| *** trisq has joined #openstack-nova | 00:03 | |
| *** baoli has joined #openstack-nova | 00:04 | |
| *** baoli_ has joined #openstack-nova | 00:05 | |
| *** ducttape_ has joined #openstack-nova | 00:06 | |
| *** mlavalle has quit IRC | 00:06 | |
| *** baoli has quit IRC | 00:09 | |
| *** yamahata has quit IRC | 00:14 | |
| *** ducttape_ has quit IRC | 00:16 | |
| *** sdake has joined #openstack-nova | 00:16 | |
| *** zengine has quit IRC | 00:18 | |
| *** gouthamr has joined #openstack-nova | 00:20 | |
| *** thorst has joined #openstack-nova | 00:21 | |
| *** sdake has quit IRC | 00:21 | |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: WIP: Use admin context to create neutron ports https://review.openstack.org/357540 | 00:22 |
|---|---|---|
| *** Swami__ has joined #openstack-nova | 00:24 | |
| *** sdake has joined #openstack-nova | 00:25 | |
| *** mingyu has quit IRC | 00:25 | |
| *** raunak has quit IRC | 00:26 | |
| *** Swami has quit IRC | 00:28 | |
| *** sdague has joined #openstack-nova | 00:32 | |
| *** mingyu has joined #openstack-nova | 00:33 | |
| *** tonytan4ever has quit IRC | 00:33 | |
| *** lixiaoy1 has joined #openstack-nova | 00:33 | |
| *** rfolco has joined #openstack-nova | 00:33 | |
| *** mtanino has quit IRC | 00:34 | |
| *** ducttape_ has joined #openstack-nova | 00:35 | |
| *** thorst_ has joined #openstack-nova | 00:37 | |
| *** hoangcx has joined #openstack-nova | 00:37 | |
| *** thorst has quit IRC | 00:39 | |
| *** wangqun has joined #openstack-nova | 00:41 | |
| *** mingyu has quit IRC | 00:41 | |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: WIP: Use admin context to create neutron ports https://review.openstack.org/357540 | 00:45 |
| *** mingyu has joined #openstack-nova | 00:47 | |
| *** trisq has quit IRC | 00:47 | |
| *** yamahata has joined #openstack-nova | 00:47 | |
| *** delattec has quit IRC | 00:56 | |
| *** raunak has joined #openstack-nova | 00:58 | |
| *** gouthamr_ has joined #openstack-nova | 01:01 | |
| *** pwnall has quit IRC | 01:01 | |
| *** salv-orlando has joined #openstack-nova | 01:03 | |
| *** phuongnh has joined #openstack-nova | 01:04 | |
| *** gouthamr has quit IRC | 01:05 | |
| *** Apoorva has quit IRC | 01:06 | |
| *** gongysh has joined #openstack-nova | 01:06 | |
| mriedem | dansmith: i meant to ping you *here* | 01:07 |
| dansmith | figured :) | 01:07 |
| openstackgerrit | melanie witt proposed openstack/nova: Make req spec and RT aware of volume-backed instances https://review.openstack.org/355091 | 01:08 |
| *** ducttape_ has quit IRC | 01:08 | |
| openstackgerrit | Sean Dague proposed openstack/nova: create placement API wsgi entry point https://review.openstack.org/363333 | 01:09 |
| *** Apoorva has joined #openstack-nova | 01:09 | |
| *** gouthamr_ is now known as gouthamr | 01:09 | |
| mriedem | ooo even sdague is working late | 01:10 |
| *** salv-orlando has quit IRC | 01:11 | |
| *** salv-orlando has joined #openstack-nova | 01:13 | |
| *** diga has joined #openstack-nova | 01:14 | |
| *** raunak has quit IRC | 01:15 | |
| sdague | mriedem: yeh, well, not a lot of days left | 01:15 |
| sdague | ok, I can drop dogpile | 01:15 |
| sdague | I've got the devstack patch that uses that as well | 01:16 |
| *** salv-orlando has quit IRC | 01:17 | |
| sdague | so, the thing that is still kind of dumb is that we're sharing nova config | 01:18 |
| sdague | and nova logs are set for colorizing | 01:18 |
| sdague | and placement is under apache | 01:18 |
| sdague | so it's a giant eyesore of escapes | 01:18 |
| sdague | pretty much only a devstack problem, but still annoying | 01:18 |
| openstackgerrit | Sean Dague proposed openstack/nova: create placement API wsgi entry point https://review.openstack.org/363333 | 01:20 |
| sdague | mriedem: look better? | 01:20 |
| *** ducttape_ has joined #openstack-nova | 01:20 | |
| auggy | one mox test remaining to be fixed! https://review.openstack.org/#/c/357540 | 01:22 |
| *** david-lyle_ has joined #openstack-nova | 01:23 | |
| auggy | i'm just a little stuck on how test_allocate_for_instance_ex2 was triggering its expected NeutronClientException | 01:23 |
| *** ducttape_ has quit IRC | 01:23 | |
| *** baoli_ has quit IRC | 01:24 | |
| *** ducttape_ has joined #openstack-nova | 01:26 | |
| *** akshai has quit IRC | 01:26 | |
| *** ducttape_ has quit IRC | 01:27 | |
| auggy | oh snap, i just fixed it | 01:27 |
| sdague | mriedem / dansmith - ok, I'm going to actually drift off to get some sleep for the night, I'll redo my logging patch on top of the wsgi stuff tomorrow, but we should get the new wsgi script in soonish, it will make a ton of stuff easier | 01:28 |
| *** rfolco has quit IRC | 01:28 | |
| *** Swami has joined #openstack-nova | 01:28 | |
| dansmith | sdague: okay we also need to get the allocations api stuff landed ASAP because otherwise I can't even fail at writing the client code for it | 01:29 |
| dansmith | cdent said he put that up but I looked earlier and didn't find it | 01:29 |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: WIP: Use admin context to create neutron ports https://review.openstack.org/357540 | 01:29 |
| *** kevinz has joined #openstack-nova | 01:30 | |
| *** Apoorva has quit IRC | 01:30 | |
| *** Swami__ has quit IRC | 01:32 | |
| mriedem | dansmith: i think that's this https://review.openstack.org/#/c/363209/ | 01:32 |
| *** sdague has quit IRC | 01:32 | |
| dansmith | oh I guess so I thought he said it was all the object stuff for allocations | 01:32 |
| dansmith | mriedem: and the api doesn't depend on that, but he said it needed to | 01:33 |
| mriedem | yeah i thought so too | 01:33 |
| *** tonytan4ever has joined #openstack-nova | 01:33 | |
| *** EinstCrazy has joined #openstack-nova | 01:33 | |
| *** rrecio has quit IRC | 01:35 | |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: tests: Adding functional tests to cover VM creation with sriov https://review.openstack.org/294780 | 01:37 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: tests: verify pci passthrough with numa https://review.openstack.org/356077 | 01:37 |
| *** thorst_ has quit IRC | 01:38 | |
| *** thorst has joined #openstack-nova | 01:38 | |
| *** tonytan4ever has quit IRC | 01:38 | |
| *** trisq has joined #openstack-nova | 01:44 | |
| mriedem | i also have jury duty tomorrow afternoon, | 01:46 |
| mriedem | unless they call it off again at the last minute | 01:46 |
| *** thorst has quit IRC | 01:46 | |
| *** lixiaoy1 has quit IRC | 01:48 | |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: inject files when config drive is not requested https://review.openstack.org/335676 | 01:53 |
| *** artom has quit IRC | 01:53 | |
| *** yulong|away is now known as liuyulong | 01:55 | |
| *** vishwanathj has quit IRC | 01:59 | |
| *** EinstCra_ has joined #openstack-nova | 02:03 | |
| *** hoangcx has quit IRC | 02:05 | |
| *** bkopilov has quit IRC | 02:05 | |
| *** EinstCrazy has quit IRC | 02:06 | |
| *** raunak has joined #openstack-nova | 02:10 | |
| *** unicell has quit IRC | 02:13 | |
| *** Kevin_Zheng has joined #openstack-nova | 02:13 | |
| *** sdake has quit IRC | 02:16 | |
| *** sdake has joined #openstack-nova | 02:17 | |
| *** lixiaoy1 has joined #openstack-nova | 02:17 | |
| *** hoangcx has joined #openstack-nova | 02:18 | |
| lixiaoy1 | mriedem: Hi I created a tempest test for this patch https://review.openstack.org/#/c/252809/ , and please have a look | 02:19 |
| *** diogogmt has quit IRC | 02:19 | |
| lixiaoy1 | mriedem: this is about "Attach/detach encryptors during swap_volume" | 02:20 |
| *** Sukhdev has joined #openstack-nova | 02:23 | |
| *** EinstCrazy has joined #openstack-nova | 02:24 | |
| *** Swami__ has joined #openstack-nova | 02:25 | |
| *** EinstCra_ has quit IRC | 02:27 | |
| *** ducttape_ has joined #openstack-nova | 02:28 | |
| *** Swami has quit IRC | 02:28 | |
| *** Swami_ has quit IRC | 02:29 | |
| *** Swami has joined #openstack-nova | 02:29 | |
| *** mriedem has quit IRC | 02:31 | |
| *** david-lyle_ has quit IRC | 02:31 | |
| *** gyee has quit IRC | 02:31 | |
| *** ducttape_ has quit IRC | 02:32 | |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: adding debug info for pinning calculation https://review.openstack.org/355150 | 02:33 |
| *** baoli has joined #openstack-nova | 02:34 | |
| *** tonytan4ever has joined #openstack-nova | 02:34 | |
| *** yamahata has quit IRC | 02:35 | |
| *** dimtruck is now known as zz_dimtruck | 02:37 | |
| *** Jeffrey4l_ has joined #openstack-nova | 02:38 | |
| *** tonytan4ever has quit IRC | 02:38 | |
| *** tonytan4ever has joined #openstack-nova | 02:38 | |
| *** baoli has quit IRC | 02:40 | |
| *** thorst has joined #openstack-nova | 02:45 | |
| *** ducttape_ has joined #openstack-nova | 02:48 | |
| *** thorst has quit IRC | 02:52 | |
| *** hemanthm|bbl is now known as hemanthm | 02:53 | |
| *** takashin has quit IRC | 02:59 | |
| openstackgerrit | huan proposed openstack/nova: XenAPI: Fix VM live-migrate with iSCSI SR volume https://review.openstack.org/359548 | 03:00 |
| *** mriedem has joined #openstack-nova | 03:00 | |
| *** mriedem has quit IRC | 03:01 | |
| *** mriedem has joined #openstack-nova | 03:01 | |
| *** mriedem has quit IRC | 03:01 | |
| *** ducttape_ has quit IRC | 03:07 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-novaclient: Updated from global requirements https://review.openstack.org/362595 | 03:09 |
| *** browne has quit IRC | 03:13 | |
| *** TravT has quit IRC | 03:13 | |
| *** TravT has joined #openstack-nova | 03:14 | |
| *** salv-orlando has joined #openstack-nova | 03:17 | |
| *** ducttape_ has joined #openstack-nova | 03:17 | |
| *** ducttape_ has quit IRC | 03:19 | |
| *** bkopilov has joined #openstack-nova | 03:22 | |
| *** salv-orlando has quit IRC | 03:24 | |
| *** ducttape_ has joined #openstack-nova | 03:25 | |
| *** dikonoor has joined #openstack-nova | 03:26 | |
| *** devkulkarni has joined #openstack-nova | 03:27 | |
| *** ayoung has quit IRC | 03:27 | |
| *** prateek_ has joined #openstack-nova | 03:27 | |
| *** diga has quit IRC | 03:27 | |
| *** ducttape_ has quit IRC | 03:28 | |
| *** prateek has joined #openstack-nova | 03:30 | |
| *** parora has quit IRC | 03:30 | |
| *** prateek_ has quit IRC | 03:33 | |
| *** mingyu has quit IRC | 03:33 | |
| *** dikonoor has quit IRC | 03:34 | |
| *** RuiChen has joined #openstack-nova | 03:35 | |
| *** TravT_ has joined #openstack-nova | 03:37 | |
| *** TravT has quit IRC | 03:37 | |
| *** woodster_ has quit IRC | 03:39 | |
| *** diogogmt has joined #openstack-nova | 03:44 | |
| *** devkulkarni has quit IRC | 03:48 | |
| *** thorst has joined #openstack-nova | 03:50 | |
| *** gouthamr has quit IRC | 03:51 | |
| *** mingyu has joined #openstack-nova | 03:53 | |
| *** mingyu has quit IRC | 03:53 | |
| *** ManishD has joined #openstack-nova | 03:54 | |
| *** takashin has joined #openstack-nova | 03:55 | |
| *** annp has joined #openstack-nova | 03:56 | |
| *** xinliang has joined #openstack-nova | 03:56 | |
| *** thorst has quit IRC | 03:57 | |
| *** lixiaoy1 has quit IRC | 04:02 | |
| *** willise has left #openstack-nova | 04:02 | |
| *** prateek has quit IRC | 04:03 | |
| *** bkopilov_ has joined #openstack-nova | 04:03 | |
| *** bkopilov_ has quit IRC | 04:03 | |
| *** prateek has joined #openstack-nova | 04:04 | |
| *** bkopilov_ has joined #openstack-nova | 04:04 | |
| *** ijw has quit IRC | 04:04 | |
| *** ijw has joined #openstack-nova | 04:08 | |
| *** ducttape_ has joined #openstack-nova | 04:10 | |
| *** ducttape_ has quit IRC | 04:15 | |
| openstackgerrit | Nguyen Phuong An proposed openstack/nova: Config logABug feature for Nova api-ref https://review.openstack.org/363386 | 04:15 |
| *** crinkle_ has joined #openstack-nova | 04:16 | |
| *** crinkle has quit IRC | 04:17 | |
| *** kaisers has quit IRC | 04:17 | |
| *** ratailor has joined #openstack-nova | 04:18 | |
| *** kaisers has joined #openstack-nova | 04:19 | |
| *** su_zhang has joined #openstack-nova | 04:26 | |
| *** salv-orlando has joined #openstack-nova | 04:27 | |
| *** ijw has quit IRC | 04:27 | |
| *** ijw has joined #openstack-nova | 04:28 | |
| *** salv-orlando has quit IRC | 04:32 | |
| *** ijw has quit IRC | 04:33 | |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: tests: Adding functional tests to cover VM creation with sriov https://review.openstack.org/294780 | 04:35 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: tests: verify pci passthrough with numa https://review.openstack.org/356077 | 04:35 |
| *** hoangcx has quit IRC | 04:36 | |
| *** coolsvap_ has joined #openstack-nova | 04:36 | |
| *** links has joined #openstack-nova | 04:36 | |
| *** Hosam has joined #openstack-nova | 04:37 | |
| *** garyk1 has joined #openstack-nova | 04:38 | |
| *** hoangcx has joined #openstack-nova | 04:39 | |
| *** su_zhang has quit IRC | 04:47 | |
| *** prateek has quit IRC | 04:48 | |
| *** su_zhang_ has joined #openstack-nova | 04:50 | |
| *** pbandark has joined #openstack-nova | 04:50 | |
| *** rmcall has quit IRC | 04:51 | |
| *** mamandle has joined #openstack-nova | 04:52 | |
| *** su_zhang_ has quit IRC | 04:52 | |
| *** su_zhang has joined #openstack-nova | 04:53 | |
| *** thorst has joined #openstack-nova | 04:54 | |
| *** ijw has joined #openstack-nova | 04:55 | |
| *** Jeffrey4l_ has quit IRC | 04:56 | |
| *** su_zhang has quit IRC | 04:57 | |
| *** anush__ has quit IRC | 04:58 | |
| *** Fdaisuke has quit IRC | 04:59 | |
| *** ijw has quit IRC | 05:00 | |
| *** Fdaisuke has joined #openstack-nova | 05:01 | |
| *** claudiub has joined #openstack-nova | 05:01 | |
| *** thorst has quit IRC | 05:01 | |
| *** prateek has joined #openstack-nova | 05:02 | |
| *** yamahata has joined #openstack-nova | 05:03 | |
| *** ecelik has joined #openstack-nova | 05:04 | |
| *** yamahata has quit IRC | 05:06 | |
| *** sdake has quit IRC | 05:10 | |
| *** sdake_ has joined #openstack-nova | 05:10 | |
| *** ducttape_ has joined #openstack-nova | 05:11 | |
| *** salv-orlando has joined #openstack-nova | 05:13 | |
| *** ducttape_ has quit IRC | 05:16 | |
| *** diogogmt has quit IRC | 05:17 | |
| *** fragatina has quit IRC | 05:19 | |
| *** jaosorior has joined #openstack-nova | 05:21 | |
| *** ekuris has joined #openstack-nova | 05:21 | |
| *** yonglihe has joined #openstack-nova | 05:22 | |
| *** mamandle has quit IRC | 05:23 | |
| *** sdake_ has quit IRC | 05:26 | |
| *** dikonoor has joined #openstack-nova | 05:27 | |
| openstackgerrit | Rajesh Tailor proposed openstack/nova: Fix notify instance usage on evacuate https://review.openstack.org/362794 | 05:27 |
| *** paul-carlton2 has joined #openstack-nova | 05:30 | |
| *** sdake has joined #openstack-nova | 05:33 | |
| *** sridharg has joined #openstack-nova | 05:38 | |
| *** lixiaoy1 has joined #openstack-nova | 05:40 | |
| openstackgerrit | Bhagyashri Shewale proposed openstack/nova: Return 400 error for non-existing snapshot_id https://review.openstack.org/360941 | 05:45 |
| *** raunak has quit IRC | 05:47 | |
| *** mtanino has joined #openstack-nova | 05:48 | |
| *** sahid has joined #openstack-nova | 05:51 | |
| *** armax has quit IRC | 05:56 | |
| *** mjura has joined #openstack-nova | 05:57 | |
| *** markvoelker has joined #openstack-nova | 05:58 | |
| *** thorst has joined #openstack-nova | 06:00 | |
| *** nkrinner_afk is now known as nkrinner | 06:01 | |
| *** oanson has joined #openstack-nova | 06:02 | |
| *** fragatina has joined #openstack-nova | 06:02 | |
| *** Swami has quit IRC | 06:04 | |
| *** Swami__ has quit IRC | 06:04 | |
| *** sahid has quit IRC | 06:04 | |
| *** sahid has joined #openstack-nova | 06:05 | |
| *** mjura has quit IRC | 06:05 | |
| *** ttrumm has joined #openstack-nova | 06:05 | |
| *** mjura has joined #openstack-nova | 06:05 | |
| *** thorst has quit IRC | 06:07 | |
| *** salv-orl_ has joined #openstack-nova | 06:07 | |
| *** fragatina has quit IRC | 06:07 | |
| *** takashin_ has joined #openstack-nova | 06:09 | |
| *** lpetrut has joined #openstack-nova | 06:09 | |
| *** salv-orlando has quit IRC | 06:10 | |
| *** takashin has quit IRC | 06:11 | |
| *** ducttape_ has joined #openstack-nova | 06:12 | |
| *** unicell has joined #openstack-nova | 06:13 | |
| *** pcaruana has joined #openstack-nova | 06:15 | |
| *** ducttape_ has quit IRC | 06:17 | |
| *** claudiub has quit IRC | 06:17 | |
| *** andreas_s has joined #openstack-nova | 06:20 | |
| *** rcernin has joined #openstack-nova | 06:21 | |
| *** sdake has quit IRC | 06:22 | |
| *** sheel has joined #openstack-nova | 06:24 | |
| *** salv-orl_ has quit IRC | 06:24 | |
| *** ttrumm has quit IRC | 06:25 | |
| *** ttrumm has joined #openstack-nova | 06:25 | |
