| *** mlavalle has quit IRC | 00:06 | |
| *** browne has quit IRC | 00:09 | |
| *** jerrygb has joined #openstack-nova | 00:17 | |
| *** jerrygb_ has joined #openstack-nova | 00:20 | |
| *** jerrygb has quit IRC | 00:23 | |
| *** ducttape_ has joined #openstack-nova | 00:28 | |
| *** lamt has joined #openstack-nova | 00:31 | |
| *** mtanino has quit IRC | 00:35 | |
| *** smatzek has joined #openstack-nova | 00:42 | |
| *** kiran-r has quit IRC | 00:43 | |
| *** smatzek_ has joined #openstack-nova | 00:44 | |
| *** baoli has joined #openstack-nova | 00:47 | |
| *** kiran-r has joined #openstack-nova | 00:48 | |
| *** smatzek has quit IRC | 00:48 | |
| *** baoli has quit IRC | 00:48 | |
| *** baoli has joined #openstack-nova | 00:50 | |
| *** baoli_ has joined #openstack-nova | 00:51 | |
| *** jerrygb_ has quit IRC | 00:52 | |
| *** baoli has quit IRC | 00:55 | |
| *** baoli_ has quit IRC | 00:56 | |
| *** tuanluong has joined #openstack-nova | 00:59 | |
| *** tbachman has quit IRC | 01:00 | |
| *** hoangcx has joined #openstack-nova | 01:01 | |
| *** tovin07 has joined #openstack-nova | 01:05 | |
| *** huanxie has joined #openstack-nova | 01:06 | |
| *** aignatov has quit IRC | 01:08 | |
| *** Zer0Byte__ has quit IRC | 01:10 | |
| *** ducttape_ has quit IRC | 01:10 | |
| *** aignatov has joined #openstack-nova | 01:11 | |
| *** rmcall has quit IRC | 01:11 | |
| *** ducttape_ has joined #openstack-nova | 01:16 | |
| *** mbirru has joined #openstack-nova | 01:17 | |
| *** ducttape_ has quit IRC | 01:19 | |
| *** erbadr has joined #openstack-nova | 01:23 | |
| *** ducttape_ has joined #openstack-nova | 01:23 | |
| *** Apoorva has quit IRC | 01:24 | |
| *** ducttape_ has quit IRC | 01:28 | |
| *** Daisy has joined #openstack-nova | 01:31 | |
| *** rmcall has joined #openstack-nova | 01:33 | |
| *** baoli has joined #openstack-nova | 01:33 | |
| *** rmcall has quit IRC | 01:33 | |
| *** tbachman has joined #openstack-nova | 01:34 | |
| *** rmcall has joined #openstack-nova | 01:34 | |
| *** maticue has quit IRC | 01:36 | |
| *** mtanino has joined #openstack-nova | 01:42 | |
| *** Apoorva has joined #openstack-nova | 01:43 | |
| *** masayukig_ has joined #openstack-nova | 01:45 | |
| *** stewie925 has quit IRC | 01:45 | |
| *** Apoorva has quit IRC | 01:47 | |
| *** Daisy has quit IRC | 01:49 | |
| *** masayukig has quit IRC | 01:50 | |
| *** masayukig_ is now known as masayukig | 01:52 | |
| *** zhangyufei has joined #openstack-nova | 01:54 | |
| *** smatzek_ has quit IRC | 01:54 | |
| *** Daisy has joined #openstack-nova | 01:55 | |
| *** Daisy has quit IRC | 01:55 | |
| *** Daisy has joined #openstack-nova | 01:55 | |
| *** thorst has quit IRC | 01:55 | |
| openstackgerrit | zhaolihui proposed openstack/nova: Fix failed to boot vm with pci_alias's name including space https://review.openstack.org/392077 | 01:55 |
|---|---|---|
| *** thorst has joined #openstack-nova | 01:56 | |
| *** baoli has quit IRC | 01:59 | |
| *** baoli has joined #openstack-nova | 01:59 | |
| *** mtanino has quit IRC | 02:01 | |
| *** baoli has quit IRC | 02:04 | |
| *** thorst has quit IRC | 02:04 | |
| *** jerrygb has joined #openstack-nova | 02:05 | |
| *** jvgrant has quit IRC | 02:14 | |
| *** jvgrant has joined #openstack-nova | 02:14 | |
| *** zhangyufei has quit IRC | 02:15 | |
| *** tbachman has quit IRC | 02:18 | |
| *** ducttape_ has joined #openstack-nova | 02:24 | |
| *** armax has quit IRC | 02:26 | |
| *** ducttape_ has quit IRC | 02:29 | |
| *** zhangyufei has joined #openstack-nova | 02:29 | |
| *** tlian has quit IRC | 02:37 | |
| *** sree has joined #openstack-nova | 02:40 | |
| *** tbachman has joined #openstack-nova | 02:42 | |
| *** sree has quit IRC | 02:45 | |
| *** tbachman_ has joined #openstack-nova | 02:45 | |
| *** tbachman has quit IRC | 02:47 | |
| *** tbachman_ is now known as tbachman | 02:47 | |
| *** ljxiash has joined #openstack-nova | 02:55 | |
| *** Daisy has quit IRC | 02:55 | |
| *** Daisy has joined #openstack-nova | 02:56 | |
| *** dave-mccowan has quit IRC | 02:57 | |
| *** FL1SK has joined #openstack-nova | 03:00 | |
| *** shaohe_feng has quit IRC | 03:00 | |
| *** thorst has joined #openstack-nova | 03:02 | |
| *** rmcall has quit IRC | 03:03 | |
| *** ssurana has joined #openstack-nova | 03:04 | |
| *** Jeffrey4l has joined #openstack-nova | 03:07 | |
| *** thorst has quit IRC | 03:10 | |
| *** ducttape_ has joined #openstack-nova | 03:20 | |
| openstackgerrit | prashkre proposed openstack/nova: Separate CRUD policy for server_groups https://review.openstack.org/391113 | 03:32 |
| *** mtanino has joined #openstack-nova | 03:33 | |
| *** Daisy has quit IRC | 03:38 | |
| *** ducnc has joined #openstack-nova | 03:39 | |
| *** markvoelker has quit IRC | 03:41 | |
| *** lifeless has quit IRC | 03:41 | |
| *** lifeless has joined #openstack-nova | 03:42 | |
| *** ducttape_ has quit IRC | 03:44 | |
| *** phuongnh has joined #openstack-nova | 03:44 | |
| *** Daisy has joined #openstack-nova | 03:44 | |
| *** lamt has quit IRC | 03:48 | |
| *** Daisy has quit IRC | 03:49 | |
| *** ljxiash has quit IRC | 03:57 | |
| *** ljxiash has joined #openstack-nova | 03:57 | |
| *** ljxiash has quit IRC | 04:02 | |
| *** coolsvap has joined #openstack-nova | 04:08 | |
| *** thorst has joined #openstack-nova | 04:09 | |
| *** Dinesh_Bhor has joined #openstack-nova | 04:12 | |
| *** thorst has quit IRC | 04:15 | |
| *** jerrygb has quit IRC | 04:19 | |
| *** jerrygb has joined #openstack-nova | 04:24 | |
| *** erbadr has quit IRC | 04:28 | |
| *** Daisy has joined #openstack-nova | 04:29 | |
| *** raunak has joined #openstack-nova | 04:40 | |
| *** markvoelker has joined #openstack-nova | 04:42 | |
| *** markvoelker has quit IRC | 04:48 | |
| *** mtanino has quit IRC | 04:48 | |
| *** ducttape_ has joined #openstack-nova | 04:54 | |
| *** hoangcx has quit IRC | 04:54 | |
| *** erbadr has joined #openstack-nova | 04:54 | |
| *** mriedem has quit IRC | 04:54 | |
| *** ducttape_ has quit IRC | 05:00 | |
| *** raunak has quit IRC | 05:00 | |
| *** jerrygb has quit IRC | 05:03 | |
| *** sree has joined #openstack-nova | 05:04 | |
| *** huanxie has quit IRC | 05:04 | |
| *** jerrygb has joined #openstack-nova | 05:06 | |
| *** sree has quit IRC | 05:09 | |
| *** ssurana has quit IRC | 05:11 | |
| *** thorst has joined #openstack-nova | 05:12 | |
| *** pc_m has quit IRC | 05:16 | |
| *** thorst has quit IRC | 05:20 | |
| *** prateek has joined #openstack-nova | 05:23 | |
| *** ljxiash has joined #openstack-nova | 05:24 | |
| *** sheel has joined #openstack-nova | 05:33 | |
| *** huanxie has joined #openstack-nova | 05:37 | |
| *** pc_m has joined #openstack-nova | 05:43 | |
| *** jerrygb has quit IRC | 05:53 | |
| *** jerrygb has joined #openstack-nova | 05:57 | |
| *** garyk1 has joined #openstack-nova | 05:58 | |
| *** garyk1 has quit IRC | 05:58 | |
| *** bkopilov has joined #openstack-nova | 06:04 | |
| *** hshiina has joined #openstack-nova | 06:04 | |
| *** garyk1 has joined #openstack-nova | 06:05 | |
| *** garyk1 has quit IRC | 06:07 | |
| *** garyk1 has joined #openstack-nova | 06:09 | |
| *** garyk1 has quit IRC | 06:11 | |
| *** garyk1 has joined #openstack-nova | 06:11 | |
| *** garyk1 has quit IRC | 06:11 | |
| *** erbadr has quit IRC | 06:19 | |
| *** shaohe_feng has joined #openstack-nova | 06:19 | |
| *** erbadr has joined #openstack-nova | 06:19 | |
| *** thorst has joined #openstack-nova | 06:19 | |
| *** garyk1 has joined #openstack-nova | 06:20 | |
| *** sridharg has joined #openstack-nova | 06:20 | |
| *** nkrinner_afk has quit IRC | 06:21 | |
| *** jerrygb has quit IRC | 06:22 | |
| *** nkrinner has joined #openstack-nova | 06:23 | |
| *** thorst has quit IRC | 06:25 | |
| *** jerrygb has joined #openstack-nova | 06:27 | |
| *** namnh has joined #openstack-nova | 06:27 | |
| openstackgerrit | huyupeng proposed openstack/nova: Fix type error https://review.openstack.org/392426 | 06:31 |
| *** yuntongjin has joined #openstack-nova | 06:34 | |
| *** prashkre has joined #openstack-nova | 06:38 | |
| *** moshele has joined #openstack-nova | 06:40 | |
| *** zhangyufei has quit IRC | 06:42 | |
| *** zhangyufei has joined #openstack-nova | 06:43 | |
| *** markvoelker has joined #openstack-nova | 06:44 | |
| *** pbandark has joined #openstack-nova | 06:44 | |
| *** moshele has quit IRC | 06:46 | |
| *** dikonoor has joined #openstack-nova | 06:47 | |
| *** dikonoor has quit IRC | 06:48 | |
| *** markvoelker has quit IRC | 06:50 | |
| *** rcernin has joined #openstack-nova | 06:50 | |
| openstackgerrit | huyupeng proposed openstack/nova: Fix typo error https://review.openstack.org/392426 | 06:50 |
| *** noama has quit IRC | 06:51 | |
| *** dikonoor has joined #openstack-nova | 06:53 | |
| *** noama has joined #openstack-nova | 06:53 | |
| *** noama_ has joined #openstack-nova | 06:53 | |
| *** noama has quit IRC | 06:54 | |
| *** noama_ has quit IRC | 06:54 | |
| *** noama has joined #openstack-nova | 06:54 | |
| *** noama_ has joined #openstack-nova | 06:54 | |
| *** noama_ has quit IRC | 06:54 | |
| *** noama has quit IRC | 06:55 | |
| *** trisq has joined #openstack-nova | 06:55 | |
| *** noama has joined #openstack-nova | 06:55 | |
| *** noama has quit IRC | 06:57 | |
| *** noama has joined #openstack-nova | 06:57 | |
| *** noama has quit IRC | 07:00 | |
| *** noama has joined #openstack-nova | 07:01 | |
| *** ekuris has joined #openstack-nova | 07:04 | |
| *** mjura has joined #openstack-nova | 07:04 | |
| *** jerrygb has quit IRC | 07:06 | |
| *** ekuris has quit IRC | 07:07 | |
| *** ekuris has joined #openstack-nova | 07:07 | |
| *** belmoreira has joined #openstack-nova | 07:10 | |
| *** jerrygb has joined #openstack-nova | 07:10 | |
| *** mjura has quit IRC | 07:11 | |
| *** mjura has joined #openstack-nova | 07:11 | |
| *** nkrinner has quit IRC | 07:13 | |
| *** nherciu has joined #openstack-nova | 07:15 | |
| *** tesseract has joined #openstack-nova | 07:17 | |
| *** tesseract is now known as Guest48712 | 07:17 | |
| *** nkrinner has joined #openstack-nova | 07:19 | |
| *** thorst has joined #openstack-nova | 07:23 | |
| *** nkrinner has quit IRC | 07:23 | |
| *** nkrinner has joined #openstack-nova | 07:24 | |
| *** tobasco has quit IRC | 07:27 | |
| *** tobasco has joined #openstack-nova | 07:28 | |
| *** avolkov has joined #openstack-nova | 07:29 | |
| *** huanxie has quit IRC | 07:29 | |
| *** huanxie has joined #openstack-nova | 07:29 | |
| *** thorst has quit IRC | 07:30 | |
| *** fragatina has joined #openstack-nova | 07:30 | |
| *** andreas_s has joined #openstack-nova | 07:31 | |
| *** liverpooler has joined #openstack-nova | 07:36 | |
| *** lpetrut has joined #openstack-nova | 07:36 | |
| *** fragatina has quit IRC | 07:39 | |
| *** tobasco has quit IRC | 07:39 | |
| *** fragatina has joined #openstack-nova | 07:39 | |
| *** tobasco has joined #openstack-nova | 07:41 | |
| *** gmann_ has joined #openstack-nova | 07:42 | |
| *** jerrygb has quit IRC | 07:42 | |
| *** sree has joined #openstack-nova | 07:46 | |
| *** jerrygb has joined #openstack-nova | 07:46 | |
| *** openstackgerrit has quit IRC | 07:48 | |
| *** openstackgerrit has joined #openstack-nova | 07:48 | |
| *** jerrygb_ has joined #openstack-nova | 07:50 | |
| *** jerrygb has quit IRC | 07:52 | |
| *** moshele has joined #openstack-nova | 07:53 | |
| *** ducttape_ has joined #openstack-nova | 07:55 | |
| *** jerrygb has joined #openstack-nova | 07:56 | |
| *** edand has joined #openstack-nova | 07:56 | |
| *** jerrygb_ has quit IRC | 07:58 | |
| *** ljxiash_ has joined #openstack-nova | 07:59 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** ducttape_ has quit IRC | 08:00 | |
| *** zzzeek has joined #openstack-nova | 08:01 | |
| *** ljxiash has quit IRC | 08:03 | |
| *** ihrachys has joined #openstack-nova | 08:05 | |
| *** yamahata has quit IRC | 08:07 | |
| *** jerrygb has quit IRC | 08:08 | |
| *** pcaruana has joined #openstack-nova | 08:09 | |
| *** markvoelker has joined #openstack-nova | 08:13 | |
| *** prashkre has quit IRC | 08:13 | |
| *** jerrygb has joined #openstack-nova | 08:13 | |
| *** fandi has joined #openstack-nova | 08:14 | |
| *** markus_z has joined #openstack-nova | 08:16 | |
| *** markvoelker has quit IRC | 08:19 | |
| *** yuntongjin has quit IRC | 08:21 | |
| openstackgerrit | Huan Xie proposed openstack/nova: xenapi: support the hotplug of a neutron port https://review.openstack.org/365505 | 08:24 |
| *** andreas_s has quit IRC | 08:24 | |
| *** xinliang has quit IRC | 08:25 | |
| *** ihrachys has quit IRC | 08:25 | |
| *** oanson has joined #openstack-nova | 08:26 | |
| *** jerrygb_ has joined #openstack-nova | 08:27 | |
| *** ducnc has quit IRC | 08:27 | |
| *** thorst has joined #openstack-nova | 08:27 | |
| *** ducnc has joined #openstack-nova | 08:28 | |
| *** xek__ is now known as xek | 08:28 | |
| *** jerrygb has quit IRC | 08:29 | |
| openstackgerrit | Moshe Levi proposed openstack/nova: pci: remove pci device from claims and allocations when freeing it https://review.openstack.org/369180 | 08:29 |
| *** prashkre has joined #openstack-nova | 08:33 | |
| *** thorst has quit IRC | 08:35 | |
| *** andreas_s has joined #openstack-nova | 08:37 | |
| *** mgoddard has joined #openstack-nova | 08:37 | |
| *** xinliang has joined #openstack-nova | 08:38 | |
| *** avolkov has quit IRC | 08:40 | |
| *** jaosorior has joined #openstack-nova | 08:45 | |
| *** lpetrut has quit IRC | 08:45 | |
| *** roeyc has joined #openstack-nova | 08:46 | |
| *** lucas-afk is now known as lucasagomes | 08:51 | |
| *** haplo37 has quit IRC | 08:51 | |
| *** haplo37_ is now known as haplo37 | 08:51 | |
| *** haplo37_ has joined #openstack-nova | 08:54 | |
| *** sandanar has joined #openstack-nova | 08:54 | |
| *** GB21 has joined #openstack-nova | 08:55 | |
| *** noama has quit IRC | 08:57 | |
| *** noama has joined #openstack-nova | 08:58 | |
| snikitin | garyk1: Hi! I want to ask you one question about Diagnostic info spec: https://review.openstack.org/#/c/357884/9/specs/ocata/approved/restore-vm-diagnostics.rst | 08:58 |
| snikitin | Matt said that you may know why during previous implementations we didn't implement driver_private_data field. Do you have some information about it? | 08:58 |
| *** matrohon has joined #openstack-nova | 09:00 | |
| *** yassine has joined #openstack-nova | 09:04 | |
| *** yassine is now known as Guest87642 | 09:04 | |
| pkholkin | johnthetubaguy: bauzas: hey folks! please return to the spec when you have time https://review.openstack.org/#/c/381912/, thanks! | 09:07 |
| *** jerrygb_ has quit IRC | 09:08 | |
| *** Daisy has quit IRC | 09:11 | |
| *** Daisy has joined #openstack-nova | 09:11 | |
| *** jerrygb has joined #openstack-nova | 09:12 | |
| *** pilgrimstack has joined #openstack-nova | 09:13 | |
| *** jaosorior has quit IRC | 09:14 | |
| *** jaosorior has joined #openstack-nova | 09:14 | |
| *** GB21 has quit IRC | 09:15 | |
| *** ihrachys has joined #openstack-nova | 09:18 | |
| bauzas | pkoniszewski: roger, roger. | 09:20 |
| bauzas | oops | 09:20 |
| bauzas | pkholkin: ditto. | 09:20 |
| *** kiran-r has quit IRC | 09:23 | |
| *** derekh has joined #openstack-nova | 09:26 | |
| *** ducttape_ has joined #openstack-nova | 09:26 | |
| garyk1 | snikitin: hi | 09:26 |
| *** dencaval has joined #openstack-nova | 09:27 | |
| garyk1 | snikitin: i think that it was there to enable each driver to provide additional information. say driver foo has feature bar. this could be very useful for diagnostics. the private fields could be used as a means to display this until other drivers support the same data | 09:28 |
| *** sree has quit IRC | 09:29 | |
| *** sree has joined #openstack-nova | 09:30 | |
| *** ducttape_ has quit IRC | 09:30 | |
| *** brault|a_ is now known as brault|away | 09:30 | |
