| *** pumaranikar has joined #openstack-nova | 00:00 | |
| *** sacharya has quit IRC | 00:00 | |
| *** aginwala has quit IRC | 00:01 | |
| *** mrkz has quit IRC | 00:01 | |
| *** zhangjn has joined #openstack-nova | 00:02 | |
| *** apoorvad has quit IRC | 00:03 | |
| *** aginwala has joined #openstack-nova | 00:04 | |
| *** otter768 has joined #openstack-nova | 00:04 | |
| *** apoorvad has joined #openstack-nova | 00:08 | |
| *** akshai has quit IRC | 00:08 | |
| *** sfinucan has quit IRC | 00:09 | |
| *** smurke_ has quit IRC | 00:09 | |
| *** EinstCrazy has joined #openstack-nova | 00:09 | |
| *** boris-42 has joined #openstack-nova | 00:10 | |
| *** signed8bit is now known as signed8bit_ZZZzz | 00:10 | |
| openstackgerrit | Andrew Laski proposed openstack/nova: Add persistence to the RequestSpec object https://review.openstack.org/211753 | 00:11 |
|---|---|---|
| alaski | bauzas: thanks, cleaned up a few things | 00:11 |
| boris-42 | alaski: bauzas dansmith hi there | 00:12 |
| *** pumaranikar has quit IRC | 00:13 | |
| boris-42 | alaski: bauzas dansmith could you please a take a look at OSProfiler spec: I have updated it and overall make it much simpler & clear | 00:13 |
| boris-42 | alaski: bauzas dansmith here is the link https://review.openstack.org/#/c/103825/ | 00:13 |
| boris-42 | thank you | 00:13 |
| *** signed8bit_ZZZzz is now known as signed8bit | 00:15 | |
| *** openstackgerrit has quit IRC | 00:16 | |
| *** ptm_away has quit IRC | 00:17 | |
| *** openstackgerrit has joined #openstack-nova | 00:17 | |
| *** andrearosa has quit IRC | 00:17 | |
| *** ptm_away has joined #openstack-nova | 00:17 | |
| *** andrearosa has joined #openstack-nova | 00:18 | |
| *** zhangjn has quit IRC | 00:19 | |
| *** EinstCrazy has quit IRC | 00:19 | |
| *** jang has quit IRC | 00:19 | |
| openstackgerrit | Tracy Jones proposed openstack/nova: Cleanup of compute api reboot method https://review.openstack.org/127427 | 00:19 |
| *** tjones has joined #openstack-nova | 00:19 | |
| tjones | melwitt: i've finally addressed your comments https://review.openstack.org/#/c/127427/48. thanks for the review | 00:21 |
| *** sacharya has joined #openstack-nova | 00:21 | |
| *** diegows has quit IRC | 00:22 | |
| *** jang has joined #openstack-nova | 00:22 | |
| *** otter768 has quit IRC | 00:23 | |
| *** sacharya has quit IRC | 00:23 | |
| *** aginwala has quit IRC | 00:25 | |
| *** otter768 has joined #openstack-nova | 00:26 | |
| tjones | edleafe: hi i am going to start helping with the config option consolidation. I'll just grab a section give it a try. any preference on which i grab? | 00:30 |
| *** nic has quit IRC | 00:31 | |
| *** aginwala has joined #openstack-nova | 00:33 | |
| *** FL1SK has quit IRC | 00:33 | |
| *** vladikr has joined #openstack-nova | 00:36 | |
| *** aginwala has quit IRC | 00:37 | |
| *** signed8bit is now known as signed8bit_ZZZzz | 00:37 | |
| *** stackdump1 has joined #openstack-nova | 00:37 | |
| *** electroc_ has joined #openstack-nova | 00:37 | |
| *** stackdump has quit IRC | 00:39 | |
| *** electroc_ has quit IRC | 00:39 | |
| *** jwcroppe has joined #openstack-nova | 00:40 | |
| *** jinxing has joined #openstack-nova | 00:40 | |
| *** electroc_ has joined #openstack-nova | 00:41 | |
| *** salv-orlando has quit IRC | 00:43 | |
| *** aginwala has joined #openstack-nova | 00:43 | |
| *** otter768 has quit IRC | 00:45 | |
| *** electroc_ has quit IRC | 00:47 | |
| *** electroc_ has joined #openstack-nova | 00:54 | |
| *** r-daneel has quit IRC | 00:55 | |
| *** xyang1 has quit IRC | 00:58 | |
| openstackgerrit | Tracy Jones proposed openstack/nova: config options: centralize section "guestfs" https://review.openstack.org/248963 | 01:00 |
| *** zhangjn has joined #openstack-nova | 01:00 | |
| *** zhangjn has quit IRC | 01:00 | |
| *** EinstCrazy has joined #openstack-nova | 01:00 | |
| *** rushil has joined #openstack-nova | 01:01 | |
| *** zhangjn has joined #openstack-nova | 01:01 | |
| *** ssurana has quit IRC | 01:02 | |
| *** ssurana has joined #openstack-nova | 01:02 | |
| *** EinstCra_ has joined #openstack-nova | 01:02 | |
| *** aginwala has quit IRC | 01:05 | |
| *** aginwala has joined #openstack-nova | 01:06 | |
| *** ssurana has quit IRC | 01:06 | |
| *** EinstCrazy has quit IRC | 01:06 | |
| *** aginwala_ has joined #openstack-nova | 01:06 | |
| *** nelsnelson has quit IRC | 01:09 | |
| *** jerrygb has quit IRC | 01:09 | |
| *** mc_nair has quit IRC | 01:09 | |
| *** jerrygb has joined #openstack-nova | 01:10 | |
| *** aginwala has quit IRC | 01:10 | |
| openstackgerrit | Tracy Jones proposed openstack/nova: config options: centralize section "guestfs" https://review.openstack.org/248963 | 01:11 |
| *** vilobhmm1 has joined #openstack-nova | 01:12 | |
| *** vilobhmm has quit IRC | 01:12 | |
| *** zhangjn has quit IRC | 01:12 | |
| *** eliqiao_ is now known as eliqiao | 01:12 | |
| *** jerrygb has quit IRC | 01:14 | |
| *** ijw_ has quit IRC | 01:15 | |
| *** jhesketh has joined #openstack-nova | 01:15 | |
| *** ljxiash has joined #openstack-nova | 01:15 | |
| *** jhesketh has quit IRC | 01:16 | |
| *** apoorvad has quit IRC | 01:17 | |
| *** stackdump1 has quit IRC | 01:18 | |
| *** pratikmallya has joined #openstack-nova | 01:20 | |
| *** zhangjn has joined #openstack-nova | 01:20 | |
| *** mc_nair has joined #openstack-nova | 01:22 | |
| openstackgerrit | Tracy Jones proposed openstack/nova: config options: centralize section "guestfs" https://review.openstack.org/248963 | 01:22 |
| *** electroc_ has quit IRC | 01:23 | |
| *** stackdump has joined #openstack-nova | 01:26 | |
| *** stackdump has quit IRC | 01:27 | |
| *** zhangjn_ has joined #openstack-nova | 01:27 | |
| *** pixelbeat has quit IRC | 01:29 | |
| *** zhangjn has quit IRC | 01:29 | |
| *** pixelbeat has joined #openstack-nova | 01:29 | |
| *** jgriffith_away is now known as jgriffith | 01:30 | |
| *** zhenguo has joined #openstack-nova | 01:38 | |
| *** baoli has quit IRC | 01:38 | |
| *** bzhou has joined #openstack-nova | 01:38 | |
| *** baoli has joined #openstack-nova | 01:39 | |
| *** thorst has joined #openstack-nova | 01:40 | |
| *** thorst has quit IRC | 01:40 | |
| *** thorst has joined #openstack-nova | 01:41 | |
| *** pixelbeat has quit IRC | 01:42 | |
| *** pixelbeat has joined #openstack-nova | 01:44 | |
| *** sbezverk has joined #openstack-nova | 01:45 | |
| *** sbezverk has left #openstack-nova | 01:46 | |
| *** LiuNanke has joined #openstack-nova | 01:49 | |
| *** edtubill has joined #openstack-nova | 01:49 | |
| *** shihanzhang has joined #openstack-nova | 01:50 | |
| *** aginwala_ has quit IRC | 01:52 | |
| *** ijw has joined #openstack-nova | 01:52 | |
| *** aginwala has joined #openstack-nova | 01:55 | |
| *** suro-patz has joined #openstack-nova | 01:57 | |
| *** Kennan has quit IRC | 01:57 | |
| openstackgerrit | li,chen proposed openstack/nova: Add unit tests for nova.virt.firewall.IpTablesFirewallDriver https://review.openstack.org/246274 | 01:57 |
| *** tjones has quit IRC | 01:58 | |
| *** jhesketh has joined #openstack-nova | 01:58 | |
| *** LiuNanke has quit IRC | 01:58 | |
| *** 17SAD4YI2 has joined #openstack-nova | 02:00 | |
| *** 17SAD4YI2 has quit IRC | 02:01 | |
| *** haomaiwa_ has joined #openstack-nova | 02:01 | |
| *** jhesketh has quit IRC | 02:03 | |
| *** ijw has quit IRC | 02:03 | |
| *** thorst has quit IRC | 02:04 | |
| *** jhesketh has joined #openstack-nova | 02:06 | |
| *** thorst has joined #openstack-nova | 02:08 | |
| *** suro-patz has quit IRC | 02:10 | |
| *** suro-patz has joined #openstack-nova | 02:11 | |
| *** sripriya_ has quit IRC | 02:16 | |
| *** Kennan has joined #openstack-nova | 02:17 | |
| *** gcb has joined #openstack-nova | 02:18 | |
| *** signed8bit_ZZZzz is now known as signed8bit | 02:19 | |
| *** thorst has quit IRC | 02:22 | |
| *** apoorvad has joined #openstack-nova | 02:23 | |
| *** vilobhmm1 has quit IRC | 02:23 | |
| *** vilobhmm has joined #openstack-nova | 02:24 | |
| openstackgerrit | Tang Chen proposed openstack/nova-specs: Support triggering crash dump in a server. https://review.openstack.org/229255 | 02:28 |
| *** signed8bit has quit IRC | 02:32 | |
| *** FL1SK has joined #openstack-nova | 02:33 | |
| *** Marga_ has quit IRC | 02:33 | |
| *** david-ly_ has quit IRC | 02:36 | |
| *** mylu has joined #openstack-nova | 02:38 | |
| *** aginwala has quit IRC | 02:39 | |
| *** dims has joined #openstack-nova | 02:40 | |
| *** Kennan has quit IRC | 02:42 | |
| *** mylu has quit IRC | 02:42 | |
| *** jerrygb has joined #openstack-nova | 02:44 | |
| *** otter768 has joined #openstack-nova | 02:46 | |
| *** salv-orlando has joined #openstack-nova | 02:47 | |
| *** vilobhmm has quit IRC | 02:48 | |
| *** pixelbeat has quit IRC | 02:48 | |
| *** mc_nair has quit IRC | 02:49 | |
| *** otter768 has quit IRC | 02:50 | |
| *** sacharya has joined #openstack-nova | 02:53 | |
| *** suro-patz has quit IRC | 02:54 | |
| *** aginwala has joined #openstack-nova | 02:56 | |
| *** salv-orlando has quit IRC | 02:59 | |
| *** haomaiwa_ has quit IRC | 03:01 | |
| *** haomaiwang has joined #openstack-nova | 03:01 | |
| *** venkat_p has joined #openstack-nova | 03:02 | |
| *** venkat_p has left #openstack-nova | 03:02 | |
| *** jwcroppe has quit IRC | 03:05 | |
| *** jwcroppe has joined #openstack-nova | 03:06 | |
| *** suro-patz has joined #openstack-nova | 03:06 | |
| *** Sree has joined #openstack-nova | 03:08 | |
| *** Marga_ has joined #openstack-nova | 03:09 | |
| oomichi | johnthetubaguy: sdague: tangchen_: hi, about https://review.openstack.org/#/c/229255/ | 03:11 |
| *** jwcroppe has quit IRC | 03:11 | |
| *** jerrygb has quit IRC | 03:11 | |
| *** mylu has joined #openstack-nova | 03:11 | |
| *** klkumar has joined #openstack-nova | 03:11 | |
| oomichi | johnthetubaguy: sdague: tangchen_: the api action name(triggerCrashDump) fits to current other actions. but the name is against api-wg guideline. | 03:12 |
| *** jerrygb has joined #openstack-nova | 03:13 | |
| *** Kennan has joined #openstack-nova | 03:13 | |
| oomichi | johnthetubaguy: sdague: tangchen_: I guess better to fit to api-wg guideline(snake_case) at this time, and nice to find a chance to make the other actions fit to api-wg guideline. | 03:13 |
| *** Marga_ has quit IRC | 03:14 | |
| Kennan | hi alex_xu: | 03:14 |
| Kennan | one question, I have openstack env(all-in-one), which have 8 CPU, and I try to boot one instance with flavor vCPU 5, I found the instance run some program very busy | 03:16 |
| Kennan | while the openstack node still very idle | 03:16 |
| Kennan | What's the meaning for VCPU in nova ? | 03:17 |
| *** dims has quit IRC | 03:17 | |
| Kennan | how can make sure the booted instance really use CPU in the host ? | 03:17 |
| Kennan | alex_xu | 03:18 |
| Kennan | what do you think ? | 03:18 |
| *** flyingtt has quit IRC | 03:20 | |
| gcb | Kennan: what virt driver are you using , kvm ? | 03:21 |
| *** dims has joined #openstack-nova | 03:22 | |
| *** pcaruana has quit IRC | 03:27 | |
| *** apoorvad has quit IRC | 03:27 | |
| *** klkumar has quit IRC | 03:28 | |
| *** vilobhmm has joined #openstack-nova | 03:30 | |
| *** ccard__ has joined #openstack-nova | 03:31 | |
| *** ccard_ has quit IRC | 03:34 | |
| *** jinxing has quit IRC | 03:35 | |
| *** dims has quit IRC | 03:35 | |
| *** jinxing has joined #openstack-nova | 03:36 | |
| *** klkumar has joined #openstack-nova | 03:44 | |
| *** mylu has quit IRC | 03:48 | |
| *** lixiaoy1 has joined #openstack-nova | 03:48 | |
| *** mylu has joined #openstack-nova | 03:49 | |
| *** mylu has quit IRC | 03:53 | |
| *** Kevin_Zheng has joined #openstack-nova | 03:56 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: [WIP]Expose quiesce/unquiesce API https://review.openstack.org/248989 | 03:57 |
| *** pratikmallya has quit IRC | 03:57 | |
| *** mylu has joined #openstack-nova | 03:58 | |
| *** haomaiwang has quit IRC | 04:01 | |
| *** haomaiwa_ has joined #openstack-nova | 04:01 | |
| *** zhangjn_ has quit IRC | 04:02 | |
| *** jwcroppe has joined #openstack-nova | 04:08 | |
| *** david-lyle has joined #openstack-nova | 04:08 | |
| *** aginwala has quit IRC | 04:09 | |
| *** akshai has joined #openstack-nova | 04:10 | |
| *** thorst has joined #openstack-nova | 04:11 | |
| *** akshai_ has joined #openstack-nova | 04:11 | |
| *** vilobhmm has quit IRC | 04:12 | |
| tangchen__ | oomichi: Hi, I saw your comment, thx. :) Will fix it in the next version. | 04:12 |
| *** otter768 has joined #openstack-nova | 04:13 | |
| tangchen__ | oomichi: BTW, please also help to review https://review.openstack.org/#/c/229282/ | 04:13 |
| tangchen__ | oomichi: Thx. :) | 04:13 |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: [WIP]Expose quiesce/unquiesce API https://review.openstack.org/248989 | 04:14 |
| *** akshai has quit IRC | 04:15 | |
| *** thorst has quit IRC | 04:15 | |
| *** sacharya has quit IRC | 04:18 | |
| *** zhangjn has joined #openstack-nova | 04:22 | |
| *** edtubill has quit IRC | 04:22 | |
| *** jerrygb has quit IRC | 04:24 | |
| *** edtubill has joined #openstack-nova | 04:24 | |
| *** jwcroppe has quit IRC | 04:27 | |
| *** edtubill has quit IRC | 04:29 | |
| oomichi | tangchen__: ok, done on https://review.openstack.org/#/c/229282/ | 04:31 |
| oomichi | tangchen__: some nit comments, please see it again | 04:31 |
| *** pratikmallya has joined #openstack-nova | 04:33 | |
| *** bzhou has quit IRC | 04:37 | |
| *** anish_ is now known as anish | 04:40 | |
| *** sacharya has joined #openstack-nova | 04:40 | |
| *** bzhou has joined #openstack-nova | 04:43 | |
| *** zhangjn has quit IRC | 04:44 | |
| *** suro-patz has quit IRC | 04:45 | |
| *** hparekh has joined #openstack-nova | 04:50 | |
| *** moshele has joined #openstack-nova | 04:53 | |
| *** klkumar has quit IRC | 04:54 | |
| *** smurke has joined #openstack-nova | 04:58 | |
| *** moshele has quit IRC | 04:58 | |
| *** jmccrory has quit IRC | 04:58 | |
| *** otter768 has quit IRC | 04:59 | |
| *** electroc_ has joined #openstack-nova | 04:59 | |
| *** jmccrory has joined #openstack-nova | 05:00 | |
| *** akshai_ has quit IRC | 05:00 | |
| *** haomaiwa_ has quit IRC | 05:01 | |
| *** haomaiwang has joined #openstack-nova | 05:01 | |
| *** jwcroppe has joined #openstack-nova | 05:04 | |
| *** klkumar has joined #openstack-nova | 05:07 | |
| alex_xu | Kennan: for kvm, one vcpu is one thread in the linux system | 05:14 |
| Kennan | hi alex_xu: how can make sure the booted instance really hard-occupy 1 physical core or CPU | 05:15 |
| Kennan | I did not want the instance seems like many VCPUS, but still very busy, while host very idle | 05:15 |
| *** rcernin has joined #openstack-nova | 05:15 | |
| Kennan | alex_xu: I remembered nova have such setting | 05:16 |
| *** Marga_ has joined #openstack-nova | 05:16 | |
| alex_xu | Kennan: yea, I'm try to find the link for you | 05:16 |
| alex_xu | Kennan: http://specs.openstack.org/openstack/nova-specs/specs/kilo/implemented/virt-driver-cpu-pinning.html I guess this is thing you are looking for | 05:17 |
| alex_xu | Kennan: and this may helpful http://specs.openstack.org/openstack/nova-specs/specs/juno/implemented/virt-driver-numa-placement.html | 05:17 |
| Kennan | alex_xu: this seesm a spec, does nova support this now ? how do you usually make instance performance good when find CPU very busy for instance ? seems only increase VCPU for instance not work | 05:19 |
| *** haomaiwang has quit IRC | 05:19 | |
| tangchen__ | oomichi: Thx. Will update it today.:) | 05:20 |
| *** deepakcs has joined #openstack-nova | 05:21 | |
| openstackgerrit | Tang Chen proposed openstack/nova-specs: Support triggering crash dump in a server. https://review.openstack.org/229255 | 05:22 |
| *** zhangjn has joined #openstack-nova | 05:23 | |
| *** gcb has quit IRC | 05:24 | |
| openstackgerrit | Tang Chen proposed openstack/nova-specs: Support soft reboot and poweroff in nova ironic driver. https://review.openstack.org/229282 | 05:26 |
| *** haomaiwa_ has joined #openstack-nova | 05:27 | |
| *** smurke has quit IRC | 05:27 | |
| *** achanda has joined #openstack-nova | 05:29 | |
| tangchen__ | alex_xu: hi alex, please also help to review the two specs above. Thx. :) | 05:30 |
| *** deepakcs has quit IRC | 05:32 | |
| *** rcernin has quit IRC | 05:34 | |
| *** EinstCra_ has quit IRC | 05:37 | |
| *** EinstCrazy has joined #openstack-nova | 05:37 | |
| openstackgerrit | Yang Wei proposed openstack/nova-specs: Add maxcpus support for qemu/kvm https://review.openstack.org/249012 | 05:38 |
| *** irenab has quit IRC | 05:38 | |
| *** achanda has quit IRC | 05:41 | |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: db: adding columns to PciDevice table https://review.openstack.org/249013 | 05:41 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: db: querry to retrieve all pci device by parent address https://review.openstack.org/249014 | 05:41 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: objects: adding a parent_addr field to the PciDevice object https://review.openstack.org/249015 | 05:41 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: pci: filter out any unavailable SRIOV Physical Functions https://review.openstack.org/249016 | 05:41 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: pci: adding support to specify a dev_type in pci requests https://review.openstack.org/249017 | 05:41 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: pci: changing the claiming and allocation logic for PF/VF assignment https://review.openstack.org/249018 | 05:41 |
| *** gcb has joined #openstack-nova | 05:41 | |
| *** deepakcs has joined #openstack-nova | 05:43 | |
| *** klkumar has quit IRC | 05:46 | |
| *** pcaruana has joined #openstack-nova | 05:49 | |
| *** deepakcs has quit IRC | 05:53 | |
| *** electroc_ has quit IRC | 05:54 | |
| openstackgerrit | Yingxin Cheng proposed openstack/nova: Use stevedore for scheduler host manager https://review.openstack.org/246476 | 05:55 |
| *** mylu has quit IRC | 05:56 | |
| openstackgerrit | Yingxin Cheng proposed openstack/nova: Use stevedore for scheduler host manager https://review.openstack.org/246476 | 05:56 |
| *** salv-orlando has joined #openstack-nova | 05:56 | |
| *** mylu has joined #openstack-nova | 05:56 | |
| *** nelsnelson has joined #openstack-nova | 05:58 | |
| *** lpetrut has joined #openstack-nova | 06:00 | |
| *** salv-orlando has quit IRC | 06:01 | |
| *** haomaiwa_ has quit IRC | 06:01 | |
| *** mylu has quit IRC | 06:01 | |
| *** haomaiwa_ has joined #openstack-nova | 06:02 | |
| *** klkumar has joined #openstack-nova | 06:02 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Imported Translations from Zanata https://review.openstack.org/242727 | 06:09 |
| *** Marga_ has quit IRC | 06:09 | |
| *** Marga_ has joined #openstack-nova | 06:10 | |
| *** mylu has joined #openstack-nova | 06:10 | |
| *** wznoinsk has quit IRC | 06:11 | |
| *** Vek has quit IRC | 06:12 | |
| *** wznoinsk has joined #openstack-nova | 06:12 | |
| *** sulo has quit IRC | 06:12 | |
| *** thorst has joined #openstack-nova | 06:13 | |
| *** sulo has joined #openstack-nova | 06:13 | |
| *** thorst has quit IRC | 06:17 | |
| *** rcernin has joined #openstack-nova | 06:18 | |
| *** liverpooler has joined #openstack-nova | 06:19 | |
| *** electroc_ has joined #openstack-nova | 06:21 | |
| *** electroc_ has quit IRC | 06:24 | |
| *** vladikr has quit IRC | 06:27 | |
| *** electroc_ has joined #openstack-nova | 06:27 | |
| openstackgerrit | Yingxin Cheng proposed openstack/nova: Use stevedore for scheduler host manager https://review.openstack.org/246476 | 06:28 |
| *** electroc_ has quit IRC | 06:28 | |
| *** pcaruana has quit IRC | 06:28 | |
| *** electroc_ has joined #openstack-nova | 06:29 | |
| *** rushil has quit IRC | 06:32 | |
| *** mylu has quit IRC | 06:32 | |
| *** irenab has joined #openstack-nova | 06:32 | |
| *** mylu has joined #openstack-nova | 06:32 | |
| *** mylu has quit IRC | 06:37 | |
| *** e0ne has joined #openstack-nova | 06:40 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: [WIP]Expose quiesce/unquiesce API https://review.openstack.org/248989 | 06:41 |
| *** deepakcs has joined #openstack-nova | 06:43 | |
| *** EinstCrazy has quit IRC | 06:45 | |
| *** cfriesen__ has quit IRC | 06:45 | |
| *** electroc_ has quit IRC | 06:46 | |
| *** suro-patz has joined #openstack-nova | 06:46 | |
| openstackgerrit | Yingxin Cheng proposed openstack/nova: Use stevedore for scheduler host manager https://review.openstack.org/246476 | 06:46 |
| *** lpetrut has quit IRC | 06:47 | |
| *** boris-42 has quit IRC | 06:48 | |
| *** EinstCrazy has joined #openstack-nova | 06:49 | |
| *** ankit_ag has joined #openstack-nova | 06:49 | |
| *** aginwala has joined #openstack-nova | 06:50 | |
| *** chinmaya has joined #openstack-nova | 06:50 | |
| *** takashin has joined #openstack-nova | 06:51 | |
| *** mjura has joined #openstack-nova | 06:51 | |
| *** tpatil has joined #openstack-nova | 06:54 | |
| *** garyk1 has joined #openstack-nova | 06:55 | |
| *** haomaiwa_ has quit IRC | 07:01 | |
| *** paul-carlton1 has joined #openstack-nova | 07:01 | |
| *** takashin has quit IRC | 07:01 | |
