repo and need to link to that in the official docs | 20:28 |
| alaski | done | 20:29 |
| mriedem | <3 | 20:29 |
| *** mvle has joined #openstack-nova | 20:29 | |
| sdague | mikal: I did a round of feedback on your dynamic vendor data spec | 20:30 |
| sdague | if you want to take a look | 20:30 |
| *** ljxiash has quit IRC | 20:31 | |
| mikal | sdague: I just noticed those | 20:31 |
| *** shaohe_feng has quit IRC | 20:31 | |
| *** sahid has quit IRC | 20:32 | |
| mikal | sdague: nothing too major I don't think | 20:32 |
| openstackgerrit | Chris Dent proposed openstack/nova: Extend ResourceProvider to have resources and aggregates https://review.openstack.org/284963 | 20:32 |
| mikal | sdague: seems like there might be some debate over if more than one external REST service should be supported | 20:32 |
| *** vishwanathj has quit IRC | 20:32 | |
| *** shaohe_feng has joined #openstack-nova | 20:32 | |
| *** roeyc has joined #openstack-nova | 20:32 | |
| *** vishwanathj has joined #openstack-nova | 20:33 | |
| *** cdelatte has joined #openstack-nova | 20:33 | |
| *** e0ne has joined #openstack-nova | 20:34 | |
| sdague | mikal: for real, or is it just something in the wind? | 20:36 |
| mikal | sdague: ? | 20:36 |
| sdague | I can't imagine for complexity reasons why you wouldn't want to build a single proxy for all those things | 20:36 |
| mikal | sdague: as in have an external REST thing which aggregates external REST things? | 20:36 |
| sdague | honestly, I think you have to do that anyway, otherwise you are going to start baking in a bunch of semantics here | 20:37 |
| *** rook has quit IRC | 20:37 | |
| mikal | So the rest I'd proposed supporting a list was because we're already making the deployment more complicated than previously, and I was trying to stop people from having to write a bunch of external software as well. | 20:38 |
| sdague | I expect in all cases people are going to need to write these new services, just like they were previously writing their own vendor data plugin | 20:38 |
| mikal | Sorry, reason | 20:38 |
| sdague | right, but, those things don't exist today | 20:38 |
| mikal | Agreed | 20:38 |
| mikal | But an aggregating proxy is more complicated for them to write than a simple values dumper | 20:39 |
| bauzas | alaski: mriedem: FWIW, I was wondering if check-destinations BP was good for a sched priority given the operators and the live-mig guys wanting it merged soon | 20:39 |
| bauzas | but that's not something really bad if not | 20:39 |
| bauzas | I just need to make sure that I'm done by n-2 then | 20:40 |
| bauzas | (given I'm hoping to help jaypipes's stuff for Newton too) | 20:40 |
| *** doug-fish has joined #openstack-nova | 20:42 | |
| *** shaohe_feng has quit IRC | 20:42 | |
| *** Sukhdev has joined #openstack-nova | 20:42 | |
| sdague | mikal: maybe, except it means they get to do it once, and don't have to also manage the content in these encoded fields | 20:42 |
| mriedem | bauzas: it's listed as medium priority, i'm not sure how 'on the side' it is for live migration or scheduler priorities for newton | 20:42 |
| mriedem | bauzas: i see 3 open patches | 20:42 |
| *** mvle has quit IRC | 20:42 | |
| mriedem | bauzas: how much else is there? i don't have a good idea in my head of the work involved and how much is left | 20:42 |
| sdague | it would be curious if the bloomberg folks and nate have an opinion there | 20:42 |
| mikal | sdague: I'm not too fussed either way to be honest. It was a request from Edmund. We could take it to the operators list and ask what people think? | 20:42 |
| sdague | I think they just want any hook they can get to do their own thing | 20:42 |