| *** ijw has joined #openstack-nova | 06:28 | |
| *** amotoki has joined #openstack-nova | 06:29 | |
| *** mtanino has quit IRC | 06:32 | |
| openstackgerrit | wanghongtaozz proposed openstack/nova: test A bug ignore it https://review.openstack.org/363425 | 06:34 |
| openstackgerrit | wanghongtaozz proposed openstack/nova: test ignore https://review.openstack.org/363426 | 06:34 |
| *** markvoelker has quit IRC | 06:34 | |
| *** garyk1 has quit IRC | 06:38 | |
| openstackgerrit | Cao ShuFeng proposed openstack/nova-specs: Add support booting baremetals from Cinder volume https://review.openstack.org/311696 | 06:39 |
| *** abalutoiu has joined #openstack-nova | 06:39 | |
| *** moshele has joined #openstack-nova | 06:46 | |
| *** ijw has quit IRC | 06:53 | |
| *** mingyu has joined #openstack-nova | 06:54 | |
| *** liverpooler has quit IRC | 06:56 | |
| *** amotoki has quit IRC | 06:56 | |
| *** mingyu has quit IRC | 06:59 | |
| *** bjolo has joined #openstack-nova | 07:00 | |
| *** paul-carlton2 has quit IRC | 07:01 | |
| *** lpetrut has quit IRC | 07:03 | |
| *** sudipto has joined #openstack-nova | 07:04 | |
| *** sudipto_ has joined #openstack-nova | 07:04 | |
| *** thorst has joined #openstack-nova | 07:05 | |
| *** liuyulong has quit IRC | 07:06 | |
| *** liuyulong has joined #openstack-nova | 07:07 | |
| *** rmart04 has joined #openstack-nova | 07:08 | |
| *** tesseract- has joined #openstack-nova | 07:08 | |
| *** sahid has quit IRC | 07:08 | |
| *** sahid has joined #openstack-nova | 07:09 | |
| *** karthiks has quit IRC | 07:09 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: fix incorrect host cpus giving to emulator threads when RT https://review.openstack.org/356383 | 07:09 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: fix serial console not correctly defined after live-migration https://review.openstack.org/356335 | 07:10 |
| *** thorst has quit IRC | 07:12 | |
| *** kaisers_ has joined #openstack-nova | 07:13 | |
| *** ducttape_ has joined #openstack-nova | 07:13 | |
| *** ihrachys has joined #openstack-nova | 07:14 | |
| *** paul-carlton2 has joined #openstack-nova | 07:16 | |
| *** amotoki has joined #openstack-nova | 07:16 | |
| openstackgerrit | liuxiuli proposed openstack/nova: Modify workflow of rebuild_instance when schedule fails https://review.openstack.org/348111 | 07:18 |
| *** markus_z has joined #openstack-nova | 07:18 | |
| *** sameo has joined #openstack-nova | 07:18 | |
| *** ducttape_ has quit IRC | 07:19 | |
| openstackgerrit | Andrey Volkov proposed openstack/nova: WIP Map instances to cell0 on failed schedule https://review.openstack.org/267827 | 07:22 |
| *** paul-carlton2 has quit IRC | 07:28 | |
| *** jlanoux has joined #openstack-nova | 07:29 | |
| *** markvoelker has joined #openstack-nova | 07:30 | |
| *** ljxiash has joined #openstack-nova | 07:30 | |
| *** sudipto_ has quit IRC | 07:31 | |
| *** sudipto has quit IRC | 07:31 | |
| *** markvoelker has quit IRC | 07:35 | |
| *** kaisers_ has quit IRC | 07:35 | |
| *** sudipto has joined #openstack-nova | 07:36 | |
| *** sudipto_ has joined #openstack-nova | 07:36 | |
| *** edand has joined #openstack-nova | 07:37 | |
| *** salv-orlando has joined #openstack-nova | 07:39 | |
| *** sudipto has quit IRC | 07:41 | |
| *** sudipto_ has quit IRC | 07:41 | |
| *** paul-carlton2 has joined #openstack-nova | 07:41 | |
| *** sudipto has joined #openstack-nova | 07:44 | |
| *** sudipto has quit IRC | 07:46 | |
| *** jlanoux has quit IRC | 07:47 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: Treat 'body_map' as a mapping of HTTP request bodies https://review.openstack.org/363457 | 07:48 |
| *** paul-carlton2 has quit IRC | 07:49 | |
| *** Oku_OS-away is now known as Oku_OS | 07:50 | |
| *** mnestratov has joined #openstack-nova | 07:51 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: Treat 'body_map' as a mapping of HTTP request bodies https://review.openstack.org/363457 | 07:51 |
| *** Sukhdev has quit IRC | 07:52 | |
| *** cdent has joined #openstack-nova | 07:56 | |
| *** bkopilov_ has quit IRC | 07:57 | |
| *** bkopilov has quit IRC | 07:57 | |
| *** liverpooler has joined #openstack-nova | 07:57 | |
| *** priteau has joined #openstack-nova | 07:57 | |
| *** roeyc has joined #openstack-nova | 07:57 | |
| *** liverpooler has quit IRC | 07:58 | |
| *** liverpooler has joined #openstack-nova | 07:58 | |
| *** mgoddard_ has joined #openstack-nova | 07:58 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** praneshp has quit IRC | 08:00 | |
| *** zenoway has joined #openstack-nova | 08:00 | |
| *** bkopilov has joined #openstack-nova | 08:01 | |
| *** markvoelker has joined #openstack-nova | 08:01 | |
| *** zzzeek has joined #openstack-nova | 08:01 | |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP Add support for allocations in placement API https://review.openstack.org/339874 | 08:04 |
| *** hoangcx has quit IRC | 08:04 | |
| *** hoangcx has joined #openstack-nova | 08:05 | |
| *** matrohon has joined #openstack-nova | 08:05 | |
| *** markvoelker has quit IRC | 08:07 | |
| *** thorst has joined #openstack-nova | 08:09 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: [WIP] Testing latest u-c https://review.openstack.org/303345 | 08:10 |
| *** claudiub has joined #openstack-nova | 08:11 | |
| *** ljxiash has quit IRC | 08:12 | |
| *** ducttape_ has joined #openstack-nova | 08:16 | |
| *** thorst has quit IRC | 08:17 | |
| *** lucas-dinner is now known as lucasagomes | 08:20 | |
| *** ducttape_ has quit IRC | 08:20 | |
| *** ljxiash has joined #openstack-nova | 08:21 | |
| *** paul-carlton2 has joined #openstack-nova | 08:24 | |
| *** lixiaoy1 has quit IRC | 08:24 | |
| *** CristinaPauna has joined #openstack-nova | 08:24 | |
| *** lixiaoy1 has joined #openstack-nova | 08:24 | |
| *** gongysh has quit IRC | 08:26 | |
| *** oomichi has quit IRC | 08:27 | |
| *** oomichi has joined #openstack-nova | 08:29 | |
| *** markvoelker has joined #openstack-nova | 08:29 | |
| *** diga has joined #openstack-nova | 08:30 | |
| *** auggy has quit IRC | 08:30 | |
| *** rkrum has quit IRC | 08:30 | |
| *** derekh has joined #openstack-nova | 08:31 | |
| *** auggy has joined #openstack-nova | 08:31 | |
| *** gongysh_ has quit IRC | 08:31 | |
| *** markvoelker has quit IRC | 08:34 | |
| *** edand has quit IRC | 08:34 | |
| *** mingyu has joined #openstack-nova | 08:35 | |
| *** jaypipes has joined #openstack-nova | 08:35 | |
| *** ihrachys has quit IRC | 08:37 | |
| *** zenoway has quit IRC | 08:39 | |
| *** zenoway has joined #openstack-nova | 08:39 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Optional separate database for placement API https://review.openstack.org/362766 | 08:42 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] add a placement_aggregates table to api_db https://review.openstack.org/362863 | 08:42 |
| *** kashyap_ has joined #openstack-nova | 08:42 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Add {get_,set_}aggregates to objects.ResourceProvider https://review.openstack.org/355263 | 08:42 |
| *** kashyap_ has quit IRC | 08:43 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Implement get and set aggregates in the placement API https://review.openstack.org/357993 | 08:43 |
| *** kashyap_ has joined #openstack-nova | 08:43 | |
| *** gongysh_ has joined #openstack-nova | 08:44 | |
| openstackgerrit | wanghongtaozz proposed openstack/nova: this is a test bug please ignore it https://review.openstack.org/363489 | 08:47 |
| openstackgerrit | wanghongtaozz proposed openstack/nova: test https://review.openstack.org/363490 | 08:47 |
| *** xinliang has quit IRC | 08:48 | |
| *** ttrumm has quit IRC | 08:49 | |
| *** xinliang has joined #openstack-nova | 08:49 | |
| *** ttrumm has joined #openstack-nova | 08:50 | |
| *** sudipto has joined #openstack-nova | 08:51 | |
| *** kashyap_ has quit IRC | 08:51 | |
| *** roeyc has quit IRC | 08:52 | |
| *** ttrumm has quit IRC | 08:55 | |
| *** lixiaoy1 has quit IRC | 08:55 | |
| openstackgerrit | liuxiuli proposed openstack/nova: Modify workflow of rebuild_instance when schedule fails https://review.openstack.org/348111 | 08:56 |
| *** pilgrimstack has quit IRC | 08:56 | |
| *** takashin_ has left #openstack-nova | 08:56 | |
| *** lpetrut has joined #openstack-nova | 08:58 | |
| *** markvoelker has joined #openstack-nova | 08:59 | |
| *** pilgrimstack has joined #openstack-nova | 08:59 | |
| *** andreykurilin_ has joined #openstack-nova | 09:00 | |
| *** karthiks has joined #openstack-nova | 09:00 | |
| *** moshele has quit IRC | 09:01 | |
| *** moshele has joined #openstack-nova | 09:02 | |
| *** lpetrut has quit IRC | 09:02 | |
| *** lpetrut1 has joined #openstack-nova | 09:02 | |
| *** markvoelker has quit IRC | 09:04 | |
| *** lpetrut1 is now known as lpetrut | 09:05 | |
| *** houming has joined #openstack-nova | 09:06 | |
| *** andreykurilin has quit IRC | 09:09 | |
| cdent | jaypipes: you got your ears on? Just want to clarify something on the robustification of the resource tracker stuff that still seems to be sticking in your craw: your code was conflating two failure modes | 09:12 |
| cdent | failure to find a placement api and failure to talk to the placement api on any given request | 09:12 |
| cdent | further: the reason sean added the keystone auth code was because without it, all requests to the api would have no auth | 09:13 |
| cdent | and thus would fail | 09:13 |
| *** sean-k-mooneyAFK has joined #openstack-nova | 09:13 | |
| openstackgerrit | Rajesh Tailor proposed openstack/nova: Fix notify instance usage on evacuate https://review.openstack.org/362794 | 09:14 |
| *** ihrachys has joined #openstack-nova | 09:14 | |
| *** thorst has joined #openstack-nova | 09:14 | |
| sean-k-mooneyAFK | jaypipes: hi jay. could you take a look at this https://review.openstack.org/#/c/334048/ i was hoping we could swap ovs with vhost-user over to os-vif for portbinding this release | 09:15 |
| *** sean-k-mooneyAFK is now known as sean-k-mooney | 09:15 | |
| *** ducttape_ has joined #openstack-nova | 09:17 | |
| *** kevinz has quit IRC | 09:19 | |
| *** jlanoux has joined #openstack-nova | 09:20 | |
| *** ducttape_ has quit IRC | 09:21 | |
| *** thorst has quit IRC | 09:22 | |
| *** andreykurilin_ has quit IRC | 09:25 | |
| *** Cristina_ has joined #openstack-nova | 09:26 | |
| *** andreykurilin has joined #openstack-nova | 09:26 | |
| *** andreykurilin has left #openstack-nova | 09:26 | |
| *** markvoelker has joined #openstack-nova | 09:29 | |
| *** CristinaPauna has quit IRC | 09:29 | |
| *** andreykurilin has joined #openstack-nova | 09:29 | |
| *** andreykurilin has left #openstack-nova | 09:33 | |
| *** markvoelker has quit IRC | 09:33 | |
| *** andreykurilin has joined #openstack-nova | 09:34 | |
| *** zenoway has quit IRC | 09:35 | |
| *** zenoway has joined #openstack-nova | 09:35 | |
| openstackgerrit | viswesuwara nathan proposed openstack/nova: fixed flavor image listing if project_id is None https://review.openstack.org/363530 | 09:36 |
| *** andreykurilin__ has quit IRC | 09:36 | |
| *** zenoway has quit IRC | 09:38 | |
| *** zenoway has joined #openstack-nova | 09:39 | |
| openstackgerrit | Li Yingjun proposed openstack/nova: Adds view builders for hypervisors controller https://review.openstack.org/335282 | 09:39 |
| openstackgerrit | Dao Cong Tien proposed openstack/nova: Ironic Driver: override get_serial_console() https://review.openstack.org/328157 | 09:41 |
| openstackgerrit | viswesuwara nathan proposed openstack/nova: fixed flavor image listing if project_id is None https://review.openstack.org/363530 | 09:42 |
| *** ljxiash has quit IRC | 09:44 | |
| *** code-R has joined #openstack-nova | 09:45 | |
| *** salv-orlando has quit IRC | 09:45 | |
| openstackgerrit | viswesuwara nathan proposed openstack/nova: fixed flavor image listing if project_id is None https://review.openstack.org/363530 | 09:46 |
| *** hoangcx has quit IRC | 09:47 | |
| *** code-R_ has joined #openstack-nova | 09:48 | |
| *** gongysh_ has quit IRC | 09:49 | |
| openstackgerrit | viswesuwara nathan proposed openstack/nova: fixed flavor image listing if project_id is None https://review.openstack.org/363530 | 09:49 |
| *** code-R has quit IRC | 09:51 | |
| openstackgerrit | Claudiu Belu proposed openstack/nova: Fixes python 3 unit tests https://review.openstack.org/278810 | 09:52 |
| openstackgerrit | Claudiu Belu proposed openstack/nova: compute: fixes python 3 related unit tests https://review.openstack.org/300262 | 09:52 |
| *** markvoelker has joined #openstack-nova | 09:57 | |
| openstackgerrit | Maxim Nestratov proposed openstack/nova-specs: libvirt: reintroduce runtime image type detection https://review.openstack.org/363547 | 09:59 |
| *** paul-carlton2 has quit IRC | 10:01 | |
| *** Hosam_ has joined #openstack-nova | 10:01 | |
| openstackgerrit | viswesuwara nathan proposed openstack/nova: fixed flavor image listing if project_id is None https://review.openstack.org/363530 | 10:01 |
| *** markvoelker has quit IRC | 10:02 | |
| *** salv-orlando has joined #openstack-nova | 10:02 | |
| *** Hosam has quit IRC | 10:04 | |
| *** roeyc has joined #openstack-nova | 10:07 | |
| *** tbachman has quit IRC | 10:09 | |
| openstackgerrit | liuxiuli proposed openstack/nova-specs: Clearing memory will be needed after deleting vm https://review.openstack.org/343150 | 10:10 |
| openstackgerrit | Jay Pipes proposed openstack/nova: rt: ensure resource provider records exist from RT https://review.openstack.org/358797 | 10:10 |
| *** Hosam has joined #openstack-nova | 10:12 | |
| *** anush__ has joined #openstack-nova | 10:15 | |
| *** Hosam_ has quit IRC | 10:15 | |
| openstackgerrit | Claudiu Belu proposed openstack/nova: libvirt: fixes python 3 related unit tests https://review.openstack.org/278834 | 10:16 |
| openstackgerrit | Claudiu Belu proposed openstack/nova: numa: Fixes NUMA topology related unit tests https://review.openstack.org/278835 | 10:16 |
| *** ducttape_ has joined #openstack-nova | 10:17 | |
| *** wangqun has quit IRC | 10:19 | |
| openstackgerrit | sahid proposed openstack/nova: live-mig: Add claims and proper resource tracking https://review.openstack.org/244489 | 10:19 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: pass libvirt reference methods to migration https://review.openstack.org/328997 | 10:19 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: live-migrate updates NUMA and cpus in the XML https://review.openstack.org/286744 | 10:19 |
| *** thorst has joined #openstack-nova | 10:20 | |
| johnthetubaguy | cdent: you got a second about that resource provider chian? | 10:22 |
| *** ducttape_ has quit IRC | 10:22 | |
| cdent | johnthetubaguy: yeah, what's up? | 10:22 |
| johnthetubaguy | cdent: I am wondering about set_aggregates, and how you actually remove an aggregate association | 10:23 |
| johnthetubaguy | I was just thinking about this bit: https://review.openstack.org/#/c/355263/7/nova/objects/resource_provider.py@442 | 10:23 |
| cdent | johnthetubaguy: as I recall you set it to an empty list | 10:24 |
| cdent | which was the subject of some discussion back at spec time | 10:25 |
| cdent | there probably needs to be a test for that case | 10:25 |
| *** markvoelker has joined #openstack-nova | 10:26 | |
| cdent | johnthetubaguy: the comment at line 432 explains how they all get flushed and only some kept | 10:26 |
| *** thorst has quit IRC | 10:26 | |
| johnthetubaguy | so I think I miss read an and as an or on line 416 | 10:26 |
| johnthetubaguy | thats impossible to read, & vs | | 10:27 |
| *** sudipto has quit IRC | 10:27 | |
| cdent | and with regard to you keep finding "out of date things" that's because we had a "oh shit let's do a hangout call" yesterday evening that resulted in this etherpad and lots of wips to so that various things on the resource tracker can move forward: https://etherpad.openstack.org/p/placement-next | 10:27 |
| cdent | johnthetubaguy: yeah, it is a bit hard, borrowed from where jay used it in set_inventory | 10:28 |
| johnthetubaguy | I am going around a few loops in my head, not convinced yet... | 10:28 |
| cdent | it could very well be wrong, and need more testing, if you can write down there which additional tests are neded that would be great | 10:29 |
| johnthetubaguy | cdent: so I am not convinced by target_aggregates, I am now thinking it should be to_keep + to_add | 10:29 |
| *** markvoelker has quit IRC | 10:30 | |
| cdent | to_keep is those aggreates which need to have an association but may or may not exist | 10:30 |
| cdent | to_add is those aggregates which need to exist (in the placementaggregate model) | 10:30 |