| *** haomaiwang has joined #openstack-nova | 07:01 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: [WIP]Expose quiesce/unquiesce API https://review.openstack.org/248989 | 07:02 |
| *** josecastroleon has joined #openstack-nova | 07:06 | |
| *** e0ne has quit IRC | 07:08 | |
| *** lpetrut has joined #openstack-nova | 07:08 | |
| *** nkrinner has joined #openstack-nova | 07:10 | |
| *** otter768 has joined #openstack-nova | 07:11 | |
| *** vilobhmm has joined #openstack-nova | 07:12 | |
| *** vilobhmm1 has joined #openstack-nova | 07:14 | |
| *** jgriffith is now known as jgriffith_away | 07:15 | |
| *** jaosorior has joined #openstack-nova | 07:15 | |
| *** otter768 has quit IRC | 07:15 | |
| *** aginwala has quit IRC | 07:16 | |
| *** vilobhmm has quit IRC | 07:17 | |
| *** yonglihe has quit IRC | 07:21 | |
| *** zenoway has joined #openstack-nova | 07:25 | |
| *** aginwala has joined #openstack-nova | 07:26 | |
| *** lpetrut has quit IRC | 07:27 | |
| *** sacharya has quit IRC | 07:27 | |
| *** suro-patz has quit IRC | 07:29 | |
| *** suro-patz has joined #openstack-nova | 07:29 | |
| openstackgerrit | Wei Yang proposed openstack/nova-specs: Add maxcpus support for qemu/kvm https://review.openstack.org/249012 | 07:30 |
| *** suro-patz has quit IRC | 07:30 | |
| *** sahid has joined #openstack-nova | 07:34 | |
| openstackgerrit | Juan Antonio Osorio Robles proposed openstack/nova: Set personality/injected_files to empty list if not specified https://review.openstack.org/248662 | 07:34 |
| *** jerrygb has joined #openstack-nova | 07:37 | |
| *** jerrygb has quit IRC | 07:42 | |
| *** ajo has quit IRC | 07:42 | |
| *** scheuran has joined #openstack-nova | 07:43 | |
| *** moshele has joined #openstack-nova | 07:43 | |
| *** rotbeard has joined #openstack-nova | 07:44 | |
| *** jaosorior has quit IRC | 07:45 | |
| openstackgerrit | Tang Chen proposed openstack/nova: Improve inject_nmi() in libvirt driver and add tests. https://review.openstack.org/208794 | 07:45 |
| tangchen__ | oomichi: Hi oomichi, Would you please help to review the patch above ? | 07:46 |
| tangchen__ | oomichi: It has been a long time since it was posted. | 07:46 |
| *** ajo has joined #openstack-nova | 07:46 | |
| tangchen__ | oomichi: It was part of NMI patch-set, but actually it has nothing to do with the spec. Just an improvement for the existing libvirt API. | 07:47 |
| tangchen__ | oomichi: So, I wish we can merge it first. | 07:48 |
| oomichi | tangchen__: "the existing libvirt API" is not used yet, right? | 07:48 |
| *** matrohon has joined #openstack-nova | 07:49 | |
| oomichi | tangchen__: it means new REST API will use/call it, IMO | 07:49 |
| *** vilobhmm1 has quit IRC | 07:53 | |
| *** gongysh_ has joined #openstack-nova | 07:54 | |
| oomichi | tangchen__: if so, we need to wait for the spec merger as the development manner | 07:54 |
| *** sacharya has joined #openstack-nova | 07:56 | |
| tangchen__ | oomichi: Yes, and OK, just wait for the spec to merge. :) | 07:57 |
| tangchen__ | oomichi: Thx. | 07:57 |
| *** otter768 has joined #openstack-nova | 08:00 | |
| *** haomaiwang has quit IRC | 08:01 | |
| *** jyuso1 has joined #openstack-nova | 08:01 | |
| *** haomaiwa_ has joined #openstack-nova | 08:02 | |
| *** sacharya has quit IRC | 08:03 | |
| *** rha has quit IRC | 08:03 | |
| *** hparekh has quit IRC | 08:04 | |
| *** pratikmallya has quit IRC | 08:04 | |
| *** matrohon1 has joined #openstack-nova | 08:05 | |
| *** matrohon1 has quit IRC | 08:05 | |
| *** hparekh has joined #openstack-nova | 08:05 | |
| *** otter768 has quit IRC | 08:05 | |
| *** rdopiera has joined #openstack-nova | 08:09 | |
| *** Vek has joined #openstack-nova | 08:09 | |
| *** hparekh2 has joined #openstack-nova | 08:10 | |
| *** hparekh has quit IRC | 08:10 | |
| *** lpetrut has joined #openstack-nova | 08:10 | |
| *** ildikov has quit IRC | 08:11 | |
| *** paul-carlton1 has quit IRC | 08:11 | |
| *** jerrygb has joined #openstack-nova | 08:12 | |
| *** salv-orlando has joined #openstack-nova | 08:13 | |
| *** wuhg has joined #openstack-nova | 08:13 | |
| *** jaosorior has joined #openstack-nova | 08:15 | |
| openstackgerrit | yuntongjin proposed openstack/nova-specs: report more live migration progress detail https://review.openstack.org/248472 | 08:16 |
| *** mpavone has joined #openstack-nova | 08:16 | |
| *** jerrygb has quit IRC | 08:17 | |
| *** jichen has joined #openstack-nova | 08:17 | |
| *** salv-orlando has quit IRC | 08:21 | |
| *** jaosorior has quit IRC | 08:21 | |
| *** tpatil has quit IRC | 08:22 | |
| *** jaosorior has joined #openstack-nova | 08:22 | |
| *** smurke has joined #openstack-nova | 08:24 | |
| *** bzhou has quit IRC | 08:26 | |
| *** paul-carlton1 has joined #openstack-nova | 08:26 | |
| *** markmc has quit IRC | 08:28 | |
| *** salv-orlando has joined #openstack-nova | 08:29 | |
| *** hparekh2 has quit IRC | 08:31 | |
| *** salv-orlando has quit IRC | 08:33 | |
| bauzas | good morning Nova | 08:35 |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: [WIP]Expose quiesce/unquiesce API https://review.openstack.org/248989 | 08:35 |
| *** subscope has joined #openstack-nova | 08:36 | |
| *** aginwala has quit IRC | 08:40 | |
| *** hparekh has joined #openstack-nova | 08:41 | |
| *** ildikov has joined #openstack-nova | 08:43 | |
| *** smurke has quit IRC | 08:44 | |
| tangchen__ | bauzas, alex_xu, johnthetubaguy, oomichi: I have updated this spec. https://review.openstack.org/#/c/229255/ | 08:45 |
| *** markus_z has joined #openstack-nova | 08:45 | |
| tangchen__ | bauzas, alex_xu, johnthetubaguy, oomichi: Please help to review whenever you have time. :) | 08:45 |
| tangchen__ | bauzas, alex_xu, johnthetubaguy, oomichi: Thanks. | 08:45 |
| *** jlanoux has joined #openstack-nova | 08:48 | |
| *** subscope has quit IRC | 08:48 | |
| oomichi | tangchen__: thanks for updating, LGTM | 08:48 |
| *** jccnd has quit IRC | 08:48 | |
| *** e0ne has joined #openstack-nova | 08:49 | |
| *** subscope has joined #openstack-nova | 08:52 | |
| alex_xu | oomichi: hi | 08:54 |
| alex_xu | tangchen__: will take look at | 08:54 |
| *** hparekh2 has joined #openstack-nova | 08:54 | |
| *** hparekh2 has quit IRC | 08:54 | |
| *** hparekh2 has joined #openstack-nova | 08:55 | |
| oomichi | alex_xu: hi | 08:55 |
| gibi | alex_xu: hi! you had concerns in https://review.openstack.org/#/c/243562/ Remove onSharedStorage from evacuate API. I answered them in the review. | 08:55 |
| alex_xu | oomichi: one question, can we remove a feature with microversion? | 08:55 |
| *** hparekh has quit IRC | 08:56 | |
| oomichi | alex_xu: I think so | 08:56 |
| alex_xu | gibi: yea, thanks, will take a look at it when I got time | 08:56 |
| oomichi | alex_xu: microversion should allow us backwards incompatible changes | 08:56 |
| gibi | alex_xu: thanks! | 08:56 |
| *** sahid has quit IRC | 08:56 | |
| oomichi | alex_xu: including removing a feature | 08:56 |
| oomichi | alex_xu: do you want to remove some features? ;) | 08:57 |
| alex_xu | oomichi: ok, thanks, just want to ensure your opionion, actually I get concern in my spec, bauzas suggest we should deprecate a feature first, before we remove it from api, there is the spec https://review.openstack.org/#/c/245543/7/specs/mitaka/approved/optional_block_migration_and_host_flags_in_live_migrate.rst | 08:57 |
| alex_xu | gibi: np | 08:57 |
| oomichi | alex_xu: yeah, deprecation period is user-friendly | 08:58 |
| *** Marga_ has quit IRC | 08:58 | |
| oomichi | alex_xu: but I'm not sure we need it really on microversion | 08:59 |
| alex_xu | oomichi: yea, but the old version give a time for user to ready to upgrade? | 08:59 |
| oomichi | alex_xu: because users can specify favorite microversion on client side | 08:59 |
| alex_xu | oomichi: yea | 08:59 |
| oomichi | alex_xu: yeah, that is the same as me :) | 09:00 |
| alex_xu | oomichi: thanks :) | 09:00 |
| oomichi | alex_xu: actually, we cannot bump minimum microversion forever | 09:00 |
| alex_xu | oomichi: yea, that's true | 09:00 |
| *** haomaiwa_ has quit IRC | 09:01 | |
| oomichi | alex_xu: for supporting the exiting clouds and new clouds | 09:01 |
| *** haomaiwa_ has joined #openstack-nova | 09:01 | |
| alex_xu | oomichi: yea | 09:01 |
| *** Marga_ has joined #openstack-nova | 09:02 | |
| *** hparekh2 has quit IRC | 09:05 | |
| openstackgerrit | Oleg Bondarev proposed openstack/nova: Live migration: wait for vif-plugged event on pre live migration https://review.openstack.org/246910 | 09:06 |
| *** eglynn has joined #openstack-nova | 09:07 | |
| tangchen__ | oomichi, alex_xu: Thx. :) | 09:08 |
| *** Sree has quit IRC | 09:08 | |
| *** Sree has joined #openstack-nova | 09:09 | |
| *** subscope has quit IRC | 09:09 | |
| *** hparekh has joined #openstack-nova | 09:09 | |
| *** subscope has joined #openstack-nova | 09:10 | |
| *** derekh has joined #openstack-nova | 09:10 | |
| *** yamahata has quit IRC | 09:11 | |
| *** sahid has joined #openstack-nova | 09:11 | |
| *** openstackgerrit has quit IRC | 09:16 | |
| *** openstackgerrit has joined #openstack-nova | 09:17 | |
| bauzas | alex_xu: oomichi: I possibly missed the outcome, you are saying that's okay to directly remove the attr by bumping a microversion ? | 09:18 |
| *** yassine__ has joined #openstack-nova | 09:18 | |
| oomichi | bauzas: yeah, I am thinking so, because api users can select microversion when sending a request | 09:18 |
| bauzas | oomichi: okay, fair enough | 09:19 |
| *** claudiub has joined #openstack-nova | 09:19 | |
| bauzas | alex_xu: please nevermind my comment then | 09:19 |
| bauzas | alex_xu: updating the spec | 09:20 |
| *** rha has joined #openstack-nova | 09:22 | |
| *** rha has joined #openstack-nova | 09:22 | |
| bauzas | johnthetubaguy: morning | 09:24 |
| openstackgerrit | yuntongjin proposed openstack/nova-specs: consistent nova instance progress https://review.openstack.org/249086 | 09:24 |
| bauzas | johnthetubaguy: I stupidly forgot to provide a reno change for the client... | 09:24 |
| bauzas | johnthetubaguy: so... https://review.openstack.org/#/c/248893/ | 09:24 |
| *** garyk1 has quit IRC | 09:24 | |
| *** garyk1 has joined #openstack-nova | 09:25 | |
| *** ndipanov has quit IRC | 09:25 | |
| *** ndipanov has joined #openstack-nova | 09:26 | |
| *** mdbooth has joined #openstack-nova | 09:27 | |
| *** lucas-afk is now known as lucasagomes | 09:31 | |
| *** jinxing has quit IRC | 09:32 | |
| andrearosa | johnthetubaguy: if you have time today, can you look at the latest comments for https://review.openstack.org/#/c/184537/ | 09:33 |
| *** claudiub has quit IRC | 09:33 | |
| *** markmc has joined #openstack-nova | 09:33 | |
| *** oomichi has quit IRC | 09:34 | |
| *** claudiub has joined #openstack-nova | 09:34 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: [WIP]Expose quiesce/unquiesce API https://review.openstack.org/248989 | 09:35 |
| *** achanda has joined #openstack-nova | 09:35 | |
| *** subscope has quit IRC | 09:37 | |
| *** paul-carlton1 has quit IRC | 09:41 | |
| *** jinxing has joined #openstack-nova | 09:42 | |
| *** ociuhandu has quit IRC | 09:42 | |
| *** jistr has joined #openstack-nova | 09:44 | |
| *** Marga_ has quit IRC | 09:46 | |
| openstackgerrit | Matthew Booth proposed openstack/nova-specs: Add detail to libvirt-storage-pools https://review.openstack.org/248705 | 09:46 |
| *** danpb has joined #openstack-nova | 09:47 | |
| johnthetubaguy | bauzas: oops, good point | 09:47 |
| *** zhangjn has quit IRC | 09:48 | |
| *** EinstCrazy has quit IRC | 09:48 | |
| johnthetubaguy | bauzas: I added a question | 09:49 |
| bauzas | johnthetubaguy: just answered, it's just autogenerated | 09:51 |
| bauzas | johnthetubaguy: I just have one open point in mind about the client versions | 09:51 |
| bauzas | johnthetubaguy: we could probably change the notes/*.rst file to make sure we can have a specific file for each version | 09:52 |
| bauzas | johnthetubaguy: for the moment, I prefer to leave a stable/liberty and an unreleased files | 09:53 |
| bauzas | johnthetubaguy: because it doesn't seem that reno supports tagging | 09:53 |
| johnthetubaguy | andrearosa: thanks for the ping, I just don't see those updates in my inbox any more :( | 09:54 |
| *** andymaier has joined #openstack-nova | 09:54 | |
| johnthetubaguy | andrearosa: I am totally agreed with it being a nova-manage command | 09:54 |
| *** Marga_ has joined #openstack-nova | 09:54 | |
| andrearosa | johnthetubaguy: np thanks for your time | 09:55 |
| johnthetubaguy | andrearosa: I just think that code would be easier to maintain if it was just an entry point in compute.api.py, but seems like thats not the consensus on there | 09:55 |
| *** paul-carlton1 has joined #openstack-nova | 09:55 | |
| *** alexschm has joined #openstack-nova | 09:55 | |
| johnthetubaguy | andrearosa: seems my wording on my comments was very unclear, sorry about that | 09:55 |
| openstackgerrit | Michal Dubiel proposed openstack/nova: libvirt, vhostuser: Add support for OpenContrail vRouter https://review.openstack.org/241652 | 09:56 |
| johnthetubaguy | bauzas: ah, interesting, we do have a branch at least, so thats cool, maybe master just gets bumped into other files as we apply the tags? | 10:00 |
| *** haomaiwa_ has quit IRC | 10:01 | |
| bauzas | johnthetubaguy: mmm, sec, I think that pbr checks the tags, right? | 10:01 |
| johnthetubaguy | bauzas: unsure | 10:01 |
| *** otter768 has joined #openstack-nova | 10:01 | |
| bauzas | johnthetubaguy: because we're using pbr for saying which novaclient version it is | 10:01 |
| johnthetubaguy | unreleased and liberty is almost accurate at least | 10:01 |
| bauzas | johnthetubaguy: checking locally | 10:01 |
| *** haomaiwa_ has joined #openstack-nova | 10:02 | |
| bauzas | johnthetubaguy: lemme clarify | 10:02 |
| bauzas | johnthetubaguy: so, we have 2 branches and multiple tags | 10:02 |
| *** zhenguo has quit IRC | 10:02 | |
| bauzas | johnthetubaguy: one for liberty and one for master + tags | 10:02 |
| johnthetubaguy | yep, thats correct | 10:02 |
| bauzas | johnthetubaguy: it *seems* that reno needs to know the branch but doesn't need to know the tags because pbr does | 10:02 |
| bauzas | johnthetubaguy: so the "unreleased" note could cover all the existing tags in the branch | 10:03 |
| johnthetubaguy | that sounds wrong though | 10:03 |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Use version convert methods from oslo_utils.versionutils https://review.openstack.org/203322 | 10:03 |
| johnthetubaguy | would need release notes for each tag | 10:03 |
| johnthetubaguy | ideally | 10:03 |
| johnthetubaguy | unreleased is still good though | 10:03 |
| johnthetubaguy | maybe liberty should be called 1.26 or whatever tag that branch is at? | 10:04 |
| bauzas | johnthetubaguy: so, we're good | 10:04 |
| bauzas | johnthetubaguy: just tested | 10:04 |
| bauzas | pasting the result | 10:04 |
| johnthetubaguy | I mean its needs to work for ironic, so I suspect it works for us | 10:05 |
| *** aix has joined #openstack-nova | 10:05 | |
| bauzas | johnthetubaguy: http://paste.openstack.org/show/479826/ | 10:05 |
| bauzas | johnthetubaguy: so we find the tags using pbr | 10:05 |
| bauzas | johnthetubaguy: so the "unreleased" page will show all the tags | 10:06 |
| *** otter768 has quit IRC | 10:06 | |
| johnthetubaguy | OK | 10:06 |
| bauzas | that needs to be confirmed but I think it's okay | 10:06 |
| bauzas | http://docs.openstack.org/developer/reno/sphinxext.html#examples can provide a version number | 10:07 |
| bauzas | in case we need to pin a specific version for the relnotes | 10:07 |
| johnthetubaguy | ah, that makes sense | 10:08 |
| johnthetubaguy | so we probably should add one for the release of python-novaclient we already made, I guess? | 10:08 |
| *** sacharya has joined #openstack-nova | 10:08 | |
| *** Marga_ has quit IRC | 10:09 | |
| *** Marga_ has joined #openstack-nova | 10:09 | |
| bauzas | johnthetubaguy: nope, because any change we could provide for adding a reno file would not be having that tag | 10:10 |
| bauzas | johnthetubaguy: since we can't backport a patch to a specific tag, right? | 10:10 |
| bauzas | johnthetubaguy: we can untag and tag tho but I don't feel it's good | 10:10 |
| bauzas | it's only forward-patching IIUC | 10:11 |
| *** gcb has quit IRC | 10:11 | |
| bauzas | because reno looks at the git repo, not at the files | 10:11 |
| johnthetubaguy | bauzas: oh I see your point, the release notes were not present in that tag, so boo hoo, no release notes | 10:11 |
| bauzas | zactly | 10:11 |
| bauzas | that said, it works for the next tags we could do | 10:12 |
| *** sacharya has quit IRC | 10:12 | |
| johnthetubaguy | so that was educational, my head hadn't quite got how that all works, so thats cools now | 10:13 |
| bauzas | johnthetubaguy: it means we need to be very clear when someone provides a note | 10:13 |
| bauzas | johnthetubaguy: if that's for an older version | 10:13 |
| bauzas | (not a stable version, of course) | 10:13 |
| *** achanda has quit IRC | 10:14 | |
| *** thorst has joined #openstack-nova | 10:15 | |
| openstackgerrit | Juan Antonio Osorio Robles proposed openstack/nova: Set personality/injected_files to empty list if not specified https://review.openstack.org/248662 | 10:15 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: Use instance in setup_networks_on_host https://review.openstack.org/249110 | 10:15 |
| paul-carlton1 | johnthetubaguy: re https://review.openstack.org/#/c/228828, see Kenichi's comments and my reply, what do you think? | 10:15 |
| paul-carlton1 | Also, Sean Dague had issues with API, where did we land on that? | 10:16 |
| *** tdurakov has quit IRC | 10:16 | |
| *** tdurakov has joined #openstack-nova | 10:16 | |
| *** Guest67865 has joined #openstack-nova | 10:16 | |
| paul-carlton1 | My thinking is that this should go forward as is, for alignment with pause vm during live migration spec (https://review.openstack.org/#/c/229040/14) and per Jay's comment we can fix API issues when we do tasks API? | 10:18 |
| *** ildikov has quit IRC | 10:19 | |
| *** Marga_ has quit IRC | 10:19 | |
| *** thorst has quit IRC | 10:20 | |
| *** jinxing has quit IRC | 10:20 | |
| *** hparekh has quit IRC | 10:23 | |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Added server tags support in nova-api https://review.openstack.org/128940 | 10:27 |
| lxsli | morning | 10:28 |
| *** jinxing has joined #openstack-nova | 10:28 | |
| *** jlanoux_ has joined #openstack-nova | 10:29 | |
| *** hparekh has joined #openstack-nova | 10:30 | |
| *** paul-carlton1 has left #openstack-nova | 10:30 | |
| *** subscope has joined #openstack-nova | 10:30 | |
| bauzas | lxsli: morning | 10:30 |
| *** lixiaoy1 has quit IRC | 10:30 | |
| bauzas | lxsli: that's a bit sad that you made a dependency for https://review.openstack.org/#/c/219761/ | 10:31 |
| lxsli | bauzas: about check-dest-on-migs... request_spec is only in the API DB now | 10:31 |
| lxsli | bauzas: yeah I did that by accident, just noticed - going to undo it | 10:31 |
| *** jlanoux has quit IRC | 10:32 | |
| lxsli | lost +2 jaypipes though >.< | 10:32 |
| bauzas | lxsli: okay, because I was about to +2/+W but since you made that mistake, you'd need to get another +2 :) | 10:32 |
| *** ildikov has joined #openstack-nova | 10:32 | |
| bauzas | lxsli: and Jenkins wasn't happy since there was a pep8 failure for the dependent change | 10:32 |
| *** jinxing has quit IRC | 10:32 | |
| openstackgerrit | Alexis Lee proposed openstack/nova: Unify on _schedule_instances https://review.openstack.org/219761 | 10:33 |
| *** lixiaoy1 has joined #openstack-nova | 10:33 | |
| bauzas | lxsli: re: your question about how we can persist the Spec object | 10:33 |
| bauzas | lxsli: so I discussed with alaski and dansmith | 10:33 |
| *** josecastroleon1 has joined #openstack-nova | 10:34 | |
| bauzas | lxsli: and we're okay to use the conductor for that | 10:34 |
| bauzas | lxsli: since cells v2 will also have an api conductor | 10:34 |
| openstackgerrit | Alexis Lee proposed openstack/nova: Only consider requested_destination https://review.openstack.org/248833 | 10:35 |
| *** rk4n has joined #openstack-nova | 10:35 | |
| lxsli | bauzas: but that isn't there yet? | 10:35 |
| bauzas | lxsli: not persisted | 10:35 |
| lxsli | bauzas: the api conductor I mean | 10:35 |
| bauzas | lxsli: just hydrated | 10:35 |
| bauzas | lxsli: no, not really | 10:35 |
| bauzas | lxsli: it's been approved tho | 10:35 |
| lxsli | bauzas: I wonder if I should just wait a little bit | 10:35 |
| bauzas | lxsli: nope | 10:35 |
| lixiaoy1 | berrange: this is lixiaoy1 from cinder. and I would like to talk with you about "copy encryptors from nova to os-brick" | 10:35 |
| bauzas | lxsli: it's okay | 10:35 |
| *** josecastroleon has quit IRC | 10:35 | |
| lxsli | lixiaoy1: try danpb | 10:36 |
| lixiaoy1 | lxsli: thanks | 10:36 |
| openstackgerrit | xiexs proposed openstack/python-novaclient: Introduce a sort function for "flavor list" https://review.openstack.org/249115 | 10:36 |
| lixiaoy1 | danpb: this is lixiaoy1 from cinder. and I would like to talk with you about "copy encryptors from nova to os-brick" | 10:36 |
| lxsli | bauzas: can you link me the API Conductor review please? | 10:36 |