| *** shaohe_feng has joined #openstack-nova | 20:43 | |
| sdague | mikal: yeh, that would be fine with me | 20:43 |
| mikal | Ok, I will do that today | 20:43 |
| sdague | I'm not super strongly opinionated either way | 20:43 |
| bauzas | mriedem: no worries, I only implemented the conductor things and I need to now implement the API related microversion | 20:43 |
| bauzas | that's fair tho | 20:43 |
| alaski | bauzas: I think it falls outside of that effort, mostly because it doesn't affect the scheduler api | 20:43 |
| alaski | I also think it would be good to get in soon | 20:43 |
| bauzas | yeah, fair enough | 20:44 |
| *** salv-orlando has joined #openstack-nova | 20:44 | |
| bauzas | anyway, it's just me needing to upload the next changes soon | 20:44 |
| bauzas | so I could help the other scheduler changes | 20:44 |
| bauzas | thanks, it clarifies :) | 20:44 |
| sdague | anyway, I need to drop for dinner. But I'll watch on the ops list | 20:44 |
| mikal | sdague: cool, have a good evening | 20:45 |
| *** ducttape_ has quit IRC | 20:45 | |
| mriedem | fyi, nova-specific release milestones patch https://review.openstack.org/#/c/312245/ | 20:45 |
| *** doug-fish has quit IRC | 20:46 | |
| bauzas | oh gosh, it's f*** expensive to go to PDX in July ! | 20:47 |
| bauzas | 2000 bucks... | 20:47 |
| *** ducttape_ has joined #openstack-nova | 20:47 | |
| mikal | Yeah, I looked the other day and its going to be $5,000 AUD or something | 20:47 |
| bauzas | what the heck is in Portland by this time ? | 20:47 |
| *** gyee has joined #openstack-nova | 20:47 | |
| bauzas | fun, SEA is just 1300 dollars... | 20:48 |
| mikal | bauzas: the price for me went down a lot of I stayed for 7 days. You could cafe hack for a few days? | 20:49 |
| bauzas | dansmith: is it hard to drive from SEA to PDX ? | 20:49 |
| dansmith | bauzas: 3-4 hours on the highway | 20:49 |
| mikal | bauzas: I just looked, its 3 hours on a freeway | 20:49 |
| *** salv-orlando has quit IRC | 20:50 | |
| mikal | i.e. not scary | 20:50 |
| bauzas | mikal: it's worst if I'm asking to leave on Friday... | 20:50 |
| mikal | dansmith: is it an interesting drive? Anything to see on the way? | 20:50 |
| dansmith | mikal: not really | 20:50 |
| mikal | Mt Rainier! | 20:50 |
| dansmith | our mountain is prettier anyway | 20:50 |
| mikal | Thems fighting words | 20:51 |
| bauzas | given the price, I should think of paying a pedalo | 20:51 |
| bauzas | (a pedal boat) | 20:51 |
| *** tbachman has quit IRC | 20:51 | |
| *** Sukhdev has quit IRC | 20:51 | |
| *** shaohe_feng has quit IRC | 20:52 | |
| *** shaohe_feng has joined #openstack-nova | 20:53 | |
| *** Swami has quit IRC | 20:53 | |
| *** smatzek has joined #openstack-nova | 20:54 | |
| *** salv-orlando has joined #openstack-nova | 20:54 | |
| *** baoli has quit IRC | 20:54 | |
| *** lucas-dinner has quit IRC | 20:54 | |
| *** aysyd has quit IRC | 20:56 | |
| cfriesen | looking at LibvirtDriver.get_available_resource(), don't we have a huge race condition? It's called without holding COMPUTE_RESOURCE_SEMAPHORE, and yet it's calculating vcpus_used, local_gb_used, memory_mb_used, and disk_available_least, all of which are only valid if you're holding the semaphore. | 20:57 |
| *** lucasagomes has joined #openstack-nova | 20:58 | |
| bauzas | cfriesen: all the drivers are not locking that AFAIK | 20:58 |
| bauzas | cfriesen: lemme just verify this | 20:58 |
| bauzas | that's not a synchronous call | 20:58 |
| *** smatzek has quit IRC | 20:59 | |
| cfriesen | yeah, but ResourceTracker takes the semaphore for _update_available_resource(), using the data returned from calling self.driver.get_available_resource() | 20:59 |