| cdent | so not the same type of thing | 10:31 |
| cdent | (yes, it's impossible to read) | 10:31 |
| johnthetubaguy | hmm... | 10:31 |
| * johnthetubaguy wonders if sql alchemy has a thing for this | 10:31 | |
| cdent | johnthetubaguy: this whole file is a confused combination of sql alchemy and jay's hate for orm | 10:32 |
| cdent | there's a lot of cleanup that can and should be done | 10:33 |
| *** chrichip has joined #openstack-nova | 10:34 | |
| *** EinstCrazy has quit IRC | 10:35 | |
| cdent | This resource provider thing is a good example of software always being bigger and more complicated than planned/expected. lots of rush and confusion. | 10:35 |
| *** pbandark has quit IRC | 10:36 | |
| cdent | That's been compounded by the fact that though code has been up in wip form for a long time, review has come later. | 10:36 |
| cdent | which I guess is just the nature of the present situation | 10:36 |
| johnthetubaguy | well, we need to do better at admitting a feature needs as much producing resource as review resource to complete, and ideally need them working together more cloesly | 10:37 |
| *** trisq has quit IRC | 10:37 | |
| * cdent nods | 10:38 | |
| johnthetubaguy | we should try something new for priority stuff next cycle, to try fix that, actually do the runways thing, but just for priorities | 10:39 |
| *** pbandark has joined #openstack-nova | 10:41 | |
| *** Hosam_ has joined #openstack-nova | 10:42 | |
| *** ansiwen has quit IRC | 10:42 | |
| *** paul-carlton2 has joined #openstack-nova | 10:43 | |
| *** Hosam has quit IRC | 10:45 | |
| *** sdague has joined #openstack-nova | 10:46 | |
| *** sdake has joined #openstack-nova | 10:49 | |
| *** chrichip has quit IRC | 10:50 | |
| *** chrichip has joined #openstack-nova | 10:51 | |
| sdague | alex_xu: check experimental isn't going to work on https://review.openstack.org/#/c/363333/ because we need the devstack change after it | 10:54 |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP Add create_all and delete_all to AllocationList https://review.openstack.org/363209 | 10:54 |
| *** code-R_ has quit IRC | 10:54 | |
| *** chrichip has quit IRC | 10:56 | |
| *** markvoelker has joined #openstack-nova | 10:56 | |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP Add support for allocations in placement API https://review.openstack.org/339874 | 10:56 |
| *** chrichip has joined #openstack-nova | 10:57 | |
| jaypipes | cdent: I personally find the non-ORM stuff easier to read, but I'm weird, I know. | 10:58 |
| cdent | jaypipes: I hate it all because I hate everything so your way isn't any worse or better than anything else :) | 10:59 |
| jaypipes | cdent: heh | 10:59 |
| *** phuongnh has quit IRC | 11:00 | |
| *** markvoelker has quit IRC | 11:00 | |
| johnthetubaguy | I thought I could read it, but my reading says its totally broken, just trying to work out which is correct, me or the code (I am personally betting on the code) | 11:02 |
| *** code-R has joined #openstack-nova | 11:03 | |
| *** salv-orlando has quit IRC | 11:04 | |
| cdent | johnthetubaguy: it's entirely possible | 11:05 |
| sdague | jaypipes: I think you are the only one that finds the non orm stuff easier to read :) | 11:08 |
| openstackgerrit | Jay Pipes proposed openstack/nova: rt: ensure resource provider records exist from RT https://review.openstack.org/358797 | 11:08 |
| openstackgerrit | Jay Pipes proposed openstack/nova: rt: write inventory records from resource tracker https://review.openstack.org/363583 | 11:08 |
| openstackgerrit | Claudiu Belu proposed openstack/nova-specs: WIP: discoverable API endpoints https://review.openstack.org/363584 | 11:08 |
| *** amotoki has quit IRC | 11:09 | |
| *** chrichip has quit IRC | 11:10 | |
| jaypipes | sdague: yeah, I know :) sorry... | 11:10 |
| sdague | jaypipes: so https://review.openstack.org/#/c/358797/12/nova/scheduler/client/report.py@53 - we shouldn't disable there, that's litterally just a network blib event | 11:10 |
| sdague | ConnectFailure just means that http request didn't work | 11:11 |
| jaypipes | sdague: BTW, pushed unit tests and pep8 fixes (and cleaned up some log message stuff in your revision to "rt: ensure resource provider records exist from RT" patch. thank you very much for hopping on this. as mentioned in my ML response, BT Internet has been spotty at best :( | 11:11 |
| jaypipes | sdague: ah, ok... | 11:11 |
| *** chrichip has joined #openstack-nova | 11:11 | |
| sdague | jaypipes: I think otherwise, this seems good, and thanks for running with those unit tests | 11:11 |
| *** bkopilov has quit IRC | 11:11 | |
| jaypipes | sdague: please feel free to change it. unfortunatel I need to run right now and perform holiday things with the family :( | 11:11 |
| jaypipes | will try to be back online later this evening. | 11:12 |
| *** jaypipes has quit IRC | 11:12 | |
| openstackgerrit | jichenjc proposed openstack/nova: no instance info cache update if instance deleted https://review.openstack.org/363585 | 11:12 |
| sdague | yep, no worries | 11:12 |
| sdague | I'll just change that one disabled, then I think we're good | 11:13 |
| *** slaweq has joined #openstack-nova | 11:14 | |
| slaweq | hello | 11:14 |
| *** amotoki has joined #openstack-nova | 11:14 | |
| slaweq | is there anybody familiar with live_snapshots in nova (with libvirt)? | 11:14 |
| openstackgerrit | Sean Dague proposed openstack/nova: rt: ensure resource provider records exist from RT https://review.openstack.org/358797 | 11:15 |
| sdague | ok, done | 11:15 |
| sdague | johnthetubaguy / alex_xu you want to hit my wsgi_scripts patch? | 11:15 |
| sdague | https://review.openstack.org/#/c/363333/ | 11:16 |
| johnthetubaguy | sdague: did want to add a reno node on how to run that? | 11:16 |
| sdague | honestly, we're going to need other docs around that entirely | 11:18 |
| *** chrichip has quit IRC | 11:18 | |
| *** wznoinsk has quit IRC | 11:18 | |
| *** ducttape_ has joined #openstack-nova | 11:18 | |
| sdague | johnthetubaguy: the point is, today, the deploy story is copy a file manually out of the source tree. This puts us back into the common pattern of the rest of openstack. | 11:19 |
| sdague | as this will actually get installed normally | 11:19 |
| *** chrichip has joined #openstack-nova | 11:19 | |
| *** ManishD has quit IRC | 11:20 | |
| *** bvanhav has joined #openstack-nova | 11:21 | |
| openstackgerrit | Chen Fan proposed openstack/nova: pci: nova fail to boot when wrong setting in pci_whitelist https://review.openstack.org/357604 | 11:23 |
| *** ducttape_ has quit IRC | 11:23 | |
| *** markvoelker has joined #openstack-nova | 11:25 | |
| *** chrichip has quit IRC | 11:26 | |
| *** salv-orlando has joined #openstack-nova | 11:26 | |
| *** chrichip has joined #openstack-nova | 11:27 | |
| *** rtheis has joined #openstack-nova | 11:28 | |
| *** markvoelker has quit IRC | 11:29 | |
| bauzas | sdague: so I took a look on https://review.openstack.org/#/c/358797/10 and was about to rebase for the few nits | 11:31 |
| bauzas | sdague: I'm working on adding tests for https://review.openstack.org/#/c/363061/2 | 11:32 |
| johnthetubaguy | cdent: so what is the expected behaviour again, for existing [a,b] and set on [b,c]? | 11:32 |
| sdague | bauzas: ok, do the nits really need to be in there? | 11:32 |
| bauzas | sdague: but we have a problem with all the tests given a returned excepting | 11:32 |
| bauzas | exception | 11:32 |
| bauzas | sdague: I was about to fix your unittest problems | 11:33 |
| sdague | bauzas: what unit test problem? | 11:33 |
| sdague | bauzas: please explain more before we reset the clock on tests again | 11:33 |
| *** wznoinsk has joined #openstack-nova | 11:33 | |
| bauzas | sdague: http://logs.openstack.org/97/358797/10/check/gate-nova-python27-db-ubuntu-xenial/11139f2/testr_results.html.gz | 11:33 |
| sdague | bauzas: ok, jay pushed an update after that | 11:34 |
| cdent | johnthetubaguy: if [a, b] is current, and the request is for it to be [b, c] then: the c aggregate needs to be created, all the rp<->aggregagates are removed, and they are recreated for [b, c] | 11:34 |
| johnthetubaguy | cdent: yeah, thats what I thought, thats good | 11:34 |
| sdague | bauzas: we're up to patch 13 now | 11:34 |
| johnthetubaguy | cdent: I found the line that was confusing me | 11:34 |
| bauzas | sdague: okay, I see https://review.openstack.org/#/c/358797/10..13/nova/tests/unit/scheduler/client/test_report.py | 11:35 |
| bauzas | okay, thanks I'll update | 11:36 |
| bauzas | (I was out for lunch before PS11 :) ) | 11:36 |
| *** baoli has joined #openstack-nova | 11:37 | |
| sdague | np | 11:37 |
| *** xyang1 has quit IRC | 11:39 | |
| *** trisq has joined #openstack-nova | 11:39 | |
| *** ttrumm has joined #openstack-nova | 11:41 | |
| *** ttrumm has quit IRC | 11:41 | |
| *** ttrumm has joined #openstack-nova | 11:42 | |
| cdent | johnthetubaguy: those comments make sense. I'll try to address them a) after I push up this allocations stuff b) have some lunch | 11:42 |
| *** lucasagomes is now known as lucas-hungry | 11:43 | |
| *** thorst has joined #openstack-nova | 11:43 | |
| sdague | johnthetubaguy so... https://review.openstack.org/#/c/363333/ ? | 11:43 |
| *** chrichip has quit IRC | 11:44 | |
| *** vishwanathj has joined #openstack-nova | 11:44 | |
| *** chrichip has joined #openstack-nova | 11:45 | |
| bauzas | sdague: for the moment, we only run it thru mod_wsgi, right? | 11:46 |
| *** baoli_ has joined #openstack-nova | 11:46 | |
| cdent | bauzas: yes | 11:47 |
| *** rfolco has joined #openstack-nova | 11:47 | |
| bauzas | okay I see why this is needed then | 11:47 |
| cdent | or rather, more correctly, as a wsgi app needing wsgi server to host it | 11:47 |
| *** coolsvap_ is now known as coolsvap | 11:47 | |
| bauzas | yup | 11:47 |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP Add support for allocations in placement API https://review.openstack.org/339874 | 11:47 |
| cdent | sdague: that ^ is writing and deleting allocations, with some caveats | 11:48 |
| cdent | (on the api side) | 11:48 |
| bauzas | cdent: could you please remind me why the consensus is not to use eventlet for ? | 11:48 |
| cdent | bauzas: just a sec | 11:48 |
| *** baoli has quit IRC | 11:49 | |
| cdent | sdague: I've updated the etherpad with some comments on what needs to happen on those | 11:50 |
| cdent | bauzas: because the long term consenses is that wsgi apps that are hosted by wsgi servers are more scalable, less error prone, less complex and more in keeping with the standard way python web apps are managed. For placement in particular, since it is setting trends for the long term it was decided that setting the precedent now was the right thing to do. | 11:51 |
| sdague | bauzas: regardless as of why, the decision is done. Now it's just about making that look like the rest of openstack services that do this | 11:52 |
| bauzas | okay | 11:52 |
| sdague | there is zero value in discussing that decision at this point in the cycle | 11:53 |
| bauzas | I'm fine | 11:53 |
| bauzas | just wanted to understand | 11:53 |
| cdent | I'm going to eat and take a little break. When I come back I'll fix up the aggregates stuff that johnthetubaguy pointed out and then check in to see what's next. If there's feedback on the allocations stuff by then, I'll do that. | 11:54 |
| cdent | biab | 11:54 |
| *** tblakes has joined #openstack-nova | 11:54 | |
| openstackgerrit | Michael Still proposed openstack/nova: Do not merge: quick sketch of a service user for vd2 https://review.openstack.org/363606 | 11:55 |
| slaweq | kashyap: hello, can You help me with live_snapshot issue in Nova? | 11:56 |
| *** markvoelker has joined #openstack-nova | 11:56 | |
| sdague | cdent: ok, I'm trying to sort out a few things to make the devstack setup nicer here, I'll look at patches to review after | 11:57 |
| sdague | bauzas / johnthetubaguy : ok, so anyway, if we want to speed up people making progress on placement api, we really need to make devstack standup not be manual cherry picking of changes and hacking in logging | 11:58 |
| *** artom has joined #openstack-nova | 11:59 | |
| *** Kevin_Zheng has quit IRC | 11:59 | |
| *** chrichip has quit IRC | 12:00 | |
| *** chrichip has joined #openstack-nova | 12:00 | |
| *** markvoelker has quit IRC | 12:01 | |
| openstackgerrit | sahid proposed openstack/nova: live-mig: Add claims and proper resource tracking https://review.openstack.org/244489 | 12:01 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: pass libvirt reference methods to migration https://review.openstack.org/328997 | 12:01 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: live-migrate updates NUMA and cpus in the XML https://review.openstack.org/286744 | 12:01 |
| *** tongli has quit IRC | 12:03 | |
| *** Kevin_Zheng has joined #openstack-nova | 12:04 | |
| *** salv-orlando has quit IRC | 12:04 | |
| *** jaosorior has quit IRC | 12:04 | |
| *** jaosorior has joined #openstack-nova | 12:06 | |
| kashyap | slaweq: What is the issue? | 12:06 |
| kashyap | slaweq: Please write the question here, making some quick lunch, back in a few min | 12:07 |
| slaweq | kashyap: thx, so I'm using libvirt 1.3.1 and sometimes snapshot made with live_snapshot is corrupted | 12:07 |
| *** roeyc has quit IRC | 12:07 | |
| slaweq | kashyap: I described what I found in bug: https://bugs.launchpad.net/nova/+bug/1530275 | 12:08 |
| openstack | Launchpad bug 1530275 in OpenStack Compute (nova) "Live snapshot is corrupted (possibly race condition?)" [Undecided,Incomplete] - Assigned to Slawek Kaplonski (slaweq) | 12:08 |
| slaweq | kashyap: comment 7 and 8 in this bug | 12:08 |
| *** roeyc has joined #openstack-nova | 12:09 | |
| slaweq | so basically I'm not sure if it is normal case that value of status.end returned from libvirt is 0 | 12:09 |
| *** rodrigods has quit IRC | 12:09 | |
| *** rodrigods has joined #openstack-nova | 12:09 | |
| *** gouthamr has joined #openstack-nova | 12:09 | |
| slaweq | maybe You know it better and could help me :) | 12:09 |
| *** Jeffrey4l_ has joined #openstack-nova | 12:10 | |
| *** annp has quit IRC | 12:11 | |
| *** salv-orlando has joined #openstack-nova | 12:12 | |
| sdague | bauzas / johnthetubaguy easy approve (unrelated to placement) - https://review.openstack.org/#/c/363160/ | 12:13 |
| *** ducttape_ has joined #openstack-nova | 12:14 | |
| *** sheel has quit IRC | 12:16 | |
| openstackgerrit | Andrey Volkov proposed openstack/nova: WIP Map instances to cell0 on failed schedule https://review.openstack.org/267827 | 12:17 |
| *** devkulkarni has joined #openstack-nova | 12:18 | |
| *** pradk has joined #openstack-nova | 12:20 | |
| *** kevinz has joined #openstack-nova | 12:20 | |
| *** gszasz has joined #openstack-nova | 12:22 | |
| *** devkulkarni has quit IRC | 12:23 | |
| *** devkulkarni has joined #openstack-nova | 12:24 | |
| *** markvoelker has joined #openstack-nova | 12:24 | |
| kashyap | slaweq: Hey, back here. So, still reading the bug | 12:25 |
| slaweq | kashyap: ok, I'm waiting | 12:26 |
| kashyap | slaweq: I presume you've created with something like: `nova image-create vm1 "snap1-of-vm1" --poll` | 12:26 |
| slaweq | exactly | 12:26 |
| kashyap | slaweq: I've just got a fresh DevStack instance setup locally. Let me see if I can reproduce. But I'm using: | 12:26 |
| kashyap | libvirt-1.3.3.1-4.fc24.x86_64 | 12:26 |
| kashyap | qemu-system-x86-2.6.0-4.fc24.x86_64 | 12:26 |
| kashyap | slaweq: Do you have a libvirt daemon debug log? That'll help me trace the libvirt APIs. Matt has asked on the bug | 12:27 |
| *** kevinz has quit IRC | 12:27 | |
| slaweq | I don't have libvirtd debug log but I can have it | 12:27 |
| *** kevinz has joined #openstack-nova | 12:27 | |
| kashyap | slaweq: Do you know how to setup libvirt log filters? If not, I can quickly write down the instructions here | 12:27 |
| slaweq | one more thing, I'm using still old Juno release but this part of code (logic) is not changed in master as I checked | 12:28 |
| slaweq | kashyap: I don't know how to write such filters | 12:28 |
| kashyap | slaweq: Here we go: | 12:28 |
| kashyap | slaweq: Oh, Juno...Please try with current stable. (Gentle reminder This channel is usually for devel discussions) | 12:29 |
| *** markvoelker has quit IRC | 12:29 | |
| kashyap | slaweq: In your /etc/libvirt/libvirtd.conf | 12:29 |
| kashyap | Add these two entries: | 12:29 |
| kashyap | log_filters="1:libvirt 1:conf 1:security 3:event 3:json 3:file 3:object 1:util 1:command 1:qemu" | 12:29 |
| kashyap | log_outputs="1:file:/var/log/libvirt/libvirtd.log" | 12:29 |