| bauzas | lxsli: http://specs.openstack.org/openstack/nova-specs/specs/mitaka/approved/cells-scheduling-interaction.html | 10:37 |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Use is_supported() to check microversion https://review.openstack.org/249116 | 10:37 |
| bauzas | lxsli: that wouldn't change how it's done | 10:37 |
| lxsli | bauzas: thank you | 10:37 |
| *** salv-orlando has joined #openstack-nova | 10:38 | |
| bauzas | lxsli: so, to be clear, you just need to get/save in the conductor now | 10:38 |
| bauzas | lxsli: when we will provide the cells v2 method, that will still okay because the table is already there | 10:38 |
| *** rk4n has quit IRC | 10:40 | |
| lxsli | bauzas: not sure I'm quite following... so in nova/conductor/manager.py I can use objects.RequestSpec.get_by_instance_uuid and save that object? Without worrying about cells at all? | 10:40 |
| bauzas | yup | 10:40 |
| bauzas | why should you be worried? | 10:40 |
| bauzas | because it's not in the current DB ? | 10:40 |
| lxsli | bauzas: it's my natural state | 10:40 |
| danpb | lixiaoy1: hi ? | 10:40 |
| bauzas | lxsli: the main problem would be if the Spec was persisted in the 'normal' DB | 10:41 |
| *** subscope has quit IRC | 10:41 | |
| bauzas | lxsli: it would mean that we would need a migration | 10:41 |
| *** tdurakov has quit IRC | 10:41 | |
| lxsli | bauzas: aha I think I caught up a bit - the API conductor will use the same code as the cells conductor? | 10:41 |
| *** paul-carlton2 has left #openstack-nova | 10:41 | |
| bauzas | lxsli: see | 10:42 |
| bauzas | http://specs.openstack.org/openstack/nova-specs/specs/mitaka/approved/cells-scheduling-interaction.html#work-items | 10:42 |
| *** salv-orlando has quit IRC | 10:42 | |
| *** tdurakov_ has joined #openstack-nova | 10:42 | |
| *** paul-carlton1 has joined #openstack-nova | 10:42 | |
| *** gszasz has joined #openstack-nova | 10:44 | |
| bauzas | lxsli: at least that's my understanding | 10:44 |
| paul-carlton1 | test | 10:44 |
| lxsli | hi paul-carlton1 | 10:44 |
| johnthetubaguy | paul-carlton1: I think the best thing is going to be asking in the API meeting today, sdague and others are usually there | 10:44 |
| bauzas | lxsli: like we could/should write the Spec in the API | 10:44 |
| johnthetubaguy | lxsli: bauzas: my understand is its more about the Rabbit queue the conductor is listening to, and its default database connection, its the same RPCAPI I thought | 10:46 |
| bauzas | johnthetubaguy: yup that | 10:46 |
| bauzas | lxsli: so the biggest question is where we should write the Spec (when booting) and getting the Spec (when moving) | 10:47 |
| bauzas | lxsli: but that really doesn't change where it's persisted | 10:47 |
| *** markus_z has quit IRC | 10:47 | |
| lxsli | I think I get it now... it seems a little dangerous to have a common RPC API when the API and cells DBs have very different schemas, but okay | 10:47 |
| bauzas | lxsli: my take on that would be to get/persit on the conductor, because that wouldn't require to change the conductor API now | 10:47 |
| bauzas | lxsli: because if you hydrate the Spec on the API side, you need to pass it thru the conductor | 10:48 |
| lxsli | bauzas: yep I think that helps, let me see how I get on today | 10:48 |
| lxsli | bauzas: if you have time +2 on 219761 and a look at 248833 would be great | 10:49 |
| *** njirap has joined #openstack-nova | 10:50 | |
| openstackgerrit | jichenjc proposed openstack/nova: Remove stub_compute config options for console https://review.openstack.org/234217 | 10:50 |
| bauzas | lxsli: I was holding for Jenkins :) | 10:51 |
| *** bzhou has joined #openstack-nova | 10:51 | |
| bauzas | lxsli: I could +1 awaiting Jenkins but that souns picky | 10:51 |
| openstackgerrit | jichenjc proposed openstack/nova: Use instance objection instead of db record https://review.openstack.org/234218 | 10:51 |
| *** bzhou has quit IRC | 10:52 | |
| lxsli | bauzas: ok np - I tend to consider it the +W person's responsibility to check Jenkins but whatever you're comfortable with :) | 10:54 |
| bauzas | lxsli: I'm actually just verifying your change again :) | 10:54 |
| bauzas | lxsli: okay, no diff | 10:56 |
| *** lixiaoy1 has left #openstack-nova | 10:57 | |
| *** lixiaoy1_ has joined #openstack-nova | 10:58 | |
| bauzas | heh, Gerrit is in weeds \o/ | 10:59 |
| * bauzas going to take coffee :p | 10:59 | |
| *** lixiaoy1_ is now known as lixiaoy1 | 10:59 | |
| bauzas | it was actually really slow, so restarting it seems good :) | 10:59 |
| lxsli | yep yep | 10:59 |
| *** haomaiwa_ has quit IRC | 11:01 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: [WIP]Expose quiesce/unquiesce API https://review.openstack.org/248989 | 11:01 |
| *** haomaiwang has joined #openstack-nova | 11:01 | |
| *** dims has joined #openstack-nova | 11:06 | |
| *** rk4n has joined #openstack-nova | 11:08 | |
| *** dgonzalez has joined #openstack-nova | 11:08 | |
| *** rk4n has quit IRC | 11:08 | |
| *** EinstCrazy has joined #openstack-nova | 11:08 | |
| *** EinstCrazy has quit IRC | 11:09 | |
| *** EinstCrazy has joined #openstack-nova | 11:09 | |
| openstackgerrit | jichenjc proposed openstack/python-novaclient: Print current nova default nova API microversion https://review.openstack.org/249132 | 11:09 |
| *** aysyd has joined #openstack-nova | 11:10 | |
| openstackgerrit | Claudiu Belu proposed openstack/nova: Hyper-V: refines the exceptions raised in the driver https://review.openstack.org/212102 | 11:12 |
| *** Sree has quit IRC | 11:12 | |
| *** Sree has joined #openstack-nova | 11:13 | |
| *** rcernin is now known as rcernin|lunch | 11:13 | |
| *** aix has quit IRC | 11:14 | |
| *** tdurakov_ has quit IRC | 11:14 | |
| *** aix has joined #openstack-nova | 11:15 | |
| tangchen__ | johnthetubaguy: Hi John, I saw you just abandoned the host maintenance mode thread. I think improving novaclient will be like this: | 11:15 |
| tangchen__ | johnthetubaguy: 1. Add message to tell caller that the host maintenance mode has been deprecated, in nova host-update command. | 11:16 |
| *** tdurakov_ has joined #openstack-nova | 11:16 | |
| tangchen__ | johnthetubaguy: 2. Maybe in nova side, when the API is called, we should also add the message. | 11:16 |
| *** Sree has quit IRC | 11:17 | |
| tangchen__ | johnthetubaguy: This doesn't need BP and spec, is that OK ? | 11:17 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: remove todo note not useful anymore https://review.openstack.org/249137 | 11:18 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: replace direct libvirt's call AbortJobBlock https://review.openstack.org/249138 | 11:18 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: move find_disk to Guest's object https://review.openstack.org/249139 | 11:18 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: introduce freeze filesystems https://review.openstack.org/249140 | 11:18 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: introduce thaw filesystems https://review.openstack.org/249141 | 11:18 |
| tangchen__ | johnthetubaguy: I cannot think of anything that we need to improve. | 11:18 |
| johnthetubaguy | tangchen__: thats a different spec, deprecate host maintenance mode. I think a very small spec would be best for that, just to get the official vote on it. | 11:19 |
| *** Daisy_ has joined #openstack-nova | 11:19 | |
| tangchen__ | johnthetubaguy: OK, will post one and wish it will catch up the spec deadline. :) | 11:20 |
| tangchen__ | johnthetubaguy: Thx. | 11:20 |
| *** ociuhandu has joined #openstack-nova | 11:21 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/nova: Ironic: Add support for microversion https://review.openstack.org/221848 | 11:23 |
| *** jlanoux_ has quit IRC | 11:27 | |
| *** pixelbeat has joined #openstack-nova | 11:27 | |
| *** jlanoux has joined #openstack-nova | 11:28 | |
| *** jwcroppe has quit IRC | 11:29 | |
| alex_xu | johnthetubaguy: sdague, I want to write an example to show what api ref doc is we expected https://review.openstack.org/248534, but you know my English and grammar...so appreciate you can help me review it. | 11:31 |
| lxsli | alex_xu: I can help | 11:31 |
| alex_xu | lxsli: thanks a lot :) | 11:31 |
| *** jwcroppe has joined #openstack-nova | 11:32 | |
| lxsli | paul-carlton1: when live migrating, if you're not doing a block migration you're doing a ____ migration? | 11:33 |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: [WIP]Expose quiesce/unquiesce API https://review.openstack.org/248989 | 11:36 |
| *** jwcroppe has quit IRC | 11:36 | |
| lxsli | paul-carlton1: nvm | 11:36 |
| *** ptm_away is now known as PaulMurray | 11:36 | |
| *** tdurakov_ has quit IRC | 11:36 | |
| *** tdurakov_ has joined #openstack-nova | 11:37 | |
| *** Daisy_ has quit IRC | 11:37 | |
| *** Daisy has joined #openstack-nova | 11:38 | |
| *** daemontool has joined #openstack-nova | 11:38 | |
| *** achanda has joined #openstack-nova | 11:39 | |
| *** daemontool has quit IRC | 11:40 | |
| *** daemontool has joined #openstack-nova | 11:40 | |
| *** Daisy has quit IRC | 11:41 | |
| *** Daisy has joined #openstack-nova | 11:41 | |
| *** josecastroleon1 has quit IRC | 11:42 | |
| *** Daisy has quit IRC | 11:43 | |
| *** markus_z has joined #openstack-nova | 11:43 | |
| *** hparekh2 has joined #openstack-nova | 11:44 | |
| *** Daisy has joined #openstack-nova | 11:46 | |
| *** hparekh has quit IRC | 11:46 | |
| *** jlanoux_ has joined #openstack-nova | 11:46 | |
| *** ildikov has quit IRC | 11:48 | |
| openstackgerrit | Alexis Lee proposed openstack/nova: Fix the error of duplicate words. https://review.openstack.org/216646 | 11:48 |
| lxsli | markus_z: trivial fix ^^, would you add it to the list? (I can't as I just updated the commit) | 11:49 |
| *** jlanoux has quit IRC | 11:49 | |
| markus_z | lxsli: sure, one sec | 11:50 |
| lxsli | markus_z: cheers | 11:50 |
| openstackgerrit | Alexis Lee proposed openstack/nova: Double 'an' in message https://review.openstack.org/216646 | 11:50 |
| *** hparekh has joined #openstack-nova | 11:50 | |
| *** Daisy has quit IRC | 11:51 | |
| *** Daisy has joined #openstack-nova | 11:51 | |
| *** hparekh2 has quit IRC | 11:52 | |
| *** gcb has joined #openstack-nova | 11:53 | |
| *** Daisy has quit IRC | 11:53 | |
| *** Daisy has joined #openstack-nova | 11:53 | |
| lxsli | BobBall: too fast, commit message is fixed now :) | 11:53 |
| *** jinxing has joined #openstack-nova | 11:55 | |
| openstackgerrit | Paul Murray proposed openstack/nova: docs: describe migration and other movement concepts https://review.openstack.org/247016 | 11:57 |
| *** oomichi has joined #openstack-nova | 11:57 | |
| BobBall | lxsli: Serves me right for checking my email :D | 11:57 |
| *** baoli has quit IRC | 11:58 | |
| *** gmann_ has joined #openstack-nova | 11:58 | |
| *** shaohe_feng has joined #openstack-nova | 11:58 | |
| *** markus_z is now known as markus_z_meeting | 11:59 | |
| *** subscope has joined #openstack-nova | 12:00 | |
| *** haomaiwang has quit IRC | 12:01 | |
| *** haomaiwa_ has joined #openstack-nova | 12:02 | |
| *** otter768 has joined #openstack-nova | 12:02 | |
| *** markmc has quit IRC | 12:02 | |
| *** alexschm has quit IRC | 12:04 | |
| openstackgerrit | ShaoHe Feng proposed openstack/nova-specs: report more live migration progress detail https://review.openstack.org/248472 | 12:05 |
| *** Daisy has quit IRC | 12:05 | |
| *** alexschm has joined #openstack-nova | 12:05 | |
| *** Daisy has joined #openstack-nova | 12:06 | |
| *** otter768 has quit IRC | 12:06 | |
| *** mpavone has quit IRC | 12:07 | |
| *** sahid has quit IRC | 12:07 | |
| openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova-specs: Fix request body of pause-vm-during-live-migration https://review.openstack.org/249167 | 12:08 |
| *** daemontool has quit IRC | 12:08 | |
| *** josecastroleon has joined #openstack-nova | 12:08 | |
| jaosorior | gibi: Hey, I think that's all we really need, but honestly I didn't figure out where to put a test for that :/ | 12:09 |
| *** daemontool has joined #openstack-nova | 12:09 | |
| gibi | jaosorior: don't we have some unit test on that API? Can duplicate that, add a hooka and check if the hook can modify the API call? | 12:11 |
| *** Daisy has quit IRC | 12:11 | |
| *** hparekh has quit IRC | 12:11 | |
| *** subscope has quit IRC | 12:13 | |
| *** subscope has joined #openstack-nova | 12:14 | |
| *** sacharya has joined #openstack-nova | 12:15 | |
| gibi | jaosorior: for example https://github.com/openstack/nova/blob/master/nova/tests/unit/compute/test_compute_api.py#L156 this test calls nova.compute.api.create and as far as I understand the problematic hook in your patch tries to modifiy the params of that call | 12:15 |
| *** thorst has joined #openstack-nova | 12:15 | |
| openstackgerrit | jichenjc proposed openstack/python-novaclient: Print current nova default nova API microversion https://review.openstack.org/249132 | 12:16 |
| *** salv-orlando has joined #openstack-nova | 12:17 | |
| jaosorior | problematic hook? | 12:18 |
| openstackgerrit | Roman Podoliaka proposed openstack/nova: servicegroup: ensure state reporting thread does not die prematurely https://review.openstack.org/247552 | 12:18 |
| *** zhangjn has joined #openstack-nova | 12:18 | |
| *** zhangjn has quit IRC | 12:18 | |
| *** venkat_p has joined #openstack-nova | 12:19 | |
| *** zhangjn has joined #openstack-nova | 12:19 | |
| *** zhangjn has quit IRC | 12:20 | |
| *** mpavone has joined #openstack-nova | 12:20 | |
| *** sacharya has quit IRC | 12:20 | |
| *** thorst has quit IRC | 12:20 | |
| *** zhangjn has joined #openstack-nova | 12:20 | |
| *** zhangjn has quit IRC | 12:20 | |
| *** zhangjn has joined #openstack-nova | 12:21 | |
| gibi | jaosorior: I mean the problematic hook in the bug report | 12:21 |
| *** zhangjn has quit IRC | 12:22 | |
| *** salv-orlando has quit IRC | 12:22 | |
| jaosorior | gibi: So yeah, there could be a test added there, though I'm not sure if there it's appropriate to mock the hooking mechanism there, since the hooks are loaded via stevedore, which I'm not even sure is being done in that test suite | 12:22 |
| *** zhangjn has joined #openstack-nova | 12:22 | |
| *** zhangjn has quit IRC | 12:23 | |
| openstackgerrit | Raga lahari proposed openstack/nova: Support for new --status parameter to "nova-manage vm list" command https://review.openstack.org/242370 | 12:24 |
| *** tdurakov_ has quit IRC | 12:25 | |
| gibi | jaosorior: I'm not sure how can you test through the stevedore mechanism but I feel that we did break the hook before and we didn't noticed because we did'nt have test for it so this would be a good time to add some test to avoid the next regression | 12:26 |
| *** EinstCrazy has quit IRC | 12:26 | |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova: [WIP]Expose quiesce/unquiesce API https://review.openstack.org/248989 | 12:27 |
| jaypipes | *yawn* morning all | 12:28 |
| *** EinstCrazy has joined #openstack-nova | 12:28 | |
| *** zhangjn has joined #openstack-nova | 12:29 | |
| gibi | morning jay | 12:29 |
| *** deepakcs has quit IRC | 12:31 | |
| *** lucasagomes is now known as lucas-hungry | 12:31 | |
| jaosorior | gibi: Well, there would need to be a better definition of what hooks should and shouldn't be doing first, then have a better defined and documented interface for the functions where the hooks are, and then there could be proper tests :/ | 12:32 |
| gibi | jaosorior: what about this tests https://github.com/openstack/nova/blob/master/nova/tests/unit/test_hooks.py | 12:32 |
| *** smatzek has joined #openstack-nova | 12:32 | |
| jaosorior | gibi: But I agree, would be nice to avoid further regressions | 12:32 |
| *** jwcroppe has joined #openstack-nova | 12:33 | |
| *** zhangjn has quit IRC | 12:33 | |
| *** zhangjn has joined #openstack-nova | 12:34 | |
| jaosorior | gibi: That seems to test the functionality of the hooking mechanism, but not how it's used or that one can still manipulate what arguments go through them | 12:34 |
| *** alejandrito has joined #openstack-nova | 12:34 | |
| gibi | gibi: but from the previous test and from this test you might be able to put togheter a test for your case | 12:35 |
| *** EinstCra_ has joined #openstack-nova | 12:35 | |
| openstackgerrit | Claudiu Belu proposed openstack/nova: Adds HostCapabilities object model and updates ComputeNode object https://review.openstack.org/228280 | 12:35 |
| openstackgerrit | Claudiu Belu proposed openstack/nova: db: Adds host_capabilities column in compute_nodes https://review.openstack.org/228279 | 12:35 |
| gibi | I know that it is not well defined how this hooks are used, but we can start first with your test case to define one of the usage | 12:35 |
| *** EinstCrazy has quit IRC | 12:35 | |
| *** rcernin|lunch is now known as rcernin | 12:36 | |
| *** sahid has joined #openstack-nova | 12:38 | |
| jaosorior | gibi: Alright, I'll do this. I'll leave this patch as it is and I'll upload a subsequent patch with better defined testing for the function and it's respective hooks | 12:39 |
| *** diegows has joined #openstack-nova | 12:40 | |
| *** wolsen has quit IRC | 12:41 | |
| *** ildikov has joined #openstack-nova | 12:42 | |
| *** zenoway has quit IRC | 12:42 | |
| gibi | jaosorior: cool | 12:43 |
| afazekas | I wonder is it really required to have blueprint for 18 line patch ? https://review.openstack.org/#/c/134134/12 | 12:44 |
| *** subscope has quit IRC | 12:44 | |
| *** subscope has joined #openstack-nova | 12:46 | |
| *** shaohe_feng has quit IRC | 12:47 | |
| edleafe | tjones: ideally grab the section you know the most about. Most of the time is writing new docs for each config option. | 12:49 |
| *** shaohe_feng has joined #openstack-nova | 12:49 | |
| *** shaohe_feng has quit IRC | 12:49 | |
| *** jwcroppe has quit IRC | 12:50 | |
| *** shaohe_feng has joined #openstack-nova | 12:50 | |
| *** shaohe_feng has quit IRC | 12:51 | |
| *** gcb has quit IRC | 12:51 | |
| *** markmc has joined #openstack-nova | 12:53 | |
| *** gmann_ has quit IRC | 12:54 | |
| *** shaohe_feng has joined #openstack-nova | 12:55 | |
| *** baoli has joined #openstack-nova | 12:55 | |
| *** MVenesio has joined #openstack-nova | 12:56 | |
| *** vladikr has joined #openstack-nova | 12:58 | |
| *** zenoway has joined #openstack-nova | 12:58 | |
| *** andymaier has quit IRC | 12:59 | |
| *** mpavone has quit IRC | 13:00 | |
| *** cfriesen__ has joined #openstack-nova | 13:00 | |
| *** zenoway has quit IRC | 13:00 | |
| *** haomaiwa_ has quit IRC | 13:01 | |
| *** zenoway has joined #openstack-nova | 13:01 | |
| *** markus_z_meeting is now known as markus_z | 13:01 | |
| *** haomaiwang has joined #openstack-nova | 13:01 | |
| *** zhenguo has joined #openstack-nova | 13:01 | |
| *** oomichi has quit IRC | 13:02 | |
| *** smatzek has quit IRC | 13:03 | |
| *** gcb has joined #openstack-nova | 13:04 | |
| *** salv-orl_ has joined #openstack-nova | 13:05 | |
| *** marcusvrn_ has quit IRC | 13:05 | |
| *** ljxiash has quit IRC | 13:06 | |
| openstackgerrit | Daniel Berrange proposed openstack/nova: hacking: check for common double word typos https://review.openstack.org/249193 | 13:06 |
| *** jhesketh has quit IRC | 13:06 | |
| *** jhesketh has joined #openstack-nova | 13:07 | |
| *** tdurakov_ has joined #openstack-nova | 13:10 | |
| *** haomaiwang has quit IRC | 13:11 | |
| *** LukeHinds has joined #openstack-nova | 13:12 | |
| *** mpavone has joined #openstack-nova | 13:12 | |
| PaulMurray | tdurakov_, looking at https://review.openstack.org/#/c/225910/ | 13:13 |
| tdurakov_ | PaulMurray, hi, need to update it | 13:13 |
| PaulMurray | tdurakov_, I was going to say there hasn't been an update for three weeks | 13:13 |
| *** Luigi_ has joined #openstack-nova | 13:13 | |
| tdurakov_ | focused on job for live-migration | 13:13 |
| tdurakov_ | yes | 13:13 |
| *** lucas-hungry is now known as lucasagomes | 13:14 | |
| PaulMurray | it needs to be done this week really | 13:14 |
| *** wolsen has joined #openstack-nova | 13:14 | |
| tdurakov_ | PaulMurray, spec or job? | 13:14 |
| PaulMurray | the spec for splitting migraiton types | 13:14 |
| PaulMurray | the spec needs to be approved next week or it misses the deadline | 13:15 |
| tdurakov_ | PaulMurray, acked, will ping you tomorrow morning with update | 13:15 |
| PaulMurray | thanks | 13:15 |
| tdurakov_ | PaulMurray, live-migration sub team meeting will start in 45 minutes? | 13:16 |
| PaulMurray | yes | 13:16 |
| tdurakov_ | ok, will be there | 13:16 |
| tdurakov_ | btw, have you seen sdague online recently? | 13:16 |
| tdurakov_ | PaulMurray, ^ | 13:17 |
| lucasagomes | hi all, if you have time mind taking a look at https://review.openstack.org/#/c/221848/ ? It's not a big patch and it's important to ironic to get microversion support in the driver | 13:17 |
| lucasagomes | thank you | 13:17 |
| PaulMurray | bauzas, I was looking back over https://review.openstack.org/#/c/245543/7 and saw your comment about microversions and dropping things | 13:18 |
| lxsli | alex_xu: did you push a new version of that live migrate api doc yet? | 13:19 |
| PaulMurray | bauzas, is it ok to drop a parameter without deprecation and rely on microversions to make up for it? | 13:19 |