| cfriesen | if that data is stale, then the output of _update_available_resource() is garbage | 21:00 |
| cfriesen | I can see querying the driver for the total available resources without holding the sema, but any calculations involving "used" or "free" values should hold the sema. | 21:00 |
| *** Sukhdev has joined #openstack-nova | 21:02 | |
| *** shaohe_feng has quit IRC | 21:02 | |
| *** shaohe_feng has joined #openstack-nova | 21:03 | |
| *** matrohon has quit IRC | 21:03 | |
| *** roeyc has quit IRC | 21:03 | |
| *** moshele has quit IRC | 21:06 | |
| *** roeyc has joined #openstack-nova | 21:06 | |
| *** doug-fish has joined #openstack-nova | 21:07 | |
| *** Sukhdev has quit IRC | 21:07 | |
| *** maurosr has quit IRC | 21:07 | |
| *** doug-fis_ has joined #openstack-nova | 21:08 | |
| cfriesen | bauzas: I kind of think we actually need to hold COMPUTE_RESOURCE_SEMAPHORE for all of ResourceTracker.update_available_resource() (Technically we could break up self.driver.get_available_resource() into two pieces, one that queries the total resources outside of the sema, and one that calculates free/used while holding the sema, but that's an optimization.) | 21:08 |
| *** mylu has joined #openstack-nova | 21:09 | |
| bauzas | that's what I'm verifying, if the driver call is in a synchronous section | 21:09 |
| *** doug-fi__ has joined #openstack-nova | 21:09 | |
| *** mvk has joined #openstack-nova | 21:09 | |
| *** maurosr has joined #openstack-nova | 21:09 | |
| cfriesen | I don't think it is, currently. Only ResourceTracker._update_available_resource() is synchronized | 21:09 |
| *** Swami has joined #openstack-nova | 21:09 | |
| bauzas | right | 21:10 |
| *** ijw has quit IRC | 21:10 | |
| *** mylu has quit IRC | 21:10 | |
| bauzas | cfriesen: and the driver call is in the main update_avail_res() right? | 21:10 |
| cfriesen | but we feed it the resources from calling self.driver.get_available_resource() outside the sema | 21:10 |
| cfriesen | yes | 21:10 |
| bauzas | cfriesen: so it's a regression | 21:10 |
| cfriesen | well, it's been like this for a long time | 21:10 |
| *** Sukhdev has joined #openstack-nova | 21:10 | |
| bauzas | cfriesen: because IIRC, update_avail_res() was using the semaphore in Juno | 21:11 |
| bauzas | and then we splitted it in twice | 21:11 |
| cfriesen | could be...in kilo at least it's not | 21:11 |
| bauzas | one being synchronized, the other not | 21:11 |
| *** doug-fish has quit IRC | 21:11 | |
| bauzas | I actually remember we made such split for a good related reason that I don't remind | 21:12 |
| bauzas | I need to find the commit | 21:12 |
| * bauzas opens his delorean | 21:12 | |
| bauzas | and runs for 88mph | 21:12 |
| *** shaohe_feng has quit IRC | 21:12 | |
| *** doug-fis_ has quit IRC | 21:13 | |
| cfriesen | git commit 84ff466: reduce the scope of rt work while holding the big lock | 21:13 |
| bauzas | heh, that was fast, me too | 21:13 |
| cfriesen | Dan Smith | 21:13 |
| *** e0ne has quit IRC | 21:13 | |
| cfriesen | hmm... | 21:13 |
| sneti | Hi auggy...I'm working on your bug: https://bugs.launchpad.net/nova/+bug/1544720. I tried to reproduce it. | 21:14 |
| openstack | Launchpad bug 1544720 in OpenStack Compute (nova) "In response template processing, some values are set to u'' instead of None" [Medium,Confirmed] - Assigned to Sujitha (sujitha-neti) | 21:14 |
| sneti | I changed OS-EXT-SRV-ATTR:kernel_id to some value in .tpl file, and then ran the test, test fails since the value in template and response are different for OS-EXT-SRV-ATTR:kernel_id. This is expected. | 21:14 |