| *** sree_ has joined #openstack-nova | 09:31 | |
| *** sree_ is now known as Guest76512 | 09:32 | |
| *** thorst has joined #openstack-nova | 09:32 | |
| dencaval | hey guys, I have configured cache=none by virsh edit command for a guest vm, but the guest seems to keep using cache from host. Do you have any ideas why it is not working to disable the cache? | 09:33 |
| *** sree has quit IRC | 09:34 | |
| *** edand has quit IRC | 09:34 | |
| *** hshiina has quit IRC | 09:35 | |
| *** markus_z has quit IRC | 09:35 | |
| *** coreywright has quit IRC | 09:36 | |
| *** Guest76512 has quit IRC | 09:36 | |
| *** thorst has quit IRC | 09:39 | |
| *** markvoelker has joined #openstack-nova | 09:41 | |
| openstackgerrit | Gleb Stepanov proposed openstack/nova: Nested Quota: Set default values to subproject https://review.openstack.org/383728 | 09:42 |
| openstackgerrit | Gleb Stepanov proposed openstack/nova: Add nested quota support https://review.openstack.org/383729 | 09:42 |
| openstackgerrit | Gleb Stepanov proposed openstack/nova: Add child_hard_limits column https://review.openstack.org/383727 | 09:42 |
| openstackgerrit | Gleb Stepanov proposed openstack/nova: Add Hierarchy info https://review.openstack.org/383726 | 09:42 |
| *** markvoelker has quit IRC | 09:46 | |
| openstackgerrit | Gleb Stepanov proposed openstack/nova: Add _rebuild_instance helper method https://review.openstack.org/376004 | 09:46 |
| *** jaosorior has quit IRC | 09:48 | |
| *** ijw has joined #openstack-nova | 09:49 | |
| *** jaosorior has joined #openstack-nova | 09:49 | |
| *** ijw has quit IRC | 09:49 | |
| *** ijw has joined #openstack-nova | 09:50 | |
| *** ihrachys has quit IRC | 09:50 | |
| *** ihrachys has joined #openstack-nova | 09:51 | |
| *** jerrygb has quit IRC | 09:54 | |
| *** jerrygb has joined #openstack-nova | 09:58 | |
| *** markus_z has joined #openstack-nova | 09:58 | |
| *** Daisy has quit IRC | 09:58 | |
| *** ducnc has quit IRC | 09:59 | |
| *** Daisy has joined #openstack-nova | 09:59 | |
| *** sree has joined #openstack-nova | 10:03 | |
| *** sree has quit IRC | 10:03 | |
| mdbooth | johnthetubaguy: If you have a sec, could you take a look at this mechanical change danpb asked for whilst reviewing my nfs patch: https://review.openstack.org/#/c/392231/ | 10:03 |
| *** sree has joined #openstack-nova | 10:04 | |
| mdbooth | It's trivial but it touches lots of files, so it's likely to attract merge conflicts | 10:04 |
| snikitin | garyk1: I meant that in previous V3 diagnostic info implementation this field was *not* included in diagnostic info object https://github.com/openstack/nova/blob/master/nova/virt/diagnostics.py#L93 | 10:04 |
| snikitin | It was in spec but it was *not* implemented. And Matt said that you may khow why | 10:04 |
| openstackgerrit | Jianghua Wang proposed openstack/nova-specs: XenAPI: add common os-xenapi libary https://review.openstack.org/379279 | 10:04 |
| *** sahid has joined #openstack-nova | 10:05 | |
| *** roeyc has quit IRC | 10:05 | |
| openstackgerrit | Arvinder Singh proposed openstack/python-novaclient: Code to remove excetion for scrub command. https://review.openstack.org/390243 | 10:05 |
| *** tovin07 has quit IRC | 10:06 | |
| *** Daisy_ has joined #openstack-nova | 10:06 | |
| johnthetubaguy | mdbooth: oh, outch, let me hit that right now | 10:06 |
| mdbooth | johnthetubaguy: Thanks | 10:07 |
| *** matrohon has quit IRC | 10:08 | |
| *** jaosorior is now known as jaosorior_lunch | 10:08 | |
| *** sree has quit IRC | 10:08 | |
| *** phuongnh has quit IRC | 10:08 | |
| kashyap | mdbooth: I see that's pre-req for the NFS vol mount / unmount patch -- https://review.openstack.org/#/c/383859/ | 10:09 |
| kashyap | "that" == https://review.openstack.org/#/c/392231/ | 10:09 |
| *** Daisy has quit IRC | 10:10 | |
| mdbooth | kashyap: Yeah, danpb didn't like that I was complicating an interface he was trying to clean up | 10:10 |
| kashyap | mdbooth: Yeah, noticed your summary (thanks for that) which clearly outlined the above | 10:11 |
| *** matrohon has joined #openstack-nova | 10:12 | |
| *** zhangyufei has quit IRC | 10:14 | |
| *** erbadr has quit IRC | 10:14 | |
| *** tuanluong has quit IRC | 10:16 | |
| *** ishant|away has quit IRC | 10:17 | |
| *** danpb has joined #openstack-nova | 10:17 | |
| *** sambetts|afk is now known as sambetts | 10:18 | |
| openstackgerrit | jichenjc proposed openstack/python-novaclient: Add version pin for image related function https://review.openstack.org/392523 | 10:20 |
| *** harbor_log is now known as potdirect | 10:22 | |
| *** ishant|away has joined #openstack-nova | 10:24 | |
| *** cdent has joined #openstack-nova | 10:24 | |
| *** ociuhandu has quit IRC | 10:29 | |
| *** guano has joined #openstack-nova | 10:31 | |
| *** cdelatte has quit IRC | 10:32 | |
| *** sree has joined #openstack-nova | 10:34 | |
| *** liusheng has quit IRC | 10:34 | |
| *** portdirect has joined #openstack-nova | 10:34 | |
| *** liusheng has joined #openstack-nova | 10:36 | |
| *** thorst has joined #openstack-nova | 10:37 | |
| *** potdirect has quit IRC | 10:37 | |
| garyk1 | snikitin: ok, now i recall. jay pipes did not like it. so we removed it | 10:39 |
| cdent | bauzas: we need to arrange a hangout to talk with dan, jay, ed and whoever else about what to do about the scheduler interface to the placement api. When do you think would be a good time. I'm available today or tomorrow, pretty much whenever, but not friday. | 10:39 |
| *** pilgrimstack has quit IRC | 10:40 | |
| *** pilgrimstack has joined #openstack-nova | 10:41 | |
| *** trisq has quit IRC | 10:41 | |
| *** avolkov has joined #openstack-nova | 10:44 | |
| *** tbachman has quit IRC | 10:44 | |
| *** lpetrut has joined #openstack-nova | 10:45 | |
| *** thorst has quit IRC | 10:45 | |
| *** roeyc has joined #openstack-nova | 10:47 | |
| *** namnh has quit IRC | 10:48 | |
| *** portdirect has quit IRC | 10:48 | |
| bauzas | cdent: sure, I was in PTO the last 2 days, but I'm there now | 10:49 |
| *** al_indigo has joined #openstack-nova | 10:50 | |
| cdent | bauzas: any particular time work better for you? | 10:52 |
| cdent | I think edleafe is still away, but he may be back tomorrow | 10:52 |
| cdent | but I can probably proxy his point of view (which is not quite the same as mine) if required | 10:52 |
| bauzas | cdent: I don't have a precise agenda | 10:54 |
| bauzas | cdent: I'm just separating my WIP change into 2 changes, one that will be the object method, and one being the REST API call | 10:54 |
| cdent | bauzas: if I remember rightly from friday the main goal was to resolve the details of the rest api call | 10:55 |
| cdent | since jay wasn't in the room we said "let's hangout next week" | 10:55 |
| bauzas | yup | 10:55 |
| bauzas | yup | 10:55 |
| bauzas | the REST API call is just the interface for the scheduler | 10:56 |
| bauzas | but we also need a backend method for getting the narrowed list of RPs | 10:56 |
| *** ducttape_ has joined #openstack-nova | 10:56 | |
| bauzas | cdent: and given we need to have an agreement on the REST API, I prefer to split those two so we can first merge the DB call | 10:57 |
| *** rtheis has joined #openstack-nova | 10:57 | |
| cdent | makes sense | 10:57 |
| moshele | danpb: hi can you review User-controlled SR-IOV ports allocation - https://review.openstack.org/#/c/182242 | 10:58 |
| *** ducttape_ has quit IRC | 11:01 | |
| *** gszasz has joined #openstack-nova | 11:01 | |
| *** guano has quit IRC | 11:04 | |
| *** haplo37 has quit IRC | 11:04 | |
| *** ociuhandu has joined #openstack-nova | 11:06 | |
| *** haplo37 has joined #openstack-nova | 11:07 | |
| johnthetubaguy | pkholkin: thats for those spec updates, its looking close, I just have questions about the last column in table 2 | 11:10 |
| *** portdirect has joined #openstack-nova | 11:11 | |
| *** cdelatte has joined #openstack-nova | 11:11 | |
| *** Oku_OS is now known as Oku_OS-away | 11:12 | |
| *** tojuvone_ has quit IRC | 11:12 | |
| *** rtheis_ has joined #openstack-nova | 11:13 | |
| *** rtheis_ has quit IRC | 11:13 | |
| *** roeyc has quit IRC | 11:13 | |
| *** cdent has quit IRC | 11:13 | |
| *** rtheis_ has joined #openstack-nova | 11:13 | |
| *** rtheis has quit IRC | 11:14 | |
| *** GB21 has joined #openstack-nova | 11:15 | |
| *** matrohon has quit IRC | 11:16 | |
| *** sree has quit IRC | 11:16 | |
| *** sree has joined #openstack-nova | 11:17 | |
| *** markus_z has quit IRC | 11:20 | |
| *** jaosorior_lunch is now known as jaosorior | 11:21 | |
| *** sree has quit IRC | 11:21 | |
| *** sridharg has quit IRC | 11:22 | |
| *** sree has joined #openstack-nova | 11:22 | |
| *** hshiina has joined #openstack-nova | 11:23 | |
| *** roeyc has joined #openstack-nova | 11:23 | |
| *** daemontool has joined #openstack-nova | 11:25 | |
| *** sridharg has joined #openstack-nova | 11:27 | |
| *** sree has quit IRC | 11:27 | |
| *** Daisy_ has quit IRC | 11:28 | |
| *** Daisy has joined #openstack-nova | 11:28 | |
| *** smatzek_ has joined #openstack-nova | 11:29 | |
| *** Oku_OS-away is now known as Oku_OS | 11:30 | |
| *** sree has joined #openstack-nova | 11:32 | |
| openstackgerrit | Stephen Finucane proposed openstack/nova: conf: Standardize formatting of virt https://review.openstack.org/392564 | 11:32 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: conf: Detail the 'injected_network_template' opt https://review.openstack.org/392565 | 11:32 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: conf: Move virt options to a group https://review.openstack.org/392566 | 11:32 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: conf: Add 'preallocate_images' to 'virt' group https://review.openstack.org/392567 | 11:32 |
| *** gouthamr has quit IRC | 11:33 | |
| *** GB21 has quit IRC | 11:33 | |
| *** Daisy has quit IRC | 11:33 | |
| sahid | johnthetubaguy: we have an important spec related to the work on realtime worklow, do you have a moment to look at it ? https://review.openstack.org/#/c/284094/7/specs/ocata/approved/libvirt-emulator-threads-policy.rst | 11:34 |
| *** dave-mccowan has joined #openstack-nova | 11:35 | |
| *** bvanhav_ has joined #openstack-nova | 11:36 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: WIP: Provide an object facade for a subset of RPs https://review.openstack.org/386242 | 11:36 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: WIP: Expose a REST API method for a specific list of RPs https://review.openstack.org/392569 | 11:36 |
| *** sree has quit IRC | 11:36 | |
| *** baoli has joined #openstack-nova | 11:38 | |
| openstackgerrit | Stephen Finucane proposed openstack/nova: hardware: Split '_add_cpu_pinning_constraint' https://review.openstack.org/367469 | 11:40 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: hardware: Flatten functions https://review.openstack.org/367470 | 11:40 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: Standardize '_get_XXX_constraints' functions https://review.openstack.org/385071 | 11:40 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: Rename '_numa_get_constraints_XXX' functions https://review.openstack.org/385072 | 11:40 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: De-duplicate _numa_get_flavor_XXX_map_list https://review.openstack.org/385074 | 11:40 |
| *** sree has joined #openstack-nova | 11:41 | |
| *** baoli_ has joined #openstack-nova | 11:42 | |
| *** markvoelker has joined #openstack-nova | 11:42 | |
| sfinucan | macsz: Could you take a look at this again? https://review.openstack.org/#/c/356611/ | 11:43 |
| macsz | sfinucan: ah, ok | 11:44 |
| *** nkrinner has quit IRC | 11:44 | |
| *** baoli has quit IRC | 11:45 | |
| *** markvoelker has quit IRC | 11:46 | |
| *** mtanino has joined #openstack-nova | 11:46 | |
| *** nkrinner has joined #openstack-nova | 11:47 | |
| *** cdelatte has quit IRC | 11:47 | |
| *** thorst has joined #openstack-nova | 11:49 | |
| *** liusheng has quit IRC | 11:50 | |
| *** sree has quit IRC | 11:50 | |
| *** sree has joined #openstack-nova | 11:51 | |
| *** liusheng has joined #openstack-nova | 11:52 | |
| *** mtanino has quit IRC | 11:52 | |
| openstackgerrit | Maciej Szankin proposed openstack/nova: config options: Improve help for SPICE https://review.openstack.org/317046 | 11:53 |
| *** nkrinner has quit IRC | 11:54 | |
| *** gouthamr has joined #openstack-nova | 11:54 | |
| *** sree has quit IRC | 12:00 | |
| openstackgerrit | sahid proposed openstack/nova: object: add new field cpu_emulator_threads_policy https://review.openstack.org/385022 | 12:00 |
| openstackgerrit | sahid proposed openstack/nova: virt: add numa constraint for emulator threads policy https://review.openstack.org/385023 | 12:00 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: return overhead if isolate emulator threads requested https://review.openstack.org/385364 | 12:00 |
| openstackgerrit | sahid proposed openstack/nova: virt: update overhead to take into account vCPUs https://review.openstack.org/385024 | 12:00 |
| *** cdelatte has joined #openstack-nova | 12:00 | |
| *** cdent has joined #openstack-nova | 12:00 | |
| *** paul-carlton2 has joined #openstack-nova | 12:00 | |
| *** sree has joined #openstack-nova | 12:01 | |
| *** prateek has quit IRC | 12:01 | |
| *** krtaylor has joined #openstack-nova | 12:03 | |
| *** sridharg has quit IRC | 12:03 | |
| *** trisq has joined #openstack-nova | 12:03 | |
| *** moshele has quit IRC | 12:03 | |
| johnthetubaguy | sahid: sfinucan: danpb: we keep adding lots of fine grained control the on the NUMA / real time front, do we have the overall direction documented somwhere? I am just worried we are making something thats really hard for deployers to use. Feel we are passing through virt concepts, rather than abstractions. Maybe thats what they need. | 12:05 |
| *** lucasagomes is now known as lucas-hungry | 12:05 | |
| *** tbachman has joined #openstack-nova | 12:06 | |
| *** ducttape_ has joined #openstack-nova | 12:07 | |
| *** nkrinner has joined #openstack-nova | 12:07 | |
| sahid | johnthetubaguy: from what i know we do not have such document, it's really related to isolate a cpu from external noise I think we want to continue the work added with realtime | 12:08 |
| johnthetubaguy | yeah, it totally seems useful, I didn't like how we did that the first time, but was important to make forward progress, and the old way was simpler. Its just I don't feel I have all the context around this | 12:10 |
| danpb | johnthetubaguy: the general vision is that we expose ability to express usage policies, but not direct controls over host placement/allocation | 12:10 |
| johnthetubaguy | sfinucan: did you do some admin docs at one point, maybe that would help me? | 12:10 |
| danpb | johnthetubaguy: eg we have the assumption that the tenant user / cloud admin have zero knowledge of the host | 12:10 |
| johnthetubaguy | yeah, that makes sense, to avoid config hell | 12:11 |
| sfinucan | johnthetubaguy: Sure - one sec | 12:11 |
| openstackgerrit | Daniel Berrange proposed openstack/os-vif: host_info: add ability to filter list of supported vifs https://review.openstack.org/392591 | 12:11 |
| sfinucan | johnthetubaguy: http://docs.openstack.org/admin-guide/compute-cpu-topologies.html | 12:11 |
| *** liusheng has quit IRC | 12:13 | |
| johnthetubaguy | sfinucan: perfect, that helps get my head straight again | 12:14 |
| *** cdent has quit IRC | 12:14 | |
| *** moshele has joined #openstack-nova | 12:14 | |
| *** sree has quit IRC | 12:14 | |
| *** sridharg has joined #openstack-nova | 12:15 | |
| *** hshiina has quit IRC | 12:16 | |
| johnthetubaguy | sfinucan: danpb: sahid: if we added hw:realtime_optimised=true/false, as an alias for multiple (changing between releases) keys, would that help for all the pinning options? | 12:19 |