| PaulMurray | alex_xu, same question ^^^ | 13:19 |
| alex_xu | lxsli: sorry, not yet | 13:19 |
| lxsli | alex_xu: ok just checking :) | 13:20 |
| alex_xu | lxsli: will try later, just finish the api meeting, and prepare the livemigrate meeting | 13:20 |
| alex_xu | PaulMurray: yes, the old version always keep the old behavior | 13:20 |
| * alex_xu have meeting night | 13:20 | |
| *** KLB has joined #openstack-nova | 13:20 | |
| lxsli | alex_xu: meeting night goes better with beer | 13:21 |
| alex_xu | lxsli: heh | 13:21 |
| PaulMurray | alex_xu, does the client go with the minimum version or the max? or a set version for the client? I can't remember | 13:21 |
| *** jyuso1 has quit IRC | 13:21 | |
| johnthetubaguy | tdurakov_: I haven't seem him actually, but it is a big holiday week in the US right now | 13:22 |
| *** daemontool has quit IRC | 13:22 | |
| alex_xu | PaulMurray: the default behavior of python-novaclient CLI is newest version between client and server. But we suggest people use client with specfic version. | 13:22 |
| tdurakov_ | johnthetubaguy, ok, thanks, maybe you could help me with merging: https://review.openstack.org/#/c/247081/ | 13:23 |
| *** Sree has joined #openstack-nova | 13:23 | |
| KLB | Hi! Does nova take into consideration whether a guest disk is put on a ram disk or not? Dealing with resource allocation for example, putting block storage on ram could eat up more ram than you initially allocated for if it's not being considered | 13:23 |
| PaulMurray | alex_xu, thanks | 13:23 |
| alex_xu | PaulMurray: so people should keep at the old version before people think they are ready to upgrade | 13:23 |
| alex_xu | PaulMurray: np | 13:23 |
| alex_xu | PaulMurray: if you agree on that, next thing I want to discuss is the api behavior after we removed the disk_over_commit. I added https://review.openstack.org/245543 in the api agenda, is it ok? | 13:24 |
| alex_xu | s/api agenda/live migration meeting agenda/.... | 13:25 |
| *** yuntongjin has joined #openstack-nova | 13:25 | |
| PaulMurray | alex_xu, yes, but not sure danpb will be there - but mdbooth should | 13:28 |
| johnthetubaguy | tdurakov_: I could, I just wasn't sure what the plan was there | 13:28 |
| tdurakov_ | johnthetubaguy, well the plan describe in a hook | 13:29 |
| alex_xu | PaulMurray: emm...yea, hard to catch all the people at same time | 13:29 |
| tdurakov_ | check file please | 13:29 |
| *** josh6627 has joined #openstack-nova | 13:29 | |
| *** josh6627 has quit IRC | 13:29 | |
| mdbooth | KLB: No, to the best of my knowledge no part of Nova is aware of ram disks. | 13:30 |
| *** doug-fish has joined #openstack-nova | 13:30 | |
| *** jhesketh has quit IRC | 13:30 | |
| danpb | PaulMurray: if we already registered for midcycle on eventbrite, are you expecting people to register again on that HP event URL you had the other day ? | 13:32 |
| openstackgerrit | ShaoHe Feng proposed openstack/nova-specs: Attach/detach SR-IOV interface https://review.openstack.org/139910 | 13:33 |
| PaulMurray | danpb, you don't have to, we will keep track of both - but you do need to if you are going to book that hotel on our rate - its the first step booking the hotel | 13:34 |
| danpb | oh, is that a special discounted rate ? | 13:34 |
| *** markvoelker has quit IRC | 13:35 | |
| KLB | mdbooth: is this an active decision letting it be up to the administrator to handle that, or is it not something that has been very prioritized? | 13:35 |
| *** pradk has joined #openstack-nova | 13:35 | |
| PaulMurray | danpb, is that sarcasm? | 13:35 |
| *** thorst has joined #openstack-nova | 13:36 | |
| mdbooth | KLB: It's a pretty weird requirement, tbh. You might have to do some advocacy if you want something like that. | 13:36 |
| danpb | heh no, i just didn't realize there was a different rate | 13:36 |
| mdbooth | I expect it would make some things considerably more complex if RAM and storage were linked. | 13:36 |
| PaulMurray | turns out they all advertise "from £45 a night" but if yo utry to book one its more like £145 | 13:37 |
| danpb | yeah, i'm just seeing that now :-) | 13:37 |
| KLB | mdbooth: yea, but I guess that it is when you're using RAM disks in your setup? | 13:37 |
| *** thorst_ has joined #openstack-nova | 13:37 | |
| lxsli | KLB: you could set reserved_host_memory_mb but that's constant, you can't reserve an amount per instance | 13:37 |
| *** Sree__ has joined #openstack-nova | 13:37 | |
| KLB | lxsli: right, it's just that the numbers isn't really accurate then when scheduling and allocating the resources for the instance | 13:38 |
| *** daemontool has joined #openstack-nova | 13:39 | |
| KLB | I guess the scheduler just looks at the host state when it places the vms, but the quotas won't be accurate, right? | 13:39 |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Use version convert methods from oslo_utils.versionutils https://review.openstack.org/203322 | 13:39 |
| lxsli | KLB: best I can offer is to feed this usecase into the extra_specs work | 13:39 |
| danpb | rather than giving the guest 1 GB of RAM and a 1 GB RAMdisk, why not just give the guest 2 GB of RAM and let the guest OS configure its own ramdisk from that alllocation | 13:39 |
| danpb | it avoids nova needing to care about this at all and gives the guest more flexibility in how it uses ram & disk | 13:39 |
| *** Sree has quit IRC | 13:40 | |
| mdbooth | danpb: It wouldn't be persistent across a guest reboot, I guess. | 13:41 |
| *** thorst has quit IRC | 13:41 | |
| KLB | danpb: I'm currently working on supporting http://wiki.qemu.org/Features/BlockReplication in openstack/libvirt. The ideal setup is to put the disk buffer on RAM disk to keep the latency to a minimum | 13:42 |
| *** thorst_ has quit IRC | 13:44 | |
| *** thorst_ has joined #openstack-nova | 13:45 | |
| danpb | ok, well that's much more complex than just supporting a ramdisk in nova | 13:46 |
| *** claudiub|2 has joined #openstack-nova | 13:46 | |
| *** claudiub has quit IRC | 13:46 | |
| danpb | i'd strongly advise you to submit a blueprint + spec proposal for continuous replication before investing time to writing the code for nova | 13:47 |
| *** pkoniszewski has joined #openstack-nova | 13:47 | |
| *** haomaiwang has joined #openstack-nova | 13:49 | |
| *** thorst_ has quit IRC | 13:49 | |
| johnthetubaguy | KLB: +1 to danpb's comment on getting a spec up for review | 13:49 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/nova: Nested Quota: Set default values to subproject https://review.openstack.org/242514 | 13:49 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/nova: Nested Quota Driver: Create Column Allocated https://review.openstack.org/242500 | 13:49 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/nova: Nested Quota Driver: Get Hierarchy info https://review.openstack.org/242568 | 13:49 |
| KLB | danpb: yes, we're planning on doing that. My work is part of a bigger project however, so I got a deadline to consider as well I'm afraid! | 13:50 |
| *** pkoniszewski has quit IRC | 13:51 | |
| KLB | A blueprint will be created before trying to get it upstream though of course | 13:51 |
| KLB | I'll create a note of the ram disk deal for now, and I guess that it can be discussed further if this leads anywhere later on | 13:52 |
| *** markvoelker has joined #openstack-nova | 13:53 | |
| danpb | KLB: oh, well be warned that if you implement it first and then submit a blueprint once its "ready" | 13:55 |
| danpb | KLB: then there's a good chance you'll end up having to re-design and/or rewrite it | 13:55 |
| danpb | based on feedback on the spec/blueprint | 13:55 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/nova: Nested Quota Driver: Get Hierarchy info https://review.openstack.org/242568 | 13:55 |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/nova: Nested Quota Driver: Get Hierarchy info https://review.openstack.org/242568 | 13:56 |
| *** daemontool has quit IRC | 13:56 | |
| KLB | danpb: I count on it, the qemu/kernel work is still in progress as well, so it's going to be changed over time anyway | 13:56 |
| *** pkoniszewski has joined #openstack-nova | 13:57 | |
| *** ljxiash has joined #openstack-nova | 13:57 | |
| *** daemontool has joined #openstack-nova | 13:57 | |
| openstackgerrit | jichenjc proposed openstack/python-novaclient: Print current nova default nova API microversion https://review.openstack.org/249132 | 13:57 |
| PaulMurray | Live migration meeting on #openstack-meeting-3 in 4 minutes: PaulMurray paul-carlton2 eliqiao shaohe_feng pkoniszewski jlanoux yuntongjin mdbooth | 13:57 |
| *** tdurakov_ is now known as tdurakov | 13:57 | |
| *** achanda has quit IRC | 13:57 | |
| ildikov | danpb: hi | 13:58 |
| *** daemontool has quit IRC | 13:58 | |
| ildikov | danpb: I addressed your comment in https://review.openstack.org/#/c/212508 | 13:58 |
| *** daemontool has joined #openstack-nova | 13:59 | |
| ildikov | danpb: could you please take a look? :) | 13:59 |
| shaohe_feng | PaulMurray: thank you. will join | 13:59 |
| *** salv-orl_ has quit IRC | 14:00 | |
| *** zhangjn has quit IRC | 14:00 | |
| *** subscope has quit IRC | 14:00 | |
| *** haomaiwang has quit IRC | 14:01 | |
| *** haomaiwa_ has joined #openstack-nova | 14:01 | |
| *** jhesketh has joined #openstack-nova | 14:02 | |
| *** jhesketh has quit IRC | 14:03 | |
| *** rlrossit has joined #openstack-nova | 14:03 | |
| *** jaosorior has quit IRC | 14:03 | |
| *** otter768 has joined #openstack-nova | 14:03 | |
| *** gszasz has quit IRC | 14:03 | |
| *** rcernin has quit IRC | 14:03 | |
| *** jaosorior has joined #openstack-nova | 14:04 | |
| *** jhesketh has joined #openstack-nova | 14:04 | |
| *** subscope has joined #openstack-nova | 14:04 | |
| ndipanov | hey folks - can anyone remember why we thought it was a good idea to depend on pbr at runtime | 14:05 |
| ndipanov | I seem to remember this coming up a bunch of times (rightfully so) and we still seem to depend on it | 14:05 |
| *** achanda has joined #openstack-nova | 14:06 | |
| *** pilgrimstack has quit IRC | 14:07 | |
| johnthetubaguy | ndipanov: I don't remember why, honestly | 14:07 |
| bauzas | PaulMurray: so, I'm not really opiniated | 14:08 |
| bauzas | PaulMurray: re https://review.openstack.org/#/c/245543/7 | 14:08 |
| *** otter768 has quit IRC | 14:08 | |
| bauzas | PaulMurray: I mean, it's understandable that having a microversion for that means that the operator can just uses the max capping for that microversion | 14:08 |
| bauzas | PaulMurray: that said, I leave oomichi and alex_xu for that because I'm not really API seasoned | 14:09 |
| *** pilgrimstack has joined #openstack-nova | 14:10 | |
| *** subscope has quit IRC | 14:10 | |
| *** gszasz has joined #openstack-nova | 14:10 | |
| *** subscope has joined #openstack-nova | 14:11 | |
| *** ctrath has joined #openstack-nova | 14:11 | |
| *** ankit_ag has quit IRC | 14:12 | |
| *** paul___ has left #openstack-nova | 14:12 | |
| *** paul___ has joined #openstack-nova | 14:12 | |
| ndipanov | johnthetubaguy, I seem to remember there being an attempt to move the version bits of pbr into oslo.version and have no idea what happened with that | 14:12 |
| *** zenoway has quit IRC | 14:12 | |
| *** subscope has quit IRC | 14:13 | |
| *** subscope has joined #openstack-nova | 14:13 | |
| *** subscope has quit IRC | 14:16 | |
| *** zenoway has joined #openstack-nova | 14:16 | |
| *** subscope has joined #openstack-nova | 14:17 | |
| *** pilgrimstack has quit IRC | 14:18 | |
| *** pilgrimstack has joined #openstack-nova | 14:19 | |
| *** eharney has joined #openstack-nova | 14:19 | |
| openstackgerrit | Raildo Mascena de Sousa Filho proposed openstack/nova: Nested Quota Driver https://review.openstack.org/242626 | 14:21 |
| *** jerrygb has joined #openstack-nova | 14:23 | |
| *** mylu has joined #openstack-nova | 14:23 | |
| *** jichen has quit IRC | 14:24 | |
| *** venkat_p has quit IRC | 14:26 | |
| *** venkat_p has joined #openstack-nova | 14:26 | |
| openstackgerrit | Alexis Lee proposed openstack/nova: Remove SoftDeleteMixin from BandwidthUsage https://review.openstack.org/240361 | 14:28 |
| *** omnipresent has joined #openstack-nova | 14:29 | |
| *** Sree__ has quit IRC | 14:30 | |
| rlrossit | lxsli: you were right about doing 1 at a time when asking for reviews... after one got merged, the rest went into merge conflict :( | 14:31 |
| *** Sree has joined #openstack-nova | 14:32 | |
| *** mc_nair has joined #openstack-nova | 14:32 | |
| *** amotoki has joined #openstack-nova | 14:32 | |
| *** akshai has joined #openstack-nova | 14:32 | |
| chinmaya | :garyk1 Hi | 14:33 |
| *** zenoway has quit IRC | 14:33 | |
| *** mriedem_away is now known as mriedem | 14:33 | |
| *** mrkz has joined #openstack-nova | 14:33 | |
| chinmaya | garyk1: was thinking to take up considering dest param in migration / allow_resize_to_same_host = true case for vmware, | 14:34 |
| chinmaya | should it be taken as a bp or a bug ? | 14:35 |
| lxsli | rlrossit: doh! | 14:35 |
| chinmaya | looks like it may lead into a big churn | 14:35 |
| lxsli | rlrossit: chain them? | 14:35 |
| *** willsama has joined #openstack-nova | 14:36 | |
| chinmaya | :garyk1: allow_resize_to_same_host = false * case | 14:36 |
| openstackgerrit | Claudiu Belu proposed openstack/nova: Hyper-V: refines the exceptions raised in the driver https://review.openstack.org/212102 | 14:36 |
| *** Sree has quit IRC | 14:36 | |
| *** mtanino has joined #openstack-nova | 14:37 | |
| rlrossit | lxsli: nah I'm just a lazy bum. I'll do one at a time now... | 14:38 |
| openstackgerrit | Pavel Kholkin proposed openstack/nova-specs: keypairs-pagination support https://review.openstack.org/239663 | 14:39 |
| beagles | vladikr, I've just been going over some of the SR-IOV PF stuff. I'm thinking I misunderstood something. Are we looking at dynamic management of VFs/PFs in the first iteration of this? | 14:39 |
| garyk1 | chinmaya: at the moment the migration code is not supported at all, unless you consider resize as migrations | 14:39 |
| openstackgerrit | Ryan Rossiter proposed openstack/nova: Use o.vo FlexibleBooleanField https://review.openstack.org/245336 | 14:39 |
| garyk1 | i also think that you are addressing a real edge case | 14:39 |
| *** zenoway has joined #openstack-nova | 14:40 | |
| vladikr | beagles, hmm.. are you referring to this? http://specs.openstack.org/openstack/nova-specs/specs/mitaka/approved/sriov-physical-function-passthrough.html | 14:42 |
| chinmaya | @garyk1 : yeah, its a edge case for resize. Since method migrate_disk_and_power_off handles both cold migrate and resize , i was thinking its good to handle "dest" and clone the vm to the other compute. | 14:42 |
| beagles | vladikr, yup | 14:43 |
| chinmaya | :garyk1: are we planning to divide resize and migrate into a separate methods ? | 14:43 |
| *** goodygum_ is now known as noqa_v_qoovnie | 14:44 | |
| *** omnipresent has quit IRC | 14:44 | |
| *** sacharya has joined #openstack-nova | 14:45 | |
| vladikr | beagles, yea. the idea is to pass a PF, that has no assigned VFs , once it's assigned , the related VFs should not be available for assignment.. this is the level of management we are looking at | 14:45 |
| garyk1 | chinmaya: for the cloning we should leverage the 6.0 features which enable | 14:45 |
| garyk1 | this should also support live migrations. | 14:45 |
| vladikr | beagles, not to reconfigure the driver on the host or anything like that, just to manage the resources | 14:46 |
| beagles | vladikr, okay. The reason I ask is that in some cases there is a neutron agent for SR-IOV running on that host - there are implications when the case is more dynamic. | 14:47 |
| vladikr | beagles, did I answer the question or just confused you more :) | 14:47 |
| vladikr | oh | 14:47 |
| beagles | vladikr, nope I think that clarified the situation :) | 14:47 |
| chinmaya | :garyk1: but that can be handled in a seperate method : live_migration , right ? | 14:47 |
| *** jwcroppe has joined #openstack-nova | 14:47 | |
| *** jgriffith_away is now known as jgriffith | 14:47 | |
| garyk1 | chinmaya: the pans are that we add the live migration support. my guess is that we will need a BP for that. the idea was to get involved in the efforts led by PaulMurray | 14:48 |
| vladikr | beagles, I think this spec (sriov-physical-function-passthrough) doesn't rely on any networking awareness | 14:48 |
| garyk1 | chinmaya: rado is going to try and work on this. he is just out this week and beginning of next | 14:48 |
| beagles | vladikr, well.. in the intel case I don't think there is an issue, but I think there are two things when there is an agent: 1. data consistency on VF allocations and 2. the agent probably needs to get involved with manipulating the NIC | 14:48 |
| *** madhaviy has joined #openstack-nova | 14:48 | |
| garyk1 | if possible can you connect up with him next week | 14:48 |
| *** sacharya has quit IRC | 14:49 | |
| chinmaya | :garyk1 sure gary. Cheers! | 14:49 |
| madhaviy | hi there, is there any way to retrieve servers list(vm list) based on user_id? | 14:49 |
| beagles | vladikr, right.. which is probably sensible. In any case, the VNIC stuff is more-or-less underway. I'll need to loop in moshele, baoli, irenab to get some advice on the agents | 14:50 |
| vladikr | beagles, yea, moshele pointed me to the patch today. sounds great | 14:51 |
| *** irenab has quit IRC | 14:51 | |
| * beagles totally forgot about the sr-iov meeting this morning :( | 14:51 | |
| chinmaya | :madhaviy : nova list --user <userid> , will do | 14:52 |
| *** yuntongjin has quit IRC | 14:52 | |
| *** achanda has quit IRC | 14:53 | |
| *** achanda has joined #openstack-nova | 14:54 | |
| *** pumaranikar has joined #openstack-nova | 14:54 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Add code-review devref for release notes https://review.openstack.org/247775 | 14:55 |
| bauzas | mriedem: so, I made a respin for the reno code-review change | 14:55 |
| bauzas | mriedem: ^ | 14:56 |
| *** achanda_ has joined #openstack-nova | 14:56 | |
| bauzas | mriedem: if you could just rant my change, that would be nice :) | 14:56 |
| *** jerrygb has quit IRC | 14:56 | |
| bauzas | johnthetubaguy: eh, I guess you'd be interested in that too ^ | 14:56 |
| *** jerrygb has joined #openstack-nova | 14:57 | |
| *** tdurakov has quit IRC | 14:58 | |
| *** akshai has quit IRC | 14:59 | |
| mriedem | bauzas: ok, after the performance meeting, starting nowish | 14:59 |
| mriedem | #openstack-performance | 14:59 |
| *** jwcroppe has quit IRC | 14:59 | |
| bauzas | is there a meeting there? | 14:59 |
| bauzas | man | 14:59 |
| bauzas | I'm just following the meeting rooms | 14:59 |
| *** achanda has quit IRC | 14:59 | |
| *** mgoddard_ has joined #openstack-nova | 14:59 | |
| bauzas | -ETOOMANYCHANS | 15:00 |
| mriedem | should be starting in a minute | 15:00 |
| *** akshai has joined #openstack-nova | 15:00 | |
| *** jaosorior has quit IRC | 15:00 | |
| mriedem | johnthetubaguy: is there a nova meeting this week? | 15:00 |
| *** haomaiwa_ has quit IRC | 15:01 | |
| mriedem | it's on the US thanksgiving holiday | 15:01 |
| *** mylu has quit IRC | 15:01 | |
| johnthetubaguy | mriedem: so I haven't cancelled it yet, I was thinking of sending a not the ML about that | 15:01 |
| *** haomaiwa_ has joined #openstack-nova | 15:01 | |
| *** garyk1 has quit IRC | 15:02 | |
| *** mylu has joined #openstack-nova | 15:02 | |
| *** akshai has quit IRC | 15:02 | |
| *** jaosorior has joined #openstack-nova | 15:02 | |
| *** boris-42 has joined #openstack-nova | 15:02 | |
| johnthetubaguy | mriedem: I am thinking we just do it anyway, noting its likely to be quite short, and I am assuming no US folks are joining for that? or does that sound like a bad plan? | 15:02 |
| mriedem | seems fine | 15:02 |
| *** rotbeard has quit IRC | 15:03 | |
| *** ljxiash has quit IRC | 15:03 | |
| *** mgoddard has quit IRC | 15:03 | |
| *** salv-orlando has joined #openstack-nova | 15:04 | |
| lxsli | mriedem: can you look at https://review.openstack.org/#/c/248780/ when you have a minute please? | 15:04 |
| *** ljxiash has joined #openstack-nova | 15:04 | |
| *** thorst_ has joined #openstack-nova | 15:04 | |
| mriedem | lxsli: that probably needs to come up in the nova meeting | 15:05 |
| lxsli | OK | 15:05 |
| *** thorst_ has quit IRC | 15:05 | |
| *** rajesht has quit IRC | 15:05 | |
| mriedem | also, | 15:05 |
| *** thorst_ has joined #openstack-nova | 15:06 | |
| lxsli | does it look vaguely plausible? I haven't done a lot of schema changes before | 15:06 |
| mriedem | nvm | 15:06 |
| *** chinmaya has quit IRC | 15:06 | |
| mriedem | lxsli: the change itself looks ok | 15:07 |
| *** breitz has quit IRC | 15:08 | |
| *** breitz has joined #openstack-nova | 15:08 | |
| lxsli | mriedem: excellent, thanks | 15:08 |
| *** dims_ has joined #openstack-nova | 15:08 | |
| mriedem | it doesn't seem to have broken the bw usage cache periodic task in the xenserver CI http://dd6b71949550285df7dc-dda4e480e005aaa13ec303551d2d8155.r49.cf1.rackcdn.com/80/248780/1/16806/logs/screen-n-cpu.txt.gz | 15:09 |
| madhaviy | :chinmaya Thanks | 15:09 |