| *** ducttape_ has quit IRC | 12:30 | |
| slaweq | kashyap: ok, I will try on devstack with master branch and will back to You with proper logs | 12:30 |
| kashyap | slaweq: Then, restart the libvirt daemon. Perform your test, and attach the /var/log/libvirt/libvirtd.log to the bug. | 12:30 |
| kashyap | Of course, all the above assumes you're doing this on the same Compute node where you're taking the snapshot. | 12:30 |
| *** devkulkarni1 has joined #openstack-nova | 12:31 | |
| slaweq | but one more quick question, is it normal that virDomainBlockJobInfo.end is 0 in libvirt? (maybe You know) | 12:31 |
| *** sudipto has joined #openstack-nova | 12:31 | |
| *** devkulkarni has quit IRC | 12:31 | |
| *** openstackgerrit has quit IRC | 12:34 | |
| kashyap | slaweq: When there's no running block operation (like blockRebase), then it should be zero | 12:34 |
| *** openstackgerrit has joined #openstack-nova | 12:34 | |
| *** kevinz has quit IRC | 12:34 | |
| *** kevinz has joined #openstack-nova | 12:35 | |
| slaweq | ok, so can I assume that in https://github.com/openstack/nova/blob/414df1e56ea9df700756a1732125e06c5d97d792/nova/virt/libvirt/guest.py#L692 should be running operation? | 12:35 |
| slaweq | so if is 0 then it is not completed? | 12:35 |
| *** su_zhang has joined #openstack-nova | 12:36 | |
| slaweq | or my thinking is wrong? | 12:36 |
| kashyap | slaweq: Check out -- http://libvirt.org/git/?p=libvirt.git;a=commitdiff;h=4cac241 | 12:36 |
| kashyap | "Ensure to zero out the virDomainBlockJobInfo arg" | 12:36 |
| *** gszasz has quit IRC | 12:36 | |
| *** cdent has left #openstack-nova | 12:39 | |
| *** cdent has joined #openstack-nova | 12:39 | |
| *** liverpooler has quit IRC | 12:39 | |
| slaweq | kashyap: I see that libvirt is set 0 in all bytes in struct "info" | 12:40 |
| slaweq | but as I understand if some block Job is running then it shouldn't be 0 at the end, right? | 12:41 |
| kashyap | slaweq: Yeah. So, default is zero, unless an active block dev job is in progress | 12:41 |
| kashyap | Right | 12:41 |
| kashyap | slaweq: BTW, instead of doing this cur == end check | 12:41 |
| slaweq | so if in Nova there is "start_block_operation" and then "wait_for_operation_is_done" it shouldn't be treated as done if "status.end=0" right? | 12:42 |
| *** eglynn has joined #openstack-nova | 12:42 | |
| kashyap | QEMU advertizes the 'ready' flag (as part of the 'query-block-jobs' command) which is a signal that, live block operation (like a copy is completely synced to dest) is complete | 12:42 |
| kashyap | Now, one can gracefully abort it (or pivot to that copy, if need be) | 12:43 |
| *** rlrossit has joined #openstack-nova | 12:43 | |
| *** lucas-hungry is now known as lucasagomes | 12:45 | |
| *** houming has quit IRC | 12:46 | |
| *** bkopilov has joined #openstack-nova | 12:46 | |
| kashyap | slaweq: Yes, as near as I know. Let's see the logs first, to see what you're hitting | 12:47 |
| *** tongli has joined #openstack-nova | 12:47 | |
| slaweq | kashyap: ok, I will try to reproduce it but for now it happend to me only on hosts with many vms | 12:47 |
| * johnthetubaguy is back from a slightly extended lunch talking to removal company! | 12:47 | |
| slaweq | I will try on devstack with master branch also | 12:47 |
| *** woodster_ has joined #openstack-nova | 12:48 | |
| *** tongli has quit IRC | 12:48 | |
| kashyap | slaweq: I've got the master branch, so I can try that report there in a few mins | 12:48 |
| *** tongli has joined #openstack-nova | 12:49 | |
| *** liverpooler has joined #openstack-nova | 12:49 | |
| slaweq | it would be great | 12:49 |
| kashyap | slaweq: You probably want to test with the buggy version you have, and post the log | 12:49 |
| slaweq | I will do it on my host for sure | 12:49 |
| alex_xu | nova-api subteam meeting is in 10 mins at #openstack-meeting-4 | 12:49 |
| *** hemnafk has quit IRC | 12:49 | |
| *** hemnafk has joined #openstack-nova | 12:50 | |
| *** Fdaisuke has quit IRC | 12:51 | |
| *** Fdaisuke has joined #openstack-nova | 12:51 | |
| *** gszasz has joined #openstack-nova | 12:52 | |
| *** ijw has joined #openstack-nova | 12:52 | |
| *** markvoelker has joined #openstack-nova | 12:54 | |
| slaweq | kashyap: I uploaded libvirt log in bug report but it's quite big | 12:56 |
| kashyap | slaweq: Oh, it should be less than 1MB for that test case. So you were albe to reproduce the prob? | 12:56 |
| *** ijw has quit IRC | 12:57 | |
| slaweq | it is log from "create instane, create snapshot from this instance, create instance from this snapshot" | 12:57 |
| slaweq | kashyap: for me it's reproducible almost every time on host with many instances | 12:57 |
| slaweq | on empty host I havn't got this issue so far | 12:57 |
| kashyap | slaweq: So this log you uploaded, is it when you hit the bug, I assume? | 12:58 |
| slaweq | yep | 12:58 |
| *** markvoelker has quit IRC | 12:58 | |
| slaweq | on http://pastebin.com/pc22CCEh You can check my logs from nova-compute | 12:59 |
| openstackgerrit | Chris Dent proposed openstack/nova: Add {get_,set_}aggregates to objects.ResourceProvider https://review.openstack.org/355263 | 12:59 |
| openstackgerrit | Chris Dent proposed openstack/nova: Implement get and set aggregates in the placement API https://review.openstack.org/357993 | 12:59 |
| openstackgerrit | Chris Dent proposed openstack/nova: Optional separate database for placement API https://review.openstack.org/362766 | 12:59 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] add a placement_aggregates table to api_db https://review.openstack.org/362863 | 12:59 |
| slaweq | there are my logs "error SK:..." added there to log values of status from libvirt | 13:00 |
| cdent | johnthetubaguy: I _think_ that ^ gets your comments | 13:00 |
| alex_xu | nova api subteam meeting is running | 13:00 |
| *** mriedem has joined #openstack-nova | 13:00 | |
| *** pradk has quit IRC | 13:01 | |
| cdent | ah johnthetubaguy forgot something, one sec | 13:01 |
| *** ratailor has quit IRC | 13:01 | |
| *** EinstCrazy has joined #openstack-nova | 13:03 | |
| kashyap | slaweq: Okay, with my versions, for the first instance, I couldn't reproduce the error when I live-snapshotted it | 13:04 |
| *** markvoelker has joined #openstack-nova | 13:04 | |
| *** devkulkarni has joined #openstack-nova | 13:04 | |
| *** devkulkarni1 has quit IRC | 13:04 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Add {get_,set_}aggregates to objects.ResourceProvider https://review.openstack.org/355263 | 13:04 |
| kashyap | slaweq: Perhaps I have to have a bunch more instances running, then it might occur. Looking at your logs now | 13:04 |
| openstackgerrit | Chris Dent proposed openstack/nova: Implement get and set aggregates in the placement API https://review.openstack.org/357993 | 13:04 |
| openstackgerrit | Chris Dent proposed openstack/nova: Optional separate database for placement API https://review.openstack.org/362766 | 13:04 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] add a placement_aggregates table to api_db https://review.openstack.org/362863 | 13:04 |
| *** lixiaoy1 has joined #openstack-nova | 13:06 | |
| *** ttrumm has quit IRC | 13:07 | |
| *** ttrumm has joined #openstack-nova | 13:08 | |
| openstackgerrit | Timofey Durakov proposed openstack/nova: ironic_host_manager: fix population of instances info on start https://review.openstack.org/347948 | 13:09 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: ironic_host_manager: fix population of instances info on schedule https://review.openstack.org/346966 | 13:09 |
| *** markvoelker has quit IRC | 13:10 | |
| openstackgerrit | Anne Gentle proposed openstack/nova: API ref: associate floating IP requires Active status https://review.openstack.org/363642 | 13:10 |
| tdurakov | alaski, just updated patch, please review, once have a time | 13:11 |
| *** tbachman has joined #openstack-nova | 13:11 | |
| *** prateek has quit IRC | 13:12 | |
| openstackgerrit | jichenjc proposed openstack/nova: no instance info cache update if instance deleted https://review.openstack.org/363585 | 13:12 |
| *** diga has quit IRC | 13:13 | |
| *** mtanino has joined #openstack-nova | 13:15 | |
| *** anush__ has quit IRC | 13:15 | |
| *** sridharg has quit IRC | 13:16 | |
| *** sdake_ has joined #openstack-nova | 13:16 | |
| *** ducttape_ has joined #openstack-nova | 13:16 | |
| *** cleong has joined #openstack-nova | 13:17 | |
| *** Hosam_ has quit IRC | 13:18 | |
| *** mdrabe has joined #openstack-nova | 13:18 | |
| kashyap | slaweq: Interesting, my successful libvirtd log is quite different than yours. Can you please exactly specify your libvirt / QEMU / Nova versions? | 13:18 |
| *** zz_dimtruck is now known as dimtruck | 13:18 | |
| *** su_zhang has quit IRC | 13:18 | |
| mriedem | bauzas: have you gone through the competing placement inventory RT changes from jaypipes and dansmith? | 13:19 |
| mriedem | do we need to drop one or just squash them? | 13:19 |
| *** sdake has quit IRC | 13:19 | |
| kashyap | slaweq: Don't bother, found it from your logs | 13:19 |
| *** zengine has joined #openstack-nova | 13:20 | |
| *** sridharg has joined #openstack-nova | 13:20 | |
| *** devkulkarni has quit IRC | 13:21 | |
| bauzas | mriedem: yup | 13:21 |
| bauzas | mriedem: I'm actually writing UTs against jay's | 13:21 |
| mriedem | bauzas: is one more "correct" than the other? | 13:21 |
| bauzas | mriedem: because he had some better implementation I think | 13:21 |
| bauzas | mriedem: yup | 13:21 |
| mriedem | ok, running the placement job on it now | 13:22 |
| bauzas | also, he covered the returned bits, while dansmith had no time for it I guess | 13:22 |
| *** sdake_ is now known as sdake | 13:22 | |
| *** devkulkarni has joined #openstack-nova | 13:23 | |
| *** coolsvap is now known as _coolsvap_ | 13:23 | |
| sdague | mriedem: jay's patch misses at least one of the safe_connect wraps | 13:24 |
| bauzas | sdague: I just updated it for that | 13:24 |
| sdague | bauzas: ok | 13:24 |
| cdent | I also think it is too complex | 13:24 |
| * cdent left a comment to that effect | 13:24 | |
| bauzas | I like the idea to transform the computenode into list of inventories | 13:25 |
| sdague | mriedem: I moved the placement job to check to speed things up | 13:25 |
| *** zengine has quit IRC | 13:25 | |
| sdague | so we're not constantly doing run experimental on these | 13:25 |
| bauzas | cdent: sure, that's adding more calls, but here, I like the generation bit check | 13:25 |
| bauzas | sdague: \o/ | 13:25 |
| *** devkulkarni has quit IRC | 13:25 | |
| *** devkulkarni1 has joined #openstack-nova | 13:26 | |
| cdent | bauzas: not the generation check, the comparison of inventories | 13:26 |
| cdent | it duplicates what the server will do for you | 13:26 |
| mriedem | sdague: ok | 13:26 |
| bauzas | cdent: meaning you'd get a new generation bit if the inventory changed ? | 13:27 |
| slaweq | kashyap: so is this log enough for You? | 13:27 |
| sdague | mriedem: so this patch I think is ready - https://review.openstack.org/#/c/358797/ | 13:27 |
| sdague | jay fixed the unit tests on it | 13:27 |
| cdent | sdague: I'm just looking through that one to make sure it doesn't talk crazy | 13:27 |
| kashyap | slaweq: Yeah, should be sufficient for now. I see the flow there. | 13:27 |
| sdague | it definitely works in a real environment | 13:27 |
| cdent | sdague: that's good, but reality can be a trap | 13:28 |
| kashyap | slaweq: I couldn't reproduce it on Nova Git master. And, the flow is different. I have to compare & contrast with yours. | 13:28 |
| mriedem | sdague: bauzas: yeah i'll review https://review.openstack.org/#/c/358797/ | 13:28 |
| *** jichen has joined #openstack-nova | 13:28 | |
| *** devkulkarni1 has quit IRC | 13:28 | |
| *** devkulkarni has joined #openstack-nova | 13:28 | |
| openstackgerrit | Alex Szarka proposed openstack/nova: Recover instances when physical data is missing https://review.openstack.org/322341 | 13:29 |
| sdague | mriedem: the wsgi script changes, which enable logging infrastructure, are in the gate now, plus devstack changes to set that up | 13:29 |
| *** zengine has joined #openstack-nova | 13:29 | |
| *** ljxiash has joined #openstack-nova | 13:29 | |
| mriedem | sdague: did you get up at 3am or what? | 13:29 |
| sdague | that's optimistically landing within an hour | 13:30 |
| *** zul has quit IRC | 13:30 | |
| sdague | mriedem: no, the changes from last night worked :) | 13:30 |
| sdague | I'm going to retool my logging patch after breakfast. Maybe enough bits will land that I don't need to cherry pick in a ton to get devstack working :) | 13:31 |
| *** su_zhang has joined #openstack-nova | 13:31 | |
| *** eglynn has quit IRC | 13:31 | |
| *** ekuris has quit IRC | 13:33 | |
| cdent | sdague, mriedem am I correct that the reno note on https://review.openstack.org/#/c/358797 needs to be fixed? | 13:33 |
| *** prateek has joined #openstack-nova | 13:33 | |
| *** zul has joined #openstack-nova | 13:33 | |
| *** ljxiash has quit IRC | 13:34 | |
| *** ansiwen has joined #openstack-nova | 13:34 | |
| sdague | cdent: we could probably tweak the reno to describe the authentication bits needed, do those get generated in the config reference | 13:34 |
| sdague | I would rather do that as a follow on, because it's currently a pretty manual process to get all these pieces into a devstack to actually see them working | 13:35 |
| cdent | that's why I didn't vote | 13:35 |
| mriedem | cdent: i don't think so | 13:36 |
| mriedem | http://docs-draft.openstack.org/97/358797/13/check/gate-nova-docs-ubuntu-xenial/6284a96//doc/build/html/sample_config.html | 13:36 |
| mriedem | the only thing in placement is #os_region_name = <None> | 13:37 |
| mriedem | unless the auth parts are missing from the generated config | 13:37 |
| *** pwnall has joined #openstack-nova | 13:37 | |
| mriedem | which i guess they are | 13:37 |
| *** mtanino has quit IRC | 13:37 | |
| *** ducttape_ has quit IRC | 13:38 | |
| mriedem | we should just fix that later with a bug fix though | 13:39 |
| *** su_zhang has quit IRC | 13:39 | |
| *** su_zhang has joined #openstack-nova | 13:39 | |
| *** chrichip has quit IRC | 13:40 | |
| *** EinstCrazy has quit IRC | 13:40 | |
| *** pbandark has quit IRC | 13:40 | |
| *** chrichip has joined #openstack-nova | 13:41 | |
| *** pwnall has quit IRC | 13:41 | |
| sdague | mriedem: yeh, sounds right. It would mirror what's in ironic / neutron, which are also probably poorly documented | 13:42 |
| *** pwnall has joined #openstack-nova | 13:43 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Add {get_,set_}aggregates to objects.ResourceProvider https://review.openstack.org/355263 | 13:43 |
| *** ttrumm has quit IRC | 13:43 | |
| mriedem | https://github.com/openstack/nova/blob/master/nova/conf/neutron.py#L116 | 13:43 |
| mriedem | ^ is how we do it for neutron | 13:43 |
| *** ttrumm has joined #openstack-nova | 13:44 | |
| *** su_zhang has quit IRC | 13:44 | |
| *** ducttape_ has joined #openstack-nova | 13:44 | |
| *** liverpooler has quit IRC | 13:44 | |
| sdague | mriedem: ah, ok, so that's just a follow on patch then, and we're good | 13:44 |
| sdague | I can write that one now if you like | 13:44 |
| mriedem | yeah, the ironic way is actually cleaner https://github.com/openstack/nova/blob/master/nova/conf/ironic.py#L101 | 13:44 |
| *** xyang1 has joined #openstack-nova | 13:44 | |
| openstackgerrit | Chris Dent proposed openstack/nova: Implement get and set aggregates in the placement API https://review.openstack.org/357993 | 13:44 |
| dansmith | cdent: bauzas: so has anyone actually tested jay's inventory patch? | 13:45 |
| *** rlrossit has quit IRC | 13:45 | |
| dansmith | it's a lot more complicated than mine, and I know mine *actually* works | 13:45 |
| dansmith | and we know jay's first RP one wasn't even close | 13:45 |
| bauzas | dansmith: I'm just running against that | 13:45 |
| cdent | dansmith: I haven't, yet. | 13:45 |
| *** rlrossit has joined #openstack-nova | 13:45 | |
| sdague | mriedem: ok, how about I fix neutron to be like ironic | 13:45 |
| sdague | and this like that too | 13:46 |
| mriedem | sdague: sounds good to me | 13:46 |
| bauzas | dansmith: there are a few bits that I'd love to see in your patch, actually, about the generation bit comparison and the local inventory cache | 13:46 |
| *** amotoki has quit IRC | 13:46 | |
| *** chrichip has quit IRC | 13:46 | |
| bauzas | s/inventory/inventories/ | 13:46 |
| mriedem | bauzas: in dan's change or from dan's change into jay's change? | 13:46 |
| dansmith | bauzas: yeah, was just trying to get something simple that worked given the time constraints | 13:46 |