| *** jaypipes has joined #openstack-nova | 12:19 | |
| *** cdent has joined #openstack-nova | 12:19 | |
| *** ljxiash_ has quit IRC | 12:19 | |
| jaypipes | *yawn* morning folks | 12:19 |
| johnthetubaguy | sfinucan: danpb: sahid: where true is optimise (isolated) guest performance, false is optimise usage across the host | 12:20 |
| *** maticue has joined #openstack-nova | 12:20 | |
| danpb | johnthetubaguy: i don't think that'd be flexible enough | 12:20 |
| *** edmondsw has joined #openstack-nova | 12:21 | |
| sahid | johnthetubaguy: wel that seems complicated, users which wnat to use such workload really have to understand what they are going, like configuring the realtime_maks, the cpu_isolated, the pinning, hugepages... | 12:21 |
| danpb | as there's multiple definitions of what "realtime optimized" means and no single definition suits all uses | 12:21 |
| sahid | I don't see how we can handle such options in one | 12:21 |
| *** markus_z has joined #openstack-nova | 12:21 | |
| danpb | for example, some guest OS require certain vCPUs are non-realtime, while other guest OS require all vCPUs to be realtime | 12:22 |
| *** sridharg has quit IRC | 12:22 | |
| *** takedakn has joined #openstack-nova | 12:22 | |
| danpb | so you can't hide that behind a single knob, without breaking some set of guest OS | 12:22 |
| *** sridharg has joined #openstack-nova | 12:22 | |
| danpb | hence the need for hw_realtime_mask | 12:22 |
| johnthetubaguy | danpb: so I am not thinking as a replacement, just a kind of "max" setting, agreed some workloads need more detail | 12:22 |
| danpb | what you describe is really just a predefined out of the box flavour when a bunjch of settings turned on | 12:23 |
| danpb | i could see devstack, for example, creating a "realtime optimized" flavour as a convenience | 12:24 |
| *** trisq has quit IRC | 12:24 | |
| danpb | likewise for other deployment tools like triple-o | 12:24 |
| pkholkin | johnthetubaguy: John, I've looked your comments, do you see some problems with empty? | 12:24 |
| danpb | but don't think nova needs to get into the business, particularly since we only just made the decision to drop all our pre-defined flavours on the basis that they were no directly usable by anyone | 12:24 |
| johnthetubaguy | danpb: its about making it really easy for deployers and the 80% use case for real time (which may not be a thing) | 12:24 |
| pkholkin | and what list k1,k2 do you mean? there is {k1:v1} or {k1:v1, k2:v2} | 12:25 |
| danpb | johnthetubaguy: sure, that's what devstack/triple-o/fuel/whatever tool can do | 12:25 |
| sahid | johnthetubaguy: it's also really related to the guest image | 12:25 |
| johnthetubaguy | pkholkin: its just that empty aggregate actually has the exclusive key set on it right? | 12:25 |
| johnthetubaguy | danpb: but this is in the flavor, real people have to think and make decisions about that right? | 12:26 |
| danpb | johnthetubaguy: there's work being done on triple-o for exactly this, to let it deploy a realtime optimized host + default flavour to use it | 12:26 |
| johnthetubaguy | danpb: if we controlled what that was, as we add new best practices, users wouldn't need to change anything | 12:26 |
| pkholkin | johnthetubaguy: we can have aggreagate with empty meta, and we accept only empty images/flavors in strict case | 12:27 |
| *** sree has joined #openstack-nova | 12:27 | |
| johnthetubaguy | pkholkin: right, the strict case is enabled by metadata on the aggregate though, so its not really empty any more? | 12:28 |
| danpb | johnthetubaguy: i don't think nova should be in that business for exactly the reasons that we dropped all our default flavours | 12:28 |
| danpb | nova can document best practice recommendations, but i don't think nova should be implementing them - the deployment tools and/or cloud admin should be doing that | 12:29 |
| *** trisq has joined #openstack-nova | 12:29 | |
| *** xyang1 has joined #openstack-nova | 12:30 | |
| pkholkin | johnthetubaguy: empty on empty is ok, I don't think it is a problem | 12:30 |
| *** liverpooler has quit IRC | 12:31 | |
| johnthetubaguy | danpb: it just feels like the 80% deployer wants to say "best isolated guest performance" vs "make best use of host resources", right now we force them to go learn about NUMA and how it affects virt technologies, which seems bad for the 80% case, but agreed its wanted for people who need to get | 12:31 |
| pkholkin | smth on empty aggregate - is bad | 12:31 |
| johnthetubaguy | pkholkin: I mean empty actually means, in an aggregate with the isolation metadata set, which doesn't seem clear from the spec | 12:32 |
| danpb | johnthetubaguy: like i said though, deployment tools can implement this default policies, so admins don't need to understand it | 12:32 |
| *** liusheng has joined #openstack-nova | 12:32 | |
| *** liverpooler has joined #openstack-nova | 12:33 | |
| danpb | the deployment tools are in the position to then let the knowledgable admins customize it at time of deployment | 12:33 |
| danpb | which is not something that's easily done if the policy is implemented in nova code | 12:33 |
| *** baoli_ has quit IRC | 12:34 | |
| pkholkin | johnthetubaguy: do you mean that 'strict_mode=true', so aggregate meta is not empty? | 12:35 |
| johnthetubaguy | pkholkin: yes | 12:36 |
| *** baoli has joined #openstack-nova | 12:36 | |
| pkholkin | johnthetubaguy: undestand :) mhhh | 12:36 |
| *** gstepanov has joined #openstack-nova | 12:36 | |
| pkholkin | hmmm) | 12:36 |
| johnthetubaguy | pkholkin: mostly because you could also have one without that set | 12:36 |
| johnthetubaguy | one aggregate that is | 12:36 |
| pkholkin | I will add that I don't consider it | 12:36 |
| johnthetubaguy | pkholkin: the other side of this is what does strict mean for the empty aggregate, we actually have to look at all aggregates in the system and what keys they have, I believe? | 12:37 |
| johnthetubaguy | pkholkin: oops, actually, we don't, its just any key that has the known separator/prefix | 12:38 |
| shewless | Hi mdbooth and ftersin: Just wondering how this review is coming along: https://review.openstack.org/#/c/346263. | 12:38 |
| pkholkin | johnthetubaguy: our approach is 'per-aggregate', so each is individual, empty aggregate is with {} or {'strict_mode:false} meta | 12:39 |
| openstackgerrit | Daniel Berrange proposed openstack/os-vif: host_info: fix has_plugin/get_plugin methods on HostInfo https://review.openstack.org/392613 | 12:40 |
| openstackgerrit | Daniel Berrange proposed openstack/os-vif: host_info: fix has_vif/get_vif methods on HostPluginInfo https://review.openstack.org/392614 | 12:40 |
| *** ducttape_ has quit IRC | 12:40 | |
| *** markvoelker has joined #openstack-nova | 12:41 | |
| johnthetubaguy | danpb: it just feels like there is an 80% case we are forcing lots of people to understand, including general API users listing flavors in the API, I get some folks need/want the extra control and detail. I should try write that up I guess, to make my case more clearly. Its possible its a make believe thing. | 12:42 |
| *** jerrygb has quit IRC | 12:42 | |
| openstackgerrit | Manas Mandlekar proposed openstack/nova: Don't destroy evacuated instance if host didn't change https://review.openstack.org/392620 | 12:43 |
| *** ljxiash has joined #openstack-nova | 12:43 | |
| *** gouthamr has quit IRC | 12:44 | |
| openstackgerrit | Daniel Berrange proposed openstack/os-vif: host_info: add ability to filter list of supported vifs https://review.openstack.org/392591 | 12:44 |
| *** jerrygb has joined #openstack-nova | 12:45 | |
| *** gouthamr has joined #openstack-nova | 12:45 | |
| *** jheroux has joined #openstack-nova | 12:47 | |
| ftersin | shewless: hi. it's still waiting for reviewers. mdbooth, danpb, and mriedem are welcome | 12:47 |
| *** liusheng has quit IRC | 12:47 | |
| *** liusheng has joined #openstack-nova | 12:48 | |
| danpb | ftersin: i want mdbooth to ack it first | 12:49 |
| alex_xu | nova api subteam meeting is in 10 mins at #openstack-meeting-4 | 12:50 |
| *** hshiina has joined #openstack-nova | 12:51 | |
| *** trisq has quit IRC | 12:52 | |
| *** ducttape_ has joined #openstack-nova | 12:58 | |
| *** tlian has joined #openstack-nova | 12:58 | |
| *** sheel has quit IRC | 13:00 | |
| *** bvanhav_ has quit IRC | 13:06 | |
| *** haplo37_ has quit IRC | 13:07 | |
| *** READ10 has joined #openstack-nova | 13:07 | |
| *** bvanhav_ has joined #openstack-nova | 13:09 | |
| *** bvanhav_ is now known as bvanhav | 13:09 | |
| openstackgerrit | Stephen Finucane proposed openstack/nova: Ignore IOError when creating 'console.log' https://review.openstack.org/392643 | 13:09 |
| *** haplo37_ has joined #openstack-nova | 13:09 | |
| openstackgerrit | Pavel Kholkin proposed openstack/nova-specs: Strict isolation of group of hosts for image and flavor https://review.openstack.org/381912 | 13:09 |
| *** prashkre has quit IRC | 13:10 | |
| *** ducttape_ has quit IRC | 13:10 | |
| *** sudipto has joined #openstack-nova | 13:10 | |
| *** sudipto_ has joined #openstack-nova | 13:10 | |
| *** prashkre has joined #openstack-nova | 13:10 | |
| *** mtanino has joined #openstack-nova | 13:11 | |
| *** Guest81678 has quit IRC | 13:13 | |
| *** ekhugen has quit IRC | 13:13 | |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: implement ResourceClass.create() https://review.openstack.org/386844 | 13:14 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: raise KeyError unknown resource class https://review.openstack.org/390062 | 13:14 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: REST API for resource classes https://review.openstack.org/391959 | 13:14 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: adds ResourceClass.get_by_name() https://review.openstack.org/391917 | 13:14 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: implement destroy() resource class https://review.openstack.org/391918 | 13:14 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: add ResourceClass and ResourceClassList https://review.openstack.org/386843 | 13:14 |
| *** andreas_s has quit IRC | 13:14 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Make n-net refuse to start unless using CellsV1 https://review.openstack.org/392236 | 13:16 |
| *** mriedem has joined #openstack-nova | 13:16 | |
| *** coreywright has joined #openstack-nova | 13:17 | |
| *** coolsvap has quit IRC | 13:18 | |
| prashkre | mriedem: Hi, Good Morning. I have put up my first Openstack patch for review(https://review.openstack.org/391113). Could you take a look when you get a chance.Thank you ! | 13:18 |
| sahid | jaypipes: can you have a look at this fix when you have a moment? https://review.openstack.org/#/c/356383/ | 13:18 |
| *** eharney_ has joined #openstack-nova | 13:19 | |
| *** andreas_s has joined #openstack-nova | 13:20 | |
| *** ralonsoh has joined #openstack-nova | 13:20 | |
| *** mdrabe has joined #openstack-nova | 13:20 | |
| *** lucas-hungry is now known as lucasagomes | 13:22 | |
| mriedem | prashkre: done | 13:22 |
| *** raunak has joined #openstack-nova | 13:23 | |
| prashkre | mriedem: Thank you! | 13:24 |
| *** matrohon has joined #openstack-nova | 13:25 | |
| *** takashin has joined #openstack-nova | 13:28 | |
| *** hshiina has quit IRC | 13:28 | |
| *** ijw has quit IRC | 13:28 | |
| openstackgerrit | Paul Carlton proposed openstack/nova: Attach/detach root volume for shelved_offloaded https://review.openstack.org/317446 | 13:29 |
| openstackgerrit | Paul Carlton proposed openstack/nova: Adding device_present field to bdm table https://review.openstack.org/268053 | 13:29 |
| *** ssurana has joined #openstack-nova | 13:33 | |
| *** mriedem has left #openstack-nova | 13:33 | |
| *** mriedem has joined #openstack-nova | 13:33 | |
| *** tblakes has joined #openstack-nova | 13:35 | |
| *** smatzek_ has quit IRC | 13:35 | |
| mriedem | prashkre: http://docs.openstack.org/developer/nova/code-review.html?highlight=release%20notes#release-notes | 13:35 |
| mdbooth | ftersin: I'm looking now, btw | 13:36 |
| *** baoli_ has joined #openstack-nova | 13:36 | |
| *** baoli has quit IRC | 13:36 | |
| mriedem | prashkre: i'd just put something in the 'fixes' or 'other' section of a release note | 13:36 |
| *** baoli has joined #openstack-nova | 13:37 | |
| *** sandanar has quit IRC | 13:38 | |
| *** baoli__ has joined #openstack-nova | 13:38 | |
| *** ducttape_ has joined #openstack-nova | 13:38 | |
| *** baoli_ has quit IRC | 13:41 | |
| johnthetubaguy | alex_xu: for that parameter validation, maybe we should add an example API with it? like a really simple one? | 13:41 |
| *** baoli has quit IRC | 13:42 | |
| *** takedakn has quit IRC | 13:42 | |
| mdbooth | ftersin: Thanks for the fantastic ignore_bdi_for_swap comment, btw. | 13:42 |
| *** baoli__ has quit IRC | 13:43 | |
| *** roeyc has quit IRC | 13:43 | |
| prashkre | mriedem: do we need any extra access rights to edit the content in release notes. how can I put my content in it. | 13:43 |
| *** rodrigods has quit IRC | 13:43 | |
| mriedem | you generate one for your patch | 13:43 |
| *** rodrigods has joined #openstack-nova | 13:43 | |
| mriedem | prashkre: http://docs.openstack.org/developer/reno/usage.html#creating-new-release-notes | 13:43 |
| *** baoli has joined #openstack-nova | 13:44 | |
| mriedem | prashkre: so: tox -e venv -- reno new bug-1636157 | 13:44 |
| mriedem | and then edit the file created under nova/releasenotes/notes/ that starts with bug-1636157 | 13:44 |
| mriedem | the reno should point out there are now separate policies for server group actions | 13:44 |
| mriedem | and that it's backward compatible by default | 13:45 |
| dansmith | bauzas: are you around this week? | 13:47 |
| *** burt has joined #openstack-nova | 13:47 | |
| *** tblakes has quit IRC | 13:49 | |
| Dinesh_Bhor | Hi all, Can I have more eyes on this: https://review.openstack.org/#/c/392159/? | 13:49 |
| openstackgerrit | Maciej Szankin proposed openstack/nova: conf: Move api options to a group https://review.openstack.org/384490 | 13:49 |
| prashkre | mriedem: sure. will try it out and get back to you. thank you. | 13:50 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: Ignore IOError when creating 'console.log' https://review.openstack.org/392643 | 13:50 |
| *** roeyc has joined #openstack-nova | 13:50 | |
| *** trisq has joined #openstack-nova | 13:52 | |
| *** nherciu has quit IRC | 13:52 | |
| *** nherciu_ has joined #openstack-nova | 13:52 | |
| *** jaosorior has quit IRC | 13:53 | |
| *** trisq has quit IRC | 13:53 | |
| *** gmann_ has quit IRC | 13:53 | |
| ftersin | mdbooth, i hope it's clear enough | 13:54 |
| *** andreas_s has quit IRC | 13:54 | |
| mdbooth | ftersin: Yep. Do you happen to remember where we update the bdms during resize in the existing code, btw? | 13:55 |
| mdbooth | I've found it before, but I'm struggling right now. | 13:55 |
| ftersin | we do not update them, afaik | 13:56 |
| *** ducttape_ has quit IRC | 13:56 | |
| ftersin | it is the reason of https://review.openstack.org/#/c/372721/ | 13:56 |
| *** andreas_s has joined #openstack-nova | 13:56 | |
| *** tblakes has joined #openstack-nova | 13:56 | |
| mdbooth | ftersin: Well, the code which updates instance.root_gb and instance.swap_mb | 13:56 |
| lbeliveau | jaypipes: can you please have a look at this bug fix when you have a chance: https://review.openstack.org/#/c/363884/ | 13:57 |
| ftersin | when a new flavor is being set as the instance flavor | 13:57 |
| *** sree has quit IRC | 13:57 | |
| mdbooth | Ah, it's called by _finish_resize | 13:57 |
| *** baoli has quit IRC | 13:58 | |
| mdbooth | I had it in my head that we were updating the root bdm or something, but I'm going senile. | 13:58 |