| mriedem | which is good | 15:09 |
| *** dims has quit IRC | 15:09 | |
| openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova: Use version convert methods from oslo_utils.versionutils https://review.openstack.org/203322 | 15:09 |
| mriedem | the db migration must not have been too bad since the t-h testing passed | 15:09 |
| mriedem | alaski: does this cause any db migration issues for rax? https://review.openstack.org/#/c/248780/ | 15:10 |
| *** burgerk has joined #openstack-nova | 15:10 | |
| *** thorst_ has quit IRC | 15:10 | |
| mriedem | xen is the only virt driver that uses that table | 15:10 |
| *** burgerk_ has joined #openstack-nova | 15:11 | |
| *** sahid has quit IRC | 15:12 | |
| *** jinxing has quit IRC | 15:13 | |
| *** gszasz has quit IRC | 15:14 | |
| *** burgerk has quit IRC | 15:15 | |
| *** jerrygb has quit IRC | 15:15 | |
| *** jerrygb has joined #openstack-nova | 15:16 | |
| *** njirap has quit IRC | 15:16 | |
| *** gcb has quit IRC | 15:17 | |
| *** thangp has joined #openstack-nova | 15:18 | |
| *** markvoelker has quit IRC | 15:18 | |
| *** markvoelker has joined #openstack-nova | 15:18 | |
| *** sneti has joined #openstack-nova | 15:20 | |
| *** sahid has joined #openstack-nova | 15:21 | |
| *** zenoway has quit IRC | 15:21 | |
| markus_z | sdague: You showed my the shiny grafana, would be cool to have your feedback to the PoC: http://lists.openstack.org/pipermail/openstack-dev/2015-November/080381.html | 15:24 |
| *** venkat_p has quit IRC | 15:25 | |
| *** subscope has quit IRC | 15:26 | |
| *** pratikmallya has joined #openstack-nova | 15:27 | |
| johnthetubaguy | mriedem: I think add foreign key can be expensive on older versions of MySQL, but it looks OK on newer versions: http://dev.mysql.com/doc/refman/5.7/en/innodb-create-index-overview.html | 15:27 |
| mriedem | resurrect johannes bot | 15:27 |
| *** gszasz has joined #openstack-nova | 15:27 | |
| johnthetubaguy | yeah | 15:28 |
| alaski | mriedem: I'll have to actually test it later, not at a good place to do it atm, but it's very likely to cause issues. | 15:29 |
| alaski | mriedem: not a reason to not add it though | 15:29 |
| *** electrocucaracha has joined #openstack-nova | 15:30 | |
| johnthetubaguy | alaski: it looks screwed for mysql 5.5, the docs for mariaDB seem to suggest you need to ask for it to be online but I am unsure, and mysql 5.6 looks OK, but we don't use that | 15:30 |
| alaski | johnthetubaguy: yeah, I'm more worried about needing to clean up data so FK constraints aren't violated before it's run | 15:31 |
| bauzas | meh, are we talking of adding a FK on the instances table ? | 15:31 |
| bauzas | that's a big operational change AFAIK | 15:31 |
| bauzas | because it will lock the table unless I'm wrong? | 15:31 |
| johnthetubaguy | bauzas: not always, but this is bandwidth usage table (which is probably bigger) | 15:32 |
| johnthetubaguy | alaski: true | 15:32 |
| bauzas | johnthetubaguy: :/ | 15:32 |
| bauzas | I played with MySQL 5.1 and 5.5 and that kind of thing is very sensitive on large tables | 15:33 |
| mriedem | bauzas: https://review.openstack.org/#/c/248780/ | 15:33 |
| *** gszasz_ has joined #openstack-nova | 15:33 | |
| mriedem | alaski: johnthetubaguy: yeah i suspect there are a lot of orphaned bw_usage_cache records, pointing at instance uuids that aren't in the instances table anymore | 15:33 |
| *** jerrygb has quit IRC | 15:33 | |
| *** electrocucaracha has quit IRC | 15:34 | |
| bauzas | johnthetubaguy: oh I see the MySQL page you said, thanks for that | 15:35 |
| *** gszasz has quit IRC | 15:35 | |
| alaski | mriedem: yeah. So likely to cause issues there, and potentially have db downtime to run the migration. The orphaned data is our problem though and shouldn't affect the review. | 15:36 |
| *** salv-orlando has quit IRC | 15:36 | |
| *** akshai has joined #openstack-nova | 15:36 | |
| *** subscope has joined #openstack-nova | 15:36 | |
| alaski | belliott did some testing recently on adding foreign keys and db downtime so he may be able to answer the other part. | 15:36 |
| mriedem | alaski: well, we could archive/purge orphaned bw_usage_cache as part of the migration if needed | 15:37 |
| *** Marga_ has joined #openstack-nova | 15:38 | |
| alaski | purging data that can't exist with the FK would be nice | 15:38 |
| *** shaohe_feng1 has joined #openstack-nova | 15:40 | |
| *** subscope has quit IRC | 15:40 | |
| alaski | I'm not supposed to be here today so I'm going to step away again. But I'll follow up with belliott on potential downtime from that migration so we can add a data point to it | 15:40 |
| mriedem | ok, thanks | 15:40 |
| *** salv-orlando has joined #openstack-nova | 15:40 | |
| markus_z | mikal: What can I do to revoke your -1 from the virtlogd spec? https://review.openstack.org/#/c/234291/ | 15:40 |
| *** gszasz_ has quit IRC | 15:41 | |
| mriedem | jhesketh: can you tell me if any of the data sets in the t-h runs have a populated bw_usage_cache table? | 15:42 |
| *** shaohe_feng has quit IRC | 15:42 | |
| *** mylu has quit IRC | 15:43 | |
| *** tdurakov has joined #openstack-nova | 15:44 | |
| *** sfinucan has joined #openstack-nova | 15:46 | |
| *** haomaiwa_ has quit IRC | 15:49 | |
| mirrorbox | jaypipes: ping | 15:50 |
| *** EinstCra_ has quit IRC | 15:51 | |
| jaypipes | mirrorbox: gimme about 15 minutes please | 15:51 |
| *** salv-orlando has quit IRC | 15:51 | |
| openstackgerrit | Stephen Finucane proposed openstack/nova: test_fields: Remove all 'Enum' subclass tests https://review.openstack.org/244372 | 15:51 |
| *** zenoway has joined #openstack-nova | 15:51 | |
| mirrorbox | jaypipes: cool, no rush, just wanted to remind you about |https://review.openstack.org/#/c/182242/ | 15:52 |
| *** achanda_ has quit IRC | 15:53 | |
| *** pratikmallya has quit IRC | 15:54 | |
| mriedem | hmm, i have a feeling this is done https://blueprints.launchpad.net/nova/+spec/fewer-networking-token-checks | 15:57 |
| mriedem | very old bp | 15:57 |
| mriedem | but i thought with the caching of neutronclient and the session stuff that jamielennox added awhile back, that's basically done | 15:57 |
| *** subscope has joined #openstack-nova | 15:57 | |
| *** pratikmallya has joined #openstack-nova | 15:59 | |
| *** subscope has quit IRC | 15:59 | |
| *** achanda has joined #openstack-nova | 15:59 | |
| sahid | danpb: any chance you look at https://review.openstack.org/#/c/224726/ and https://review.openstack.org/#/c/197569/ (was rebase) | 16:00 |
| bauzas | jaypipes: when you're around and free, got a sec for educating me on https://review.openstack.org/#/c/211753/38/nova/objects/request_spec.py,cm ? | 16:00 |
| *** salv-orlando has joined #openstack-nova | 16:00 | |
| bauzas | jaypipes: I missed those decos, so I wonder what is needed for any DB API method | 16:00 |
| *** kentb1 has joined #openstack-nova | 16:01 | |
| bauzas | (I can read the code tho) | 16:01 |
| PaulMurray | ndipanov, hi, are you around? I want to ask about device names for volumes | 16:01 |
| *** vilobhmm has joined #openstack-nova | 16:01 | |
| *** pratikma_ has joined #openstack-nova | 16:03 | |
| *** tdurakov has quit IRC | 16:03 | |
| *** gszasz has joined #openstack-nova | 16:03 | |
| *** vilobhmm1 has joined #openstack-nova | 16:04 | |
| mriedem | lxsli: you'd do it in the migration script | 16:04 |
| lxsli | sorry, wrong channel | 16:04 |
| lxsli | thanks! | 16:04 |
| *** otter768 has joined #openstack-nova | 16:04 | |
| PaulMurray | danpb, do you have a minute, same question about device names ^^^ | 16:04 |
| mriedem | lxsli: i guess the question is do we archive (move to shadow tables?) or purge? | 16:04 |
| *** salv-orlando has quit IRC | 16:04 | |
| lxsli | We should ask a Xen person I guess but I'd rather purge | 16:05 |
| mriedem | lxsli: i don't really know how valuable that data is so i'm not sure if it should be archived, that's a question more for the rax people since they are the ones using it | 16:05 |
| danpb | PaulMurray: what about them ? | 16:05 |
| *** pratikmallya has quit IRC | 16:05 | |
| mriedem | lxsli: i guess it could check shadow tables for the instance uuid and if not there, purge, else archive, | 16:06 |
| *** vilobhmm has quit IRC | 16:06 | |
| mriedem | but maybe that's too over the top for this | 16:06 |
| PaulMurray | danpb, I'm doing a spec for detaching/attaching volumes in when shelved | 16:06 |
| PaulMurray | danpb, attaching is done in a synchronized block normally | 16:06 |
| *** suro-patz has joined #openstack-nova | 16:06 | |
| PaulMurray | but if the instance is shelved there is no compute host involved, so the is no way to synchronize | 16:07 |
| PaulMurray | It turns out the only thing that needs the sync is | 16:07 |
| PaulMurray | the part that reserves a name for the device | 16:07 |
| *** ericksonsantos has joined #openstack-nova | 16:07 | |
| PaulMurray | which calls down into the virt driver | 16:07 |
| openstackgerrit | Paul Carlton proposed openstack/nova-specs: Cancel ongoing live migrations https://review.openstack.org/228828 | 16:07 |
| *** zenoway has quit IRC | 16:07 | |
| *** klkumar has quit IRC | 16:08 | |
| PaulMurray | So I'm wondering about how to get around doing that? | 16:08 |
| PaulMurray | Are you familiar with that? | 16:08 |
| *** klkumar has joined #openstack-nova | 16:08 | |
| lxsli | mriedem: +1 over the top | 16:08 |
| jaypipes | bauzas: gonna be a while... | 16:09 |
| *** otter768 has quit IRC | 16:09 | |
| bauzas | jaypipes: nvm, just reading the code | 16:09 |
| *** whenry_ has joined #openstack-nova | 16:10 | |
| openstackgerrit | Stephen Finucane proposed openstack/nova-specs: Rework policies for virt-driver CPU thread pinning https://review.openstack.org/244198 | 16:10 |
| *** nkrinner has quit IRC | 16:12 | |
| sfinucan | danpb: Could I request some of your time for a review of the CPU thread pinning spec rework? It seems you're the main man as far as this spec goes. | 16:12 |
| *** mjura has quit IRC | 16:13 | |
| lxsli | jaypipes: https://review.openstack.org/#/c/219761/ is landable! | 16:13 |
| lxsli | should be no diff since you reviewed it last | 16:14 |
| mriedem | i wonder how many people will ping jaypipes to review a thing before his head explodes | 16:15 |
| mriedem | let's start an office pool | 16:15 |
| danpb | PaulMurray: needs nikola to answer that i think | 16:15 |
| *** madhaviy has quit IRC | 16:15 | |
| PaulMurray | danpb, ok - thanks | 16:16 |
| *** matrohon has quit IRC | 16:16 | |
| lxsli | mriedem: yeah but it's wafer thin | 16:16 |
| edleafe | mriedem: put me down for $10 on 5 | 16:16 |
| *** rcernin has joined #openstack-nova | 16:16 | |
| bauzas | paul-carlton: you should enjoy some fancy answer to your ML thread... | 16:16 |
| *** mpavone has quit IRC | 16:17 | |
| *** madhaviy has joined #openstack-nova | 16:17 | |
| paul-carlton1 | I'm already re thinking it a bit, maybe sdague is right and I should re write it with a GET /servers/{id}/migration then call DELETE on it | 16:19 |
| paul-carlton1 | I'd go for migration (singular) because only one migration can be active and if you want old migrations the is migrations-list | 16:20 |
| *** alexschm has quit IRC | 16:20 | |
| *** Guest72939 has quit IRC | 16:22 | |
| *** achanda_ has joined #openstack-nova | 16:23 | |
| mriedem | gibi: this isn't actually complete right? https://blueprints.launchpad.net/nova/+spec/remove-service-management-from-nova-manage | 16:23 |
| *** rods has joined #openstack-nova | 16:23 | |
| mriedem | that's supposed to be targeted for N | 16:23 |
| *** mgoddard_ has quit IRC | 16:24 | |
| gibi | mriedem: yes, that is supposed to be done in N | 16:24 |
| *** mgoddard_ has joined #openstack-nova | 16:24 | |
| mriedem | ok, i updated launchpad | 16:24 |
| gibi | mriedem: this was completed for Mitaka https://blueprints.launchpad.net/nova/+spec/deprecate-service-management-in-nova-manage | 16:25 |
| mriedem | yeah | 16:25 |
| *** mylu has joined #openstack-nova | 16:25 | |
| gibi | mriedem: thanks | 16:26 |
| *** achanda has quit IRC | 16:26 | |
| *** zhenguo has quit IRC | 16:29 | |
| *** dims_ has quit IRC | 16:30 | |
| *** josecastroleon has quit IRC | 16:31 | |
| *** matrohon has joined #openstack-nova | 16:32 | |
| *** jdurgin1 has joined #openstack-nova | 16:32 | |
| *** suro-patz has quit IRC | 16:35 | |
| *** gszasz has quit IRC | 16:37 | |
| *** yonglihe has joined #openstack-nova | 16:37 | |
| *** signed8bit has joined #openstack-nova | 16:40 | |
| *** aix has quit IRC | 16:40 | |
| *** salv-orlando has joined #openstack-nova | 16:41 | |
| *** salv-orlando has quit IRC | 16:41 | |
| *** salv-orlando has joined #openstack-nova | 16:41 | |
| *** daemontool has quit IRC | 16:43 | |
| johnthetubaguy | PaulMurray: bauzas: I had an idea for a slight tweak to that live-migrate doc, can I share that with you both? | 16:44 |
| *** daemontool has joined #openstack-nova | 16:44 | |
| bauzas | johnthetubaguy: I need to drop off (daughter to pick up at her english class) but feel free to comment that here | 16:44 |
| johnthetubaguy | bauzas: I am thinking we could just merge it and fix up a few things in a follow on patch? | 16:45 |
| PaulMurray | johnthetubaguy, yes, I was just going over the comments | 16:45 |
| *** tdurakov has joined #openstack-nova | 16:45 | |
| bauzas | johnthetubaguy: I can remove that | 16:45 |
| PaulMurray | johnthetubaguy, I'm good with that, but what's the fix up? | 16:46 |
| *** dims has joined #openstack-nova | 16:46 | |
| bauzas | johnthetubaguy: PaulMurray: tbc, are we okay to spin a new change providing some modifications ? | 16:46 |
| bauzas | johnthetubaguy: PaulMurray: if so, I can +W and say "fair enough" | 16:47 |
| ndipanov | PaulMurray, am now | 16:47 |
| ndipanov | what | 16:47 |
| ndipanov | up | 16:47 |
| johnthetubaguy | bauzas: one sec... | 16:47 |
| PaulMurray | ndipanov, with you in one moment | 16:47 |
| openstackgerrit | John Garbutt proposed openstack/nova: DO NOT MERGE: quick idea to refactor move servers https://review.openstack.org/249321 | 16:47 |
| tdurakov | folks, could anyone review patch https://review.openstack.org/#/c/249110/ | 16:47 |
| PaulMurray | bauzas, BTW - cloud operator and cloud user are actors - not nova API roles | 16:47 |
| bauzas | johnthetubaguy: dropped my -1 | 16:47 |
| bauzas | I really need to go | 16:48 |
| bauzas | feel free to clarify that in the channel | 16:48 |
| johnthetubaguy | bauzas: no problems, you should go | 16:48 |
| *** jdurgin1 has quit IRC | 16:48 | |
| PaulMurray | johnthetubaguy, you going to +W it? | 16:49 |
| johnthetubaguy | PaulMurray: I am curious if that makes sense: https://review.openstack.org/249321 | 16:49 |
| *** daemontool has quit IRC | 16:49 | |
| PaulMurray | johnthetubaguy, yes, that does help | 16:50 |
| *** scheuran has quit IRC | 16:50 | |
| johnthetubaguy | PaulMurray: we could leave that as a follow on patch, thats probably easiest I guess? | 16:50 |
| *** bnemec has quit IRC | 16:50 | |
| PaulMurray | johnthetubaguy, sure, works for me | 16:50 |
| *** suro-patz has joined #openstack-nova | 16:51 | |
| PaulMurray | johnthetubaguy, did you get my point above about actors vs roles? | 16:51 |
| *** EinstCrazy has joined #openstack-nova | 16:51 | |
| PaulMurray | not sure I want to complicate it | 16:51 |
| claudiub|2 | johnthetubaguy: sdague: hello. the exceptions raised by the HyperVDriver have been refined into more specific exceptions. As per our discussion last week. Can you take a look at the patch pls? https://review.openstack.org/#/c/212102/ | 16:52 |
| johnthetubaguy | PaulMurray: I think so, I think if we put migrate first, those comments about the cloud operator fit into context more easily | 16:52 |
| PaulMurray | johnthetubaguy, agree - it looks a lot more natural - comes across as use cases rather than explaining actions | 16:53 |
| *** kbyrne has quit IRC | 16:53 | |
| PaulMurray | ndipanov, did you see my question to danpb above? | 16:54 |
| ndipanov | PaulMurray, let me scorll up | 16:55 |
| *** jwcroppe has joined #openstack-nova | 16:55 | |
| ndipanov | PaulMurray, hmm | 16:56 |
| ndipanov | that's a good one | 16:56 |
| ndipanov | I wonder... | 16:56 |
| ndipanov | if we had some kind of a... | 16:56 |
| ndipanov | distributed lock manager... | 16:56 |
| *** suro-patz has quit IRC | 16:56 | |
| * ndipanov grins | 16:56 | |
| *** EinstCrazy has quit IRC | 16:56 | |
| PaulMurray | ndipanov, if you bring up DLMs and world peace again I'm going home | 16:56 |
| ndipanov | lol | 16:56 |
| ndipanov | so thinking about it | 16:57 |
| ndipanov | so I would say | 16:57 |
| ndipanov | attach and leave the device name empty | 16:57 |
| ndipanov | once the instance gets a compute host | 16:57 |
| ndipanov | decide device names | 16:57 |
| ndipanov | holding an instance wide lock like attach does | 16:58 |
| *** e0ne has quit IRC | 16:58 | |
| ndipanov | PaulMurray, does that make sense? | 16:58 |
| PaulMurray | ndipanov, its brilliant - exactly what I was thinking ! | 16:58 |
| ndipanov | that's what boot does anyway | 16:58 |
| PaulMurray | so completely ignore the device-name if the user supplies one | 16:59 |
| *** danpb has left #openstack-nova | 16:59 | |
| PaulMurray | I think it gets ignored anyway doesn't it? | 16:59 |
| *** danpb has joined #openstack-nova | 16:59 | |
| PaulMurray | in the virt driver | 16:59 |
| PaulMurray | (at least libvirt) | 17:00 |
| ndipanov | yeah I mean feel free to store it but in case of libvirt it will get overridden | 17:00 |
| *** lykinsbd has joined #openstack-nova | 17:00 | |
| PaulMurray | will it trip up anywhere If the same name is supplied for two devices? | 17:00 |
| *** diegows has quit IRC | 17:00 | |
| *** mgoddard_ has quit IRC | 17:00 | |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: review guide: add location details for config options https://review.openstack.org/245789 | 17:00 |
| ndipanov | though if we are adding a new API let's start off with no device name even possible | 17:01 |
| *** mgoddard has joined #openstack-nova | 17:01 | |
| PaulMurray | ndipanov, are we adding a new API? | 17:01 |
| mtanino | sdague: hi, | 17:02 |
| markus_z | garyk: I don't fully understand your feedback in https://review.openstack.org/#/c/242064/ Could you explain it to me please? | 17:02 |
| *** jwcroppe has quit IRC | 17:03 | |
| *** shaohe_feng1 has quit IRC | 17:03 | |
| garyk | markus_z: all code under openstack/common is code that is imported from the oslo incubator | 17:05 |
| garyk | theoretically this code should note be changed | 17:05 |
| *** shihanzhang has quit IRC | 17:05 | |
| *** liusheng has quit IRC | 17:05 | |
| *** shihanzhang has joined #openstack-nova | 17:05 | |
| garyk | that is, say a bug was fixed in the oslo policy code then that would be cherry picked and your changes would be overridden | 17:05 |
| PaulMurray | johnthetubaguy, we need to stop changes in api-guide triggering all those CI jobs | 17:05 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova: Add infra for versioned notifications https://review.openstack.org/247024 | 17:05 |
| *** liusheng has joined #openstack-nova | 17:06 | |
| *** jlanoux_ has quit IRC | 17:06 | |
| markus_z | garyk: woot? That import practice is news to me. | 17:06 |
| garyk | markus_z: yeah, it has been like this since dokeys years. | 17:07 |
| johnthetubaguy | PaulMurray: oh, true, I can never remember how thats done in gate | 17:07 |
| openstackgerrit | John Garbutt proposed openstack/nova: docs: add clarifications to move servers https://review.openstack.org/249321 | 17:07 |
| garyk | markus_z: fyi - https://github.com/openstack/nova/blob/master/openstack-common.conf | 17:07 |
| lxsli | markus_z: it's true. the incubator is nearly empty now though afaik, we've been trying to move away from it | 17:08 |
| markus_z | garyk: Hah, learning never stops. OK, got it, I will change the patch set. | 17:08 |
| garyk | markus_z: thanks. fyi i think that this code will go away when noa moves to oslo_policy | 17:09 |
| markus_z | lxsli: Good to know. I haven't done anything with the incubator yet. | 17:09 |
| openstackgerrit | John Garbutt proposed openstack/nova: docs: reorder move servers https://review.openstack.org/249331 | 17:09 |
| markus_z | garyk: Do you happen to know when this is going to happen? | 17:11 |
| *** ssurana has joined #openstack-nova | 17:11 | |
| johnthetubaguy | PaulMurray: put those two patches up, they really don't change too much, just a few tweaks, let me know how they look | 17:11 |
| PaulMurray | johnthetubaguy, will do | 17:12 |
| johnthetubaguy | markus_z: there is a blueprint trying to track the olso stuff: https://blueprints.launchpad.net/nova/+spec/oslo-for-mitaka | 17:12 |
| johnthetubaguy | markus_z: not sure all the when and how have actually been agreed yet | 17:13 |
| markus_z | johnthetubaguy: Great, thanks! Seems to be blocked right now by dansmith | 17:13 |
| openstackgerrit | Alexis Lee proposed openstack/nova: Add FK bw_usage_cache.uuid -> instances.uuid https://review.openstack.org/248780 | 17:13 |
| markus_z | johnthetubaguy: no worries, I was just curious | 17:13 |
| *** ccarmack has joined #openstack-nova | 17:13 | |
| dansmith | markus_z: say wha? | 17:14 |