| mriedem | dansmith: the placement job is running on jay's inventory patch now | 13:46 |
| bauzas | mriedem: tbh, I'm writing a bit of UTs that can be moved from one place to another | 13:46 |
| *** chrichip has joined #openstack-nova | 13:47 | |
| bauzas | dansmith: either way, I don't honestly care whether we should have all the CRUD batteries that jay did in a short time period | 13:47 |
| cdent | does this "placement job" actually verify any of the stored data, or is it just making sure things don't blow up? | 13:47 |
| mriedem | cdent: the latter | 13:48 |
| cdent | (which is great, even if just that) | 13:48 |
| mriedem | since nothing is using what's in the placement API | 13:48 |
| cdent | cool, just curious | 13:48 |
| dansmith | yeah, I've been doing it by hand | 13:48 |
| dansmith | we need to I think | 13:48 |
| dansmith | before we can be sure this is working | 13:48 |
| dansmith | so if we go with jay's, someone needs to validate it I think | 13:48 |
| cdent | brb | 13:48 |
| *** ttrumm has quit IRC | 13:48 | |
| mriedem | couldn't we have a functional test do that? | 13:48 |
| bauzas | we could | 13:48 |
| dansmith | mriedem: yes, and we need those, but given the stuff that was broken with the base patch I think we also want to hand-validate that it works end to end in a real devstack, | 13:49 |
| dansmith | until we can probe those values from tempest | 13:49 |
| mriedem | sure, in the interest of time i agree | 13:49 |
| *** devkulkarni1 has joined #openstack-nova | 13:49 | |
| bauzas | mmm, I'm f***d by my local machine ram space :/ | 13:50 |
| *** rlrossit has quit IRC | 13:50 | |
| dansmith | bauzas: so imho, we could merge my naive implementation right now (with unit tests) and then work on optimizing as bugs afterwards | 13:51 |
| dansmith | that's kinda what I was going for :) | 13:51 |
| dansmith | naive being.. "just throws a new generation every time" | 13:51 |
| bauzas | fair enough | 13:51 |
| dansmith | mriedem: have you been through jay's such that you would prefer going with his? kinda seems like you're leaning that way | 13:52 |
| johnthetubaguy | just to be clear, the complex follow up is avoid races by checking the generation? | 13:52 |
| dansmith | I need to get coffee before I can really go through it | 13:52 |
| dansmith | johnthetubaguy: no, mine does that | 13:52 |
| dansmith | johnthetubaguy: his just tries not to push a new generation if nothing has changed | 13:52 |
| mriedem | dansmith: i haven't gone through jay's inventory change yet | 13:52 |
| *** devkulkarni has quit IRC | 13:52 | |
| mriedem | i'm reviewing the bottom RT change | 13:52 |
| johnthetubaguy | dansmith: OK, that seems like more than enough then | 13:52 |
| *** ekhugen has quit IRC | 13:53 | |
| dansmith | johnthetubaguy: and I know it works, that's the important bit :D | 13:53 |
| *** amotoki has joined #openstack-nova | 13:53 | |
| *** ljxiash has joined #openstack-nova | 13:53 | |
| johnthetubaguy | +100 for working | 13:53 |
| johnthetubaguy | and being able to understand it | 13:53 |
| *** ecelik has quit IRC | 13:53 | |
| dansmith | yar | 13:54 |
| *** eharney has joined #openstack-nova | 13:54 | |
| *** ayoung has joined #openstack-nova | 13:54 | |
| dansmith | bauzas: so you have unit tests that can apply to mine already? | 13:54 |
| dansmith | else I will start on them | 13:55 |
| johnthetubaguy | its https://review.openstack.org/#/c/363061/ vs https://review.openstack.org/#/c/363583/ right? | 13:55 |
| *** jamielennox|away is now known as jamielennox | 13:55 | |
| *** liverpooler has joined #openstack-nova | 13:55 | |
| bauzas | dansmith: I have a few of them, gimme 20 mins to upload them | 13:55 |
| bauzas | since I need to adapt what I did | 13:55 |
| dansmith | okay cool | 13:55 |
| dansmith | johnthetubaguy: yeah | 13:56 |
| johnthetubaguy | cool, thanks | 13:56 |
| *** c64cosmin_ has joined #openstack-nova | 13:57 | |
| *** ekhugen has joined #openstack-nova | 13:57 | |
| *** tongli_ has joined #openstack-nova | 13:58 | |
| *** efried has quit IRC | 13:58 | |
| *** rlrossit has joined #openstack-nova | 14:00 | |
| mriedem | so to be clear, we're ok with fully disabling and not retrying to connect to the placement service if there is a missing service catalog entry? and you can't get the compute trying to conncet again w/o restarting the nova-compute service? | 14:00 |
| *** anush__ has joined #openstack-nova | 14:00 | |
| *** links has quit IRC | 14:00 | |
| *** tongli has quit IRC | 14:01 | |
| dansmith | I think that's fine, personally | 14:01 |
| mriedem | that kind of seems like a pain in the ass if you have thousands of computes and just needed to add a global entry to the SC | 14:01 |
| dansmith | meh, pssh | 14:01 |
| johnthetubaguy | mriedem: that sounds like an OK starting point, could always retry on SIG_UP as well | 14:01 |
| johnthetubaguy | HUP | 14:01 |
| *** jamielennox is now known as jamielennox|away | 14:01 | |
| sdague | mriedem: the SC should be pretty static | 14:01 |
| mriedem | we might want to note that in the release note anyway | 14:02 |
| dansmith | sure, reno it definitely | 14:02 |
| *** chrichip has quit IRC | 14:02 | |
| *** rlrossit_ has joined #openstack-nova | 14:02 | |
| *** pbandark has joined #openstack-nova | 14:02 | |
| *** jamielennox|away is now known as jamielennox | 14:02 | |
| *** pradk has joined #openstack-nova | 14:02 | |
| *** burt has joined #openstack-nova | 14:02 | |
| *** c64cosmin_ has quit IRC | 14:02 | |
| *** rfolco has quit IRC | 14:03 | |
| *** chrichip has joined #openstack-nova | 14:03 | |
| cdent | I don't understand why we shouldn't just retry every time? Where's the expense? Surely the service catalog is being queried a lot already? Forcing people to restart their computes when they suddently decide "oh hey, I wanna placement endpoint" seems quite mean. | 14:03 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: Use more specific asserts in tests https://review.openstack.org/363043 | 14:04 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: hacking: Use assertIs(Not), assert(True|False) https://review.openstack.org/363685 | 14:04 |
| *** rfolco has joined #openstack-nova | 14:04 | |
| *** dimtruck is now known as zz_dimtruck | 14:04 | |
| *** rlrossit has quit IRC | 14:04 | |
| mriedem | well, you'll have to restart nova-compute once you add the auth entries to [placement] in nova.conf on that compute anyway | 14:04 |
| sdague | yep | 14:04 |
| *** ducttape_ has quit IRC | 14:05 | |
| *** devkulkarni1 has quit IRC | 14:05 | |
| *** cdent_ has joined #openstack-nova | 14:05 | |
| *** cdent has quit IRC | 14:06 | |
| *** sdake_ has joined #openstack-nova | 14:06 | |
| mriedem | sdague: why aren't the safe_connector decorators on the get and post methods? | 14:06 |
| sdague | the only scenario in which this would be relevant is if you rolled all 1000 compute nodes with all the auth credentials restarted them all (they don't take live) but forgot to add the SC entry first | 14:06 |
| mriedem | rather than the things calling those? | 14:06 |
| sdague | mriedem: because otherwise you need error handling on resp.status_code | 14:06 |
| *** akshai has joined #openstack-nova | 14:07 | |
| sdague | because resp would be None | 14:07 |
| *** cdent_ has quit IRC | 14:07 | |
| mriedem | oh right | 14:07 |
| sdague | and, it means that you short circuit abort *all* these methods if you are disabled | 14:07 |
| *** cdent has joined #openstack-nova | 14:07 | |
| openstackgerrit | sahid proposed openstack/os-vif: Adds Windows support for OvsPlugin https://review.openstack.org/344433 | 14:07 |
| openstackgerrit | sahid proposed openstack/os-vif: add standard constants for supported platforms https://review.openstack.org/363689 | 14:07 |
| openstackgerrit | sahid proposed openstack/os-vif: rework ovs plugin in order to use platform module https://review.openstack.org/363690 | 14:07 |
| *** sdake has quit IRC | 14:07 | |
| cdent | ah okay, that makes sense. still pretty mean! :) | 14:08 |
| *** ducttape_ has joined #openstack-nova | 14:08 | |
| sdague | cdent: it's not mean, there is just no realistic operational situation where it's an issue | 14:08 |
| mriedem | the msg.format() stuff is something i havne't seen in awhile :) | 14:08 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: ironic_host_manager: fix population of instances info on start https://review.openstack.org/347948 | 14:08 |
| *** tonytan4ever has quit IRC | 14:09 | |
| * cdent switches everything to be minions talking to etcd | 14:09 | |
| *** tonytan4ever has joined #openstack-nova | 14:09 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Register keystone opts for placement sample config https://review.openstack.org/363699 | 14:10 |
| *** rfolco has quit IRC | 14:11 | |
| *** rfolco has joined #openstack-nova | 14:12 | |
| cdent | sdague: https://review.openstack.org/#/c/363333/ went through without me having a chance to look at it last night (which is fine), but I have some questions: a) you named it nova-placement-api which seems an error from the long term perspective, b) should we follow with a delete of nova/api/openstack/placement/placement-api.py ? | 14:12 |
| cdent | (I'll do (b) if nobody is on that and that's groovy) | 14:13 |
| sdague | everything nova builds for binaries is nova-, when we do a split we can do a deprecation phase | 14:13 |
| sdague | b) yes | 14:13 |
| *** kevinz has quit IRC | 14:13 | |
| *** marst has quit IRC | 14:13 | |
| sdague | but it needed - https://review.openstack.org/#/c/363335/ to land first | 14:14 |
| cdent | or we could just avoid a deprecation by using the "right" name from that start? | 14:14 |
| * cdent nods | 14:14 | |
| sdague | we don't know what the right name is | 14:14 |
| sdague | the placement project has no name | 14:14 |
| edleafe | ... or mascot | 14:14 |
| *** kevinz has joined #openstack-nova | 14:14 | |
| cdent | ah, I though we had decided we were going to stop with the toy names and just name things by their service type (in which case placement-api is right) | 14:15 |
| cdent | (not necessarily projects, but their products) | 14:15 |
| mriedem | sdague: bauzas: do we have any unit tests in https://review.openstack.org/#/c/358797/ for the case where safe_connect short circuits on self._disabled being True? | 14:15 |
| *** amotoki has quit IRC | 14:16 | |
| sdague | cdent: I don't know who thought they decided what there, but as there is *not* a placement project in governance, thinking that was decided any way is premature | 14:16 |
| *** pbandark has quit IRC | 14:16 | |
| sdague | mriedem: there are not yet | 14:17 |
| sdague | I can either work on that patch next, or the logging on the placement api | 14:17 |
| sdague | which would you like? | 14:17 |
| mriedem | probably logging | 14:17 |
| sdague | ok | 14:17 |
| mriedem | my comments are in https://review.openstack.org/#/c/358797/ | 14:18 |
| mriedem | nothing blocking, i'd -1 if it were different circumstances | 14:18 |
| cdent | edleafe: have you and sdague collided on logging or did you reach some kind of aggreement? | 14:18 |
| sdague | mriedem: if the only concern is missing unit tests, we can do those as a follow on patch | 14:18 |
| mriedem | it's not only that, but we can address my comments as follow ups i think | 14:18 |
| *** tonytan4ever has quit IRC | 14:18 | |
| edleafe | cdent: sdague really doesn't want to copy FaultWrapper now | 14:18 |
| dansmith | I'm going to start working on reviewing cdent's allocation stuff because we're still effed if we don't get that in too | 14:18 |
| sdague | mriedem: ok | 14:18 |
| sdague | dansmith: sounds great | 14:19 |
| dansmith | but when that is done I can circle back on whatever | 14:19 |
| *** ducttape_ has quit IRC | 14:19 | |
| cdent | thanks dansmith | 14:19 |
| *** pwnall has quit IRC | 14:19 | |
| edleafe | cdent: maybe mine can be a follow-up | 14:19 |
| sdague | bauzas: you were in a unit test mood right? | 14:19 |
| sdague | maybe we can draft you to test that decorator :) | 14:19 |
| cdent | dansmith: it has, as noted in the commit messages and code a significant but not totally blocking TODO that needs some discussion | 14:19 |
| *** tonytan4ever has joined #openstack-nova | 14:20 | |
| cdent | and it is fast to complete code, not nice code | 14:20 |
| *** pwnall has joined #openstack-nova | 14:20 | |
| * cdent tries to come up with some more caveats | 14:20 | |
| *** mtanino has joined #openstack-nova | 14:20 | |
| *** su_zhang has joined #openstack-nova | 14:21 | |
| mriedem | sdague: we can at least see it works in practice http://logs.openstack.org/97/358797/13/check/gate-tempest-dsvm-neutron-full-ubuntu-xenial/3ce2677/logs/screen-n-cpu.txt.gz?level=WARNING#_2016-08-31_11_37_54_356 | 14:21 |
| mriedem | apparently ksa checks the auth options before trying to check the SC, but we hit that and disable and don't check again, so that's working | 14:21 |
| sdague | yes | 14:21 |
| sdague | dansmith noticed that yesterday | 14:21 |
| sdague | the endpoint fail scenario I manually tested locally by changing the service type in the code to garbage | 14:22 |
| mriedem | ok, +W | 14:23 |
| openstackgerrit | Andrey Volkov proposed openstack/nova: WIP Map instances to cell0 on failed schedule https://review.openstack.org/267827 | 14:23 |
| sdague | \o/ - one more step forward | 14:24 |
| mriedem | https://github.com/openstack-dev/devstack/blob/master/lib/placement#L117 | 14:24 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: ironic_host_manager: fix population of instances info on start https://review.openstack.org/347948 | 14:24 |
| mriedem | that should probably be os_region_name shouldn't it? | 14:24 |
| mriedem | not that it matters, we only have one region in devstack | 14:24 |
| sdague | probably | 14:25 |
| *** mamandle has joined #openstack-nova | 14:25 | |
| *** zz_dimtruck is now known as dimtruck | 14:25 | |
| sdague | push a fix and I'll +2 | 14:25 |
| *** amotoki has joined #openstack-nova | 14:26 | |
| cdent | sdague, mriedem, dansmith we still planning a hangout in approx 1.5 hours? If so I might go do something else for an hour or so (while dan reivews my shti) to trade some time for the evening when you guys will be around. | 14:26 |
| *** marst has joined #openstack-nova | 14:26 | |
| sdague | yeh | 14:26 |
| mriedem | yup | 14:26 |
| openstackgerrit | Andrew Laski proposed openstack/nova: DNM logs for debugging https://review.openstack.org/362442 | 14:27 |
| cdent | cool, see ya | 14:27 |
| *** Guest36472 has joined #openstack-nova | 14:29 | |
| dansmith | hmm, I'm missing context on the placement_aggregates thing | 14:29 |
| dansmith | alaski: do you understand that? | 14:29 |
| *** lixiaoy1 has quit IRC | 14:30 | |
| *** tonytan4ever has quit IRC | 14:30 | |
| alaski | dansmith: link? I haven't been following the channel very well today, tracking a weird cells/service version thing | 14:31 |
| dansmith | alaski: well this and down: https://review.openstack.org/#/c/355263/10 | 14:31 |
| *** tonytan4ever has joined #openstack-nova | 14:31 | |
| dansmith | I wasn't thinking we'd have two sets of aggregates, one for regular and one for "placement" | 14:31 |
| alaski | yeah, that's a big point of unclarity | 14:32 |
| dansmith | I also don't see what that has to do with being able to set allocations from the clients either | 14:32 |
| dansmith | but he said it was related | 14:32 |
| dansmith | so I'm super confused | 14:32 |
| *** gszasz has quit IRC | 14:32 | |
| *** akshai has quit IRC | 14:32 | |
| *** ZZelle has joined #openstack-nova | 14:32 | |
| alaski | as I understand it the aggregates in Nova were not going to be moved fully to the placement db, once placement db is fully separate from Nova. So the idea was to just record aggregate uuids and their relation to a rp | 14:33 |
| *** akshai has joined #openstack-nova | 14:33 | |
| dansmith | I'm not sure how that helps | 14:33 |
| alaski | all a rp cares about is which aggregate(s) it's associated with | 14:34 |
| johnthetubaguy | helps compared to what? | 14:34 |
| alaski | not any more details about the agg | 14:34 |
| dansmith | if we record them, then we keep them all as aggregates in the same place and have a list of the ones we need to send with placement when we move it, | 14:34 |
| *** mlavalle has joined #openstack-nova | 14:34 | |
| dansmith | is that the point? | 14:35 |
| dansmith | I'm not sure why we need an API for that | 14:35 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: hacking: Use assertIs(Not), assert(True|False) https://review.openstack.org/363685 | 14:35 |