| Dinesh_Bhor | Thanks mriedem | 13:58 |
| *** smatzek has joined #openstack-nova | 13:59 | |
| *** baoli has joined #openstack-nova | 13:59 | |
| *** yamahata has joined #openstack-nova | 13:59 | |
| mdbooth | ftersin: So, during creation we're using the bdm, during resize we're using the flavor directly... | 14:00 |
| ftersin | mdbooth, it's impossible to permanently keep in mind all this spaghetty | 14:00 |
| mdbooth | Yep, that merits a big comment :) | 14:00 |
| *** sree has joined #openstack-nova | 14:01 | |
| *** GB21 has joined #openstack-nova | 14:01 | |
| *** sree has quit IRC | 14:02 | |
| ftersin | mdbooth: to be more accurate, we've started to use bdms during creation, but forgot about resize | 14:02 |
| mdbooth | ftersin: Yup | 14:02 |
| mdbooth | ffs, gerrit >:( | 14:04 |
| * mdbooth types his comment again | 14:04 | |
| *** markus_z has quit IRC | 14:04 | |
| mdbooth | ftersin: How about rebuild after resize? | 14:06 |
| *** whenry has quit IRC | 14:06 | |
| *** tongli has joined #openstack-nova | 14:06 | |
| *** sree has joined #openstack-nova | 14:06 | |
| mdbooth | Or manually specified swap which is smaller than the flavor swap size. | 14:07 |
| mdbooth | ... I'm not sure we can even handle that. | 14:07 |
| openstackgerrit | sahid proposed openstack/nova: live-mig: Add claims and proper resource tracking https://review.openstack.org/244489 | 14:07 |
| openstackgerrit | sahid proposed openstack/nova: Change RPC post_live_migration_at_destination from cast to call. https://review.openstack.org/389687 | 14:07 |
| *** mlavalle has joined #openstack-nova | 14:08 | |
| mdbooth | ftersin: Incidentally, why have you put the swap resize fix in here? | 14:08 |
| mdbooth | Ephemeral resize still doesn't work either, right? | 14:08 |
| ftersin | mdbooth: someone needed swap resize, filed the bug (mentioned in the review comments), and corresponding fix was merged | 14:09 |
| * mdbooth has a horrible suspicion he's asking to not have something somebody else specifically asked to have added... | 14:09 | |
| mdbooth | ftersin: Let me read the earlier comments first. | 14:10 |
| ftersin | i.e. swap resize works now, it is used by someone, and we have to support it | 14:10 |
| mdbooth | ftersin: 'works' | 14:10 |
| *** eharney_ has quit IRC | 14:11 | |
| openstackgerrit | sahid proposed openstack/nova: live-mig: Add claims and proper resource tracking https://review.openstack.org/244489 | 14:11 |
| openstackgerrit | sahid proposed openstack/nova: Change RPC post_live_migration_at_destination from cast to call. https://review.openstack.org/389687 | 14:11 |
| mdbooth | If they rebuild, they'll get a smaller swap disk though, right? | 14:11 |
| ftersin | i mean this: https://bugs.launchpad.net/nova/+bug/1244222 | 14:11 |
| openstack | Launchpad bug 1244222 in OpenStack Compute (nova) "nova-resize does not resize swap disk" [Low,Fix released] - Assigned to Thang Pham (thang-pham) | 14:11 |
| ftersin | mdbooth: probably they will | 14:12 |
| *** bjolo_ has joined #openstack-nova | 14:12 | |
| *** danpb has quit IRC | 14:14 | |
| *** baoli has quit IRC | 14:15 | |
| *** baoli has joined #openstack-nova | 14:15 | |
| *** danpb has joined #openstack-nova | 14:15 | |
| *** sree has quit IRC | 14:16 | |
| *** baoli_ has joined #openstack-nova | 14:16 | |
| *** baoli has quit IRC | 14:20 | |
| *** tbachman has quit IRC | 14:20 | |
| *** ducttape_ has joined #openstack-nova | 14:22 | |
| *** fandi has quit IRC | 14:24 | |
| *** sree has joined #openstack-nova | 14:25 | |
| *** mriedem1 has joined #openstack-nova | 14:25 | |
| *** tbachman has joined #openstack-nova | 14:26 | |
| openstackgerrit | Sergey Nikitin proposed openstack/nova-specs: Restore V3 Diagnostics info https://review.openstack.org/357884 | 14:26 |
| *** ducttape_ has quit IRC | 14:26 | |
| *** mriedem has quit IRC | 14:26 | |
| jaypipes | sahid, lbeliveau: yes, reviewing them now. | 14:27 |
| snikitin | jaypipes: Hi Jay! I'm trying to restore Diagnostic info API which will return some standardised information about instances from hypervisors. https://review.openstack.org/#/c/357884/ And I heard from garyk that you was against adding of field 'driver_private_data' with some driver specific information about instance. Do you still against it? | 14:28 |
| openstackgerrit | Dan Smith proposed openstack/nova: Return build_requests instead of instances https://review.openstack.org/374267 | 14:28 |
| openstackgerrit | Dan Smith proposed openstack/nova: WIP Add schedule_and_build_instances conductor method https://review.openstack.org/367557 | 14:28 |
| openstackgerrit | Dan Smith proposed openstack/nova: WIP Add schedule_and_build_instances RPC method https://review.openstack.org/379320 | 14:28 |
| openstackgerrit | Dan Smith proposed openstack/nova: WIP Move instance creation to conductor https://review.openstack.org/319379 | 14:28 |
| jaypipes | snikitin: is this a normal end-user API? | 14:28 |
| *** daemontool has quit IRC | 14:31 | |
| *** daemontool has joined #openstack-nova | 14:32 | |
| mdbooth | ftersin: create_image unit tests make me sad :( I'm going to have to rewrite them again. Not your problem, though. | 14:32 |
| mriedem1 | jaypipes: no it's admin-only | 14:32 |
| *** mriedem1 is now known as mriedem | 14:32 | |
| *** baoli_ has quit IRC | 14:32 | |
| *** sree has quit IRC | 14:33 | |
| *** akshai has joined #openstack-nova | 14:34 | |
| jaypipes | mriedem: k, I'll look at that and give my opinion. Generally I don't care to leak private implementation details out through the public HTTP API... but we'll see. | 14:34 |
| mriedem | snikitin: nothing is using that driver_private_data thing in the virt drivers today | 14:34 |
| mriedem | snikitin: so if nothing is using it, and there is no obvious thing to add there, i'd opt to just remove it from the spec | 14:34 |
| mriedem | snikitin: the Diagnostics object will be versioned, so if we need to add fields to it for some reason we can do that, and bump the version | 14:35 |
| *** anush has quit IRC | 14:36 | |
| *** eharney has joined #openstack-nova | 14:36 | |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: implement ResourceClass.create() https://review.openstack.org/386844 | 14:38 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: raise KeyError unknown resource class https://review.openstack.org/390062 | 14:38 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: REST API for resource classes https://review.openstack.org/391959 | 14:38 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: adds ResourceClass.get_by_name() https://review.openstack.org/391917 | 14:38 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: implement destroy() resource class https://review.openstack.org/391918 | 14:38 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: add ResourceClass and ResourceClassList https://review.openstack.org/386843 | 14:38 |
| *** sree has joined #openstack-nova | 14:39 | |
| *** dikonoor has quit IRC | 14:40 | |
| *** prashkre has quit IRC | 14:41 | |
| *** jaosorior has joined #openstack-nova | 14:41 | |
| *** victorhe has joined #openstack-nova | 14:42 | |
| *** dmellado is now known as dmellado|lunch | 14:42 | |
| openstackgerrit | Sergey Nikitin proposed openstack/nova-specs: Restore V3 Diagnostics info https://review.openstack.org/357884 | 14:43 |
| *** jed56 has joined #openstack-nova | 14:43 | |
| *** sree has quit IRC | 14:43 | |
| snikitin | mriedem: ok, lets remove it | 14:43 |
| *** sree has joined #openstack-nova | 14:44 | |
| mriedem | snikitin: more comments inline https://review.openstack.org/#/c/357884/11 | 14:44 |
| snikitin | looking | 14:44 |
| mriedem | it's mostly good, just needs to remove some inaccuracies | 14:44 |
| *** paw has quit IRC | 14:45 | |
| *** markus_z has joined #openstack-nova | 14:45 | |
| *** mgoddard_ has joined #openstack-nova | 14:47 | |
| *** ducttape_ has joined #openstack-nova | 14:48 | |
| *** tbachman has quit IRC | 14:49 | |
| mdbooth | ftersin: Done. I've asked for 1 mechanical test change. If you ping me later I'll be happy to +1. | 14:50 |
| *** mgoddard has quit IRC | 14:50 | |
| *** raunak has quit IRC | 14:51 | |
| *** rrecio has joined #openstack-nova | 14:52 | |
| mriedem | diana_clarke: want to fix the typo in https://review.openstack.org/#/c/386771/ ? then i'm +2 | 14:52 |
| mriedem | very nicely done spec | 14:52 |
| *** mtanino has quit IRC | 14:52 | |
| *** mgoddard_ has quit IRC | 14:53 | |
| mriedem | mdbooth: did you have plans for https://review.openstack.org/#/c/382449/ or should i hack on that a bit? | 14:53 |
| *** mtanino has joined #openstack-nova | 14:53 | |
| *** mgoddard_ has joined #openstack-nova | 14:53 | |
| mriedem | i think we had agreed to make some short-term hacky fixes in there and then save the longer-term event listening stuff for later | 14:53 |
| *** victorhe has quit IRC | 14:54 | |
| mdbooth | mriedem: Sec, let me refresh my memory | 14:54 |
| *** sai has quit IRC | 14:55 | |
| *** raunak has joined #openstack-nova | 14:56 | |
| *** raunak has left #openstack-nova | 14:56 | |
| mdbooth | mriedem: Are you happy hacking on it? | 14:56 |
| *** rrecio has quit IRC | 14:56 | |
| mdbooth | I'm probably not going to get to it imminently, tbh. | 14:57 |
| mriedem | i've also lost some context, but i think i was looking for a short-term backoff retry loop checking the xml for "ready='yes'" | 14:57 |
| mdbooth | kashyap: Do you want to hack on it? | 14:57 |
| mriedem | i'd prefer kashyap to work on it | 14:57 |
| mdbooth | So with that fix specifically I think we just want to add tests and get it in. | 14:57 |
| * kashyap clicks | 14:57 | |
| kashyap | First trying to read context | 14:57 |
| mdbooth | We also need to fix the real problem, though, which is the xml thing you mention, or the event handling | 14:58 |
| *** victorhe has joined #openstack-nova | 14:58 | |
| *** daemontool has quit IRC | 14:58 | |
| *** sai has joined #openstack-nova | 14:58 | |
| mriedem | true, we could just do this incrementally | 14:58 |
| mriedem | this is an obvious fix, add tests, get'r done | 14:58 |
| mriedem | then do the backoff retry which is backportable if we want | 14:58 |
| mriedem | then the event listener | 14:58 |
| jaypipes | sahid: currently reading through libvirt docs... :( this is going to get ugly. | 14:59 |
| mriedem | brb | 14:59 |
| kashyap | mriedem: If it's the same mount / unmount issue, to honestly admit I was not able to completely my wrap my head around the intricacies, and mdbooth was /is far more competent in figuring out the details. | 14:59 |
| bauzas | dansmith: hola, yup, since today, was just at a doctor's appointment :) | 14:59 |
| bauzas | yesterday was a bank holiday :) | 14:59 |
| *** rrecio has joined #openstack-nova | 14:59 | |
| dansmith | bauzas: okay | 14:59 |
| kashyap | Oh, this is a different issue | 14:59 |
| bauzas | dansmith: jaypipes: just lemme know when you want to have a discussion about the REST API call for the scheduler | 15:00 |
| openstackgerrit | Artom Lifshitz proposed openstack/nova: Wait for ports down during evacuate https://review.openstack.org/385086 | 15:00 |
| mdbooth | kashyap: Yeah, this is the older one where you did the libvirt sleuthing. | 15:00 |
| dansmith | bauzas: so we have a cells meeting scheduled for today, but it would just be us I thnk | 15:00 |
| kashyap | mdbooth: Sorry, my eyes were warped after 2 hours of looking logs from another downstream bug | 15:00 |
| *** _ducttape_ has joined #openstack-nova | 15:00 | |
| kashyap | Just recuperating | 15:00 |
| dansmith | bauzas: I'm trying to wrap my head around all the code currently proposed and address some of my own feedback on them | 15:01 |
| *** cdelatte has quit IRC | 15:01 | |
| pkoniszewski | hello nova, is someone aware whether conf parser has been changed recently? I've observed some issues with libvirt live migration on trunk, basically because all strings read from nova.conf are treated as unicode | 15:01 |
| dansmith | bauzas: do you think we should meet or punt for today? | 15:01 |
| *** mriedem1 has joined #openstack-nova | 15:01 | |
| bauzas | dansmith: I'm fine for today's meeting, but we have an internal meeting one hour before :) | 15:01 |
| mriedem1 | kashyap: mdbooth: ok so i'll hack on unit tests for this | 15:01 |
| mriedem1 | and play around with the backoff retry | 15:01 |
| openstackgerrit | Diana Clarke proposed openstack/nova-specs: Simple tenant usage pagination https://review.openstack.org/386771 | 15:01 |
| dansmith | bauzas: I don't, AFAICT | 15:02 |
| dansmith | bauzas: you're fine meeting or fine canceling? | 15:02 |
| diana_clarke | mriedem1: done, thanks! | 15:02 |
| bauzas | dansmith: fine for the meeting | 15:02 |
| *** mgoddard_ has quit IRC | 15:02 | |
| sahid | jaypipes: what do you mean by "this is going to get ugly" ? do you want me to refactor the change or something | 15:02 |
| mriedem1 | dansmith: +2 on the simple tenant usage paging spec https://review.openstack.org/#/c/386771/ | 15:02 |
| *** mgoddard has joined #openstack-nova | 15:02 | |
| bauzas | dansmith: 6pm UTC right? | 15:02 |
| dansmith | bauzas: I'm trying to figure out: do you have something to discuss with just me? :) | 15:02 |
| kashyap | mdbooth: Sure. (Disregard me earlier response, as that applies to a different issue -- volume attach / detach race that mdbooth was investigating) | 15:02 |
| dansmith | bauzas: 2100UTC by my calendar | 15:03 |
| mriedem1 | bauzas: are you talking about the cellsv2 meeting? | 15:03 |
| bauzas | dansmith: I thought you were asking me for the cells meeting :) | 15:03 |
| dansmith | I am | 15:03 |
| bauzas | oh okay, 2100UTC is better for me then | 15:03 |
| jaypipes | sahid: :) no, I mean reading the libvirt documentation gives me zero insight into what the emulatorpin has to do with realtime stuff. | 15:03 |
| bauzas | sorry for the confusion :) | 15:03 |
| *** Zer0Byte__ has joined #openstack-nova | 15:03 | |
| dansmith | bauzas: I'm not asking about scheduling | 15:03 |
| dansmith | bauzas: I'm asking if there is any point in having the meeting if it's just the two f us | 15:04 |
| bauzas | cellsv2 meeting is fine by me for 2100UTC | 15:04 |
| bauzas | oh ok | 15:04 |
| bauzas | :p | 15:04 |
| *** ducttape_ has quit IRC | 15:04 | |
| bauzas | grah, my brain is probably still in vacation... | 15:04 |
| *** mriedem has quit IRC | 15:04 | |
| mriedem1 | dansmith: fwiw i probably won't be there today | 15:04 |
| mriedem1 | have a conflict | 15:04 |
| bauzas | dansmith: so, good point, what could we be discussing for the cellsv2 meeting, I dunno, maybe around the last points we discussed during the meetup | 15:05 |
| bauzas | maybe about following up | 15:05 |
| dansmith | bauzas: right my point is we both just discussed things on friday, and I'm busy getting up to speed with the things proposed | 15:05 |
| bauzas | but if mriedem1 can't be there and melwitt too, then meh. | 15:05 |
| dansmith | so if nobody else is here, I don't see the point | 15:05 |
| bauzas | okay, then let's punt it | 15:05 |
| sahid | jaypipes: ohh i see :) - we just want to ensure that the threads owned by qemu to make the guest running, to be pinned on host CPUs which are not intended to be used for RT | 15:06 |
| kashyap | mriedem1 I can work on that issue, as you may have other things on your plate. | 15:06 |
| openstackgerrit | Sergey Nikitin proposed openstack/nova-specs: Restore V3 Diagnostics info https://review.openstack.org/357884 | 15:06 |
| mriedem1 | kashyap: so you'll add tests to https://review.openstack.org/#/c/382449/ ? | 15:07 |
| bauzas | dansmith: does melwitt need some kind of discussion again for the quotas ? | 15:07 |
| dansmith | bauzas: she's on PTO this week | 15:07 |
| bauzas | or do we have kind of agreement here? | 15:07 |