| sneti | Can you explain me what your bug is about. Because I see the expected behaviour here. | 21:14 |
| mriedem | 84ff466 was needed to fix a gate issue | 21:14 |
| mriedem | a pretty nasty one | 21:14 |
| bauzas | https://review.openstack.org/#/c/123774/2 | 21:14 |
| bauzas | mriedem: sure I remember we had good reason for that | 21:14 |
| cfriesen | https://bugs.launchpad.net/nova/+bug/1371677 | 21:15 |
| openstack | Launchpad bug 1371677 in OpenStack Compute (nova) "Race in resource tracker causes 500 response on deleting during verify_resize state " [Critical,Fix released] - Assigned to Dan Smith (danms) | 21:15 |
| bauzas | mriedem: I was just commenting out that it was not the case before that | 21:15 |
| *** mylu has joined #openstack-nova | 21:15 | |
| *** svenkat has quit IRC | 21:15 | |
| *** raildo is now known as raildo-afk | 21:15 | |
| *** shaohe_feng has joined #openstack-nova | 21:16 | |
| cfriesen | hmm, doesn't look directly related to that bugfix...the commit message doesn't match the bug report | 21:16 |
| bauzas | which means fixing cfriesen's bug about the driver not synchronous would then perhaps resurrect that race | 21:16 |
| *** thorst has quit IRC | 21:16 | |
| cfriesen | no discussion of locking scope on that commit | 21:17 |
| bauzas | dansmith: any chance you remember exactly what was the race condition fixed by https://review.openstack.org/#/c/123774/2 ? | 21:18 |
| dansmith | cfriesen: you mean the bug? the commit is talking about reducing the scope | 21:18 |
| *** thorst has joined #openstack-nova | 21:18 | |
| *** jichen has quit IRC | 21:18 | |
| *** ivase_ has quit IRC | 21:19 | |
| cfriesen | dansmith: so the problem I see is that LibvirtDriver.get_available_resource() is called without holding the sema, but then does calculations for used/free values by looping over instances. That data (which can get stale) is then passed into ResourceTracker.update_available_resource() which does hold the sema, but by then it's too late | 21:19 |
| *** ivase has joined #openstack-nova | 21:19 | |
| *** lucasagomes has quit IRC | 21:20 | |
| cfriesen | I think it'd be valid to query the "total" available amounts without holding the sema, but the used/free calculations need to hold the sema to prevent creation/deletion/resize/etc from changing the values | 21:20 |
| dansmith | cfriesen: but what is the problem, that we report a somewhat stale result? | 21:20 |
| *** tonytan4ever has quit IRC | 21:21 | |
| *** haplo37 has quit IRC | 21:21 | |
| cfriesen | yep. we had a case where someone deleted an instance and then tried to recreate it and couldn't | 21:21 |
| *** cleong has quit IRC | 21:21 | |
| cfriesen | or you could get the opposite, where the sched thinks it'll work but it really can't | 21:22 |
| *** lucasagomes has joined #openstack-nova | 21:22 | |
| *** thorst has quit IRC | 21:22 | |
| *** shaohe_feng has quit IRC | 21:23 | |
| *** ivase_ has joined #openstack-nova | 21:23 | |
| cfriesen | I suppose we could say that this is acceptable for the improved performance of not holding the sema over the whole resource audit | 21:23 |
| *** shaohe_feng has joined #openstack-nova | 21:24 | |
| *** ivase has quit IRC | 21:24 | |
| cfriesen | but if so we should document the race as a known issue | 21:24 |
| *** vilobhmm11 has quit IRC | 21:24 | |
| bauzas | so, it's something again about having the scheduler very quickly giving us a destination even if the dest is wrong | 21:24 |
| *** vilobhmm11 has joined #openstack-nova | 21:25 | |
| bauzas | having a global lock that would hold all of that helps the scheduler giving better accurate decisions | 21:25 |