| dansmith | and really not sure how that is related to being able to write allocations from the compute nodes, which is, IMHO the important line we need to cross | 14:35 |
| *** amotoki has quit IRC | 14:35 | |
| sdague | so, the separate database if fully a distraction for newton, right? | 14:35 |
| sdague | like, nothing breaks without it | 14:36 |
| *** oanson has quit IRC | 14:36 | |
| *** rmcall has joined #openstack-nova | 14:36 | |
| *** su_zhang has quit IRC | 14:36 | |
| dansmith | being able to have the separate one if you jump through hoops was the compromise between jay wanting a fully separate one and me thinking it was just a distraction | 14:36 |
| alaski | on a call atm, will get this in my head in a few | 14:36 |
| *** ansiwen has quit IRC | 14:36 | |
| *** su_zhang has joined #openstack-nova | 14:37 | |
| *** david-lyle_ has joined #openstack-nova | 14:37 | |
| dansmith | however, this aggregate list is, I think, just to facilitate the later split according to alaski | 14:37 |
| dansmith | seems odd to me, but... | 14:37 |
| alaski | dansmith: yes, I believe that's the case | 14:37 |
| *** ansiwen has joined #openstack-nova | 14:37 | |
| dansmith | I still don't see how it's related to allocations | 14:37 |
| moshele | dansmith, jaypipes: hi, can you review some pci patches https://review.openstack.org/#/c/349060 and https://review.openstack.org/#/c/347558/ | 14:37 |
| *** salv-orlando has quit IRC | 14:38 | |
| dansmith | moshele: not right now, sorry... and jaypipes is not around | 14:38 |
| *** ansiwen has quit IRC | 14:39 | |
| *** ansiwen has joined #openstack-nova | 14:40 | |
| edleafe | dansmith: it doesn't seem to be related. My understanding is that it's just a step toward the eventual separation, when it is needed for pooled resources | 14:40 |
| dansmith | edleafe: yeah, so it's just a complete distraction right now, but mriedem and cdent seem to think it's critical | 14:41 |
| dansmith | although I think mriedem just because cdent | 14:41 |
| *** su_zhang has quit IRC | 14:41 | |
| *** su_zhang has joined #openstack-nova | 14:41 | |
| edleafe | dansmith: let's talkabout that in the hangout then | 14:41 |
| *** kevinz has quit IRC | 14:41 | |
| dansmith | edleafe: by that time I may be reduced to just making fart noises into the mic, but.. yeah | 14:41 |
| mriedem | i don't know enough about what's required in the placement api to have an opinion on what's critical | 14:42 |
| *** kevinz has joined #openstack-nova | 14:42 | |
| edleafe | dansmith: :) | 14:42 |
| sdague | that's fine, but I'm definitely in the camp that separate db needs to just be dropped. That had until last week to merge, right now it gets in the way of moving forward | 14:42 |
| mriedem | i would have to reference the spec | 14:42 |
| sdague | mriedem: well, it's definitively optional | 14:42 |
| mriedem | unless i'm missing what it is that i think is critical? | 14:42 |
| mriedem | oh i thought you guys were talking about aggregates in the placement api | 14:42 |
| mriedem | which i know nothing about | 14:43 |
| dansmith | sdague: the completely-totally-separate DB thing is what we already reverted, yeah.. this was just setup for a later split, | 14:43 |
| mriedem | the separate placement db is not critical, i never said it was | 14:43 |
| dansmith | but agree, it's low on the prio list compared to the critical things | 14:43 |
| *** tonytan4ever has quit IRC | 14:43 | |
| dansmith | not sure cdent agrees, but we'll see when he returns | 14:43 |
| sdague | mriedem: right, the issue is right now the aggregate api is stacked on the optional db patches | 14:43 |
| *** tonytan4ever has joined #openstack-nova | 14:43 | |
| mriedem | well we said it was optional at the midcycle, and it was in the etherpad with a big fat 'decision' note, and it was probably also in my recap of the midcycle | 14:44 |
| mriedem | so yeah | 14:44 |
| mriedem | optional :) | 14:44 |
| mriedem | oh | 14:44 |
| mriedem | well, shart | 14:44 |
| *** ansiwen has quit IRC | 14:44 | |
| sdague | right, so I think we go from optional, to dead and move that bit to ocata | 14:44 |
| dansmith | there is no allocation handler in the placement api at the moment, so *that* is the thing we need to find and get mergeable before we can do it from the compute client | 14:44 |
| mriedem | dansmith: i believe that is this and down https://review.openstack.org/#/c/339874/ | 14:45 |
| *** ansiwen has joined #openstack-nova | 14:45 | |
| dansmith | mriedem: ah yeah.. last night you said it was the aggregate bit but maybe you linked the wrong thing? | 14:45 |
| *** mdrabe has quit IRC | 14:46 | |
| mriedem | last night i linked https://review.openstack.org/#/c/363209/2 | 14:46 |
| dansmith | okay well maybe I picked the wrong thing this morning thn | 14:46 |
| mriedem | btw, the placement job just completed on jay's inventory change http://logs.openstack.org/83/363583/1/experimental/gate-tempest-dsvm-neutron-placement-full-ubuntu-xenial-nv/d288f21/ | 14:46 |
| kashyap | markus_z: Hey, I think this bug can have an exception of not placing in "expired" category -- https://bugs.launchpad.net/nova/+bug/1334398 | 14:46 |
| openstack | Launchpad bug 1334398 in OpenStack Compute (nova) "libvirt live_snapshot periodically explodes on libvirt 1.2.2 in the gate" [Undecided,Expired] | 14:46 |
| mriedem | http://logs.openstack.org/83/363583/1/experimental/gate-tempest-dsvm-neutron-placement-full-ubuntu-xenial-nv/d288f21/logs/screen-n-cpu.txt.gz?level=TRACE | 14:46 |
| kashyap | markus_z: Because we have a workaround that's 'True' by default in Nova Git master ('workarounds.disable_libvirt_livesnapshot = True'), until that bug is resolved. | 14:47 |
| *** ducttape_ has joined #openstack-nova | 14:47 | |
| dansmith | mriedem: .....yeah. | 14:47 |
| *** amotoki has joined #openstack-nova | 14:48 | |
| *** gszasz_ has joined #openstack-nova | 14:48 | |
| *** tonytan_brb has joined #openstack-nova | 14:49 | |
| mriedem | i'm not seeing anything in the n-cpu logs from his _inventory_changed method | 14:50 |
| *** tonytan4ever has quit IRC | 14:50 | |
| *** Swami has joined #openstack-nova | 14:52 | |
| markus_z | kashyap: yeah, true, I'll revert the expiration and leave a comment. Good catch. | 14:52 |
| dansmith | mriedem: we're moving on with mine right? | 14:52 |
| *** markvoelker has joined #openstack-nova | 14:52 | |
| mriedem | sure - i missed the major difference, unnecessary client side optimization in jay's that we can add in later? | 14:52 |
| mriedem | well, not unnecessary, | 14:53 |
| mriedem | but not required for FF | 14:53 |
| *** ducttape_ has quit IRC | 14:53 | |
| *** fragatina has joined #openstack-nova | 14:53 | |
| kashyap | markus_z: Thanks. | 14:53 |
| *** fragatina has quit IRC | 14:53 | |
| dansmith | mriedem: http://logs.openstack.org/61/363061/2/experimental/gate-tempest-dsvm-neutron-placement-full-ubuntu-xenial-nv/780cd4a/logs/screen-n-cpu.txt.gz#_2016-08-30_23_28_58_954 | 14:53 |
| *** ducttape_ has joined #openstack-nova | 14:53 | |
| dansmith | mriedem: right | 14:53 |
| dansmith | mriedem: that log link is mine | 14:53 |
| dansmith | mriedem: and that log statement means everything was cool | 14:54 |
| *** fragatina has joined #openstack-nova | 14:54 | |
| *** sudipto has quit IRC | 14:54 | |
| *** nkrinner is now known as nkrinner_afk | 14:54 | |
| mriedem | dansmith: yeah i checked yours over last night | 14:54 |
| *** mdrabe has joined #openstack-nova | 14:54 | |
| mriedem | i still don't know why we're hitting this but it's in all of them | 14:54 |
| sdague | mriedem: something is just not right at that job - http://logs.openstack.org/83/363583/1/experimental/gate-tempest-dsvm-neutron-placement-full-ubuntu-xenial-nv/d288f21/logs/screen-n-cpu.txt.gz#_2016-08-31_13_43_04_633 | 14:54 |
| mriedem | http://logs.openstack.org/61/363061/2/experimental/gate-tempest-dsvm-neutron-placement-full-ubuntu-xenial-nv/780cd4a/logs/screen-n-cpu.txt.gz?level=TRACE#_2016-08-30_23_25_52_632 | 14:54 |
| mriedem | so just a bug to work out later | 14:55 |
| *** sudipto has joined #openstack-nova | 14:55 | |
| *** sudipto has quit IRC | 14:55 | |
| sdague | it's calling and the other side isn't responding | 14:55 |
| sdague | mriedem: configuration fail - http://logs.openstack.org/83/363583/1/experimental/gate-tempest-dsvm-neutron-placement-full-ubuntu-xenial-nv/d288f21/logs/apache/placement-api.txt.gz ? | 14:55 |
| sdague | the placement api didn't start there | 14:55 |
| mriedem | sdague: jay's change isn't rebased on the latest version of the RT change below it | 14:56 |
| mriedem | not sure if that would make a difference | 14:56 |
| *** raunak has joined #openstack-nova | 14:56 | |
| sdague | yeh, I don't know | 14:56 |
| sdague | mriedem: https://review.openstack.org/#/c/363699/ | 14:56 |
| mriedem | anyway, let's not get distracted with jay's inventory change | 14:56 |
| sdague | we don't need session options there | 14:57 |
| sdague | notice we never register them | 14:57 |
| mriedem | sdague: yeah, i know, but i think that's an oversight too | 14:57 |
| sdague | I don't think it is | 14:57 |
| mriedem | jay was originally registering those | 14:57 |
| dansmith | mriedem: yeah, dunno about that initial 404, but maybe just a sequencing thing.. we do start n-cpu before placement | 14:57 |
| sdague | right, except jay had no idea how any of this worked | 14:57 |
| mriedem | dansmith: oh yeah good point | 14:58 |
| mriedem | sdague: :) | 14:58 |
| bauzas | dansmith: I need to bail out for getting my car at the garage but you can find my PS | 14:58 |
| *** kevinz has quit IRC | 14:58 | |
| *** fragatin_ has joined #openstack-nova | 14:58 | |
| alaski | dansmith: back. so looking at the resource-providers-allocations spec I think the agg info is needed because allocations are reported against a "consumer_uuid" which is an agg or compute uuid iiuc | 14:58 |
| *** kevinz has joined #openstack-nova | 14:59 | |
| cdent | dansmith: the placementdb stuff is optional but very useful and my understanding when we evaluated this etherpad https://etherpad.openstack.org/p/placement-db that we were going to go for it. | 14:59 |
| dansmith | alaski: but for the compute node resources it should always be a compute node | 14:59 |
| dansmith | bauzas: are you pushing something? | 14:59 |
| *** david-lyle_ has quit IRC | 14:59 | |
| alaski | dansmith: yes | 14:59 |
| *** raunak has quit IRC | 14:59 | |
| bauzas | dansmith: I just did | 14:59 |
| alaski | dansmith: for everything except shared storage it would be a compute node | 14:59 |
| dansmith | cdent: no, I think most of us agree the placement db is way way way out of the running at this point | 14:59 |
| bauzas | dansmith: oh the bot seems to be off | 14:59 |
| cdent | the PlacementAggregates do not depend on it, but I put them above the db changes because of the context manager handling | 15:00 |
| dansmith | cdent: we're seriously in danger of not making this by FF, IMHO | 15:00 |
| bauzas | dansmith: https://review.openstack.org/#/c/363061/ | 15:00 |
| * cdent shrugs | 15:00 | |
| dansmith | bauzas: okay thanks I shall pick it up | 15:00 |
| *** ijw has joined #openstack-nova | 15:00 | |
| cdent | the aggregates stuff is not at all related to the allocation stuff, I'm not sure where that came from | 15:00 |
| * bauzas bailing out for the next hours, but bbiab for my evening | 15:00 | |
| dansmith | cdent: might just be me picking the wrong set pre-coffee or something | 15:00 |
| cdent | both should be linked from yesterdays etherpad: https://etherpad.openstack.org/p/placement-next | 15:00 |
| dansmith | cdent: thought you and mriedem pointed to it yesterday | 15:00 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: hacking: Always use 'assertIs(Not)None' https://review.openstack.org/363623 | 15:01 |
| openstackgerrit | Stephen Finucane proposed openstack/nova: hacking: Reword in the imperative https://review.openstack.org/363624 | 15:01 |
| *** mgoddard__ has joined #openstack-nova | 15:01 | |
| cdent | i'll rebase the set_aggregates to not use the placement db now | 15:01 |
| *** tongli_ has quit IRC | 15:01 | |
| *** ttrumm has joined #openstack-nova | 15:02 | |
| *** tongli has joined #openstack-nova | 15:02 | |
| *** fragatina has quit IRC | 15:02 | |
| mriedem | sdague: the sessions options give the ability to make secure rest calls to the placement api right? | 15:02 |
| mriedem | with a cert? | 15:02 |
| *** efried has joined #openstack-nova | 15:03 | |
| sdague | honestly, we should loop in a keystone person on that | 15:04 |
| *** mgoddard_ has quit IRC | 15:04 | |
| sdague | because, ffs, it's the dumbest named thing if that's the case :) | 15:04 |
| *** dmellado is now known as dmellado|mtg | 15:04 | |
| mriedem | jamielennox: are you around? | 15:05 |
| mriedem | or stevemar | 15:05 |
| openstackgerrit | Sivasathurappan Radhakrishnan proposed openstack/nova: Using get() method to prevent KeyError https://review.openstack.org/359424 | 15:05 |
| *** jamielennox is now known as jamielennox|away | 15:05 | |
| auggy | johnthetubaguy: got all the tests passing on https://review.openstack.org/#/c/357540 just have to fix a pep8 issue | 15:06 |
| sdague | mriedem: you've got to be kidding me..... | 15:06 |
| johnthetubaguy | auggy: awesome, those tests are super scary | 15:06 |
| *** tongli has quit IRC | 15:07 | |
| sdague | oh, wait, it's a bunch of aliases, this will take me some time sift through | 15:07 |
| auggy | johnthetubaguy: yeah... well once i figured out how mox was supposed to work (thanks melwitt!) then it was just a matter of getting the order of operations to match | 15:07 |
| sdague | mriedem: you are correct, session makes it work with certs | 15:07 |
| *** abalutoiu has quit IRC | 15:08 | |
| auggy | johnthetubaguy: on thing i did that may or may not make sense, since we are dealing with 2 neutron clients in some cases now, i made a second mock_client that i called mock_admin_client... | 15:08 |
| johnthetubaguy | auggy: yeah, those unit tests are more complex than the code right now, there are mock ones I have started adding already, with the aim of killing the mox ones | 15:08 |
| *** jaosorior has quit IRC | 15:08 | |
| johnthetubaguy | auggy: thats probably simpler | 15:08 |
| *** ijw has quit IRC | 15:09 | |
| *** armax has joined #openstack-nova | 15:09 | |
| *** _coolsvap_ is now known as coolsvap | 15:09 | |
| sdague | so... that wins the worst naming collision in keystone for this cycle | 15:09 |
| *** hemnafk is now known as hemna | 15:09 | |
| sdague | session means active connection, also means certificate configs | 15:09 |
| auggy | johnthetubaguy: ah nice, yeah well after digging around in those, i can help with moving those over to mock after we get through this FF hump | 15:09 |
| *** markus_z has quit IRC | 15:10 | |
| *** yamahata has joined #openstack-nova | 15:10 | |
| *** rmart04 has quit IRC | 15:11 | |
| mriedem | sdague: \o/ | 15:11 |
| mriedem | sdague: there is some other weirdness i found when diffing the neutron options before and after your change | 15:12 |
| mriedem | it looks like the 3 deprecated neutron options don't show up as deprecated/aliases in the config sample, i'm not sure if they'd continue to work or not though | 15:12 |
| sdague | mriedem: ok, I'll let you run with that patch, otherwise I'll never get this logging stuff sorted | 15:12 |
| mriedem | if those were deprecated in mitaka we could just drop them and not care | 15:12 |
| *** trisq has quit IRC | 15:13 | |
| *** trisq has joined #openstack-nova | 15:13 | |
| mriedem | sdague: sounds good | 15:13 |
| *** diogogmt has joined #openstack-nova | 15:14 | |
| *** pwnall has joined #openstack-nova | 15:14 | |
| *** tongli has joined #openstack-nova | 15:14 | |
| mriedem | yeah i'm just going to remove those in a pre-patch, they were deprecated since kilo | 15:15 |
| *** sudipto has joined #openstack-nova | 15:15 | |
| *** tonytan4ever has joined #openstack-nova | 15:15 | |
| *** rrecio has joined #openstack-nova | 15:15 | |
| dansmith | sdague: mriedem so I think bauzas' unit tests on my patch are good to go | 15:15 |
| dansmith | I want to make sure I get a CAB on there for him, but.. does the base patch now need some love? if so I can do that | 15:16 |
| mriedem | dansmith: the base RT one? | 15:16 |
| mriedem | that's +W | 15:16 |
| dansmith | yeah, I thought some extra test path needed doing | 15:17 |
| *** rlrossit_ has quit IRC | 15:17 | |
| mriedem | dansmith: there are several things to be cleaned up / addressed, but those are happening in later changes | 15:17 |
| dansmith | okay, shall I do those or is someone else on those already? | 15:17 |