| bauzas | oh okay, then let's punt that meeting | 15:07 |
| snikitin | mriedem1: fixed your comments | 15:07 |
| kashyap | mriedem1: I want to give it a shot, and you / MattB could review | 15:07 |
| mriedem1 | kashyap: go nuts | 15:07 |
| mriedem1 | let my -1 be your guide | 15:08 |
| kashyap | Heh | 15:08 |
| kashyap | mriedem1: Just FYI - tomm evening / Fri, I have to attend some urgent errands, just noting if you don't see me active on that this week | 15:08 |
| kashyap | (Don't assume I dropped it :-)) | 15:08 |
| jaypipes | sahid: so you are saying that in order to support the realtime contract, the QEMU emulator process must run on the same physical CPU as all of the virtual CPUs are pinned to? | 15:08 |
| mriedem1 | kashyap: ok, maybe diana_clarke would like to write the tests | 15:09 |
| mriedem1 | i'm happy to do it, but then i can't +2 | 15:09 |
| kashyap | Sure, if she has cycles right now, can go ahead | 15:09 |
| *** matrohon has quit IRC | 15:09 | |
| danpb | jaypipes: no, that's not right | 15:09 |
| jaypipes | danpb: could you explain for me please? I'm not following along this bug report unfortunately. :( | 15:09 |
| danpb | jaypipes: the emulator thread(s) must *never* run on a pCPU that is used by a guest vCPU marked as real-time | 15:09 |
| jaypipes | danpb: ah. | 15:10 |
| danpb | as that guest vCPU will starve the emulator thread | 15:10 |
| jaypipes | understood. that makes more sense. | 15:10 |
| sahid | jaypipes: that is the "normal" contract, for the realtime contract we want the QEMU emulator process pinned only on host CPUs which are not used for RT | 15:10 |
| sahid | yes as danpb said better than me | 15:10 |
| danpb | by default Nova currently requires that at least 1 vCPU be marked non-realtime, and so we've currently placed the emulator threads on the pCPU used by the non-realtime vCPU | 15:10 |
| *** paw has joined #openstack-nova | 15:10 | |
| *** mriedem1 is now known as mriedem | 15:10 | |
| jaypipes | danpb: k | 15:10 |
| danpb | jaypipes: some (non-Linux) guest OS require that *all* vCPUs are realtime | 15:11 |
| *** lpetrut has quit IRC | 15:11 | |
| danpb | jaypipes: thus, we need to enhance nova so that it can place emulator threads on a pCPU that is not used by *any* vCPU, and account for that extra pCPU usage by the guest | 15:11 |
| *** annegentle has joined #openstack-nova | 15:12 | |
| danpb | this will enable a much wider variety of real time operating systems to be used on nova | 15:12 |
| *** _ducttape_ has quit IRC | 15:12 | |
| *** mnestratov has joined #openstack-nova | 15:12 | |
| *** ducttape_ has joined #openstack-nova | 15:12 | |
| *** browne has joined #openstack-nova | 15:13 | |
| bauzas | jaypipes: FWIW, I sliced the WIP patch for the RP API into 2 separate commits, one for the object interface and one for the needs-consensus REST API call | 15:13 |
| jaypipes | danpb, sahid: I'm still confused by the cpu_realtime_mask=^0 thing. How would that value be determined by an operator (the ^0 value..) | 15:13 |
| jaypipes | bauzas: cool, thank you for that! | 15:13 |
| danpb | jaypipes: that basically tells you how many vCPUs are realtime vs non-realtime | 15:14 |
| bauzas | jaypipes: AFAIR, we now longer use indexes for RC classes, so I guess I should query the classname string instead of the index for https://review.openstack.org/#/c/386242/3 right? | 15:14 |
| sahid | danpb: hum I think you are talking about the spec "emulator threads" where jaypipes is (so far) only reviewing the fix https://review.openstack.org/#/c/356383/ | 15:14 |
| danpb | so you have to decide that based on what guest OS you're using | 15:14 |
| danpb | so most commonly you'd set that against the image meta | 15:14 |
| bauzas | jaypipes: that means the resource class name needs and is actually unique, right? | 15:14 |
| *** garyk1 has quit IRC | 15:15 | |
| jaypipes | danpb, sahid: or rather... I'm wondering what the purpose of the cpu_realtime_mask thing is at all. Can't we figure out where to place the emulator thread simply by looking at the cpuset of vcpus and choosing one not in that list? | 15:15 |
| *** tongli has quit IRC | 15:15 | |
| bauzas | jaypipes: hence the proposal to have prefixes for the custom resource classes, right, | 15:15 |
| danpb | jaypipes: if you have say 4 vcpus, we need to say how many of those vCPUs are real-time vs non-realtime - that's what cpu_realtime_mask determines - there's no other info to make that decision on | 15:16 |
| jaypipes | bauzas: the resource_classes.name column is unique and indexed. | 15:16 |
| *** lpetrut has joined #openstack-nova | 15:17 | |
| *** anush has joined #openstack-nova | 15:17 | |
| jaypipes | danpb: ah, ok, I got it now. sorry, I wasn't aware that you could have non-realtime vCPU and realtime vCPUs all on the same guest. | 15:17 |
| bauzas | jaypipes: cool then | 15:17 |
| *** diegows_ has quit IRC | 15:18 | |
| danpb | jaypipes: yeah, for linux it is in fact mandatory to have that | 15:18 |
| *** tbachman has joined #openstack-nova | 15:18 | |
| danpb | jaypipes: because its a legacy os and so many parts of userspace are not able to run real time | 15:18 |
| *** mdrabe has quit IRC | 15:18 | |
| danpb | so for linux you essentially always need at least 1 vcpu non-realtime | 15:18 |
| jaypipes | danpb: k, I think I understand this space a bit better now thanks very much for your help. you too sahid! :) | 15:19 |
| danpb | its only the dedicated realtime operating systems where you find all vCPUs real time | 15:19 |
| sahid | jaypipes: no problem, thanks to take time for reviewing it, sorry if my "explanations" confused you | 15:20 |
| * danpb thinks this is a reasonable illustration of linux real-time capabilities ;-P http://3.bp.blogspot.com/-qye_di2qj1Q/U8Cenb0RJoI/AAAAAAAAAgk/vRQaZfFDLjg/s1600/Dog-rocket.png | 15:20 | |
| kashyap | LOL | 15:21 |
| jaypipes | danpb: lol :) | 15:21 |
| jaypipes | danpb: looks like mriedem's Mortimer.. | 15:21 |
| kashyap | Sheesh, that's crazy | 15:21 |
| *** woodster_ has joined #openstack-nova | 15:21 | |
| *** lpetrut has quit IRC | 15:23 | |
| *** pbandark has quit IRC | 15:26 | |
| *** mdrabe has joined #openstack-nova | 15:26 | |
| *** tbachman has quit IRC | 15:28 | |
| *** tbachman has joined #openstack-nova | 15:30 | |
| *** moshele has quit IRC | 15:32 | |
| *** haplo37 has quit IRC | 15:33 | |
| johnthetubaguy | paul-carlton2: are you still working on this live-migration spec? https://review.openstack.org/#/c/347161/ I think raj_singh might have some folks who could help out with that | 15:36 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/392753 | 15:36 |
| *** takashin has left #openstack-nova | 15:37 | |
| *** harlowja_at_home has joined #openstack-nova | 15:38 | |
| *** mmedvede has quit IRC | 15:38 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/os-vif: Updated from global requirements https://review.openstack.org/377527 | 15:38 |
| *** roeyc has quit IRC | 15:39 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-novaclient: Updated from global requirements https://review.openstack.org/392826 | 15:40 |
| *** rama_y has joined #openstack-nova | 15:42 | |
| *** haplo37 has joined #openstack-nova | 15:42 | |
| ftersin | mdbooth: thanks | 15:44 |
| *** tblakes has quit IRC | 15:45 | |
| diana_clarke | mriedem, kashyap: sure, I'll look into tests for https://review.openstack.org/#/c/382449/ | 15:46 |
| mriedem | thanks | 15:46 |
| kashyap | diana_clarke: Thanks! | 15:46 |
| openstackgerrit | Julian proposed openstack/nova: Transform instance.snapshot notifications https://review.openstack.org/330128 | 15:46 |
| openstackgerrit | Julian proposed openstack/nova: Remove redundant VersionedObject Fields https://review.openstack.org/389840 | 15:47 |
| openstackgerrit | Julian proposed openstack/nova: Create schema generation for AddressBase https://review.openstack.org/391546 | 15:47 |
| *** dmellado|lunch is now known as dmellado | 15:47 | |
| *** Zer0Byte__ has quit IRC | 15:48 | |
| *** mmedvede has joined #openstack-nova | 15:50 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: api-ref: distinguish between 2.10 for keypair user_id in responses https://review.openstack.org/392845 | 15:50 |
| *** ekhugen_alt has joined #openstack-nova | 15:50 | |
| *** egarbade- has joined #openstack-nova | 15:51 | |
| *** krtaylor has quit IRC | 15:54 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova-specs: Restore V3 Diagnostics info https://review.openstack.org/357884 | 15:55 |
| mriedem | snikitin: +2 on the diagnostics spec now | 15:55 |
| *** rrecio_ has joined #openstack-nova | 15:56 | |
| *** andymaier has joined #openstack-nova | 15:57 | |
| *** ekuris has quit IRC | 15:57 | |
| kashyap | diana_clarke: If you have functional change related context questions, feel free to ask here or on the review... | 15:57 |
| *** rrecio has quit IRC | 15:59 | |
| *** anushkrishnamurt has joined #openstack-nova | 16:02 | |
| *** harlowja_at_home has quit IRC | 16:03 | |
| mriedem | jaypipes: dansmith: i was going to write a single session recap for the 3 placement/resource providers sessions - is that ok with you or did you want to write those up individually? | 16:04 |
| mriedem | they aren' tsuper fun to write so i figured you'd be ok with me doing it | 16:04 |
| dansmith | fine with me | 16:05 |
| mriedem | ok | 16:05 |
| *** pcaruana has quit IRC | 16:06 | |
| *** oanson has quit IRC | 16:06 | |
| *** mgoddard_ has joined #openstack-nova | 16:07 | |
| jaypipes | mriedem: fine with me, yes | 16:07 |
| bauzas | mriedem: heh, I love why we're asking you to write those | 16:07 |
| mriedem | does not compute | 16:07 |
| *** rcernin has quit IRC | 16:09 | |
| *** anush has quit IRC | 16:10 | |
| *** mgoddard has quit IRC | 16:10 | |
| openstackgerrit | John Garbutt proposed openstack/nova-specs: Live Migration of Rescued Instances https://review.openstack.org/347161 | 16:11 |
| bauzas | jaypipes: given I'm updating https://review.openstack.org/#/c/386242/3/nova/objects/resource_provider.py for now using the RC cache (for getting the class ID), I was thinking how it could be useful for operators troubleshooting resource usage to get the resource class IDs | 16:11 |
| bauzas | jaypipes: given they could be checking SQL, they need to know which class name corresponds with each RC id | 16:11 |
| bauzas | jaypipes: are we envisaging to expose those matchings at the REST API level ? | 16:12 |
| bauzas | if that's in the custom-resource-class spec, please blame me, I need to reread it | 16:12 |
| paul-carlton2 | johnthetubaguy, raj_singh should go ahead I will not have time to work on it | 16:13 |
| johnthetubaguy | paul-carlton2: OK, thanks for confirming, we certainly have a head start with all your patches that are already up there | 16:13 |
| raj_singh | paul-carlton2: Thx | 16:14 |
| *** armax has joined #openstack-nova | 16:14 | |
| johnthetubaguy | pkoniszewski: just a quick heads up, trying to kick start the work on live-migrate of rescued things | 16:14 |
| johnthetubaguy | (see above) | 16:14 |
| *** ssurana has quit IRC | 16:15 | |
| bauzas | jaypipes: well, AFAICS, that's not in https://review.openstack.org/#/c/391959/5/nova/api/openstack/placement/handlers/resource_class.py | 16:15 |
| openstackgerrit | John Garbutt proposed openstack/nova-specs: Live Migration of Rescued Instances https://review.openstack.org/347161 | 16:15 |
| jaypipes | bauzas: no plans to expose the autoincrementing ID outside of the HTTP API, no. | 16:17 |
| bauzas | jaypipes: do you think it could be something operators would need ? | 16:17 |
| bauzas | jaypipes: I'm just with my MySQL workbench tool, and I'd love to know what means "1" as a resource class index :) | 16:18 |
| bauzas | oops s/index/id | 16:18 |
| bauzas | of course, I can guess it now | 16:18 |
| bauzas | but with nested resource classes, I think it would be far more difficult :) | 16:19 |
| jaypipes | bauzas: there's no such thing as nested resource classes :) | 16:20 |
| jaypipes | only nested resource providers. | 16:20 |
| bauzas | oh man, right | 16:20 |
| pkoniszewski | johnthetubaguy: oh, thanks, i will be happy to help there, code, review, tests, anything :) | 16:20 |
| jaypipes | but anyway, I can kind of think it would be useful to add the standard resource class index/ids into the resource_classes table, but I still don't want to leak those IDs out of the API. | 16:20 |
| bauzas | jaypipes: right, that's the main problem I see | 16:21 |
| jaypipes | bauzas: if you wanted to propose a script that added the stadnard resource classes index IDs into resource_classes, I'd be cool with that. | 16:21 |
| bauzas | jaypipes: I just see that as a very specific REST call for operators wanting to troubleshoot | 16:21 |
| johnthetubaguy | pkoniszewski: tests and review would be a great help once we get that moving | 16:21 |
| bauzas | or anything but REST | 16:21 |
| jaypipes | bauzas: just so you could, say, do a join against resource_classes to get the resource class name? | 16:21 |
| bauzas | jaypipes: I'm confused, I thought we were not persisting the non-custom classes? | 16:22 |
| *** baoli has joined #openstack-nova | 16:22 | |
| jaypipes | bauzas: we aren't. | 16:23 |
| bauzas | jaypipes: so, you're proposing to persist those? | 16:23 |
| jaypipes | bauzas: but of course, we could put them in resource_classes and eventually get rid of the entire fields.ResourceClassField | 16:23 |
| jaypipes | bauzas: that, though, wouldn't be backwards-compatible which is the reason for the resource class cache stuff to begin with. | 16:24 |
| bauzas | yeah | 16:24 |
| bauzas | jaypipes: the remediation could be a "dump" tool that would store the mappings into that table, but that's ugly :) | 16:24 |
| bauzas | either way, I need to think more of that | 16:25 |
| bauzas | the fact that we're using an internal cache for the mapping as the single source means that somehow we need to assert that what could be in the DB is just a view, not the real list | 16:26 |
| *** jaosorior has quit IRC | 16:28 | |
| *** bjolo_ has quit IRC | 16:31 | |
| mriedem | cdent: something we never talked about at the summit in https://etherpad.openstack.org/p/ocata-nova-summit-resource-providers-quantitative was documenting the placement api-ref | 16:32 |
| mriedem | cdent: did you have ideas on how to generate those docs? | 16:32 |
| cdent | mriedem: sweat and tears? | 16:32 |
| mriedem | well i thought you said that gabbi could somehow be used for that | 16:32 |
| mriedem | or are we just going to do it like the api-ref we already have for the compute api? | 16:33 |
| *** bvanhav has quit IRC | 16:33 | |
| cdent | not that I recall. you may be thinking of this thing which people asked for: https://review.openstack.org/#/c/370204/ (that's now out of date) | 16:33 |
| *** zul_ has joined #openstack-nova | 16:33 | |
| *** mjura has quit IRC | 16:33 | |
| *** zul_ has quit IRC | 16:33 | |
| cdent | but there's nothing built into gabbi for dumping _good_ docs | 16:33 |
| cdent | so yeah, probably basically like api-ref | 16:33 |
| mriedem | ok | 16:33 |
| mriedem | the compute api-ref uses the api-samples for the examples | 16:34 |
| mriedem | which was basically i think how https://review.openstack.org/#/c/370204/ came up | 16:34 |
| *** Guest48712 has quit IRC | 16:34 | |
| cdent | ah, okay | 16:34 |
| mriedem | i wonder if swagger could be used | 16:35 |
| mriedem | there were reasons it wasn't used for the compute api, which i thought had to do with microversions, | 16:35 |
| mriedem | but sdague and oomichi would know most about that | 16:35 |
| *** zul has quit IRC | 16:35 | |
| cdent | given time and energy there's long termish plans for swagger<->gabbi interoperation | 16:35 |
| *** zul has joined #openstack-nova | 16:36 | |