| mriedem | hrm, untracked ceph plugin job failure http://logs.openstack.org/53/312153/1/check/gate-tempest-dsvm-full-devstack-plugin-ceph/aac456a/console.html | 21:25 |
| bauzas | but we pay the price with some longer calls | 21:25 |
| cfriesen | coming from the kernel, the locking in nova is messed up. :) | 21:25 |
| *** vilobhmm11 has quit IRC | 21:25 | |
| *** vilobhmm11 has joined #openstack-nova | 21:26 | |
| dansmith | the bug seems to have lots of detail in it that is definitely relevant | 21:26 |
| *** vilobhmm11 has quit IRC | 21:26 | |
| dansmith | us holding a semaphore while we make a call to conductor to list all instances is not a great idea | 21:26 |
| *** vilobhmm11 has joined #openstack-nova | 21:27 | |
| bauzas | I agree | 21:27 |
| *** ivase_ has quit IRC | 21:27 | |
| bauzas | hence my point, there are some cases where I prefer having wrong stats rather than locking all | 21:27 |
| dansmith | yeah, | 21:28 |
| melwitt | mriedem: looking | 21:28 |
| mriedem | not ceph related | 21:28 |
| mriedem | it's keystone | 21:28 |
| mriedem | only 3 hits in the check queue in 7 days | 21:28 |
| dansmith | we know the scheduler relies on having excess space and not being able to schedule the last available chunk | 21:28 |
| cfriesen | where do we call to the conductor for the list of instances? | 21:28 |
| melwitt | mriedem: okay | 21:28 |
| cfriesen | I see us querying the hypervisor for the list of local domains | 21:28 |
| dansmith | cfriesen: it's already in the lock-protected section there | 21:29 |
| cfriesen | right, just saw it | 21:29 |
| *** burgerk has quit IRC | 21:30 | |
| dansmith | we make a bunch of calls in there actually | 21:30 |
| *** vilobhmm11 has quit IRC | 21:30 | |
| cfriesen | right...I mean it looked like we were trying to be accurate, but then there was a race condition in the data going into there....which is what confused me | 21:30 |
| *** vilobhmm11 has joined #openstack-nova | 21:30 | |
| *** zenoway has joined #openstack-nova | 21:30 | |
| bauzas | cfriesen: dansmith: I need to drop-off, but feel free to continue that convo and I'll catch up later the logs | 21:31 |
| auggy | sneti: hey sorry, i'm not sure, i think the issue was that None values were being set to u'' instead of None | 21:31 |
| cfriesen | do we keep a local list of instances assigned to the node? maybe we could skip querying the conductor | 21:31 |
| cfriesen | bauzas: thanks for taking a look | 21:31 |
| bauzas | cfriesen: oh, please don't | 21:31 |
| auggy | but if you can't reproduce it then just document how you attempted to repro it and mark it as invalid and that you couldn't repro it | 21:31 |
| *** tbachman has joined #openstack-nova | 21:32 | |
| auggy | it might not be an issue anymore, hard to say | 21:32 |
| bauzas | cfriesen: trying to cache the list of instances seems scary to me | 21:32 |
| sneti | auggy, yes..I don't see u' ' anywhere. I got this: Template: kernel | 21:32 |
| sneti | Response: | 21:32 |
| *** shaohe_feng has quit IRC | 21:33 | |
| *** cdent has quit IRC | 21:33 | |
| *** sdake_ is now known as sdake | 21:33 | |
| cfriesen | bauzas: query it once at nova-compute startup and then update a local list after that. shouldn't be too bad. maybe we could audit the sanity of that list separately from the resource tracker | 21:33 |
| *** shaohe_feng has joined #openstack-nova | 21:34 | |
| auggy | sneti: yeah the instructions i wrote aren't very clear, if you can't repro following them, then just mark it as invalid | 21:34 |
| *** palexster has quit IRC | 21:34 | |
| sneti | auggy, sure..thanks for your response | 21:34 |
| bauzas | cfriesen: we already do that kind of things for notifying the scheduler, you should pay attention to it | 21:34 |