| *** tongli has quit IRC | 15:18 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Report compute node inventories through placement https://review.openstack.org/363061 | 15:18 |
| mriedem | i haven't heard of a volunteer, but they could wait | 15:18 |
| mriedem | sean is working on improving logging in the placement api | 15:18 |
| *** tonytan_brb has quit IRC | 15:18 | |
| cdent | dansmith: did you do a review on the AllocationsList stuff? | 15:18 |
| dansmith | okay well, I think this is ready for you to pick apart: https://review.openstack.org/#/c/363061/ | 15:18 |
| mriedem | i'm working on cleaning up the ksa options in the placement section | 15:18 |
| dansmith | cdent: no I have it open though | 15:19 |
| *** Hosam has joined #openstack-nova | 15:19 | |
| cdent | cool | 15:19 |
| openstackgerrit | Aline Bousquet proposed openstack/nova: Add Image Remote Authentication Verification https://review.openstack.org/330773 | 15:19 |
| openstackgerrit | Chris Dent proposed openstack/nova: Add {get_,set_}aggregates to objects.ResourceProvider https://review.openstack.org/355263 | 15:19 |
| openstackgerrit | Chris Dent proposed openstack/nova: Implement get and set aggregates in the placement API https://review.openstack.org/357993 | 15:19 |
| openstackgerrit | Chris Dent proposed openstack/nova: [placement] add a placement_aggregates table to api_db https://review.openstack.org/362863 | 15:19 |
| johnthetubaguy | dansmith: I think those reserved values need to come from the appropriate CONF values? | 15:19 |
| cdent | that ^ pulls the set_aggregates out from under the new database | 15:19 |
| dansmith | johnthetubaguy: oh sorry hadn't seen your comments, lemme look | 15:20 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Remove deprecated neutron options https://review.openstack.org/363746 | 15:20 |
| johnthetubaguy | ah, so I notice bauzas says that not needed, but I am not sure why | 15:20 |
| dansmith | johnthetubaguy: okay yeah I can change that.. did you see bauzas' reply on the memory one? | 15:20 |
| *** mamandle has quit IRC | 15:20 | |
| *** rlrossit has joined #openstack-nova | 15:21 | |
| johnthetubaguy | ah, yeah, I should follow up with bauzas on that one | 15:21 |
| *** pcaruana has quit IRC | 15:22 | |
| *** rcernin has quit IRC | 15:22 | |
| johnthetubaguy | bauzas that ram one worries me, because I know getting CONF.reserved_host_memory_mb wrong really breaks rackspace public cloud in horrible ways | 15:22 |
| *** mjura has quit IRC | 15:23 | |
| dansmith | johnthetubaguy: so, here's the thing | 15:23 |
| dansmith | johnthetubaguy: we're not scheduling on this until ocata, and we update this value every minute | 15:23 |
| *** sudipto has quit IRC | 15:23 | |
| dansmith | johnthetubaguy: so I think we'll be okay if we figure out later that it needs to be adjusted :) | 15:23 |
| openstackgerrit | Andrey Volkov proposed openstack/nova: Eliminate additional DB queries in nova lists https://review.openstack.org/355558 | 15:24 |
| openstackgerrit | Andrey Volkov proposed openstack/nova: Wrap servers detail API call in single transaction https://review.openstack.org/352950 | 15:24 |
| johnthetubaguy | dansmith: yeah, thats true | 15:24 |
| *** mdrabe has quit IRC | 15:24 | |
| dansmith | johnthetubaguy: that said, I also don't know why it's not right to use that value | 15:24 |
| dansmith | johnthetubaguy: so let me just make that change and we can alter it later if need be | 15:24 |
| *** mdrabe has joined #openstack-nova | 15:24 | |
| johnthetubaguy | yeah, it feels nice being consistent there, even if that turns out to be wrong. | 15:25 |
| dansmith | aye | 15:25 |
| *** rlrossit has quit IRC | 15:25 | |
| *** su_zhang has quit IRC | 15:25 | |
| dansmith | johnthetubaguy: making that change now.. anything else from you on it? | 15:25 |
| johnthetubaguy | dansmith: there was a nit of the if statement being needed in update_resource_stats, but its halmless | 15:26 |
| *** jichen has quit IRC | 15:26 | |
| *** su_zhang has joined #openstack-nova | 15:26 | |
| dansmith | johnthetubaguy: you mean the "if compute_node.uuid in self._resource_providers" ?\ | 15:27 |
| *** andreas_s has quit IRC | 15:27 | |
| johnthetubaguy | yeah | 15:27 |
| *** moshele has quit IRC | 15:27 | |
| dansmith | johnthetubaguy: it's needed because I hit it the first time I ran :) | 15:27 |
| dansmith | johnthetubaguy: because the _ensure method doesn't explode if it can't create the provider and we'll fail with keyerror when we try to fetch out the rp later | 15:27 |
| johnthetubaguy | ah... gotcha | 15:27 |
| johnthetubaguy | I missed it didn't raise | 15:27 |
| *** ttrumm has quit IRC | 15:28 | |
| *** mamandle has joined #openstack-nova | 15:28 | |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: Assume neutron port_binding extensions enabled https://review.openstack.org/357726 | 15:29 |
| johnthetubaguy | dansmith: hmm, so I actually thought we would only update this data on SIG_HUP or restart, because it only reporting the max available, its generally static right? | 15:29 |
| *** ljxiash has quit IRC | 15:29 | |
| *** gszasz_ has quit IRC | 15:29 | |
| dansmith | johnthetubaguy: this gets called every minute (or whatever the timer is set to), so we have the opportunity to update it every minute | 15:29 |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: Assume neutron port_binding extensions enabled https://review.openstack.org/357726 | 15:29 |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: Use admin context to create neutron ports https://review.openstack.org/357540 | 15:29 |
| dansmith | johnthetubaguy: and later I need to make this only do the update if something actually changed, but it will be able to diff itself against placement and make updates, potentially every minute | 15:30 |
| dansmith | johnthetubaguy: look at one of the runs and you'll see | 15:30 |
| *** su_zhang has quit IRC | 15:30 | |
| johnthetubaguy | dansmith: OK, looking | 15:30 |
| dansmith | http://logs.openstack.org/61/363061/2/experimental/gate-tempest-dsvm-neutron-placement-full-ubuntu-xenial-nv/780cd4a/logs/screen-n-cpu.txt.gz#_2016-08-30_23_28_58_954 | 15:31 |
| dansmith | johnthetubaguy: ^ for example.. search for that string in that log | 15:31 |
| johnthetubaguy | yeah, just looking for the ones mriedem pointed out in his comments | 15:31 |
| *** browne has joined #openstack-nova | 15:31 | |
| * johnthetubaguy shakes fist at snail speed internet | 15:32 | |
| *** abalutoiu has joined #openstack-nova | 15:33 | |
| *** dmellado|mtg is now known as dmellado | 15:33 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Report compute node inventories through placement https://review.openstack.org/363061 | 15:33 |
| dansmith | johnthetubaguy: ^ | 15:33 |
| *** jlanoux has quit IRC | 15:34 | |
| openstackgerrit | Andrey Volkov proposed openstack/nova: Add create cell command https://review.openstack.org/332713 | 15:34 |
| openstackgerrit | Andrey Volkov proposed openstack/nova: Trivial: add ability to define action description https://review.openstack.org/355781 | 15:34 |
| johnthetubaguy | dansmith: thanks, that sorts those reserved values out | 15:34 |
| dansmith | aye | 15:35 |
| dansmith | johnthetubaguy: thanks for catching :) | 15:35 |
| *** code-R has quit IRC | 15:35 | |
| openstackgerrit | Alex Szarka proposed openstack/nova: Additional resume_guests_state enhancements https://review.openstack.org/343542 | 15:36 |
| *** tonytan_brb has joined #openstack-nova | 15:36 | |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: Assume neutron port_binding extensions enabled https://review.openstack.org/357726 | 15:36 |
| *** tesseract- has quit IRC | 15:36 | |
| *** dikonoor has quit IRC | 15:37 | |
| *** abalutoiu_ has joined #openstack-nova | 15:37 | |
| *** abalutoiu_ has quit IRC | 15:38 | |
| *** tonytan4ever has quit IRC | 15:38 | |
| *** abalutoiu has quit IRC | 15:38 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Remove deprecated neutron options https://review.openstack.org/363746 | 15:40 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Register keystone opts for placement sample config https://review.openstack.org/363699 | 15:40 |
| mriedem | sdague: johnthetubaguy: ^ | 15:40 |
| *** salv-orlando has joined #openstack-nova | 15:40 | |
| *** TravT_ has quit IRC | 15:40 | |
| mriedem | ah crap i lost the dependency on the base RT change | 15:40 |
| *** sdake_ is now known as sdake | 15:40 | |
| johnthetubaguy | mriedem: do we want to update ironic to match? | 15:41 |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: Use admin context to create neutron ports https://review.openstack.org/357540 | 15:41 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Register keystone opts for placement sample config https://review.openstack.org/363699 | 15:41 |
| *** kevinz has quit IRC | 15:41 | |
| *** roeyc has quit IRC | 15:41 | |
| *** tbachman has quit IRC | 15:42 | |
| mriedem | oh the depends-on wasn't necessary, the RT Change is merged, that was fast | 15:42 |
| mriedem | johnthetubaguy: what about ironic? | 15:43 |
| mriedem | like the password and v2password plugins? | 15:43 |
| *** TravT has joined #openstack-nova | 15:44 | |
| johnthetubaguy | mriedem: I was just checking for consistency how we do all these things for the other services | 15:44 |
| *** mgoddard has joined #openstack-nova | 15:44 | |
| *** mgoddard__ has quit IRC | 15:45 | |
| *** armax has quit IRC | 15:45 | |
| *** paul-carlton2 has quit IRC | 15:45 | |
| johnthetubaguy | mriedem: now I get your question, I guess ironic is just missing password and v2password | 15:45 |
| auggy | mriedem: dansmith: if there's anything i can help out with lemme know | 15:46 |
| *** code-R has joined #openstack-nova | 15:46 | |
| johnthetubaguy | but I think that was intentional | 15:46 |
| *** sahid has quit IRC | 15:46 | |
| mriedem | johnthetubaguy: yeah i think that's intentional | 15:46 |
| *** salv-orlando has quit IRC | 15:46 | |
| sdague | remember, this is just example config | 15:46 |
| johnthetubaguy | mriedem: yeah, sorry, just remembering all that now | 15:46 |
| johnthetubaguy | sdague: ack | 15:46 |
| sdague | it in no way actually impacts the options available | 15:46 |
| *** EinstCrazy has joined #openstack-nova | 15:46 | |
| *** ssurana has joined #openstack-nova | 15:47 | |
| mriedem | johnthetubaguy: i'm going to handle cinder separately i think | 15:47 |
| *** armax has joined #openstack-nova | 15:47 | |
| johnthetubaguy | mriedem: np | 15:47 |
| johnthetubaguy | its probably better that way | 15:47 |
| *** EinstCrazy has quit IRC | 15:48 | |
| *** matrohon has quit IRC | 15:48 | |
| *** markvoelker has quit IRC | 15:48 | |
| *** kaisers_ has joined #openstack-nova | 15:48 | |
| mriedem | auggy: add etc/nova/policy.yaml.merged to .gitignore | 15:49 |
| mriedem | or some variant of it | 15:49 |
| auggy | mriedem: kk | 15:49 |
| *** syjulian_ has joined #openstack-nova | 15:49 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Remove deprecated neutron options https://review.openstack.org/363746 | 15:51 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Register keystone opts for placement sample config https://review.openstack.org/363699 | 15:51 |
| mriedem | auggy: and do a cinder version of this https://review.openstack.org/#/c/363746/3 - the kilo commit that deprecated the same options for cinder was 49192695423b949be0e914dee47695db30636280 | 15:51 |
| auggy | mriedem: alright | 15:52 |
| *** praneshp has joined #openstack-nova | 15:53 | |
| auggy | mriedem: do you want me to tag both of those as part of the bp/generic-resource-pools ? | 15:53 |
| mriedem | no | 15:53 |
| *** rcernin has joined #openstack-nova | 15:54 | |
| *** sheel has joined #openstack-nova | 15:57 | |
| openstackgerrit | Alex Szarka proposed openstack/nova: resume_guests_state tests enhancements https://review.openstack.org/347256 | 15:57 |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: Ignore generated merged policy files https://review.openstack.org/363768 | 15:57 |
| cdent | dansmith, mriedem, sdague, bauzas, edleafe, anyone else if you're ready to hangout here we go: https://hangouts.google.com/call/joehze42tvbkfpbcmhktxnv4vue | 15:58 |
| dansmith | was it now, or 9:30? | 15:58 |
| cdent | we started the last one now ish, ended the prior one 9:30ish | 15:58 |
| dansmith | okay then | 15:58 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Update placement config reno https://review.openstack.org/363771 | 15:59 |
| *** ansiwen has quit IRC | 15:59 | |
| * mriedem puts on his headgear | 16:00 | |
| johnthetubaguy | cdent: I have another VC just about to start, else I would join you | 16:00 |
| openstackgerrit | Claudiu Belu proposed openstack/os-vif: Adds NoopPlugin https://review.openstack.org/362876 | 16:01 |
| sdague | coming | 16:01 |
| openstackgerrit | Gábor Antal proposed openstack/nova: Fixed indentation https://review.openstack.org/363772 | 16:02 |
| dansmith | whyfor mriedem has no camera? | 16:02 |
| mriedem | idk | 16:04 |
| *** devananda|MOVING is now known as devananda | 16:05 | |
| *** ansiwen has joined #openstack-nova | 16:05 | |
| openstackgerrit | Alex Szarka proposed openstack/nova: Changed maximum code complexity to 25 https://review.openstack.org/351176 | 16:05 |
| *** ildikov has quit IRC | 16:06 | |
| *** zengine has quit IRC | 16:07 | |
| *** links has joined #openstack-nova | 16:10 | |
| openstackgerrit | Roman Podoliaka proposed openstack/nova: ironic_host_manager: fix population of instances info on start https://review.openstack.org/347948 | 16:11 |
| *** raunak has joined #openstack-nova | 16:11 | |
| *** Sukhdev has joined #openstack-nova | 16:11 | |
| *** tongli has joined #openstack-nova | 16:13 | |
| *** ianw has quit IRC | 16:13 | |
| *** fragatin_ has quit IRC | 16:14 | |
| *** gszasz has joined #openstack-nova | 16:15 | |
| *** chrichip has quit IRC | 16:16 | |
| *** Sukhdev has quit IRC | 16:17 | |
| *** chrichip has joined #openstack-nova | 16:17 | |
| *** kaisers_ has quit IRC | 16:18 | |
| *** fragatina has joined #openstack-nova | 16:18 | |
| *** akshai has quit IRC | 16:18 | |
| openstackgerrit | Claudiu Belu proposed openstack/os-vif: Adds NoopPlugin https://review.openstack.org/362876 | 16:19 |
| *** ijw has joined #openstack-nova | 16:20 | |
| *** lpetrut has quit IRC | 16:22 | |
| *** fragatina has quit IRC | 16:22 | |
| *** moshele has joined #openstack-nova | 16:22 | |
| *** mamandle has quit IRC | 16:22 | |
| *** fragatina has joined #openstack-nova | 16:22 | |
| *** Swami has quit IRC | 16:23 | |
| mriedem | alaski: what do we absolutely need to get into newton for cells v2 before FF? sounds like we're going to have some downtime waiting for jaypipes to write this allocation client code in the RT | 16:24 |
| *** ijw has quit IRC | 16:24 | |
| *** sridharg has quit IRC | 16:25 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Add basic logging to placement api https://review.openstack.org/363130 | 16:25 |
| *** eandersson has quit IRC | 16:25 | |
| *** moshele has quit IRC | 16:26 | |
| *** mamandle has joined #openstack-nova | 16:26 | |
| *** tongli_ has joined #openstack-nova | 16:26 | |
| *** nic has joined #openstack-nova | 16:28 | |
| *** tongli has quit IRC | 16:29 | |
| *** zengine has joined #openstack-nova | 16:30 | |
| *** tongli_ has quit IRC | 16:30 | |
| *** tongli has joined #openstack-nova | 16:32 | |
| *** mingyu has quit IRC | 16:32 | |
| *** mingyu has joined #openstack-nova | 16:32 | |
| *** mingyu has quit IRC | 16:32 | |
| *** ssurana has quit IRC | 16:33 | |
| *** trisq has quit IRC | 16:34 | |
| *** tongli has quit IRC | 16:36 | |
| *** tongli has joined #openstack-nova | 16:38 | |
| *** aswadr_ has joined #openstack-nova | 16:39 | |
| cdent | auggy: https://etherpad.openstack.org/p/placement-next | 16:39 |
| auggy | cdent: thanks! | 16:40 |
| *** tongli has quit IRC | 16:43 | |
| alaski | mriedem: my goal is https://review.openstack.org/#/c/356138/ and the WIP above it, but I'm tracking down an intermittent bug in there right now. something odd with service version checking | 16:43 |
| alaski | mriedem: and https://review.openstack.org/#/c/326906/ would be good to have | 16:43 |
| *** ianw has joined #openstack-nova | 16:43 | |
| sdague | mriedem: +2 on your respin of - https://review.openstack.org/#/c/363699 | 16:43 |
| *** tongli has joined #openstack-nova | 16:44 | |
| *** lucasagomes has quit IRC | 16:44 | |
| sdague | mriedem / dansmith - fyi, I become a pumkin at about 4pm EST tomorrow. We have a thing we are going to in the evening, and I'm off on friday | 16:45 |
| *** lucasagomes has joined #openstack-nova | 16:46 | |