| cdent | there's no reason why things like the api-samples couldn't be created (in anything, including gabbi) for supporting the placemen docs | 16:36 |
| *** pilgrimstack has quit IRC | 16:36 | |
| *** dikonoor has joined #openstack-nova | 16:36 | |
| cdent | I believe the issues that sdague reported with swagger were that it wasn't great for apis that already exist | 16:36 |
| cdent | as it imposes its own sets of constraints | 16:37 |
| *** mmedvede has quit IRC | 16:37 | |
| mriedem | ok, hmm, maybe we could hook something into the placement tests to just generate the api samples | 16:37 |
| mriedem | to dump those json files | 16:37 |
| mriedem | we have something like that for the nova functional api_samples tests | 16:37 |
| mriedem | anyway, don't want to rathole on it right now | 16:38 |
| cdent | it's on my radar, and I think we can come up with a way to make it relatively straightforward | 16:38 |
| *** zul has quit IRC | 16:39 | |
| *** zul has joined #openstack-nova | 16:39 | |
| *** ayoung has quit IRC | 16:40 | |
| *** pilgrimstack has joined #openstack-nova | 16:40 | |
| *** anushkrishnamurt has quit IRC | 16:40 | |
| *** nkrinner is now known as nkrinner_afk | 16:42 | |
| *** baoli has quit IRC | 16:42 | |
| *** baoli has joined #openstack-nova | 16:42 | |
| cdent | edleafe: are you still in bcn or transit or are you back in the world? | 16:44 |
| *** mmedvede has joined #openstack-nova | 16:45 | |
| *** paul-carlton2 has quit IRC | 16:45 | |
| mriedem | jaypipes: dansmith: i don't think, besides this ironic flavors extra specs thing, that we ever got into what's going to be required to upgrade to ocata wrt resource providers - in https://etherpad.openstack.org/p/ocata-nova-summit-resource-providers-quantitative | 16:45 |
| mriedem | am i missing something? | 16:45 |
| *** sridharg has quit IRC | 16:45 | |
| mriedem | like, for example, does the placement service become required in ocata? | 16:46 |
| mriedem | i thought that was the plan | 16:46 |
| dansmith | yeah, required in ocata | 16:46 |
| dansmith | so stand it up in newton, run it enough to get all your data in there before you roll to ocata | 16:46 |
| mriedem | and you break in ocata if you don't have it because the scheduler is going to rely on the placement data? | 16:47 |
| dansmith | yeah | 16:47 |
| *** artom has quit IRC | 16:48 | |
| dansmith | now that you're saying this, I'm thinking we probably weren't as forceful in our release notes about what "optional" meant and what we expect you to do in newton | 16:48 |
| jaypipes | mriedem: yup, what dansmith is saying. | 16:48 |
| mriedem | yeah the release notes don't really say anything. they just say the compute tries to communicate with it via the service catalog 'placement' entry, but if it's not there, it just logs a warning and doesn't try again | 16:49 |
| mriedem | http://docs.openstack.org/releasenotes/nova/newton.html | 16:50 |
| *** Apoorva has joined #openstack-nova | 16:51 | |
| dansmith | mriedem: I wonder if we're getting to the point where a "nova-manage ready_to_upgrade?" command would be warranted | 16:51 |
| dansmith | we have kindof a lot of irons in the fire at the moment that we could report | 16:51 |
| *** Apoorva has quit IRC | 16:51 | |
| mriedem | yeah maybe | 16:52 |
| cdent | dansmith: I bet people would like that | 16:52 |
| mriedem | no cell mappings, you're not ready, | 16:52 |
| mriedem | no placement service, you're not ready | 16:52 |
| mdbooth | danpb: Quick question about Host.get_connection() | 16:53 |
| *** belmoreira has quit IRC | 16:53 | |
| *** ljxiash has quit IRC | 16:53 | |
| *** artom has joined #openstack-nova | 16:54 | |
| *** bvanhav has joined #openstack-nova | 16:54 | |
| mdbooth | If I want to have the mount manager go up and down automatically based on the connection callback, which seems like a good idea, I need _get_connection to be reentrant | 16:54 |
| mdbooth | which it currently isn't | 16:54 |
| openstackgerrit | Drew Thorstensen (thorst) proposed openstack/nova-specs: PowerVM Initial Driver Work https://review.openstack.org/381772 | 16:55 |
| *** Apoorva has joined #openstack-nova | 16:55 | |
| mdbooth | Because _get_connection unconditionally takes _wrapped_conn_lock, so it deadlocks when it reenters | 16:55 |
| mdbooth | Naively I could set _wrapped_conn before calling the callback, and not take the lock if it's already set | 16:56 |
| mriedem | jroll: jaypipes: can you help clarify L16 in here https://etherpad.openstack.org/p/ocata-nova-summit-resource-providers-quantitative - at least wrt upgrades to ocata? i see ironic has the node.resource_class field in the ironic 1.21 microversion. so when upgrading to ocata, do i need to have a flavor extra_spec with key 'node_class' that maps a flavor to a specific ironic node.resource_class value? | 16:57 |
| mdbooth | However, that would mean exposing the connection before we've registered things against it | 16:57 |
| mdbooth | So that doesn't sound like a good idea | 16:57 |
| mriedem | jroll: jaypipes: or just, the ironic driver in nova when reporting resources will provide the node.resource_class and the resource tracker will use that to create a resource provider of that class? | 16:57 |
| *** krtaylor has joined #openstack-nova | 16:58 | |
| danpb | mdbooth: how about just spawning a greenthread from the callback to populate the mount manager | 16:59 |
| danpb | i don't think you need to populte mount manager synchronously from the callback | 17:00 |
| *** tlian has quit IRC | 17:00 | |
| mdbooth | danpb: Gimme a minute to consider that | 17:00 |
| danpb | and indeed it is probably a bad idea to do it synchronously | 17:00 |
| danpb | as if populating mount manager talks to libvirt, and libvirt drops in the middle | 17:00 |
| danpb | then you could re-enter mount manager setup code | 17:00 |
| mdbooth | MountManager locks that | 17:01 |
| jroll | mriedem: the latter, afaik, but the notes appear to say the former (I wasn't in that session) | 17:01 |
| mdbooth | ... which would also deadlock | 17:01 |
| mdbooth | fun | 17:01 |
| jroll | mriedem: I think the extra_spec thing was to support the weird "flavor matches multiple node classes" use case | 17:02 |
| *** pilgrimstack has quit IRC | 17:02 | |
| *** mgoddard has joined #openstack-nova | 17:02 | |
| *** mgoddard_ has quit IRC | 17:02 | |
| mriedem | ok i'm not privy to that weirdness | 17:03 |
| *** pilgrimstack has joined #openstack-nova | 17:03 | |
| mriedem | and https://review.openstack.org/#/c/312696/ doesn't mention it | 17:03 |
| *** oomichi_ has joined #openstack-nova | 17:04 | |
| *** sree has quit IRC | 17:06 | |
| *** andreas_s has quit IRC | 17:06 | |
| *** tlian has joined #openstack-nova | 17:07 | |
| jaypipes | sahid: k, review of the rt patch done. | 17:09 |
| jaypipes | sahid: if you can fix up those comments and doc issues, I can re-review fairly quickly and push it. | 17:10 |
| *** sree has joined #openstack-nova | 17:10 | |
| jaypipes | mriedem: yes, the resource tracker will be creating inventory/allocation records for the ironic nodes (inventory records will use the node.node_class as the custom resource class). we still need to figure out the "what to do with the flavor extraspec thing". | 17:11 |
| openstackgerrit | Chris Dent proposed openstack/nova: Add rudimentary CORS support to placement API https://review.openstack.org/392891 | 17:12 |
| *** tbachman has quit IRC | 17:12 | |
| *** ftersin has quit IRC | 17:14 | |
| *** sree has quit IRC | 17:15 | |
| openstackgerrit | Daniel Berrange proposed openstack/os-vif: host_info: fix get_common_version method on HostVIFInfo https://review.openstack.org/392892 | 17:15 |
| jroll | jaypipes: can we just not do a flavor extra spec thing | 17:15 |
| *** ssurana has joined #openstack-nova | 17:17 | |
| *** rrecio_ has quit IRC | 17:19 | |
| *** ihrachys has quit IRC | 17:20 | |
| *** tblakes has joined #openstack-nova | 17:20 | |
| *** harlowja has quit IRC | 17:21 | |
| mriedem | edleafe: you and jaypipes got this all figured out right? http://lists.openstack.org/pipermail/openstack-dev/2016-October/105965.html | 17:22 |
| *** sree has joined #openstack-nova | 17:24 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova-specs: Resource providers: custom resource classes https://review.openstack.org/312696 | 17:27 |
| openstackgerrit | Julian proposed openstack/nova: Remove redundant VersionedObject Fields https://review.openstack.org/389840 | 17:29 |
| cdent | jaypipes: on an Inventory object of the total-reserved * allocation ratio is called "capacity" what's the overarching name of the thing described by min_unit, max_unit, step_size? | 17:31 |
| cdent | Inventory constraints? | 17:31 |
| *** lucasagomes is now known as lucas-afk | 17:32 | |
| *** tlian has quit IRC | 17:33 | |
| *** bvanhav has quit IRC | 17:34 | |
| *** bvanhav has joined #openstack-nova | 17:35 | |
| *** sree has quit IRC | 17:37 | |
| *** paw has quit IRC | 17:38 | |
| *** baoli has quit IRC | 17:38 | |
| *** fragatina has quit IRC | 17:40 | |
| *** tlian has joined #openstack-nova | 17:40 | |
| *** tblakes has quit IRC | 17:41 | |
| *** tbachman has joined #openstack-nova | 17:43 | |
| *** harlowja has joined #openstack-nova | 17:44 | |
| *** tlian has quit IRC | 17:50 | |
| *** sree has joined #openstack-nova | 17:51 | |
| *** rmcall has joined #openstack-nova | 17:53 | |
| *** marst has joined #openstack-nova | 17:54 | |
| cfriesen__ | jaypipes: hopefully this has come up already, but for CPU/RAM (and probably disk) we don't want to allow an instance to overcommit against itself. we do want it to be able to overcommit against other instances though. | 17:54 |
| *** anush has joined #openstack-nova | 17:55 | |
| *** derekh has quit IRC | 17:55 | |
| openstackgerrit | Feodor Tersin proposed openstack/nova: Fix cold migration with qcow2 ephemeral disks https://review.openstack.org/346263 | 17:58 |
| *** dikonoor has quit IRC | 17:58 | |
| *** sree has quit IRC | 18:00 | |
| *** tlian has joined #openstack-nova | 18:01 | |
| *** ralonsoh has quit IRC | 18:01 | |
| *** Jeffrey4l has quit IRC | 18:02 | |
| *** haplo37_ has quit IRC | 18:05 | |
| jaypipes | mriedem: yes, we sorted out the nested resource providers modeling. | 18:06 |
| jaypipes | cfriesen__: I'm not following you.. could you elaborate please? | 18:06 |
| jaypipes | cdent: sure, that sounds like a good term. | 18:07 |
| jaypipes | jroll: I'm still thinking on it :) | 18:07 |
| jaypipes | jroll: working on the resource tracker integrations for custom resource classes right now. should have a much better idea of end game after that's done. | 18:07 |
| *** lpetrut has joined #openstack-nova | 18:08 | |
| *** haplo37_ has joined #openstack-nova | 18:08 | |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: implement ResourceClass.create() https://review.openstack.org/386844 | 18:08 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: raise KeyError unknown resource class https://review.openstack.org/390062 | 18:08 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: REST API for resource classes https://review.openstack.org/391959 | 18:08 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: adds ResourceClass.get_by_name() https://review.openstack.org/391917 | 18:08 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: implement destroy() resource class https://review.openstack.org/391918 | 18:09 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: add ResourceClass and ResourceClassList https://review.openstack.org/386843 | 18:09 |
| jaypipes | mriedem: ^^ whack-a-mole fixed. | 18:09 |
| mriedem | that poor mole | 18:09 |
| jaypipes | indeed. | 18:09 |
| jaypipes | cdent: in response to your comment about the "why microversion the aggregates stuff if either way it will return a 404", I fall back to the argument I made yesterday that the microversion history is a way to document changes to the API so just go ahead and make a microversion for it. | 18:11 |
| cfriesen__ | jaypipes: what cdent above said reminded me of it, so I thought I'd make sure it was still possible with the new resource providers. Basically there needs to be enough _total_ resources for the instance to fit (without factoring in overcommit). | 18:12 |
| cfriesen__ | jaypipes: then there needs to be enough free space (factoring in overcommit) for the instance to fit. | 18:12 |
| jaypipes | cfriesen__: ah, yes, absolutely. | 18:12 |
| jaypipes | cfriesen__: sorry for some reason I thought you were referring to the conversation about realtime pinning I had with sahid and danpb earlier :) | 18:13 |
| jaypipes | lbeliveau: ok, finally now able to get to your review! | 18:14 |
| lbeliveau | jaypipes: thanks a bunch ! | 18:14 |
| *** anush has quit IRC | 18:14 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: fix incorrect host cpus giving to emulator threads when RT https://review.openstack.org/356383 | 18:14 |
| jaypipes | lbeliveau: sorry for the delay! | 18:14 |
| lbeliveau | jaypipes: no worries, know you are busy | 18:15 |
| openstackgerrit | Gábor Antal proposed openstack/nova: Transform instance.rebuild notification https://review.openstack.org/384922 | 18:15 |
| *** jvgrant_ has joined #openstack-nova | 18:15 | |
| sahid | jaypipes: thanks for your review, i addressed some of them and reply to your question about cpu_online | 18:15 |
| jaypipes | lbeliveau: btw, do you know about using /g in the mysql client? | 18:15 |
| sahid | will do the refactor tomorrow | 18:15 |
| *** paw has joined #openstack-nova | 18:15 | |
| cdent | jaypipes: roger on the microversionoger | 18:15 |
| mnestratov | hi all | 18:15 |
| lbeliveau | jaypipes: haven't use it, let me check it out | 18:16 |
| mnestratov | trivial https://review.openstack.org/#/c/388643/ needs one more +2 | 18:16 |
| jaypipes | lbeliveau: if you use /g at the end of your SQL statement in the mysql client, it displays in columnar format instead of tabular format, making copy/paste into launchpad bugs much nicer :) | 18:16 |
| jaypipes | lbeliveau: /g instead of ; | 18:16 |
| openstackgerrit | Lee Yarwood proposed openstack/nova: encryptors: Replace mangled passphrases for LuksEncryptor volumes https://review.openstack.org/386670 | 18:16 |
| lbeliveau | jaypipes: haaa good to know, thanks !!! | 18:16 |
| jaypipes | np :) | 18:16 |
| *** openstackgerrit has quit IRC | 18:18 | |
| *** sahid has quit IRC | 18:18 | |
| *** openstackgerrit has joined #openstack-nova | 18:18 | |
| *** jvgrant has quit IRC | 18:18 | |
| lbeliveau | jaypipes: btw I would be glad to work on resource provider if you are looking for others to contribute, can do PCI and numa or anything (even working on unit tests would be good) | 18:20 |
| cdent | jaypipes, dansmith: I'm working on https://bugs.launchpad.net/nova/+bug/1623545 but just realized: https://github.com/openstack/nova/blob/74fadb4bcf272ab93a7bf8270f087adf588735d9/nova/scheduler/client/report.py#L73 | 18:22 |
| openstack | Launchpad bug 1623545 in OpenStack Compute (nova) "placement API allocation handling does not check min_unit, max_unit, step_size" [Medium,Confirmed] - Assigned to Ukesh (ukeshkumar) | 18:22 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add schedule_and_build_instances conductor method https://review.openstack.org/367557 | 18:23 |
| openstackgerrit | Dan Smith proposed openstack/nova: WIP Add schedule_and_build_instances RPC method https://review.openstack.org/379320 | 18:23 |
| openstackgerrit | Dan Smith proposed openstack/nova: WIP Move instance creation to conductor https://review.openstack.org/319379 | 18:23 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add CellDatabase fixture https://review.openstack.org/392925 | 18:23 |
| *** GB21 has quit IRC | 18:24 | |