| *** e0ne has joined #openstack-nova | 21:34 | |
| bauzas | cfriesen: but that's not a RT thing | 21:34 |
| *** roeyc has quit IRC | 21:34 | |
| cfriesen | dansmith: so are you suggesting that we should be faster-but-maybe-inaccurate in the resource audit? | 21:36 |
| cfriesen | ie status quo? | 21:36 |
| dansmith | I'm not suggesting anything | 21:37 |
| *** salv-orlando has quit IRC | 21:37 | |
| *** e0ne has quit IRC | 21:38 | |
| cfriesen | it kind of looks like https://git.openstack.org/cgit/openstack/nova/commit/?id=833357301bc80a27422f7bf081fae2d3da730a24 is the real fix for the issue raised in bug 1371677...or am I reading that wrong? | 21:38 |
| openstack | bug 1371677 in OpenStack Compute (nova) "Race in resource tracker causes 500 response on deleting during verify_resize state " [Critical,Fix released] https://launchpad.net/bugs/1371677 - Assigned to Dan Smith (danms) | 21:38 |
| *** baoli has joined #openstack-nova | 21:39 | |
| *** baoli has quit IRC | 21:39 | |
| *** baoli has joined #openstack-nova | 21:40 | |
| *** derekh has quit IRC | 21:40 | |
| *** lucasagomes has quit IRC | 21:41 | |
| *** rlrossit has quit IRC | 21:41 | |
| *** baoli_ has joined #openstack-nova | 21:42 | |
| *** palexster has joined #openstack-nova | 21:42 | |
| *** shaohe_feng has quit IRC | 21:43 | |
| *** ssurana has quit IRC | 21:43 | |
| diana_clarke | Are there any coverage reports available for nova on a per patch basis from Zuul? | 21:43 |
| *** shaohe_feng has joined #openstack-nova | 21:44 | |
| *** lucasagomes has joined #openstack-nova | 21:44 | |
| *** baoli has quit IRC | 21:45 | |
| *** ekhugen has quit IRC | 21:45 | |
| cfriesen | Looking at _get_instance_disk_info() it appears to now track volume devices, so I suspect the original issue would no longer apply. | 21:45 |
| *** ekhugen has joined #openstack-nova | 21:45 | |
| *** Sukhdev has quit IRC | 21:47 | |
| mriedem | diana_clarke: like unit test coverage? | 21:48 |
| diana_clarke | mriedem: yup | 21:48 |
| mriedem | diana_clarke: there is a coverage tox target, but it's not in the check queue | 21:49 |
| mriedem | might be in the experimental queue | 21:49 |
| mriedem | but you could just run that locally too | 21:49 |
| mriedem | at one point i thought mtreinish had a way to get api coverage from tempest runs, but required some changes to nova to instrument it? | 21:50 |
| mtreinish | mriedem: https://wiki.openstack.org/wiki/Nova/CoverageExtension and https://review.openstack.org/#/c/130855/ | 21:50 |
| mriedem | yeah those | 21:50 |
| diana_clarke | And I was just looking at: https://review.openstack.org/#/c/221494/ | 21:51 |
| *** zhenq has quit IRC | 21:51 | |
| *** kylek3h_ has quit IRC | 21:51 | |
| *** salv-orlando has joined #openstack-nova | 21:51 | |
| jroll | mriedem: diana_clarke: it's only in the post queue | 21:52 |
| jroll | https://github.com/openstack-infra/project-config/blob/master/zuul/layout.yaml#L7983 | 21:52 |
| *** zenoway has quit IRC | 21:52 | |
| diana_clarke | I'm running coverage locally to answer a question I have, just wondered if I could have checked zuul for a faster answer is all. | 21:52 |
| *** zenoway has joined #openstack-nova | 21:53 | |
| *** shaohe_feng has quit IRC | 21:53 | |
| *** shaohe_feng has joined #openstack-nova | 21:54 | |
| doffm | This mailing list thread is getting me down. There should be a new rule just like "don't read the comments". | 21:57 |
| *** ccarmack has quit IRC | 21:57 | |
| *** vilobhmm11 has quit IRC | 21:59 | |
| *** mriedem has quit IRC | 21:59 | |
| *** burt has quit IRC | 22:00 | |
| *** antonyfm has quit IRC | 22:00 | |
| *** kengo_sakai has joined #openstack-nova | 22:00 | |