| garyk | markus_z: no idea. | 17:14 |
| markus_z | dansmith: https://review.openstack.org/#/c/198065/ | 17:14 |
| *** madhaviy has quit IRC | 17:14 | |
| dansmith | markus_z: I was just -2ing it by request before the last release | 17:15 |
| *** Luigi_ has quit IRC | 17:15 | |
| markus_z | dansmith: It's a free elf now :) | 17:15 |
| markus_z | thanks | 17:15 |
| markus_z | dims: FYI: https://review.openstack.org/#/c/198065/ | 17:16 |
| *** jwcroppe has joined #openstack-nova | 17:17 | |
| *** kfarr has joined #openstack-nova | 17:18 | |
| *** dims has quit IRC | 17:18 | |
| *** dims_ has joined #openstack-nova | 17:18 | |
| *** david-ly_ has joined #openstack-nova | 17:19 | |
| *** matrohon has quit IRC | 17:19 | |
| *** nic has joined #openstack-nova | 17:20 | |
| rlrossit | gibi: I have a feeling the notifications meeting is going to suddenly become popular this week ;) | 17:20 |
| openstackgerrit | Pushkar Umaranikar proposed openstack/nova: Modify VM's updated_at field on volume actions https://review.openstack.org/247176 | 17:21 |
| *** lpetrut has quit IRC | 17:21 | |
| *** david-l__ has joined #openstack-nova | 17:23 | |
| *** david-ly_ has quit IRC | 17:23 | |
| *** david-lyle has quit IRC | 17:23 | |
| *** tdurakov has quit IRC | 17:23 | |
| *** vilobhmm1 has quit IRC | 17:25 | |
| *** mylu has quit IRC | 17:26 | |
| *** kylek3h has quit IRC | 17:26 | |
| *** mylu has joined #openstack-nova | 17:26 | |
| markus_z | dansmith: Would it make sense to merge a new patch set of https://review.openstack.org/#/c/215102/ before your LibvirtLiveMigrationData object merges? I'm thinking about a possible backport of the patch to Liberty. | 17:27 |
| dansmith | markus_z: I'd rather we not | 17:28 |
| markus_z | dansmith: pre-condition is that the new patch set doesn't screw up the update behavior | 17:28 |
| dansmith | markus_z: it's adding a new thing to an unversioned blob, that existing nodes running that release don't have | 17:29 |
| dansmith | markus_z: I'd rather just call it broken and fix it going forward | 17:29 |
| dansmith | especially since it's not a regression | 17:30 |
| *** david-l__ has quit IRC | 17:30 | |
| markus_z | dansmith: Right, it never worked really... | 17:30 |
| *** mylu has quit IRC | 17:31 | |
| markus_z | dansmith: A backport would have been great and if I understand the code correctly, it would have worked in a mixed environment. | 17:32 |
| *** rdopiera has quit IRC | 17:32 | |
| dansmith | markus_z: the current code completely breaks interacting with the current release | 17:33 |
| markus_z | dansmith: True, a new patch set is needed. | 17:33 |
| dansmith | markus_z: it could be done in a way such that it doesn't break things, but I'd rather us not | 17:33 |
| markus_z | dansmith: OK, just wanted to double-check, thans. | 17:34 |
| markus_z | *thanks | 17:34 |
| *** derekh has quit IRC | 17:35 | |
| *** Marga_ has quit IRC | 17:42 | |
| *** sayali has quit IRC | 17:42 | |
| *** burt has joined #openstack-nova | 17:44 | |
| *** david-lyle has joined #openstack-nova | 17:44 | |
| *** Marga_ has joined #openstack-nova | 17:46 | |
| *** lykinsbd has quit IRC | 17:48 | |
| *** markmc has quit IRC | 17:49 | |
| *** ijw has joined #openstack-nova | 17:50 | |
| *** yamahata has joined #openstack-nova | 17:50 | |
| *** Marga_ has quit IRC | 17:51 | |
| *** klkumar has quit IRC | 17:51 | |
| *** pratikmallya has joined #openstack-nova | 17:54 | |
| *** pratikmallya has quit IRC | 17:54 | |
| *** pratikmallya has joined #openstack-nova | 17:54 | |
| *** ccarmack has left #openstack-nova | 17:55 | |
| *** pratikma_ has quit IRC | 17:56 | |
| *** aix has joined #openstack-nova | 17:57 | |
| *** suro-patz has joined #openstack-nova | 17:57 | |
| *** stackdump has joined #openstack-nova | 17:59 | |
| *** sacharya has joined #openstack-nova | 18:00 | |
| *** e0ne has joined #openstack-nova | 18:01 | |
| *** irenab has joined #openstack-nova | 18:01 | |
| *** david-lyle has quit IRC | 18:03 | |
| *** pratikmallya has quit IRC | 18:04 | |
| *** jistr has quit IRC | 18:04 | |
| *** sacharya has quit IRC | 18:04 | |
| *** otter768 has joined #openstack-nova | 18:05 | |
| *** rotbeard has joined #openstack-nova | 18:05 | |
| *** PaulMurray is now known as ptm_away | 18:06 | |
| jroll | would y'all need a spec for this, or is the BP sufficient (when the ironic spec lands)? https://blueprints.launchpad.net/nova/+spec/ironic-rescue-mode | 18:07 |
| *** markus_z has quit IRC | 18:08 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Actually pass the migration data object down to the virt drivers https://review.openstack.org/248211 | 18:09 |
| *** jaosorior has quit IRC | 18:09 | |
| *** otter768 has quit IRC | 18:09 | |
| *** jaosorior has joined #openstack-nova | 18:10 | |
| *** rotbart has joined #openstack-nova | 18:10 | |
| *** david-lyle has joined #openstack-nova | 18:11 | |
| *** wasmum has joined #openstack-nova | 18:11 | |
| *** hemna has quit IRC | 18:12 | |
| *** rotbeard has quit IRC | 18:12 | |
| *** claudiub|2 has quit IRC | 18:12 | |
| *** achanda has joined #openstack-nova | 18:13 | |
| *** Marga_ has joined #openstack-nova | 18:14 | |
| dansmith | jroll: I think one could make an argument for a specless "feature parity" blueprint for that | 18:15 |
| dansmith | jroll: what is the crypt() piece? | 18:16 |
| dansmith | the admin password for the rescue instance? | 18:16 |
| jroll | dansmith: yeah, /etc/shadow style to avoid it in the db in plaintext | 18:16 |
| jroll | I meant to write (like /etc/shadow uses) there | 18:17 |
| *** achanda_ has quit IRC | 18:17 | |
| dansmith | jroll: but you will still install the key like normal too right? | 18:17 |
| *** mriedem has quit IRC | 18:17 | |
| dansmith | jroll: do you support admin password for regular instances? | 18:17 |
| jroll | dansmith: there were reasons I don't necessarily remember why we didn't do the ssh key downstream, but if that's what other drivers do that's fine | 18:18 |
| dansmith | jroll: you don't do the key in normal instances? | 18:18 |
| jroll | dansmith: no, we do, I meant we didn't inject ssh key for rescue | 18:18 |
| dansmith | oh | 18:19 |
| lucasagomes | dansmith, jroll not totally related but yet related we do something similar here https://review.openstack.org/#/c/245082/ | 18:19 |
| dansmith | I'm not sure about the other drivers, but libvirt doesn't implement admin pass at all, AFAIK | 18:19 |
| jroll | as for admin password on regular instances, I think so? assuming that cloud-init/metadata service/configdrive support it | 18:19 |
| jroll | orly | 18:19 |
| lucasagomes | we generate the encrypted password with openssl | 18:19 |
| dansmith | okay, we can do it if the agent is running it looks like | 18:20 |
| *** kevinbenton has quit IRC | 18:20 | |
| dansmith | anyway, I'm not saying you shouldn't do it, but I think key would be required too | 18:21 |
| jroll | dansmith: ok cool, I'll mention that in a comment on the ironic spec | 18:21 |
| jroll | thanks | 18:21 |
| *** mriedem has joined #openstack-nova | 18:21 | |
| dansmith | jroll: so I'l put that on the meeting agenda for thursday, with a comment about it being feature parity, | 18:22 |
| dansmith | and when you return from food coma week, it might be approved | 18:22 |
| *** achanda has quit IRC | 18:22 | |
| jroll | dansmith: sweet, ty sir. note the ironic spec isn't approved but I don't expect much to change | 18:23 |
| dansmith | ah, well, that'd be a thing probably | 18:23 |
| jroll | dansmith: on the ssh key thing, is that --key-name passed to nova rescue or? | 18:23 |
| jroll | just whatever key is used for that instance? | 18:23 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: introduce "snapshot" to Guest's object https://review.openstack.org/249353 | 18:23 |
| dansmith | jroll: should be whatever key the instance was built with | 18:23 |
| dansmith | yeah | 18:24 |
| mriedem | jroll: make sure your nova bp depends on the ironic bp in launchpad | 18:24 |
| jroll | dansmith: so what if their key was lost/compromised? :) | 18:24 |
| dansmith | jroll: I guess maybe cloud-init in the rescue instance will handle that? I dunno | 18:24 |
| dansmith | jroll: meaning that's the reason for the rescue? | 18:24 |
| jroll | dansmith: yeah | 18:24 |
| dansmith | jroll: I'm not sure that's a thing we care about, but if adminpass works too, then.. :) | 18:24 |
| jroll | dansmith: keeping in mind not all ironic deployments allow for console | 18:25 |
| jroll | heh, yeah. cool. | 18:25 |
| dansmith | jroll: I'm not sure that's a thing we care about, but if adminpass works too, then.. :) | 18:25 |
| dansmith | oops | 18:25 |
| jroll | :P | 18:25 |
| jroll | mriedem: done | 18:25 |
| *** pixelbeat has quit IRC | 18:27 | |
| *** liverpooler has quit IRC | 18:27 | |
| gibi | rlrossit: we will see :) | 18:27 |
| *** jaosorior has quit IRC | 18:28 | |
| *** sahid has quit IRC | 18:28 | |
| gibi | dansmith: hi! do you have a minute? | 18:31 |
| dansmith | gibi: yeah | 18:31 |
| *** tjones has joined #openstack-nova | 18:31 | |
| gibi | dansmith: so I'm trying to figure out how the Service.version field is updated | 18:31 |
| gibi | the code here https://github.com/openstack/nova/blob/master/nova/objects/service.py#L267-L271 | 18:31 |
| gibi | makes me wonder | 18:31 |
| gibi | I'm implementing a new notification for service status change and I'd like to send notification if the version changes too | 18:32 |
| *** danpb has quit IRC | 18:33 | |
| dansmith | gibi: that version never changes at runtime, so I'm not sure what you're trying to notify about | 18:33 |
| gibi | dansmith: does it changed in the db directly? | 18:34 |
| dansmith | gibi: inside a given service, it's defined here: https://github.com/openstack/nova/blob/master/nova/objects/service.py#L31 | 18:34 |
| dansmith | gibi: which we write to the database so that other services know how old or new we are | 18:34 |
| gibi | dansmith: so when somebody deploys a new version of a service and restarts the service the new code will update the service.version in the specific db row? | 18:36 |
| *** pixelbeat has joined #openstack-nova | 18:36 | |
| *** ctrath has quit IRC | 18:37 | |
| dansmith | gibi: yes | 18:37 |
| *** tdurakov has joined #openstack-nova | 18:38 | |
| gibi | dansmith: so the Service.__init__ will do the update in the object, but then somebody will call save later on to persist the change | 18:39 |
| dansmith | gibi: yes | 18:39 |
| gibi | and we assume that this save will not be a skip because other than the version field is changed | 18:40 |
| *** vilobhmm11 has joined #openstack-nova | 18:40 | |
| *** lucasagomes is now known as lucas-dinner | 18:40 | |
| dansmith | gibi: it requires us to do a .save() to update it, but we do that every $periodic on the services anyway | 18:41 |
| gibi | dansmith: I see. Thanks it makes sense. So my plan to send the notification from the .save() will work for the version change as well | 18:43 |
| *** ctrath has joined #openstack-nova | 18:44 | |
| dansmith | gibi: well, it'll always look dirty, so you probably want to filter it out somehow | 18:44 |
| dansmith | gibi: otherwise it'll keep looking changed in your notifications | 18:45 |
| gibi | if it is only the version changed then I will not send notification | 18:45 |
| gibi | as the save() will be a skip anyhow | 18:45 |
| *** rlrossit has quit IRC | 18:46 | |
| *** bnemec has joined #openstack-nova | 18:47 | |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova-specs: Enable SR-IOV physical functions assignment with Neutron port https://review.openstack.org/239875 | 18:47 |
| *** mtanino has quit IRC | 18:47 | |
| gibi | dansmith: this is the current piece of code https://review.openstack.org/#/c/245678/8/nova/objects/service.py,cm | 18:47 |
| *** jichen has joined #openstack-nova | 18:48 | |
| *** pilgrimstack has quit IRC | 18:49 | |
| *** ijw has quit IRC | 18:50 | |
| *** pratikmallya has joined #openstack-nova | 18:50 | |
| *** pratikmallya has quit IRC | 18:50 | |
| doffm | edleafe: Do you have a PoC for the alternative scheduler described in your resource-tracker spec? | 18:51 |
| dansmith | gibi: if you don't care to show the fields that changed in the parent object, then that's fine | 18:51 |
| *** pratikmallya has joined #openstack-nova | 18:51 | |
| edleafe | doffm: yes | 18:51 |
| *** garyk has quit IRC | 18:51 | |
| mriedem | i'm surprised we don't notify on interface attach | 18:52 |
| *** apoorvad has joined #openstack-nova | 18:52 | |
| edleafe | doffm: I haven't put it through any sort of heavy tests, though. Just enough to show it working | 18:52 |
| doffm | edleafe: Is it posted anywhere? What did it end up using Zookeeper? Cassandra? I can't remember. :/ | 18:52 |
| *** rlrossit has joined #openstack-nova | 18:53 | |
| edleafe | doffm: No, not posted anywhere. I showed it to a few people in Tokyo. | 18:53 |
| edleafe | doffm: since it was a hack due to not being able to swap out the resource tracker, I didn't post it | 18:54 |
| gibi | dansmith: you mean in every other save() call the version will look like as changed even if it is previously saved? | 18:54 |
| edleafe | doffm: and it used Cassandra, but really could have been Kafka or anything else | 18:54 |
| doffm | edleafe: I'd be happy to help if you're interested in taking the PoC any further. Just let me know. | 18:55 |
| dansmith | gibi: no, I'm saying it doesn't look like your top-level notification has a "these are the fields that actually changed" list | 18:55 |
| dansmith | gibi: which is fine | 18:55 |
| doffm | I guess johnthetubaguy was saying he wanted something that *could* go in-tree before adding a config option. | 18:56 |
| dansmith | gibi: if you did, you might want to figure out how to filter version so you only show it as changed if it actually did change | 18:56 |
| *** sacharya has joined #openstack-nova | 18:56 | |
| edleafe | doffm: honestly, there is such an entrenched resistance that I'm not planning on spending more time on it | 18:56 |
| edleafe | doffm: but to be in-tree, it has to support an in-tree solution | 18:57 |
| edleafe | doffm: and using anything other than the current db will *not* be supported in-tree | 18:57 |
| mriedem | i feel like the orchestration that's going to be added with this is going to be gross https://review.openstack.org/#/c/139910/ | 18:57 |
| *** achanda has joined #openstack-nova | 18:57 | |
| gibi | dansmith: OK, thanks for the help! | 18:57 |
| doffm | edleafe: Fair enough. That does seem to be the reaction. :) | 18:58 |
| edleafe | doffm: indeed. Well, at least I got people to "think different" :) | 18:59 |
| *** pradk has quit IRC | 19:00 | |
| *** garyk has joined #openstack-nova | 19:00 | |
| doffm | Tooo different. :) | 19:00 |
| *** zenoway has joined #openstack-nova | 19:01 | |
| *** jerrygb has joined #openstack-nova | 19:01 | |
| *** wuhg has quit IRC | 19:01 | |
| *** pradk has joined #openstack-nova | 19:02 | |
| *** pradk has quit IRC | 19:02 | |
| *** aginwala has joined #openstack-nova | 19:02 | |
| *** hemna has joined #openstack-nova | 19:03 | |
| *** pratikmallya has quit IRC | 19:05 | |
| *** pratikmallya has joined #openstack-nova | 19:05 | |
| *** LukeHinds has quit IRC | 19:06 | |
| *** tdurakov has quit IRC | 19:09 | |
| *** unicell has joined #openstack-nova | 19:09 | |
| *** kevinbenton has joined #openstack-nova | 19:10 | |
| *** alejandrito has quit IRC | 19:10 | |
| *** tdurakov has joined #openstack-nova | 19:10 | |
| mriedem | what is the deal with the hyper-v manual rechecks from all of their people? | 19:12 |
| mriedem | ociuhandu: ^? | 19:12 |
| mriedem | it's hard for me to filter hyper-v ci noise when it's real people manually rechecking things to run the hyper-v ci | 19:12 |
| *** signed8bit is now known as signed8bit_ZZZzz | 19:13 | |
| edleafe | mriedem: yeah, I've gotten them on things that haven't even had the hyper-v ci run yet | 19:13 |
| mriedem | i just want them to use robots like everyone else | 19:14 |
| *** mylu has joined #openstack-nova | 19:14 | |
| mriedem | i guess these could be robots | 19:14 |
| mriedem | romanian CI cyborgs | 19:14 |
| ociuhandu | mriedem: we are having at the moment some issues with some pip dependencies and have had fails on the CI due to those | 19:14 |
| mriedem | ociuhandu: so these are people rechecking on hyper-v ci fails? or just to trigger the initial run? | 19:15 |
| mriedem | ociuhandu: like pip mirror issues? | 19:15 |
| ociuhandu | mriedem: we get notifications on failed CI runs, that are looked over for reasons. If the reason is found to be in the CI run (like it failed during the environment initialization) we recheck only on our CI | 19:16 |
| mriedem | ok | 19:18 |
| *** mtanino has joined #openstack-nova | 19:18 | |
| mriedem | another example of how i wish we had like an elastic-recheck for 3rd party CI, but that's not likely going to happen | 19:18 |
| ociuhandu | mriedem: we have CI internal checks and retries for any build / run step to catch these issues and we’re working on improving those also. | 19:18 |
| vilobhmm11 | bauzas : ping | 19:20 |
| bauzas | vilobhmm11: ack | 19:20 |
| vilobhmm11 | wanted to talk regarding https://blueprints.launchpad.net/nova/+spec/isolate-scheduler-db | 19:20 |
| *** achanda has quit IRC | 19:21 | |
| ociuhandu | mriedem: for instance the latest issue we had (which just got fixed like 5 minutes ago) was with some pip packages (newer versions in oslo) that were failing due to the version not being available on the cache server and for some reason the upstream pip mirror it reached. This can’t be solved automatically by a script yet in the CI | 19:21 |
| bauzas | vilobhmm11: mmm ? | 19:21 |
| *** mylu has quit IRC | 19:21 | |
| vilobhmm11 | bauzas : the update_aggr and delete_aggr will make sure the cache is consistent right ? | 19:22 |
| *** burgerk_ has quit IRC | 19:22 | |
| bauzas | vilobhmm11: correct, you mean the RPC API ? | 19:22 |
| bauzas | if so, yes | 19:22 |
| *** penick has joined #openstack-nova | 19:23 | |
| *** liusheng has quit IRC | 19:23 | |
| *** liusheng has joined #openstack-nova | 19:23 | |
| *** signed8bit_ZZZzz is now known as signed8bit | 19:23 | |
| *** tjones has quit IRC | 19:24 | |
| vilobhmm11 | bauzas : yes the RPC api…are there any other interfaces through which the cache might get updated ? because if i understand correctly the data is fecthed from agg table and the map is contructed and worked upon by the scheduler so this blueprint has changes along API->scheduler client ->scheduler query -> then actually using the map in host manager | 19:24 |
| vilobhmm11 | bauzas : and nothing that gets updated from compute or resource tracker that the map needs to be aware of right ? | 19:25 |
| vilobhmm11 | unless i am missing something | 19:25 |
| *** achanda has joined #openstack-nova | 19:25 | |
| bauzas | vilobhmm11: the HostManager attribute for the list of Aggregates is set when the scheduler is started | 19:25 |
| bauzas | vilobhmm11: by looking at the list of aggs | 19:25 |
| vilobhmm11 | bauzas : agreed https://github.com/openstack/nova/blob/master/nova/scheduler/host_manager.py#L174 | 19:26 |
| bauzas | vilobhmm11: then, anytime an CRUD operation (well, CUD) for aggregates is done, then it calls the scheduler API to update that yes | 19:26 |
| *** tjones has joined #openstack-nova | 19:27 | |
| bauzas | vilobhmm11: rather https://github.com/openstack/nova/blob/master/nova/scheduler/host_manager.py#L337 | 19:27 |
| vilobhmm11 | bauzas : thats the place where map is constructed | 19:28 |
| bauzas | yup | 19:28 |
| vilobhmm11 | ok cool…just wanted to confirm that something like RT is not updating the cache | 19:29 |
| vilobhmm11 | bauzas : ^^ | 19:29 |
| mriedem | bauzas: i'm wondering if we can get the instance uuid from the request spec in the numa topology filter here https://review.openstack.org/#/c/202650/ but not really sure what pitfalls there are there | 19:29 |
| mriedem | ooo https://review.openstack.org/#/c/202677/ | 19:30 |
| *** omnipresent has joined #openstack-nova | 19:31 | |
| *** ijw has joined #openstack-nova | 19:31 | |
| bauzas | mriedem: yeah, I need to respin a bit... | 19:32 |
| *** ijw_ has joined #openstack-nova | 19:33 | |
| *** ijw has quit IRC | 19:33 | |
| bauzas | vilobhmm11: for the aggregates, nope | 19:34 |
| bauzas | vilobhmm11: but you should look at the instances ;) | 19:34 |
| *** zenoway has quit IRC | 19:34 | |
| *** paul-carlton1 has quit IRC | 19:34 | |
| vilobhmm11 | bauzas : can you point me to the code ? something in host manager that I should look at ? are you referring to https://github.com/openstack/nova/blob/master/nova/scheduler/host_manager.py#L614 ? | 19:35 |
| *** eglynn has quit IRC | 19:36 | |
| *** stackdump has quit IRC | 19:37 | |
| *** kentb1 has quit IRC | 19:37 | |
| *** pixelbeat has quit IRC | 19:40 | |
| *** salv-orlando has quit IRC | 19:41 | |
| *** pixelbeat has joined #openstack-nova | 19:42 | |
| openstackgerrit | melanie witt proposed openstack/python-novaclient: Fix repr of a host from a hosts.list() https://review.openstack.org/248300 | 19:42 |
| *** tpatil has joined #openstack-nova | 19:43 | |
| *** Marga_ has quit IRC | 19:45 | |