| *** andymaier has quit IRC | 18:24 | |
| *** edand has joined #openstack-nova | 18:25 | |
| *** anush has joined #openstack-nova | 18:26 | |
| openstackgerrit | Gábor Antal proposed openstack/nova: Transform instance.rebuild notification https://review.openstack.org/384922 | 18:27 |
| sneti | danpb, can you please check this: https://review.openstack.org/#/c/338553/. I've added response to your comment | 18:29 |
| jaypipes | lbeliveau: https://review.openstack.org/#/c/363884 reviewed. push up those two little fixes and I'm good +2ing. | 18:31 |
| *** _ducttape_ has joined #openstack-nova | 18:31 | |
| lbeliveau | jaypipes: thanks, looking at it ... | 18:32 |
| *** rrecio has joined #openstack-nova | 18:34 | |
| *** edand has quit IRC | 18:34 | |
| *** ducttape_ has quit IRC | 18:35 | |
| *** jed56 has quit IRC | 18:37 | |
| jaypipes | cdent: yeah, we're hard-coding the min-unit/max-unit stuff at the moment. Idea long-term is to have that stuff configurable all via the placement API of course. | 18:40 |
| cdent | jaypipes: sure, but the point is that if I fix the bug, then the resource tracker will start failing | 18:40 |
| cdent | I have code ready to push which makes it obvious, so if you give me a sec to write the commit message it'll be clear | 18:40 |
| *** sree has joined #openstack-nova | 18:41 | |
| jaypipes | cdent: yeah, max_unit for VCPU should be the total number of cores on the machine. for memory, should be the total amount of memory on the machine. | 18:42 |
| *** marst has quit IRC | 18:42 | |
| jaypipes | cdent: for disk, should be the maximum volume size. | 18:42 |
| jaypipes | s/volume/disk/ | 18:42 |
| jaypipes | cdent: for the record, my original patch that added inventory records in the resource tracker did just that: https://review.openstack.org/#/c/363583/1/nova/scheduler/client/report.py | 18:44 |
| cdent | ;) | 18:44 |
| jaypipes | cdent: I missed that when reviewing the patch that actually landed. | 18:44 |
| *** oomichi_ has quit IRC | 18:44 | |
| *** baoli has joined #openstack-nova | 18:45 | |
| *** sree has quit IRC | 18:46 | |
| *** ociuhandu has quit IRC | 18:46 | |
| *** gszasz has quit IRC | 18:48 | |
| *** matrohon has joined #openstack-nova | 18:50 | |
| *** matrohon has quit IRC | 18:50 | |
| * cdent glares at gerrit | 18:50 | |
| *** markus_z has quit IRC | 18:51 | |
| *** portdirect is now known as portdirect_away | 18:51 | |
| *** sudipto has quit IRC | 18:52 | |
| *** sudipto_ has quit IRC | 18:52 | |
| openstackgerrit | Chris Dent proposed openstack/nova: DNM [placement] Enforce min_unit, max_unit and step_size https://review.openstack.org/392933 | 18:52 |
| cdent | jaypipes ^^ . Is it worth making a bug against the resource tracker? | 18:53 |
| *** _ducttape_ has quit IRC | 18:58 | |
| *** sree has joined #openstack-nova | 18:59 | |
| jaypipes | cdent: sure, go for it. | 18:59 |
| *** ducttape_ has joined #openstack-nova | 18:59 | |
| *** sree_ has joined #openstack-nova | 19:01 | |
| *** sree_ is now known as Guest81782 | 19:02 | |
| cdent | jaypipes: https://bugs.launchpad.net/nova/+bug/1638681 | 19:03 |
| openstack | Launchpad bug 1638681 in OpenStack Compute (nova) "resource tracker sets wrong max_unit in placement Inventory" [Undecided,Triaged] | 19:03 |
| cdent | that's the end of my day, I'll check back in on that, and the enforcement code tomorrow | 19:03 |
| * cdent waves | 19:03 | |
| *** cdent has quit IRC | 19:03 | |
| *** sree has quit IRC | 19:03 | |
| *** lamt has joined #openstack-nova | 19:04 | |
| *** ducttape_ has quit IRC | 19:04 | |
| *** andymaier has joined #openstack-nova | 19:05 | |
| *** andymaier has quit IRC | 19:06 | |
| *** Guest81782 has quit IRC | 19:06 | |
| *** akshai has quit IRC | 19:08 | |
| *** Daisy has joined #openstack-nova | 19:09 | |
| openstackgerrit | Gábor Antal proposed openstack/nova: Transform instance.create notification https://review.openstack.org/384621 | 19:09 |
| *** slaweq_ has joined #openstack-nova | 19:10 | |
| *** anush has quit IRC | 19:11 | |
| *** knikolla has left #openstack-nova | 19:12 | |
| *** moshele has joined #openstack-nova | 19:13 | |
| *** knikolla has joined #openstack-nova | 19:14 | |
| *** Daisy has quit IRC | 19:14 | |
| *** ducttape_ has joined #openstack-nova | 19:14 | |
| dansmith | bauzas: still around? | 19:19 |
| *** jerrygb has quit IRC | 19:21 | |
| *** jerrygb has joined #openstack-nova | 19:22 | |
| openstackgerrit | Ludovic Beliveau proposed openstack/nova: PCI: Fix PCI with fully qualified address https://review.openstack.org/363884 | 19:22 |
| lbeliveau | jaypipes: ^ | 19:23 |
| *** marst has joined #openstack-nova | 19:23 | |
| *** jerrygb has quit IRC | 19:26 | |
| *** lamt has quit IRC | 19:27 | |
| *** annegentle has quit IRC | 19:27 | |
| *** mmedvede has quit IRC | 19:28 | |
| *** tblakes has joined #openstack-nova | 19:29 | |
| *** moshele has quit IRC | 19:30 | |
| *** ayoung has joined #openstack-nova | 19:34 | |
| *** tblakes has quit IRC | 19:35 | |
| *** oanson has joined #openstack-nova | 19:36 | |
| *** liverpooler has quit IRC | 19:37 | |
| jaypipes | lbeliveau: done. | 19:38 |
| lbeliveau | thanks | 19:38 |
| lbeliveau | jaypipes: ping me if you need cheap labor on resource provider ;) | 19:38 |
| *** READ10 has quit IRC | 19:40 | |
| *** bvanhav has quit IRC | 19:42 | |
| *** jerrygb has joined #openstack-nova | 19:42 | |
| jaypipes | lbeliveau: :) | 19:43 |
| jaypipes | vladikr: ah, doh, yeah, totally didn't think of that. thanks! | 19:44 |
| *** anush has joined #openstack-nova | 19:44 | |
| *** rcernin has joined #openstack-nova | 19:45 | |
| *** sfinucan has left #openstack-nova | 19:45 | |
| *** sfinucan has joined #openstack-nova | 19:45 | |
| *** sridharg has joined #openstack-nova | 19:46 | |
| *** dave-mcc_ has joined #openstack-nova | 19:47 | |
| *** dave-mccowan has quit IRC | 19:48 | |
| *** rtheis_ has quit IRC | 19:49 | |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: REST API for resource classes https://review.openstack.org/391959 | 19:49 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: adds ResourceClass.get_by_name() https://review.openstack.org/391917 | 19:49 |
| openstackgerrit | Jay Pipes proposed openstack/nova: placement: implement destroy() resource class https://review.openstack.org/391918 | 19:49 |
| *** dtp has joined #openstack-nova | 19:50 | |
| *** danpb has quit IRC | 19:50 | |
| *** matrohon has joined #openstack-nova | 19:54 | |
| *** openstack has joined #openstack-nova | 19:59 | |
| dansmith | bauzas: so, I'm massively confused by what was going on in this patch around schedule_instances and the request spec stuff: https://review.openstack.org/#/c/367557/11 | 20:00 |
| dansmith | bauzas: I've mostly backed it all out with a little hacky bit to make it work | 20:00 |
| dansmith | bauzas: running pep8 now, but maybe you could take a look when I push it up? | 20:01 |
| bauzas | dansmith: lemme look | 20:01 |
| dansmith | bauzas: incoming | 20:03 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add schedule_and_build_instances conductor method https://review.openstack.org/367557 | 20:03 |
| openstackgerrit | Dan Smith proposed openstack/nova: WIP Add schedule_and_build_instances RPC method https://review.openstack.org/379320 | 20:03 |
| openstackgerrit | Dan Smith proposed openstack/nova: WIP Move instance creation to conductor https://review.openstack.org/319379 | 20:03 |
| *** mmedvede has joined #openstack-nova | 20:04 | |
| bauzas | dansmith: fair enough | 20:04 |
| bauzas | dansmith: which point in particular is confusing you? | 20:04 |
| dansmith | bauzas: this: https://review.openstack.org/#/c/367557/9/nova/scheduler/utils.py | 20:05 |
| dansmith | and, | 20:05 |
| dansmith | https://review.openstack.org/#/c/367557/9/nova/conductor/manager.py@789 | 20:05 |
| *** anushkrishnamurt has joined #openstack-nova | 20:05 | |
| bauzas | sec, otp :( | 20:05 |
| *** anush has quit IRC | 20:05 | |
| dansmith | so if you can look at that, and then check out by "XXX is this right?" section here: https://review.openstack.org/#/c/367557/12/nova/conductor/manager.py@855 | 20:06 |
| dansmith | that would be helpful | 20:06 |
| dansmith | s/by/my/ | 20:06 |
| openstackgerrit | Lee Yarwood proposed openstack/nova: encryptors: Workaround mangled passphrases https://review.openstack.org/386670 | 20:07 |
| *** anushkrishnamurt has quit IRC | 20:09 | |
| *** Zara has quit IRC | 20:11 | |
| *** knikolla has left #openstack-nova | 20:11 | |
| *** knikolla has joined #openstack-nova | 20:12 | |
| *** devananda has quit IRC | 20:12 | |
| *** mriedem has quit IRC | 20:13 | |
| *** openstackgerrit has quit IRC | 20:18 | |
| *** openstackgerrit has joined #openstack-nova | 20:18 | |
| *** ihrachys has joined #openstack-nova | 20:19 | |
| *** slaweq_ has quit IRC | 20:19 | |
| bauzas | dansmith: back, sorry loooong family call :p | 20:20 |
| * dansmith is impressed with the number of calls bauzas gets | 20:20 | |
| bauzas | oh man, really ? :) | 20:21 |
| cburgess | Yeah for some reason I would think dansmith would hate being on the phone much less for a long period of time. | 20:23 |
| *** akshai has joined #openstack-nova | 20:24 | |
| bauzas | dansmith: so, about https://review.openstack.org/#/c/367557/12/nova/conductor/manager.py@855 I'm curious why we're not rather passing the RequestSpec object directly in the new schedule_and_run_instances() method | 20:24 |
| dansmith | cburgess: heh | 20:24 |
| dansmith | bauzas: we are? | 20:24 |
| bauzas | dansmith: the main problem I had when writing that series is that the scheduler utils methods are still using the legacy dicts | 20:24 |
| *** marst has quit IRC | 20:25 | |
| bauzas | dansmith: so, I wrote the to_legacy() helpers for passing the dicts | 20:25 |
| *** marst has joined #openstack-nova | 20:25 | |
| dansmith | bauzas: right, there are a bunch of TODOs everywhere, which probably need addressing | 20:25 |
| dansmith | bauzas: but that's my point -- that's why I'm passing the legacy-ified version | 20:25 |
| bauzas | dansmith: yeah, that's something I needed to cleanup, and something edleafe wanted to help | 20:25 |
| bauzas | dansmith: about your point, you want to set the instance uuid, right? | 20:26 |
| bauzas | dansmith: if so, why ? | 20:26 |
| *** akshai_ has joined #openstack-nova | 20:26 | |
| bauzas | dansmith: because the instance uuid should already be set by the API, nope ? | 20:26 |
| dansmith | bauzas: I need to else the server groups setup thing fails | 20:26 |
| cfriesen__ | possibly a weird question...but does anyone know of a heat template for setting up a multi-node devstack environment? | 20:27 |
| dansmith | bauzas: this is the patch to move instance creation out of the api.. I don't even understand what it means for a reqspec to have a uuid, and num_instances=>1 | 20:27 |
| bauzas | dansmith: mmm, I'm pretty sure that the API would provide this when calling this RPC method | 20:27 |
| dansmith | bauzas: but stuff below _schedule_instances() seems to need it there | 20:27 |
| dansmith | bauzas: where? I don't see instance uuid in the ReqSpec object | 20:28 |
| bauzas | dansmith: I mean, the request_spec dict passed as a param would have an uuid | 20:28 |
| bauzas | dansmith: in the instance props | 20:28 |
| dansmith | oh, heh, there it is | 20:28 |
| dansmith | looked a minute ago and missed it | 20:28 |
| dansmith | bauzas: well, I dont' see instance_props either | 20:28 |
| bauzas | dansmith: sec | 20:29 |
| *** smatzek has quit IRC | 20:29 | |
| *** akshai has quit IRC | 20:29 | |
| bauzas | dansmith: see the comment here https://github.com/openstack/nova/blob/master/nova/objects/request_spec.py#L282 | 20:31 |
| *** devananda has joined #openstack-nova | 20:31 | |
| bauzas | dansmith: we're not persisting all the instance object in the ReqSpec | 20:31 |
| dansmith | ...right | 20:31 |
| *** kiran-r has joined #openstack-nova | 20:31 | |
| bauzas | dansmith: so, when we're returning a legacy dict, it only gets what the ReqSpec object knows | 20:31 |
| *** sridharg has quit IRC | 20:31 | |
| openstackgerrit | Kam Nasim proposed openstack/nova: Fixes exception during a swap to no swap resize https://review.openstack.org/300616 | 20:32 |
| dansmith | bauzas: so this seems to work for me: http://pastebin.com/67bhMfF2 | 20:32 |
| dansmith | does that look right? | 20:32 |
| bauzas | dansmith: we could have been calling the instance object to get all the fields that the ReqSpec doesn't know, but that's not necessary given the goal of that ReqSpec object is just to persist what is needed by the scheduler | 20:32 |
| *** mriedem has joined #openstack-nova | 20:32 | |
| dansmith | bauzas: just using the uuid from the buildrequest in the test so it matches | 20:32 |
| bauzas | dansmith: yup, zactly | 20:33 |
| dansmith | okay | 20:33 |
| bauzas | dansmith: you need somehow to pass a right ReqSpec | 20:33 |
| bauzas | for the UTs | 20:33 |
| *** eharney has quit IRC | 20:33 | |
| dansmith | I didn't see instance_uuid in there before so I didn't think that was a thing I could do | 20:34 |
| bauzas | dansmith: so, when you're writing tests, just create a Spec object that provides fields what are looked up | 20:34 |
| dansmith | but I just missed it | 20:34 |
| bauzas | dansmith: okay | 20:34 |
| *** annegentle has joined #openstack-nova | 20:34 | |
| bauzas | dansmith: tbc, that RPC method called schedule_and_build_instances() would get called with request_spec and filter_properties that are in the API | 20:35 |
| dansmith | yeah | 20:35 |
| bauzas | dansmith: I just think we can kill in fire the filter_props parameter in that method | 20:35 |
| *** moshele has joined #openstack-nova | 20:36 | |
| bauzas | dansmith: given that if the request_spec parameter is supporting a RequestSpec object, those filter_props dict keys could be found in that object | 20:36 |
| openstackgerrit | Dan Smith proposed openstack/nova: Return build_requests instead of instances https://review.openstack.org/374267 | 20:36 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add schedule_and_build_instances conductor method https://review.openstack.org/367557 | 20:36 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add CellDatabase fixture https://review.openstack.org/392925 | 20:36 |
| openstackgerrit | Dan Smith proposed openstack/nova: WIP Add schedule_and_build_instances RPC method https://review.openstack.org/379320 | 20:36 |
| openstackgerrit | Dan Smith proposed openstack/nova: WIP Move instance creation to conductor https://review.openstack.org/319379 | 20:36 |
| dansmith | okay well, comment away ^ :) | 20:36 |
| bauzas | dansmith: sec, just thinking of something | 20:37 |
| bauzas | dansmith: are we planning to reschedule by calling that method ? | 20:37 |
| bauzas | dansmith: or would we be calling the existing conductor method ? | 20:37 |
| dansmith | the existing one, IIRC | 20:37 |
| dansmith | the spec spec-ifies that IIRC | 20:37 |
| bauzas | you mean ? https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L1826 | 20:39 |
| bauzas | dansmith: again, I need to cleanup that, my strategy being to pass down a RequestSpec object down to the compute manager, so that it can reschedule by passing that object back to the conductor | 20:40 |
| dansmith | yeah, it would be good to be able to avoid polluting this new call with history | 20:40 |
| bauzas | dansmith: but I'd appreciate a clean conductor API that would certainly accept only a RequestSpec object instead of the legacy tuple (request_spec, filter_props) | 20:41 |
| bauzas | dansmith: I said that, because you could get old history passed bottom/up to the conductor from the compute in case it's a reschedule | 20:41 |
| *** tblakes has joined #openstack-nova | 20:41 | |