| *** antonyfm has joined #openstack-nova | 22:01 | |
| *** tbachman has quit IRC | 22:01 | |
| *** rtheis has quit IRC | 22:01 | |
| *** vilobhmm11 has joined #openstack-nova | 22:02 | |
| *** vilobhmm11 has quit IRC | 22:02 | |
| *** vilobhmm11 has joined #openstack-nova | 22:03 | |
| *** shaohe_feng has quit IRC | 22:04 | |
| *** sambetts|afk has quit IRC | 22:04 | |
| *** mylu has quit IRC | 22:04 | |
| *** vilobhmm11 has quit IRC | 22:04 | |
| *** shaohe_feng has joined #openstack-nova | 22:04 | |
| *** doug-fi__ has quit IRC | 22:04 | |
| *** kengo_sakai has quit IRC | 22:05 | |
| *** mylu has joined #openstack-nova | 22:05 | |
| *** salv-orlando has quit IRC | 22:06 | |
| *** sambetts has joined #openstack-nova | 22:07 | |
| *** mageshgv has quit IRC | 22:07 | |
| *** ducttape_ has quit IRC | 22:09 | |
| *** trams_ has quit IRC | 22:09 | |
| *** trams has joined #openstack-nova | 22:09 | |
| *** mylu has quit IRC | 22:09 | |
| *** tlbr has joined #openstack-nova | 22:12 | |
| *** mylu has joined #openstack-nova | 22:12 | |
| *** ijw has joined #openstack-nova | 22:14 | |
| *** baoli_ has quit IRC | 22:14 | |
| *** shaohe_feng has quit IRC | 22:14 | |
| *** ijw has quit IRC | 22:14 | |
| *** shaohe_feng has joined #openstack-nova | 22:14 | |
| *** salv-orlando has joined #openstack-nova | 22:15 | |
| *** mdrabe has quit IRC | 22:19 | |
| *** moshele has joined #openstack-nova | 22:22 | |
| *** xyang1 has quit IRC | 22:23 | |
| *** ccarmack has joined #openstack-nova | 22:23 | |
| sneti | hi johnthetubaguy...can you please check this patch: https://review.openstack.org/#/c/290156/ | 22:23 |
| *** shaohe_feng has quit IRC | 22:24 | |
| *** ljxiash has joined #openstack-nova | 22:25 | |
| *** shaohe_feng has joined #openstack-nova | 22:25 | |
| *** zenoway has quit IRC | 22:27 | |
| *** salv-orlando has quit IRC | 22:28 | |
| *** salv-orlando has joined #openstack-nova | 22:28 | |
| *** dimtruck is now known as zz_dimtruck | 22:29 | |
| *** ljxiash has quit IRC | 22:30 | |
| *** ijw has joined #openstack-nova | 22:30 | |
| *** ijw has quit IRC | 22:32 | |
| *** ijw has joined #openstack-nova | 22:32 | |
| *** ctrath has quit IRC | 22:34 | |
| diana_clarke | jroll: Thank you *so* much! That's exactly what I was looking for. Example: http://logs.openstack.org/9b/9b2329d10021a7fa5f43b9fd8be3c833dd7212d0/post/nova-coverage-db/799bcac/cover/ | 22:34 |
| *** shaohe_feng has quit IRC | 22:34 | |
| *** diogogmt has quit IRC | 22:35 | |
| *** shaohe_feng has joined #openstack-nova | 22:37 | |
| *** pushkaru has quit IRC | 22:38 | |
| *** jamesden_ has quit IRC | 22:38 | |
| *** mylu has quit IRC | 22:38 | |
| *** moshele has quit IRC | 22:39 | |
| *** baoli has joined #openstack-nova | 22:39 | |
| *** rook has joined #openstack-nova | 22:41 | |
| *** salv-orlando has quit IRC | 22:42 | |
| *** krotscheck is now known as krotscheck_dcm | 22:44 | |
| *** shaohe_feng has quit IRC | 22:45 | |
| *** rbridgeman has quit IRC | 22:45 | |
| *** shaohe_feng has joined #openstack-nova | 22:45 | |
| *** ametts has quit IRC | 22:45 | |
| *** bwallis has quit IRC | 22:45 | |
| *** mylu has joined #openstack-nova | 22:46 | |
| *** pwnall1 has joined #openstack-nova | 22:47 | |
| *** pwnall has quit IRC | 22:47 | |
| *** mylu has quit IRC | 22:47 | |
| *** pwnall has joined #openstack-nova | 22:47 | |
| *** baoli has quit IRC | 22:48 | |
| *** baoli has joined #openstack-nova | 22:49 | |
| *** ssurana has joined #openstack-nova | 22:49 | |
| *** sdague has quit IRC | 22:50 | |
| *** ssurana has quit IRC | 22:50 | |
| *** pwnall1 has quit IRC | 22:51 | |
| *** ssurana has joined #openstack-nova | 22:53 | |