| *** claudiub has joined #openstack-nova | 19:45 | |
| openstackgerrit | Ed Leafe proposed openstack/nova: Add better help text to scheduler options https://review.openstack.org/247181 | 19:47 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Config options: centralize section "scheduler" https://review.openstack.org/245891 | 19:47 |
| sfinucan | etiquette question: Should we -1 on commit messages, or is that too nitpicky? I assume they're valuable and should make sense? | 19:48 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova: Add infra for versioned notifications https://review.openstack.org/247024 | 19:53 |
| *** dims has joined #openstack-nova | 19:53 | |
| *** omnipresent has quit IRC | 19:53 | |
| *** aginwala has quit IRC | 19:54 | |
| *** zenoway has joined #openstack-nova | 19:54 | |
| *** achanda has quit IRC | 19:54 | |
| *** dims_ has quit IRC | 19:56 | |
| *** achanda has joined #openstack-nova | 19:57 | |
| bauzas | sfinucan: in general, you should rather propose a new PS by just updating the commit msg, like the Gerrit UI helps | 19:58 |
| bauzas | sfinucan: or just leave a comment without -1ing | 19:58 |
| bauzas | but you can still -1, the vote doesn't really mean you hate the owner | 19:59 |
| edleafe | sfinucan: for spelling/grammar, don't -1. Fix it as bauzas suggested | 19:59 |
| sfinucan | bauzas: OK. I didn't want to do that because it changes the author and the assignee for any bugs in Launchpad | 19:59 |
| bauzas | sfinucan: leave a +0 then | 19:59 |
| *** aginwala has joined #openstack-nova | 20:00 | |
| edleafe | sfinucan: but if the commit message is missing something important (DocImpact, etc), or doesn't properly describe the change, -1 can be appropriate | 20:00 |
| bauzas | in general, people hate receiving -1s while I feel it's just a note | 20:00 |
| *** rfolco_ has quit IRC | 20:00 | |
| sfinucan | edleafe: So +0 for nitpick. -1 for important changes. Got it | 20:01 |
| bauzas | if that needs a respin, the note doesn't change that | 20:01 |
| sfinucan | bauzas: Yes, people can be protective about their code (and commit messages) :) | 20:01 |
| *** penick has quit IRC | 20:01 | |
| bauzas | sfinucan: meh, johnthetubaguy made a very nice session in Tokyo explaining why people should care more about problems and not implementations :) | 20:02 |
| rlrossit | for anyone that was interested in the notifications ML thread, we have the notifications subteam meeting going on in #openstack-meeting-alt | 20:02 |
| *** sripriya has joined #openstack-nova | 20:02 | |
| sfinucan | bauzas: Oh really? :) I'd love to share that with some folks, should you ever stumble upon a video of it | 20:03 |
| bauzas | sfinucan: https://www.youtube.com/watch?v=ffPRfMUmsFM&feature=youtu.be | 20:03 |
| sfinucan | bauzas: Excellent. Time to go find my comfy spot to watch it. Thank you. | 20:04 |
| *** otter768 has joined #openstack-nova | 20:06 | |
| *** sacharya has quit IRC | 20:06 | |
| *** omnipresent has joined #openstack-nova | 20:08 | |
| openstackgerrit | Stephen Finucane proposed openstack/nova: tests_fields: Remove most enum-type field tests https://review.openstack.org/244370 | 20:10 |
| *** otter768 has quit IRC | 20:10 | |
| *** mylu has joined #openstack-nova | 20:10 | |
| *** zenoway has quit IRC | 20:11 | |
| *** changbl has quit IRC | 20:11 | |
| *** ociuhandu has quit IRC | 20:13 | |
| *** signed8bit is now known as signed8bit_ZZZzz | 20:14 | |
| mriedem | jroll: i've gone through https://review.openstack.org/#/c/237067/ again | 20:14 |
| mriedem | jroll: let me know if you want to chit chat in the next hour or so before i'm out for the week | 20:14 |
| *** mylu has quit IRC | 20:15 | |
| *** signed8bit_ZZZzz is now known as signed8bit | 20:16 | |
| mriedem | sfinucan: if the commit message just has typos, point them out and ask thye are fixed if they respin, but not a -1 | 20:17 |
| mriedem | sfinucan: if it's a complicated change and the commit message makes 0 sense or doesn't explain anything, that's a -1 | 20:18 |
| mriedem | if it's an upgrade impact and the UpgradeImpact tag isn't in there, that's a -1 | 20:18 |
| *** otter768 has joined #openstack-nova | 20:18 | |
| mriedem | and if they say 'the details are in the bug report, go read that', then you stop reviewing their code and move on | 20:19 |
| *** thangp has quit IRC | 20:19 | |
| sfinucan | mriedem: :D understood | 20:19 |
| openstackgerrit | jichenjc proposed openstack/nova: Change some wording on server_concepts.rst https://review.openstack.org/248063 | 20:20 |
| mriedem | whatever you do, don't ask this question in the -dev ML | 20:20 |
| *** apoorvad has quit IRC | 20:21 | |
| edleafe | mriedem: heh | 20:22 |
| *** zenoway has joined #openstack-nova | 20:23 | |
| *** thangp has joined #openstack-nova | 20:24 | |
| tpatil | mriedem: Hi | 20:25 |
| *** omnipresent has quit IRC | 20:26 | |
| *** pixelbeat has quit IRC | 20:26 | |
| tpatil | mriedem: Want to discuss with you about specs : https://review.openstack.org/#/c/135387 | 20:26 |
| *** stackdump has joined #openstack-nova | 20:27 | |
| *** pixelbeat has joined #openstack-nova | 20:27 | |
| tpatil | mriedem: I have replied to your comments on the above patch. Would appreciate if you can have a look at it and give your feedback | 20:28 |
| *** ccarmack1 has joined #openstack-nova | 20:28 | |
| *** ccarmack1 has left #openstack-nova | 20:29 | |
| *** changbl has joined #openstack-nova | 20:31 | |
| *** mylu has joined #openstack-nova | 20:31 | |
| openstackgerrit | jichenjc proposed openstack/nova: Use SpawnIsSynchronousFixture in most unit tests https://review.openstack.org/230953 | 20:32 |
| openstackgerrit | jichenjc proposed openstack/nova: Replace fake_utils by using Fixture https://review.openstack.org/230954 | 20:32 |
| *** edtubill has joined #openstack-nova | 20:34 | |
| rlrossit | jaypipes: if you have a chance (you liked the last one in this group): https://review.openstack.org/#/c/245336/ :) | 20:34 |
| mriedem | tpatil: ok....but i'll be honest that a lot of that spec is lost on me | 20:34 |
| *** pixelbeat has quit IRC | 20:35 | |
| *** dane-fichter has joined #openstack-nova | 20:35 | |
| *** ociuhandu has joined #openstack-nova | 20:35 | |
| tpatil | mriedem: We have requested Andrew Laski to review it since he is the one who knows it from the beginning | 20:36 |
| tpatil | mriedem: We will add this topic in stuck review in the next weekly meeting to discuss it further | 20:36 |
| mriedem | tpatil: i don't think adding it to the nova meeting every week is going to help | 20:37 |
| *** omnipresent has joined #openstack-nova | 20:37 | |
| mriedem | and alaski is out of the office i believe | 20:37 |
| mriedem | i saw it was posted to the ops list, but no replies | 20:37 |
| tpatil | mriedem: then, what do you suggest how we should proceed on this. I'm worried if we can make it before freeze date 3rd of Dec. | 20:37 |
| tpatil | mriedem: Is it possible to get approval on non-priority specs post 3rd of Dec for the Mitaka release? | 20:38 |
| *** e0ne has quit IRC | 20:38 | |
| *** zenoway has quit IRC | 20:40 | |
| *** penick has joined #openstack-nova | 20:40 | |
| *** zenoway has joined #openstack-nova | 20:40 | |
| *** yamahata has quit IRC | 20:42 | |
| *** omnipresent has quit IRC | 20:43 | |
| *** achanda has quit IRC | 20:43 | |
| mriedem | tpatil: no | 20:43 |
| *** sacharya has joined #openstack-nova | 20:44 | |
| *** aginwala has quit IRC | 20:44 | |
| *** mylu has quit IRC | 20:44 | |
| tpatil | mriedem: Ok, so we have just handful of days remaining to get the approval. We will followup with John and Andrew to see if they agree on it. Thanks for your time. | 20:45 |
| *** cfriesen__ has quit IRC | 20:46 | |
| *** ctrath has quit IRC | 20:47 | |
| *** ctrath has joined #openstack-nova | 20:50 | |
| *** markmc has joined #openstack-nova | 20:51 | |
| *** mylu has joined #openstack-nova | 20:51 | |
| *** aysyd has quit IRC | 20:52 | |
| jroll | mriedem: replied on https://review.openstack.org/#/c/237067/ | 20:53 |
| jroll | mriedem: happy to chat if that doesn't all make sense, otherwise I can get a new version up if you want to +2 before you disappear | 20:53 |
| mriedem | tpatil: i don't think i can get over why we need a new config option for this, | 20:54 |
| mriedem | tpatil: isn't the point that if i have shared storage, then i don't want to offload the disk? | 20:55 |
| *** stackdump has quit IRC | 20:55 | |
| mriedem | shouldn't that just be automatic w/o the config option telling me? | 20:55 |
| mriedem | tpatil: or is that b/c nova doesn't know how you've configured glance? | 20:56 |
| *** stackdump has joined #openstack-nova | 20:56 | |
| tpatil | mriedem: this new config option was suggested by Andrew | 20:56 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Add code-review devref for release notes https://review.openstack.org/247775 | 20:59 |
| bauzas | mriedem: Vek had some good comments for my bad English grammar so... ^ | 20:59 |
| bauzas | mriedem: in case you're not yet seated and waiting for the turkey | 21:00 |
| *** salv-orlando has joined #openstack-nova | 21:00 | |
| dansmith | bauzas: turkey on thursday | 21:01 |
| mriedem | bauzas: just reading through this unshelve api spec again | 21:01 |
| *** salv-orl_ has joined #openstack-nova | 21:01 | |
| *** raildo is now known as raildo-afk | 21:01 | |
| bauzas | dansmith: meh, I'm sad, I was expecting US folks to eat turkey for the whole week | 21:01 |
| mriedem | bauzas: that's the week after | 21:01 |
| mriedem | leftovers | 21:01 |
| bauzas | mriedem: a-ha no worries | 21:01 |
| tpatil | mriedem: even if compute nodes are aware of glance configuration it will take lot of time for downloading/copying snapshot during unshelving process. If you get instance disk files directly from the instance path it will be fastest | 21:01 |
| *** baoli has quit IRC | 21:02 | |
| *** sripriya_ has joined #openstack-nova | 21:03 | |
| *** sripriya has quit IRC | 21:04 | |
| *** salv-orlando has quit IRC | 21:05 | |
| dansmith | heh | 21:05 |
| *** EinstCrazy has joined #openstack-nova | 21:06 | |
| tpatil | mriedem: I'm checking Andrew's comments to point out why exactly he recommended to add a new config option. | 21:07 |
| *** pixelbeat has joined #openstack-nova | 21:08 | |
| *** otter768 has quit IRC | 21:08 | |
| *** mylu has quit IRC | 21:09 | |
| *** omnipresent has joined #openstack-nova | 21:09 | |
| mriedem | tpatil: more commends inline | 21:10 |
| tpatil | mriedem: https://review.openstack.org/#/c/135387/12/specs/liberty/approved/improve-unshelve-performance.rst | 21:10 |
| tpatil | mriedem: let me check | 21:10 |
| mriedem | tpatil: andrew said basically 'i don't like using -2 as a flag, if we're going to flag like this, i'd rather use a boolean flag' | 21:11 |
| mriedem | tpatil: that's not the same as saying, 'we should really have a config option for this' | 21:11 |
| mriedem | i would need to see something in the spec that explicitly says in what case i'd be using shared storage and would leave shelved_offload_time=True | 21:11 |
| mriedem | b/c if we have to document that if you're using shared storage, set that to False, we should do it automatically (with no config option) | 21:12 |
| *** zenoway has quit IRC | 21:12 | |
| *** EinstCrazy has quit IRC | 21:12 | |
| tpatil | mriedem: during shelving process, we can determine whether the instance disk files are stored on shared storage or not. so we need to introduce this new config option. | 21:13 |
| tpatil | mriedem: we cannot determine | 21:13 |
| tpatil | mriedem: during shelving process, we cannot determine whether the instance disk files are stored on shared storage or not. so we need to introduce this new config option. | 21:13 |
| *** penick has quit IRC | 21:14 | |
| mriedem | tpatil: huh? the spec says, and i quote, | 21:14 |
| mriedem | "then call Virt driver instance_on_disk method to check whether instance files exists in the instance path. If yes, that means the instance path is mounted on the shared storage." | 21:14 |
| mriedem | it says that for unshelve, but what's the difference? | 21:14 |
| openstackgerrit | Clinton Knight proposed openstack/nova-specs: Manila support in metadata service https://review.openstack.org/248301 | 21:15 |
| *** zenoway has joined #openstack-nova | 21:16 | |
| mriedem | tpatil: also, evacuate is also being updated to detect shared storage and do the right thing, so how would that work if we didn't know we were on shared storage | 21:16 |
| mriedem | migrate also checks that | 21:16 |
| *** RuiChen has quit IRC | 21:16 | |
| *** RuiChen has joined #openstack-nova | 21:16 | |
| *** dane-fichter has quit IRC | 21:16 | |
| tpatil | mriedem: Unshelve instance is the next step and it's a separate API call | 21:17 |
| *** ctrath has quit IRC | 21:17 | |
| * mriedem finds gibi's bp for evacuate + shared storage | 21:17 | |
| mriedem | tpatil: what does that matter? | 21:17 |
| mriedem | if you can tell if you're on shared storage with unshelve, why couldn't you in the shelve operation? | 21:17 |
| openstackgerrit | Dan Smith proposed openstack/nova: Actually pass the migration data object down to the virt drivers https://review.openstack.org/248211 | 21:17 |
| dansmith | mriedem: well, | 21:18 |
| dansmith | mriedem: once you're unshelved (offloaded) you don't have a compute node | 21:18 |
| tpatil | mriedem: When you shelve instance, how will the compute node know if the instance disk files are mounted on the shared storage or not. Let's say it's not, then we have to take the snapshot so that user can unshelve instance using snapshot | 21:19 |
| dansmith | which means you don't have config | 21:19 |
| dansmith | but that means we need to stash it with the instance, | 21:19 |
| *** aginwala has joined #openstack-nova | 21:19 | |
| dansmith | not have a config | 21:19 |
| dansmith | but this also means there is probably something we need to do to make sure you land on a compute node with the same storage config | 21:19 |
| dansmith | when you get rescheduled | 21:19 |
| dansmith | like, if you have some computes that use ceph storage and some that use local | 21:19 |
| *** omnipresent has quit IRC | 21:19 | |
| openstackgerrit | Pushkar Umaranikar proposed openstack/nova: Modify VM's updated_at field on volume actions https://review.openstack.org/247176 | 21:20 |
| dansmith | live migration handles that by passing crazy dicts of crap around to make that call | 21:20 |
| tpatil | mriedem:If yes, then no need to take snapshot as we can assume when the newly selected compute node will contain instance disk files in the cache (shared storage) | 21:20 |
| tpatil | mriedem:If yes, then no need to take snapshot during shelve api as we can assume when the newly selected compute node will contain instance disk files in the cache (shared storage) when an instance will be unshelved | 21:21 |
| dansmith | mriedem: so, I could be remembering wrong, but I thought at the summit the room kinda felt like this was not a worthwhile thing to do | 21:21 |
| dansmith | mriedem: did you read the minutes? I'd have to go back and re-read to see if it was captured | 21:21 |
| mriedem | dansmith: nope | 21:21 |
| *** yonglihe has quit IRC | 21:21 | |
| dansmith | but that basically shelving makes less sense in a shared storage world, IIRC | 21:22 |
| *** akshai has quit IRC | 21:22 | |
| *** omnipresent has joined #openstack-nova | 21:22 | |
| *** ctrath has joined #openstack-nova | 21:23 | |
| *** penick has joined #openstack-nova | 21:23 | |
| *** akshai has joined #openstack-nova | 21:23 | |
| mriedem | tpatil: so i guess the config option is to handle the shared storage vs non-shared storage computes? | 21:23 |
| tpatil | mriedem: yes | 21:23 |
| bauzas | meh, just catching the convo | 21:24 |
| mriedem | dansmith: but if people shelve to keep their uuid and IP, then the storage backend doesn't really matter, right? | 21:24 |
| tpatil | mriedem: we are going to use host aggregate group. Each host aggregate group compute nodes will be use it's own shared storage server | 21:24 |
| bauzas | tpatil: is it you who just wrote an ShelveFilter? | 21:24 |
| tpatil | bauzas: another team member from NTT group | 21:25 |
| *** sripriya_ has quit IRC | 21:25 | |
| bauzas | okay, I need to scroll back then | 21:25 |
| dansmith | mriedem: yeah, it's just less of a deal because they're still paying for live storage, instead of the shelve case where they're only paying for glance storage | 21:25 |
| *** sripriya has joined #openstack-nova | 21:25 | |
| bauzas | https://review.openstack.org/#/c/234013/ FWIW | 21:25 |
| *** zenoway has quit IRC | 21:26 | |
| *** lpetrut has joined #openstack-nova | 21:26 | |
| dansmith | mriedem: I mean, in the typical shelve case you go down to just paying for a snapshot image and some overhead for the uuid | 21:26 |
| bauzas | to be clear, I feel those 2 specs should somehow be discussed at the same time, since I'm pretty -1 for the latter (scheduler filter) | 21:26 |
| *** zenoway has joined #openstack-nova | 21:26 | |
| tpatil | mriedem : if shelved_offload_disk parameter is configured correctly, then you can guarantee instance disk files will be available on the newly select compute node in the unshelve call | 21:27 |
| dansmith | mriedem: but in the case of this shared storage shelve, you still want to charge them for the ceph storage | 21:27 |
| *** suro-patz has quit IRC | 21:27 | |
| mriedem | tpatil: so rather than the config option, can we, like dan said, stash that info off in the instance system_metadata and then when unshelving, we make sure it lands on a host that satisfies the storage requirement? | 21:28 |
| mriedem | maybe that's where the filter comes in | 21:28 |
| bauzas | mriedem: yup, that's why they wrote that | 21:28 |
| *** suro-patz has joined #openstack-nova | 21:28 | |
| mriedem | i'd much rather store info in the db and re-use that later when scheduling than force config option deployment everywhere | 21:29 |
| bauzas | mriedem: but I'm pretty against having a very specific filter for just one usecase | 21:29 |
| bauzas | mriedem: +1 | 21:29 |
| *** otter768 has joined #openstack-nova | 21:29 | |
| mriedem | bauzas: why? | 21:29 |
| mriedem | i mean, why against the filter | 21:29 |
| bauzas | I fear a big-bang here | 21:30 |
| *** omnipresent has quit IRC | 21:30 | |
| bauzas | while it's pretty easy to have custom filters out-of-tree | 21:30 |
| bauzas | for specific deployments | 21:30 |
| mriedem | bauzas: yeah, but then if we're just adding this new system_metadata to the instance in the shelve case for shared storage, | 21:30 |
| mriedem | and the only consumer of that data is in a scheduler filter, | 21:31 |
| mriedem | then i think that scheduler filter should be in tree | 21:31 |
| tpatil | mriedem: If instance disk files are not present on the new compute node, then the instance will be set to error status or may be we can start the instance using snapshot | 21:31 |
| bauzas | mriedem: some existing filters can have that logic | 21:31 |
| bauzas | mriedem: I'm not against adding some logic, just not adding a new filter for that specific stuff | 21:31 |
| bauzas | mriedem: we actually already filter on flavors and aggregates | 21:32 |
| tpatil | mriedem: the new config option is introduced so that cleanup of instance disk files from shelved_host and new compute node is easier | 21:32 |
| bauzas | hence me wondering why all of that is really needed | 21:32 |
| tpatil | mriedem:when you unshelve instance, if the shelve_offload_disk=False is set on the new compute node, then we will check if instance disk files are present or not. If yes, no problem. If not, set the instance status to "error". | 21:34 |
| *** tdurakov has quit IRC | 21:35 | |
| *** omnipresent has joined #openstack-nova | 21:35 | |
| mriedem | tpatil: but can't we just call the virt driver's instance_on_disk method? | 21:35 |
| mriedem | to check for shared storage during unshelve? | 21:35 |
| mriedem | i think the spec even says it would do that | 21:35 |
| mriedem | this really just seems like: | 21:36 |
| mriedem | 1. shelve offload - if on shared storage, don't offload disk files | 21:36 |
| dansmith | tpatil: that's not how unshelve is supposed to work though | 21:36 |
| mriedem | 2. unshelve - if on shared storage, re-use disk files, else use snapshot | 21:36 |
| dansmith | tpatil: unshelve is supposed to work regardless of those details, because it's just redeploying a snapshot from glance | 21:36 |
| mriedem | right | 21:36 |
| mriedem | if i shelve an instance and unshelve fails b/c it landed on the wrong node or the operator didn't have the config option set properly everywhere, i'd be pissed | 21:37 |
| mriedem | especially b/c i'm shelving and unshelving my pet | 21:37 |
| mriedem | and i need madame sparkles to come back to me | 21:37 |
| tpatil | mriedem: 1. shelve_offload, how to know whether instance disk files are on shared storage during shelving process? | 21:37 |
| *** Marga_ has joined #openstack-nova | 21:37 | |
| *** dims has quit IRC | 21:37 | |
| mriedem | virt_driver.instance_on_disk | 21:38 |
| mriedem | same as unshelve | 21:38 |
| *** zenoway has quit IRC | 21:38 | |