| dansmith | https://review.openstack.org/#/c/381275/4/specs/ocata/approved/cells-scheduling-interaction.rst@73 | 20:42 |
| *** moshele has quit IRC | 20:42 | |
| bauzas | dansmith: if we're clear with the path that this new "api_db conductor" method will just be top/down and then calling the regular build_instances conductor method that can be called in a reschedule, I'm fine | 20:42 |
| dansmith | I think the spec is clear there | 20:42 |
| *** matrohon has quit IRC | 20:43 | |
| bauzas | dansmith: yeah, I agree, the spec is clear, I just wanted to stress that point at the implementation phase :) | 20:43 |
| bauzas | having two conductor processes could be confusing, but I'm not | 20:43 |
| *** matrohon has joined #openstack-nova | 20:43 | |
| *** marst_ has joined #openstack-nova | 20:44 | |
| *** oanson has quit IRC | 20:46 | |
| *** marst has quit IRC | 20:47 | |
| *** eharney has joined #openstack-nova | 20:48 | |
| *** edmondsw has quit IRC | 20:50 | |
| *** matrohon has quit IRC | 20:51 | |
| *** matrohon has joined #openstack-nova | 20:51 | |
| *** matrohon has quit IRC | 20:52 | |
| *** matrohon has joined #openstack-nova | 20:52 | |
| dtp | is there a cells v2 meeting this week? | 20:52 |
| *** matrohon has quit IRC | 20:53 | |
| *** matrohon has joined #openstack-nova | 20:53 | |
| eandersson | btw cells! anyone with insight into cells that could take a look at https://review.openstack.org/#/c/390984/ ? | 20:53 |
| bauzas | dtp: nope, see http://lists.openstack.org/pipermail/openstack-dev/2016-November/106578.html | 20:55 |
| dtp | thanks | 20:55 |
| *** lpetrut has quit IRC | 20:55 | |
| *** krtaylor has quit IRC | 20:55 | |
| *** Sukhdev has joined #openstack-nova | 20:57 | |
| *** akshai_ has quit IRC | 20:57 | |
| *** matrohon has quit IRC | 21:01 | |
| *** slaweq_ has joined #openstack-nova | 21:01 | |
| *** matrohon has joined #openstack-nova | 21:02 | |
| dansmith | dtp: no, we canceled it (on the ML) | 21:03 |
| *** akshai has joined #openstack-nova | 21:03 | |
| dansmith | heh, yeah, late to that convo | 21:03 |
| *** esberglu has quit IRC | 21:03 | |
| dtp | :) i've joined the ML now | 21:04 |
| *** esberglu has joined #openstack-nova | 21:04 | |
| *** sach-m has joined #openstack-nova | 21:05 | |
| *** sree has joined #openstack-nova | 21:07 | |
| *** esberglu has quit IRC | 21:08 | |
| *** sach-m has quit IRC | 21:08 | |
| *** oomichi_online has joined #openstack-nova | 21:09 | |
| *** sach-m has joined #openstack-nova | 21:09 | |
| *** sree has quit IRC | 21:12 | |
| *** xyang1 has quit IRC | 21:12 | |
| *** sach-m_ has joined #openstack-nova | 21:13 | |
| *** sach-m has quit IRC | 21:13 | |
| mriedem | you know what's terrible? 3 5 year old girls all yelling at alexa to play them songs. | 21:15 |
| * mriedem head explodes | 21:15 | |
| *** tblakes has quit IRC | 21:18 | |
| *** sach-m_ has quit IRC | 21:19 | |
| *** sach-m has joined #openstack-nova | 21:20 | |
| *** lpetrut has joined #openstack-nova | 21:20 | |
| sach-m | Hi Everyone, I am sure you all watched the hybrid cloud demo @Barcelona keynote. To take the effort further, we are proposing a nova driver for AWS. Blueprints are here: https://blueprints.launchpad.net/nova/+spec/nova-aws. Thanks! | 21:20 |
| *** sambetts is now known as sambetts|afk | 21:23 | |
| tbachman | mriedem: fight back. Blare ESPN on all possible displays. | 21:23 |
| JayF | mriedem: sounds like the only worst thing would be only one of them doing it, and you having to listen to some frozen song for the 900th time | 21:24 |
| *** dave-mcc_ has quit IRC | 21:24 | |
| mriedem | ffs, now there is fighting and crying over which song is picked | 21:24 |
| *** lpetrut has quit IRC | 21:24 | |
| *** nherciu_ has quit IRC | 21:25 | |
| mriedem | i do have the ability to play some slayer and drown them out... | 21:25 |
| tbachman | reintroduce them to Headbanger’s Ball | 21:26 |
| tbachman | (er, “introduce”) | 21:27 |
| jlk | mriedem: welcome to our glorious future. | 21:30 |
| *** sach-m has quit IRC | 21:30 | |
| *** sach-m has joined #openstack-nova | 21:30 | |
| *** thorst has quit IRC | 21:31 | |
| *** maticue has quit IRC | 21:31 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Transform instance.shutdown notifications https://review.openstack.org/329214 | 21:32 |
| *** esberglu has joined #openstack-nova | 21:34 | |
| *** sach-m has quit IRC | 21:34 | |
| bauzas | heh, I was just about replying to sach-m | 21:34 |
| bauzas | https://twitter.com/sylvainbauza/status/791197312051732480 | 21:35 |
| *** sach-m has joined #openstack-nova | 21:36 | |
| *** esberglu has quit IRC | 21:38 | |
| *** fandi has joined #openstack-nova | 21:39 | |
| *** zzelle has joined #openstack-nova | 21:40 | |
| *** browne has quit IRC | 21:41 | |
| mriedem | bauzas: that omni thing is using the openstack api, it's just using a virt driver that's talking to aws | 21:42 |
| mriedem | it's the same idea as jacket | 21:42 |
| mriedem | except with code | 21:42 |
| mriedem | same as fusion compute | 21:42 |
| mriedem | same as powervc | 21:42 |
| bauzas | mriedem: the real problem I see with that is that's a cloud-ish driver | 21:43 |
| oomichi_online | mriedem: can you help https://review.openstack.org/#/c/387644/ ? tempest patch (https://review.openstack.org/#/c/381362/) requires for the merger | 21:44 |
| bauzas | mriedem: for example, how can we know which AWS resources do we still have unless by saying quotas ? | 21:44 |
| *** akshai has quit IRC | 21:46 | |
| *** sach-m has quit IRC | 21:46 | |
| *** jheroux has quit IRC | 21:46 | |
| *** sach-m has joined #openstack-nova | 21:47 | |
| *** akshai has joined #openstack-nova | 21:47 | |
| *** maticue has joined #openstack-nova | 21:48 | |
| oomichi_online | bauzas: will nova become a proxy of aws?? | 21:48 |
| *** sach-m has quit IRC | 21:49 | |
| bauzas | oomichi: if you mean that OpenStack will be the controlplane and AWS the dataplane, then yes if using that AWS drivre | 21:49 |
| *** sach-m has joined #openstack-nova | 21:49 | |
| *** sach-m has quit IRC | 21:49 | |
| *** sach-m has joined #openstack-nova | 21:49 | |
| oomichi_online | bauzas: oh, interesting idea | 21:50 |
| bauzas | oomichi_online: well, sure, but I'm not sure a nova driver should be fine for that | 21:50 |
| openstackgerrit | Erik Olof Gunnar Andersson proposed openstack/nova: Re-worked ram filter and weights to support overcommit https://review.openstack.org/390984 | 21:50 |
| oomichi_online | bauzas: heh, yeah. I have the same question for that | 21:51 |
| bauzas | oomichi_online: the abstraction means that the operator knows the resources their have | 21:51 |
| bauzas | oomichi_online: while here, this is not | 21:51 |
| *** sach-m has quit IRC | 21:51 | |
| *** sach-m has joined #openstack-nova | 21:52 | |
| bauzas | here, that would mean that there would only be one compute node with all the resources being the AWS quotas, I guess | 21:52 |
| oomichi_online | bauzas: not sure that "nova should be abstraction layer for all resources including AWS" | 21:52 |
| oomichi_online | interesting anyways :) | 21:53 |
| mriedem | oomichi_online: done | 21:53 |
| bauzas | oomichi_online: if you see their repos, it's not only nova | 21:53 |
| oomichi_online | mriedem: thanks :) | 21:53 |
| *** eharney has quit IRC | 21:53 | |
| bauzas | oomichi_online: nova is just proxying calls to EC2 | 21:53 |
| bauzas | the other AWS resources being drived by Cinder or Neutron | 21:53 |
| bauzas | driven* | 21:54 |
| mriedem | sure i wouldn't actually use it | 21:54 |
| mriedem | but it's not a new concept | 21:54 |
| oomichi_online | bauzas: oh, they has some implementation already? cool | 21:55 |
| bauzas | mriedem: I agree, I just wonder if that concept is something we still love :) | 21:55 |
| mriedem | i don't love it | 21:56 |
| mriedem | personally | 21:56 |
| mriedem | some customer might, until they run into resource sync issues | 21:56 |
| mriedem | a service provider would probably hate dealing with it | 21:56 |
| bauzas | mriedem: that's my point | 21:56 |
| mriedem | but red hat doesn't have to support it so meh :) | 21:57 |
| bauzas | shhhht | 21:57 |
| *** smatzek has joined #openstack-nova | 21:57 | |
| bauzas | some might have ideas | 21:57 |
| *** tblakes has joined #openstack-nova | 21:58 | |
| sach-m | mriedem: which resource sync issues? | 21:59 |
| mriedem | sach-m: you do something with your resources running in AWS via ec2 directly and things are out of sync in the nova db | 21:59 |
| mriedem | nova doesn't deal with out of band changes very well | 21:59 |
| sach-m | mriedem: We already do that for VMware and KVM in our deployments through some extensions added for openstack. We can do the same for hybrid work as well | 22:00 |
| mriedem | sure you can, and some do | 22:01 |
| mriedem | it's not fun | 22:01 |
| mriedem | powervc does it also | 22:01 |
| *** diana_clarke has quit IRC | 22:01 | |
| sach-m | mriedem: Isn't this a generic problem with any driver though? | 22:02 |
| *** harlowja has quit IRC | 22:02 | |
| cfriesen__ | is it expected that for stable/mitaka when doing a "nova evacuate" we lose files specified with --file at boot? | 22:02 |
| bauzas | sach-m: that's not really a problem, that's rather than nova isn't just an hypervisor layer | 22:04 |
| bauzas | sach-m: meaning that nova is there for consuming all the resources, not just proxying calls to libvirt | 22:04 |
| bauzas | (or any other driver) | 22:04 |
| sach-m | bauzas: It does that for VMware already | 22:05 |
| bauzas | sach-m: where we have lots of problems | 22:05 |
| bauzas | sach-m: because for example, any operation that would be made out of Nova is hard to reconcile | 22:06 |
| cfriesen__ | sach-m: if you go mucking around with libvirt directly behind nova's back you could get things into a bad state too. | 22:06 |
| *** mriedem is now known as mriedem_away | 22:06 | |
| bauzas | sach-m: in your proposal, what if the user does some resize operation in AWS directly? | 22:07 |
| bauzas | also, what would mean a "live-migration" operation from a Nova standpoint if we don't know how many computes are running behind our AWS tenant? | 22:08 |
| sach-m | bauzas: I do agree in general with the assessment that resource sync is going to be an issue. But with AWS, there's a chance to reconcile state in a better way -- everything comes in standard t-shirt sizes (flavors). | 22:08 |
| *** mdrabe has quit IRC | 22:09 | |
| sach-m | bauzas: I've seen weird issues in KVM/VMware where the instance sizes have no counterpart in nova flavors | 22:09 |
| sach-m | bauzas: My understanding is that AWS and most public cloud providers do not support live migration feature today | 22:10 |
| bauzas | that really depends on your policy | 22:10 |
| bauzas | either way, I was thinking when I saw the keynote that there was an upper service that would do that proxy call, and not nova | 22:11 |
| *** lpetrut has joined #openstack-nova | 22:11 | |
| bauzas | sach-m: tbc, I certainly understand the usecase | 22:12 |
| bauzas | sach-m: and I think it's very valid to propose that | 22:12 |
| bauzas | sach-m: I'm more concerned by the implementation that leaves a huge number of open questions we know we have | 22:12 |
| sach-m | bauzas: understood | 22:13 |
| *** victorhe has quit IRC | 22:15 | |
| *** matrohon has quit IRC | 22:18 | |
| *** matrohon has joined #openstack-nova | 22:19 | |
| *** zzelle has quit IRC | 22:19 | |
| *** burt has quit IRC | 22:21 | |
| *** browne has joined #openstack-nova | 22:23 | |
| *** sach-m has quit IRC | 22:27 | |
| *** ducttape_ has quit IRC | 22:28 | |
| *** sach-m has joined #openstack-nova | 22:28 | |
| *** flaper87 has quit IRC | 22:29 | |
| *** smatzek has quit IRC | 22:30 | |
| jlk | At summit, didn't Nova reject a driver that sat in front of a cloud like thing, FusionSphere? | 22:30 |
| jlk | the sentiment was that the VMWare vsphere driver was a mistake, one that shouldn't be repeated with fusionsphere, which would apply to AWS too | 22:30 |
| *** rrecio has quit IRC | 22:31 | |
| *** matrohon has quit IRC | 22:32 | |
| sach-m | jlk: What's the goal of nova though? Just be a virtualization layer to manage KVM? | 22:32 |
| *** matrohon has joined #openstack-nova | 22:32 | |
| jlk | that's... a difficult question. | 22:33 |
| jlk | Obviously not, given that there is a driver for Xen | 22:33 |
| jlk | given that KVM support is behind libvirt, I'm not know how much is KVM specific, or "anything libvirt supports". | 22:34 |
| sach-m | because many other platforms -- hyper-v, vmware, and so on are going to have their own management server. | 22:34 |
| sach-m | Essentially, we are saying -- use nova as virtual machine manager not a cloud API | 22:35 |
| jlk | that's correct | 22:35 |
| sach-m | I don't agree with that. OpenStack cloud API brings in a lot more capabilities like self service, key management over virtualization | 22:36 |
| *** marst_ has quit IRC | 22:36 | |
| jlk | but the nova part is simply VM management | 22:36 |
| *** huanxie has quit IRC | 22:36 | |
| jlk | with cooperation from Neutron for networking, glance for images, keystone for authentication | 22:36 |
| sach-m | What about quotas, access to images, flavors, etc? There's a valuable delegation & self service that nova provides which a platform like VMware does not | 22:37 |
| *** larainema has quit IRC | 22:37 | |
| jlk | Are you sure there's not a higher price VMWare offering that provides that? | 22:39 |
| sach-m | that's basically the reason people put openstack on top of VMware | 22:39 |
| *** larainema has joined #openstack-nova | 22:39 | |
| jlk | IIRC people put OpenStack on top of vmware to make use of vmware as the HV, and to save money on higher $$ VMware products | 22:39 |
| sach-m | VMware has its own cloud offering, yes. But it is not open. | 22:39 |
| jlk | sure, so if you want those things, buy VMWare | 22:40 |
| jlk | don't try to slap openstack on top of it | 22:40 |
| *** pradk has quit IRC | 22:41 | |
| *** matrohon has quit IRC | 22:42 | |
| *** matrohon has joined #openstack-nova | 22:42 | |
| *** mlavalle has quit IRC | 22:44 | |
| *** ihrachys has quit IRC | 22:45 | |
| *** gouthamr has quit IRC | 22:46 | |
| *** kiran-r has quit IRC | 22:49 | |
| *** ducttape_ has joined #openstack-nova | 22:53 | |
| *** matrohon has quit IRC | 22:54 | |
| bauzas | jaypipes: silly question, but should we accept "name" and "uuid" as resource class names ? :p | 22:57 |
| *** tblakes has quit IRC | 22:58 | |
| *** ducttape_ has quit IRC | 22:58 | |
| *** vladikr has quit IRC | 23:01 | |
| *** sach-m has quit IRC | 23:03 | |
| *** david-lyle has quit IRC | 23:04 | |
| *** harlowja has joined #openstack-nova | 23:08 | |
| *** sach-m has joined #openstack-nova | 23:09 | |
| *** Guest87642 has quit IRC | 23:12 | |
| *** thorst has joined #openstack-nova | 23:13 | |
| *** owalsh has quit IRC | 23:14 | |
| *** owalsh has joined #openstack-nova | 23:15 | |
| *** ducttape_ has joined #openstack-nova | 23:16 | |
| *** thorst has quit IRC | 23:18 | |
| *** mtanino has quit IRC | 23:20 | |
| *** annegentle has quit IRC | 23:22 | |
| *** sach-m has quit IRC | 23:30 | |
| *** david-lyle has joined #openstack-nova | 23:30 | |
| *** sach-m has joined #openstack-nova | 23:31 | |
| *** dave-mccowan has joined #openstack-nova | 23:31 | |
| *** rcernin has quit IRC | 23:31 | |
| *** lpetrut has quit IRC | 23:32 | |
| *** sach-m has quit IRC | 23:37 | |
| *** sach-m has joined #openstack-nova | 23:38 | |
| *** ducttape_ has quit IRC | 23:38 | |
| *** sach-m has quit IRC | 23:43 | |
| *** sach-m has joined #openstack-nova | 23:43 | |
| *** ducttape_ has joined #openstack-nova | 23:50 | |
| *** esberglu has joined #openstack-nova | 23:53 | |
| *** gouthamr has joined #openstack-nova | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