| *** ssurana has quit IRC | 22:53 | |
| *** baoli has quit IRC | 22:54 | |
| *** dmk0202 has quit IRC | 22:54 | |
| *** sdake has quit IRC | 22:55 | |
| *** sdake has joined #openstack-nova | 22:55 | |
| *** shaohe_feng has quit IRC | 22:55 | |
| *** shaohe_feng has joined #openstack-nova | 22:55 | |
| *** akshai has quit IRC | 22:57 | |
| *** ssurana has joined #openstack-nova | 22:57 | |
| *** baoli has joined #openstack-nova | 22:58 | |
| *** akshai has joined #openstack-nova | 22:58 | |
| *** ssurana has quit IRC | 22:59 | |
| *** akshai has quit IRC | 23:00 | |
| siva_krishnan | Hi Guys! | 23:01 |
| siva_krishnan | I was looking at this bug https://bugs.launchpad.net/nova/+bug/1554226. Stack trace shows nova.db.sqlalachemy.api.py has get _engine() and get_api_engine() using get_legacy_facade.get_engine which are deprecated in oslo_db.sqlalchemy. Should we remove these methods ? can someone help me in this one | 23:05 |
| openstack | Launchpad bug 1554226 in OpenStack Compute (nova) "Clean up warnings about enginefacade" [Low,In progress] - Assigned to Jeffrey Augustine (ja224e) | 23:05 |
| *** shaohe_feng has quit IRC | 23:05 | |
| *** shaohe_feng has joined #openstack-nova | 23:05 | |
| *** baoli has quit IRC | 23:06 | |
| *** baoli has joined #openstack-nova | 23:06 | |
| *** ssurana has joined #openstack-nova | 23:11 | |
| *** ssurana has quit IRC | 23:12 | |
| *** baoli has quit IRC | 23:13 | |
| *** pushkaru has joined #openstack-nova | 23:15 | |
| *** shaohe_feng has quit IRC | 23:15 | |
| *** shaohe_feng has joined #openstack-nova | 23:16 | |
| *** markvoelker_ has quit IRC | 23:16 | |
| *** rbridgeman has joined #openstack-nova | 23:20 | |
| *** antonyfm has joined #openstack-nova | 23:23 | |
| *** ccarmack has quit IRC | 23:24 | |
| *** BjoernT has quit IRC | 23:24 | |
| *** mriedem has joined #openstack-nova | 23:25 | |
| *** shaohe_feng has quit IRC | 23:26 | |
| *** shaohe_feng has joined #openstack-nova | 23:26 | |
| *** diogogmt has joined #openstack-nova | 23:28 | |
| *** pwnall has quit IRC | 23:28 | |
| *** nelsnels_ has joined #openstack-nova | 23:30 | |
| *** nelsnelson has quit IRC | 23:31 | |
| openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova: Remove legacy v2 unit tests[q-v] https://review.openstack.org/311656 | 23:31 |
| openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova: Remove legacy v2 unit tests[f-n] https://review.openstack.org/311658 | 23:31 |
| openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova: Remove legacy v2 unit tests[a-e] https://review.openstack.org/311659 | 23:31 |
| openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova: WIP: Remove legacy v2 API code completely https://review.openstack.org/311653 | 23:32 |
| *** shaohe_feng has quit IRC | 23:36 | |
| *** shaohe_feng has joined #openstack-nova | 23:36 | |
| *** antonyfm has quit IRC | 23:37 | |
| *** rfolco has quit IRC | 23:38 | |
| *** tjones has joined #openstack-nova | 23:39 | |
| *** pushkaru has quit IRC | 23:39 | |
| *** pushkaru has joined #openstack-nova | 23:40 | |
| *** Swami has quit IRC | 23:43 | |
| *** pumarani__ has joined #openstack-nova | 23:44 | |
| *** pushkaru has quit IRC | 23:44 | |
| *** efried1 has joined #openstack-nova | 23:44 | |
| *** rbridgeman has quit IRC | 23:45 | |
| *** shaohe_feng has quit IRC | 23:46 | |
| *** efried has quit IRC | 23:47 | |
| *** shaohe_feng has joined #openstack-nova | 23:47 | |
| *** shaohe_feng has quit IRC | 23:56 | |
| *** shaohe_feng has joined #openstack-nova | 23:57 | |
| *** pumarani__ has quit IRC | 23:57 | |
| *** efried has joined #openstack-nova | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!