| mriedem | this all really seems like a nice to have optimization | 21:38 |
| *** dims has joined #openstack-nova | 21:38 | |
| mriedem | but fallback to spawn from snapshot if the shared storage files aren't ofund | 21:39 |
| mriedem | *found | 21:39 |
| dansmith | well, that defeats the whole thing | 21:39 |
| mriedem | and i can really only see landing the unshelve on the right host with a filter change | 21:39 |
| dansmith | because there is no real snapshot in this shared case right? | 21:39 |
| bauzas | that's a sharding problem | 21:39 |
| bauzas | which is a capacity problem | 21:39 |
| bauzas | that's not related to the shelve operation by itself | 21:40 |
| tpatil | mriedem: instance_on_disk needs two host to configure whether they share same storage or not. | 21:40 |
| *** gjayavelu has joined #openstack-nova | 21:40 | |
| mriedem | dansmith: there isn't a snapshot image in glance? | 21:40 |
| dansmith | mriedem: not in this shared case I don't think so | 21:40 |
| tpatil | mriedem: during shelving process, we don't know whether the instance will be unshelved | 21:40 |
| *** pratikmallya has quit IRC | 21:40 | |
| dansmith | mriedem: that's the point of this work, right? to avoid the snapshot if you're on shared storage? | 21:40 |
| bauzas | mriedem: if that's pointing out a right compute having the same shared storage, that's a deployment issue | 21:41 |
| bauzas | mriedem: you could just make use of AZs like I mentioned | 21:41 |
| tpatil | dansmith: Correct, if instance disk files are retained, then no need to take snapshot | 21:41 |
| *** sacharya has quit IRC | 21:41 | |
| dansmith | mriedem: see, this is really taking shelve and making it something different | 21:41 |
| dansmith | mriedem: which is kindof the problem, IMHO | 21:41 |
| tpatil | dansmith: but from user's point of view, it's an API change . So we are just wondering how to solve this problem | 21:42 |
| *** Marga_ has quit IRC | 21:42 | |
| dansmith | tpatil: right exactly | 21:42 |
| mriedem | dansmith: if we don't take the snapshot, then yeah it's an api change | 21:42 |
| mriedem | as was pointed out in the spec at some point | 21:42 |
| tpatil | dansmith: mriedem, recommended to return actual snapshot id or None in the shelve api response. | 21:42 |
| *** pixelbeat has quit IRC | 21:42 | |
| doffm | Why wouldn't we take the snapshot? Is that the thing that isn't performing. I thought it was the *unshelve* that has performance issues? | 21:43 |
| *** sacharya has joined #openstack-nova | 21:43 | |
| dansmith | doffm: if you never take the snapshot then unshelve becomes SUPER fast :P | 21:43 |
| doffm | dansmith: No I mean take the snapshot anyway, even if you end up using the old image. | 21:44 |
| doffm | That way if it starts up on another host you use the snapshot that was created. | 21:44 |
| doffm | And delete the old image. | 21:44 |
| dansmith | doffm: but the whole point is to never start from the image | 21:44 |
| dansmith | doffm: these are pets, remember | 21:44 |
| dansmith | doffm: restarting from the image is like a clone.. it's like the original instance, but it's not the actual instance | 21:44 |
| dansmith | doffm: it would be like saying "is it cool if we sometimes don't boot from your volume, but instead clone it and fork and make you not a boot-from-volume instance?" | 21:45 |
| mriedem | i guess i was thinking you use a snapshot as a fallback in case you land on the wrong node | 21:45 |
| dansmith | no | 21:45 |
| dansmith | that's not the thing | 21:45 |
| dansmith | that's like saying you have a boot-from-volume instance, and we'll just snapshot your volume and make you a regular instance if we can't do that bfv thing you asked for | 21:45 |
| dansmith | that would be a real problem :) | 21:45 |
| mriedem | ok | 21:46 |
| doffm | dansmith: Ok. Understand. Only boot from node where image is available. | 21:46 |
| *** pratikmallya has joined #openstack-nova | 21:46 | |
| mriedem | so, don't snapshot, hope you land on the right node that uses the same shared storage, and if not, error | 21:46 |
| dansmith | doffm: right, and that's a thing, but it's different from what unshelve is | 21:46 |
| openstackgerrit | Tracy Jones proposed openstack/nova: Cleanup of compute api reboot method https://review.openstack.org/127427 | 21:46 |
| *** jaypipes has quit IRC | 21:47 | |
| mriedem | and part of hoping you land on the right node is the scheduler filter, which bauzas is -1 on | 21:47 |
| dansmith | it's a little better if you filter in the scheduler first, but it's still not the same thing, IMHO | 21:47 |
| mriedem | and having hte config options correct everywhere | 21:47 |
| tpatil | mriedem: It's certainly possible to unshelve instance from snapshot if instance files are not present on the shared storage. But in this case we don't know whether instance disk files are still there on the shelved_host, If yes, how to cleaned them up. | 21:47 |
| bauzas | mriedem: the problem is because some nodes are not using the same shared storage, right? | 21:47 |
| *** sacharya has quit IRC | 21:47 | |
| mriedem | bauzas: yeah i guess | 21:48 |
| mriedem | some could be using ceph, some using lvm | 21:48 |
| dansmith | I think it's worse than that | 21:48 |
| doffm | Can't this be done with host aggregates? | 21:48 |
| bauzas | mriedem: because I guess they treat some computes with various levels of support | 21:48 |
| dansmith | I think tpatil wants to make it fail if those files are not already cached | 21:48 |
| doffm | Ensure that particular aggregates have shared storage. | 21:48 |
| *** otter768 has quit IRC | 21:48 | |
| bauzas | doffm: that's my point, it just sounds a deployment issue | 21:48 |
| *** boris-42 has quit IRC | 21:48 | |
| doffm | bauzas: Yeah. If deployed properly aggregates work. | 21:49 |
| mriedem | dansmith: yeah, so NoValidHost | 21:49 |
| tpatil | bauzas,mriedem: we want provider to configure shelve_offload_disk new config parameter only in case whether all compute nodes are using shared storage or each host aggregate group compute node has it's own shared storage server | 21:49 |
| dansmith | mriedem: what I mean is, not just ceph vs. lvm, but "ceph and already has that image cached" | 21:49 |
| *** omnipresent has quit IRC | 21:50 | |
| mriedem | right | 21:50 |
| mriedem | yuck | 21:50 |
| mriedem | so.... | 21:50 |
| dansmith | tpatil: so you know that we're getting ready to have coordinated glance-and-ceph snapshots, right? | 21:50 |
| dansmith | which makes a snapshot of a ceph instance essentially take zero time | 21:50 |
| dansmith | because we don't have to upload it | 21:50 |
| mriedem | so as an operator, i need to align the planets to make bursting of unshelve API calls faster | 21:50 |
| dansmith | we just tell glance that the snapshot is "this thing in ceph" | 21:50 |
| *** yamahata has joined #openstack-nova | 21:50 | |
| dansmith | tpatil: seems like regular shelve then becomes very fast and not required to have these extra cases | 21:51 |
| mriedem | dansmith: they want to use NFS | 21:51 |
| mriedem | not ceph | 21:51 |
| openstackgerrit | Tracy Jones proposed openstack/nova: config options: centralize section "guestfs" https://review.openstack.org/248963 | 21:51 |
| dansmith | mriedem: then I *REALLY* don't care :P | 21:51 |
| mriedem | ha | 21:51 |
| dansmith | that's the way to solve this, IMHO | 21:51 |
| *** zenoway has joined #openstack-nova | 21:52 | |
| *** ijw_ has quit IRC | 21:52 | |
| tpatil | dansmith: I understand ceph is another option, but right now, we want to use NFS | 21:52 |
| dansmith | tpatil: right, but ceph is built for doing this, NFS is not, | 21:54 |
| dansmith | and asking us to take on this really weird feature special case thing is not cool | 21:54 |
| * dansmith runs off for a bit | 21:55 | |
| *** Marga_ has joined #openstack-nova | 21:55 | |
| tpatil | dansmith: I understand your point and also know why it's taking more time to get the approval on this spec. If there are any issues in spec we are ready to solve them | 21:56 |
| *** Marga_ has quit IRC | 21:57 | |
| tpatil | mriedem: What do you suggest about new config option after all this discussion? IMO, if service provider configures it correctly, then lot of things will be simpler especially cleaning up instance disk files on the shelved_host | 21:57 |
| *** Marga_ has joined #openstack-nova | 21:58 | |
| *** aginwala has quit IRC | 21:58 | |
| *** suro-patz has quit IRC | 21:59 | |
| *** akshai has quit IRC | 21:59 | |
| *** suro-patz has joined #openstack-nova | 22:00 | |
| mriedem | i don't think i even undersatnd the cleaning up part | 22:01 |
| *** suro-patz has quit IRC | 22:01 | |
| *** akshai has joined #openstack-nova | 22:01 | |
| *** aginwala has joined #openstack-nova | 22:02 | |
| *** suro-patz has joined #openstack-nova | 22:02 | |
| mriedem | so when deleting a shelved_offloaded instance, we use shelved_host to go back to the original host and cleanup files | 22:02 |
| *** ijw has joined #openstack-nova | 22:02 | |
| *** gjayavelu has quit IRC | 22:02 | |
| mriedem | what does the config option do in this case? if we're on the shelved_host, we check shelved_offload_disk and if it's True, we cleanup local disk files, else we don't | 22:03 |
| mriedem | again, it seems like a thing you could figure out on the original host if it's using shared storage or not | 22:03 |
| *** gjayavelu has joined #openstack-nova | 22:03 | |
| *** aginwala has quit IRC | 22:03 | |
| *** MVenesio has quit IRC | 22:04 | |
| mriedem | honestly, i've tried to wrap my head around this and we've talked about it for an hour and a half, and i'm kind of done reviewing/thinking about this corner case technical debt thing that a single provider wants to add to nova | 22:04 |
| mriedem | there seems to be 0 interest from other operators on this | 22:04 |
| *** aginwala has joined #openstack-nova | 22:05 | |
| *** zenoway has quit IRC | 22:05 | |
| *** nic has quit IRC | 22:06 | |
| tpatil | mriedem: We will again consider ceph as a option and see if my customer is ok with it.Thanks for your time. | 22:06 |
| *** zenoway has joined #openstack-nova | 22:07 | |
| mriedem | tpatil: do you guys by chance have some proprietary NFS backend or something? | 22:07 |
| *** lpetrut has quit IRC | 22:07 | |
| *** apoorvad has joined #openstack-nova | 22:08 | |
| tpatil | mriedem: Evaluating some commercial NFS storage providers to decide on one | 22:11 |
| *** rlrossit has left #openstack-nova | 22:12 | |
| *** smurke has joined #openstack-nova | 22:12 | |
| *** baoli has joined #openstack-nova | 22:12 | |
| *** edtubill has quit IRC | 22:13 | |
| mriedem | hmm | 22:13 |
| mriedem | just tell your senior architect to drop that idea and use ceph :) | 22:13 |
| mriedem | i'm sure red hat wouldn't mind selling you a ceph solution | 22:14 |
| *** jerrygb has quit IRC | 22:14 | |
| mriedem | jroll: i think i'm ok on https://review.openstack.org/#/c/237067/ if you clean it up quick | 22:14 |
| *** pratikmallya has quit IRC | 22:14 | |
| *** paul-carlton has left #openstack-nova | 22:15 | |
| *** jerrygb has joined #openstack-nova | 22:15 | |
| *** Piet has quit IRC | 22:16 | |
| *** lucas-dinner has quit IRC | 22:18 | |
| *** ZZelle_ has joined #openstack-nova | 22:18 | |
| tpatil | mriedem: Sure, we will think of some solution that would benefit both sides | 22:18 |
| *** gjayavelu has quit IRC | 22:19 | |
| *** mylu has joined #openstack-nova | 22:19 | |
| *** baoli has quit IRC | 22:19 | |
| *** signed8b_ has joined #openstack-nova | 22:20 | |
| *** signed8bit has quit IRC | 22:20 | |
| *** thorst_ has joined #openstack-nova | 22:21 | |
| *** zaina has joined #openstack-nova | 22:21 | |
| *** kfarr has quit IRC | 22:22 | |
| *** thorst__ has joined #openstack-nova | 22:22 | |
| *** mylu has quit IRC | 22:23 | |
| *** vladikr has quit IRC | 22:25 | |
| *** omnipresent has joined #openstack-nova | 22:25 | |
| tjones | melwitt: sdague: dims: i have updated https://review.openstack.org/#/c/127427/ to address melwitt's comments. thanks for reviews if you have time | 22:25 |
| *** thorst_ has quit IRC | 22:25 | |
| *** lucasagomes has joined #openstack-nova | 22:26 | |
| smurke | https://bugs.launchpad.net/nova/+bug/1514550 I am working on this bug and have difficulty in converting dict to instance obj in messaging.py | 22:26 |
| openstack | Launchpad bug 1514550 in OpenStack Compute (nova) "nova.cells.messaging.instance_update_at_top is assuming it gets an Instance object" [Low,Triaged] - Assigned to Sumant Murke (sumant-murke) | 22:26 |
| *** thorst__ has quit IRC | 22:26 | |
| smurke | instance = objects_base.NovaObject.obj_from_primitive(instance) | 22:27 |
| *** thorst_ has joined #openstack-nova | 22:27 | |
| smurke | this is not working where instance passed to the function is a dictionary | 22:27 |
| *** rotbart has quit IRC | 22:27 | |
| *** mylu has joined #openstack-nova | 22:28 | |
| smurke | I even tried to create instance by instance = objects.Instance(**instance) | 22:28 |
| smurke | ...but this also dosent work | 22:28 |
| smurke | i am curious to know how to convert dictionary to object | 22:28 |
| *** zenoway has quit IRC | 22:28 | |
| *** suro-patz has quit IRC | 22:30 | |
| *** aginwala has quit IRC | 22:30 | |
| bauzas | smurke: so you want to hydrate an Instance object from a dict, but do you know how the dict was set? | 22:30 |
| *** mylu has quit IRC | 22:30 | |
| *** aginwala has joined #openstack-nova | 22:30 | |
| *** thorst_ has quit IRC | 22:31 | |
| bauzas | smurke: oh sec, saw the bug | 22:31 |
| *** achanda has joined #openstack-nova | 22:31 | |
| *** thangp has quit IRC | 22:31 | |
| smurke | @bauzas: i am setting it manually for testing purpose {'_orig_system_metadata': {'key1': u'value1', 'key2': u'value2'}, '_obj_info_cache': InstanceInfoCache(created_at=<?>,deleted=<?>,deleted_at=<?>,instance_uuid=fake-uuid,network_info=<?>,updated_at=<?>), '_orig_metadata': {}, '_context': None, '_changed_fields': set(['cell_name', 'system_metadata', 'uuid', 'info_cache']), '_obj_uuid': 'b48316c5-71e8-45e4-9884-6c78055b9b13', | 22:31 |
| smurke | '_obj_system_metadata': {'key1': u'value1', 'key2': u'value2'}, '_obj_cell_name': u'fake'} | 22:31 |
| *** unicell1 has joined #openstack-nova | 22:32 | |
| *** signed8b_ has quit IRC | 22:32 | |
| *** achanda has quit IRC | 22:33 | |
| *** mylu has joined #openstack-nova | 22:34 | |
| *** unicell has quit IRC | 22:34 | |
| *** mylu has quit IRC | 22:36 | |
| smurke | bauzas: Do I have to use NovaObjectSerializer class to hydrate from dict to object ...and not oslo_versionedobjects method obj_from_primitive ? | 22:36 |
| bauzas | smurke: mmm, the main problem is that base.obj_to_primitive is not idempotent | 22:37 |
| bauzas | smurke: it provides a dict from the fields, but it doesn't create a real primitive | 22:38 |
| *** signed8bit has joined #openstack-nova | 22:38 | |
| bauzas | like the one you could use as a DB primitive | 22:38 |
| *** mriedem has quit IRC | 22:38 | |
| *** ctrath has quit IRC | 22:38 | |
| smurke | but then do i have to create an obj_from_primitive method (i.e inherit from base class ) ? | 22:38 |
| smurke | bauzus: but then do i have to create an obj_from_primitive method (i.e inherit from base class ) ? | 22:39 |
| bauzas | smurke: unfortunately, I really need to go :( | 22:40 |
| smurke | ok np | 22:40 |
| openstackgerrit | jichenjc proposed openstack/python-novaclient: Not transform to str on potential unicode fields https://review.openstack.org/248626 | 22:42 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify Scheduler RPC API select_dests() https://review.openstack.org/199205 | 22:43 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify left filters for RequestSpec https://review.openstack.org/202678 | 22:43 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify conductor to use RequestSpec object https://review.openstack.org/203078 | 22:43 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify metric-related filters for RequestSpec https://review.openstack.org/202676 | 22:43 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Use RequestSpec in the ChanceScheduler https://review.openstack.org/199051 | 22:43 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify NUMA, PCI and num_instances filters for RequestSpec https://review.openstack.org/202677 | 22:43 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify Aggregate filters for RequestSpec https://review.openstack.org/202674 | 22:43 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Modify default filters for RequestSpec https://review.openstack.org/202675 | 22:43 |
| *** aginwala has quit IRC | 22:43 | |
| *** zaina has quit IRC | 22:43 | |
| *** zaina has joined #openstack-nova | 22:45 | |
| *** zaina has quit IRC | 22:45 | |
| *** jichen has quit IRC | 22:45 | |
| *** mylu has joined #openstack-nova | 22:46 | |
| *** pratikmallya has joined #openstack-nova | 22:46 | |
| *** mc_nair has quit IRC | 22:46 | |
| *** mylu has quit IRC | 22:48 | |
| *** mylu has joined #openstack-nova | 22:48 | |
| *** unicell1 has quit IRC | 22:49 | |
| *** zenoway has joined #openstack-nova | 22:49 | |
| *** doug-fis_ has joined #openstack-nova | 22:49 | |
| *** unicell has joined #openstack-nova | 22:49 | |
| *** mylu_ has joined #openstack-nova | 22:50 | |
| *** suro-patz has joined #openstack-nova | 22:50 | |
| *** zenoway has quit IRC | 22:50 | |
| *** mc_nair has joined #openstack-nova | 22:51 | |
| *** omnipresent has quit IRC | 22:52 | |
| *** mylu has quit IRC | 22:52 | |
| *** doug-fish has quit IRC | 22:53 | |
| *** mc_nair has quit IRC | 22:53 | |
| *** doug-fis_ has quit IRC | 22:53 | |
| *** aginwala has joined #openstack-nova | 22:55 | |
| *** hemna is now known as hemnafk | 22:55 | |
| *** smurke has quit IRC | 22:57 | |
| *** wuhg has joined #openstack-nova | 22:57 | |
| *** Marga_ has quit IRC | 22:57 | |
| *** sacharya has joined #openstack-nova | 22:57 | |
| *** Marga_ has joined #openstack-nova | 22:59 | |
| *** mylu_ has quit IRC | 22:59 | |
| *** mylu has joined #openstack-nova | 23:00 | |
| *** sneti has quit IRC | 23:00 | |
| openstackgerrit | Tracy Jones proposed openstack/nova: config options: centralize section "guestfs" https://review.openstack.org/248963 | 23:02 |
| openstackgerrit | Jim Rollenhagen proposed openstack/nova-specs: Tenant networking support for Ironic driver. https://review.openstack.org/237067 | 23:04 |
| jroll | mrie<tab> noooo too late. | 23:04 |
| *** doug-fish has joined #openstack-nova | 23:04 | |
| openstackgerrit | Tracy Jones proposed openstack/nova: Cleanup of compute api reboot method https://review.openstack.org/127427 | 23:05 |
| *** burt has quit IRC | 23:07 | |
| *** otter768 has joined #openstack-nova | 23:08 | |
| *** doug-fish has quit IRC | 23:08 | |
| *** vilobhmm11 has quit IRC | 23:10 | |
| *** vilobhmm11 has joined #openstack-nova | 23:10 | |
| *** boris-42 has joined #openstack-nova | 23:10 | |
| *** baoli has joined #openstack-nova | 23:11 | |
| *** jwcroppe has quit IRC | 23:11 | |
| *** jwcroppe has joined #openstack-nova | 23:13 | |
| *** baoli has quit IRC | 23:13 | |
| *** baoli has joined #openstack-nova | 23:14 | |
| *** baoli has quit IRC | 23:16 | |
| *** jwcroppe has quit IRC | 23:17 | |
| *** ZZelle_ has quit IRC | 23:17 | |
| *** jwcroppe has joined #openstack-nova | 23:18 | |
| *** aginwala has quit IRC | 23:19 | |
| *** signed8bit is now known as signed8bit_ZZZzz | 23:20 | |
| *** jwcroppe has quit IRC | 23:22 | |
| *** pumaranikar has quit IRC | 23:24 | |
| *** apoorvad has quit IRC | 23:25 | |
| *** dims_ has joined #openstack-nova | 23:28 | |
| openstackgerrit | Ed Leafe proposed openstack/nova: Add better help text to scheduler options https://review.openstack.org/247181 | 23:28 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Config options: centralize section "scheduler" https://review.openstack.org/245891 | 23:28 |
| *** vilobhmm11 has quit IRC | 23:28 | |
| *** vilobhmm11 has joined #openstack-nova | 23:28 | |
| *** dims has quit IRC | 23:29 | |
| *** thorst has joined #openstack-nova | 23:29 | |
| *** willsama has quit IRC | 23:31 | |
| *** mylu_ has joined #openstack-nova | 23:33 | |
| *** claudiub has quit IRC | 23:33 | |
| *** ijw has quit IRC | 23:35 | |
| *** signed8bit_ZZZzz is now known as signed8bit | 23:36 | |
| *** mylu has quit IRC | 23:36 | |
| *** thorst has quit IRC | 23:36 | |
| *** pixelbeat has joined #openstack-nova | 23:38 | |
| *** akshai has quit IRC | 23:39 | |
| *** aginwala has joined #openstack-nova | 23:39 | |
| *** akshai has joined #openstack-nova | 23:40 | |
| *** rcernin has quit IRC | 23:42 | |
| *** suro-patz has quit IRC | 23:43 | |
| *** akshai_ has joined #openstack-nova | 23:45 | |
| *** akshai has quit IRC | 23:46 | |
| *** vilobhmm11 has quit IRC | 23:49 | |
| *** vilobhmm11 has joined #openstack-nova | 23:50 | |
| *** apoorvad has joined #openstack-nova | 23:51 | |
| *** ijw has joined #openstack-nova | 23:52 | |
| *** otter768 has quit IRC | 23:53 | |
| *** jwcroppe has joined #openstack-nova | 23:53 | |
| *** EinstCrazy has joined #openstack-nova | 23:56 | |
| *** mrkz has quit IRC | 23:57 | |
| *** signed8bit is now known as signed8bit_ZZZzz | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!