| dansmith | sdague: okay | 16:46 |
| dansmith | when is the hard deadline for N3? I assume it's EOD tomorrow? | 16:46 |
| sdague | yeh, something like that | 16:47 |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: Remove deprecated cinder options https://review.openstack.org/363803 | 16:47 |
| *** tongli has quit IRC | 16:48 | |
| *** Apoorva has joined #openstack-nova | 16:48 | |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: Remove deprecated cinder options https://review.openstack.org/363803 | 16:48 |
| melwitt | sdague: does https://review.openstack.org/#/c/363295/ look reasonable to you for getting rabbit userid from devstack for use in grenade runs? we are already sourcing lib/rpc_backend in grenade | 16:49 |
| mriedem | oh f yes, jury duty was cancelled again | 16:49 |
| mriedem | it's better than a snow day | 16:49 |
| *** daemontool has joined #openstack-nova | 16:50 | |
| mriedem | dansmith: yeah EOD tomorrow | 16:50 |
| mriedem | i have to have the n-3 tag in EOD tomorrow | 16:50 |
| *** akshai has joined #openstack-nova | 16:50 | |
| dansmith | okay | 16:51 |
| auggy | mriedem: got https://review.openstack.org/#/c/363768 and https://review.openstack.org/363803 done, just waiting for it to clear the gate | 16:51 |
| mriedem | sdague: ruh roh | 16:51 |
| mriedem | https://review.openstack.org/#/c/363613/1/zuul/layout.yaml | 16:51 |
| mriedem | sdague: you moved the placement job to devstack's check queue, not nova's | 16:51 |
| mriedem | sdague: want me to fix that so you can +W? | 16:52 |
| *** matias has joined #openstack-nova | 16:53 | |
| sdague | mriedem: yeh, oops | 16:53 |
| mriedem | i'm going to move the job back to experimental for devstack | 16:54 |
| sdague | mriedem: sure, fine with that | 16:56 |
| *** Guest36472 has quit IRC | 16:56 | |
| *** unicell has quit IRC | 16:57 | |
| *** salv-orlando has joined #openstack-nova | 16:57 | |
| *** tongli has joined #openstack-nova | 16:57 | |
| openstackgerrit | Andrew Laski proposed openstack/nova: DNM logs for debugging https://review.openstack.org/362442 | 16:58 |
| openstackgerrit | melanie witt proposed openstack/nova: Move cell message queue switching and add caching https://review.openstack.org/326906 | 16:58 |
| alaski | super quick cells meeting in 1 minute | 16:58 |
| *** mgoddard_ has joined #openstack-nova | 17:00 | |
| *** tongli has quit IRC | 17:02 | |
| *** mgoddard has quit IRC | 17:03 | |
| *** rlrossit has joined #openstack-nova | 17:04 | |
| *** yamahata has quit IRC | 17:05 | |
| *** code-R_ has joined #openstack-nova | 17:06 | |
| mriedem | sdague: does FF apply to devstack? | 17:07 |
| mriedem | and grenade | 17:07 |
| *** akshai has quit IRC | 17:07 | |
| mriedem | b/c alaski has some changes up for cells v2 and i'm wondering what the priority is on those | 17:07 |
| sdague | mriedem: no | 17:07 |
| mriedem | ok | 17:08 |
| sdague | but, bandwidth and risk are what we go by post freeze | 17:08 |
| sdague | remembering that bw remains low there | 17:08 |
| *** derekh has quit IRC | 17:09 | |
| *** code-R has quit IRC | 17:09 | |
| *** akshai has joined #openstack-nova | 17:12 | |
| *** tongli has joined #openstack-nova | 17:12 | |
| mriedem | sdague: we need this https://review.openstack.org/#/c/361436/ and the one below it to land a nova change for FF | 17:13 |
| openstackgerrit | Dan Smith proposed openstack/nova: Report compute node inventories through placement https://review.openstack.org/363061 | 17:13 |
| *** akshai has quit IRC | 17:14 | |
| *** su_zhang has joined #openstack-nova | 17:15 | |
| *** markvoelker has joined #openstack-nova | 17:15 | |
| sdague | mriedem: ok, those need to link to release notes showing where we wrote those steps down for operators in the commit message | 17:16 |
| mriedem | k | 17:16 |
| sdague | melwitt: on the devstack change, can you move the rabbit host as well? +2 after that | 17:16 |
| *** tongli has quit IRC | 17:17 | |
| *** akshai has joined #openstack-nova | 17:17 | |
| *** lucasagomes is now known as lucas-dinner | 17:19 | |
| *** rama_y has joined #openstack-nova | 17:19 | |
| sdague | dansmith: I'm super confused that you are getting a result thing that's None/False | 17:19 |
| sdague | it should be a response object, or an exception, I really didn't think none was an option there | 17:19 |
| dansmith | sdague: this was on an earlier version, but you clearly returned none from the helpers | 17:19 |
| dansmith | lemme look | 17:19 |
| mriedem | alaski: ok https://review.openstack.org/#/c/355602/ is rebased on top of https://review.openstack.org/#/c/362154/ now | 17:19 |
| melwitt | sdague: there was a reason I didn't do that, one sec | 17:19 |
| sdague | dansmith: that's when I had the try block deeper | 17:20 |
| dansmith | oh yeah | 17:20 |
| dansmith | gone now | 17:20 |
| dansmith | sdague: fixed.. otherwise good? | 17:20 |
| alaski | mriedem: you are a gentleman | 17:20 |
| alaski | mriedem: I think that invalidates your -1 on there | 17:20 |
| mriedem | it does | 17:20 |
| melwitt | sdague: I noticed grenade sets the rabbit host manually, does that matter? https://github.com/openstack-dev/grenade/blob/master/devstack.localrc.base#L21 | 17:20 |
| *** code-R_ has quit IRC | 17:21 | |
| sdague | melwitt: it shouldn't | 17:22 |
| sdague | you can still set it | 17:22 |
| *** su_zhang has quit IRC | 17:23 | |
| melwitt | sdague: okay, let me move rabbit host too then | 17:23 |
| sdague | dansmith: why are you returning json here - https://review.openstack.org/#/c/363061/6/nova/scheduler/client/report.py@266 - given that nothing uses the output of this function | 17:23 |
| mriedem | i'll update https://review.openstack.org/#/c/362310/ with the link to the nova change that adds the reno | 17:23 |
| *** su_zhang has joined #openstack-nova | 17:23 | |
| *** akshai has quit IRC | 17:23 | |
| *** su_zhang has quit IRC | 17:24 | |
| *** tongli has joined #openstack-nova | 17:24 | |
| dansmith | sdague: I needed something truthy for an earlier iteration where the 409 retry was in the parent (like the original one). I still need something truthy for the allocation stuff which shouldn't complete if the inventory one didn't | 17:24 |
| sdague | dansmith: it feels like it would be good to figure out if the 409 retry could be extacted a bit because we have this odd double fault thing | 17:24 |
| dansmith | sdague: so I can return True there if you want | 17:24 |
| dansmith | sdague: well, I did have it extracted initially but it was uglier, IMHO | 17:25 |
| sdague | dansmith: ok, I guess I just didn't see it getting used anywhere | 17:25 |
| *** akshai has joined #openstack-nova | 17:25 | |
| *** su_zhang has joined #openstack-nova | 17:25 | |
| *** akshai has quit IRC | 17:26 | |
| sean-k-mooney | hi any chance that https://review.openstack.org/#/c/334048/ can be merged to swap vhost-user port binding to use os-vif in the ovs case as part of n3? | 17:26 |
| sdague | dansmith: yeh, I guess I can see that. It feels odd to not have it live in a retry decorator that could be tuned to retry counts | 17:26 |
| mriedem | sean-k-mooney: is there any 3rd party ci, maybe intel, that runs with that? | 17:27 |
| *** mamandle has quit IRC | 17:27 | |
| sdague | dansmith: ok, well, I think if you just dump the None check at 239-242 I can live with it | 17:27 |
| *** ssurana has joined #openstack-nova | 17:27 | |
| sean-k-mooney | mriedem: well yes intels will test it | 17:27 |
| dansmith | sdague: the way this works, I don't think there is much point in retrying more than once.. we'll just run this again in 59 more seconds | 17:27 |
| sdague | dansmith: ok, is there a reason to bother with the 409 catch at all then? | 17:28 |
| *** akshai has joined #openstack-nova | 17:28 | |
| mriedem | sean-k-mooney: which one? | 17:28 |
| mriedem | http://intel-openstack-ci-logs.ovh/48/334048/5/check/tempest-dsvm-ovsdpdk-nfv-networking/127e4cb/logs/ ? | 17:28 |
| dansmith | sdague: yes because we have to invalidate our cache specifically | 17:28 |
| *** tongli has quit IRC | 17:28 | |
| dansmith | sdague: with the re-fetch of the resource provider | 17:28 |
| sean-k-mooney | mriedem: yes | 17:28 |
| dansmith | sdague: so that we get a new generation | 17:28 |
| sean-k-mooney | mriedem: we have not got our xenial job working yet but we are trying to get it in place in the next few days | 17:29 |
| sdague | ok | 17:29 |
| sdague | dansmith: ok, 3 more comments on the patch | 17:29 |
| dansmith | sdague: I changed the result to True instead of json and added a docstring about it | 17:30 |
| mriedem | 2016-08-26 14:43:58.909 30272 DEBUG nova.network.os_vif_util [req-db698e71-8d4c-42ad-94d1-e9dc4012828e tempest-TestNetworkAdvancedServerOps-850235663 tempest-TestNetworkAdvancedServerOps-850235663] Converted object VIFVHostUser(active=False,address=fa:16:3e:62:b7:50,has_traffic_filtering=False,id=734fac50-5934-41d8-961f-ec1c38993438,mode='client',network=Network(537ae646-23fc-449f-8898-3e73e8aca9c8),path='/var/run/openvswitch/vhu734fac50 | 17:30 |
| *** tongli has joined #openstack-nova | 17:30 | |
| sdague | dansmith: cool | 17:30 |
| *** ttrumm has joined #openstack-nova | 17:31 | |
| mriedem | sean-k-mooney: i need to do a few things but then i can maybe come back to this | 17:31 |
| alaski | mriedem: if you're updating https://review.openstack.org/#/c/362310 let's try removing those two lines | 17:31 |
| *** tongli_ has joined #openstack-nova | 17:31 | |
| mriedem | alaski: will do | 17:31 |
| sean-k-mooney | mriedem: no worries. i need to look into the output of this ci job too. it looks like some files i expect to be generated are not copied over | 17:31 |
| mriedem | alaski: i was going to link to this change for the release note ack https://review.openstack.org/#/c/274885/ | 17:32 |
| dansmith | sdague: on that last one, bauzas added that I think to specifically log something if we got a successful but non-200 result | 17:32 |
| dansmith | sdague: not really sure why, TBH | 17:32 |
| sdague | dansmith: that doesn't make any sense to me | 17:32 |
| dansmith | sdague: I can just collapse the last two | 17:32 |
| dansmith | yeah | 17:32 |
| sdague | and seems just convoluted | 17:32 |
| sdague | as long as the error message records the status code, it should be fine to work out what went wrong | 17:33 |
| *** mingyu has joined #openstack-nova | 17:33 | |
| dansmith | sdague: and actually, we'd never get there anyway | 17:33 |
| dansmith | not sure why he did that | 17:33 |
| dansmith | sdague: because of the "if not result"above | 17:33 |
| sdague | dansmith: also, this should probably be LOG.warning not error | 17:33 |
| sdague | because it's something that will recover next time | 17:34 |
| sdague | honestly, if it's expected to happen, it probably should be an info | 17:34 |
| *** tongli_ has quit IRC | 17:34 | |
| dansmith | sdague: that part shouldn't though | 17:34 |
| *** tongli has quit IRC | 17:35 | |
| *** tongli_ has joined #openstack-nova | 17:35 | |
| dansmith | sdague: that's my point.. we'd fail above if we get 404 for example, but I will change that one to a warning | 17:35 |
| alaski | dansmith: yeah, looks like a wrong binary name lookup. Should be nova-os_compute rather than nova-api which is what I assumed it would be. I think | 17:35 |
| sdague | dansmith: right, but you could end up in your final error because of 2 409s | 17:36 |
| alaski | mriedem: cool, that looks like a good one to use | 17:36 |
| sdague | which you just said aren't the end of the world | 17:36 |
| dansmith | sdague: ah true, so let me call that one out and only info it, okay? | 17:37 |
| sdague | sure | 17:37 |
| *** bradjones has quit IRC | 17:38 | |
| *** bradjones has joined #openstack-nova | 17:38 | |
| *** tonytan_brb has quit IRC | 17:38 | |
| *** mingyu has quit IRC | 17:38 | |
| *** pcaruana has joined #openstack-nova | 17:39 | |
| *** tongli_ has quit IRC | 17:39 | |
| mriedem | alaski: done and done https://review.openstack.org/#/q/topic:setup_cell0_before_migrations | 17:39 |
| *** bradjones is now known as Guest30374 | 17:39 | |
| *** unicell has joined #openstack-nova | 17:39 | |
| alaski | mriedem: much thanks | 17:40 |
| *** tongli has joined #openstack-nova | 17:40 | |
| *** mgoddard_ has quit IRC | 17:42 | |
| *** yamahata has joined #openstack-nova | 17:42 | |
| sdague | mriedem: so https://review.openstack.org/#/c/274885/25/releasenotes/notes/cell-id-db-sync-nova-manage-8504b54dd115a2e9.yaml is pretty weak language | 17:43 |
| sdague | "This command only needs to be called when upgrading to CellsV2." | 17:43 |
| sdague | but, we're saying it's mandatory before newton? | 17:43 |
| *** ttrumm has quit IRC | 17:43 | |
| *** tongli has quit IRC | 17:43 | |
| *** markvoelker has quit IRC | 17:43 | |
| mriedem | alaski: are we going to get cell0 in for newton? is it required? | 17:44 |
| *** tongli has joined #openstack-nova | 17:44 | |
| dansmith | required during newton | 17:44 |
| dansmith | hence the soft language | 17:44 |
| *** ttrumm has joined #openstack-nova | 17:44 | |
| dansmith | before ocata | 17:44 |
| alaski | yep, that | 17:44 |
| sdague | ok, then we're not going to unconditionally blast it through in grenade | 17:44 |
| mriedem | ok, so getting cell0 in newton devstack is enough | 17:44 |
| mriedem | right? | 17:44 |
| dansmith | sdague: what does that mean? | 17:45 |
| mriedem | and nova in newton has to be able to handle that cell0 doesn't exist | 17:45 |
| sdague | https://review.openstack.org/#/c/362310/2 | 17:45 |
| alaski | Nova in N handles that cell0 does not exist, unless you run the simple_cell_setup command after which it needs to exist | 17:45 |
| mriedem | yeah so https://review.openstack.org/#/c/362310/ says you have ot have cell0 before you can get to newton | 17:46 |
| *** tongli has quit IRC | 17:46 | |
| mriedem | and we have sibling changes for this in devstack, to create cell0 and run simple_cell_setup | 17:46 |
| alaski | okay. the language can be change to say that the cell0 db must exist before running simple-cell_setup | 17:46 |
| dansmith | we need to test rolling to it with newton code before the newton release so that we know we can require people to do it | 17:46 |
| *** tongli has joined #openstack-nova | 17:46 | |
| mgagne | jroll: ping | 17:46 |
| mgagne | jroll: re OOM errors you mentioned in Nova scheduler with Ironic | 17:47 |
| openstackgerrit | Dan Smith proposed openstack/nova: Report compute node inventories through placement https://review.openstack.org/363061 | 17:47 |
| dansmith | sdague: ^ | 17:47 |
| mgagne | jroll: I think I found the root cause | 17:47 |
| sdague | dansmith: but we just said we weren't requiring people to do it, right? | 17:47 |
| mriedem | i.e. https://review.openstack.org/#/q/status:open+project:openstack-dev/devstack+branch:master+topic:cell0_db | 17:47 |
| dansmith | sdague: required to do it *in* newton | 17:47 |
| mriedem | dansmith: which is ^ | 17:47 |
| dansmith | sdague: newton will run without it, but it has to be done before ocata | 17:47 |
| mriedem | yeah, so we don't need these grenade changes | 17:47 |
| sdague | ok, then we don't do it with mitaka -> newton | 17:47 |
| dansmith | sdague: so we need to test that it works on newton code | 17:47 |
| alaski | sdague: the migration to cellsv2, and cell0, must happen after moving to N but before going to O | 17:47 |
| sdague | we do it with newton -> mitaka | 17:47 |
| jroll | mgagne: the InstanceList.get_all_for_compute() or whatever call? | 17:47 |
| sdague | sorry newton -> ocata | 17:48 |
| mgagne | jroll: this line: https://github.com/openstack/nova/blob/kilo-eol/nova/scheduler/host_manager.py#L590 | 17:48 |
| mriedem | right, so we drop https://review.openstack.org/#/q/topic:setup_cell0_before_migrations | 17:48 |
| dansmith | sdague: but if we don't test it now, we have no idea if it works *on* newton | 17:48 |
| *** zenoway has quit IRC | 17:48 | |
| jroll | mgagne: yeah, that's what I'd pinpointed in the past | 17:48 |
| mgagne | jroll: should be objects.InstanceList.get_by_host_and_node | 17:48 |
| *** ttrumm has quit IRC | 17:48 | |
| alaski | dansmith: right | 17:48 |
| mriedem | dansmith: alaski: sdague: doesn't https://review.openstack.org/#/q/status:open+project:openstack-dev/devstack+branch:master+topic:cell0_db test that cell0 is on newton? | 17:48 |
| *** zenoway has joined #openstack-nova | 17:48 | |
| mgagne | jroll: otherwise in Ironic context, it will load ALL instances since you often have 1 nova-compute anyway | 17:48 |
| sdague | dansmith: but if we land it, then the entire universe is over there for newton and we never test them not doing it | 17:49 |
| jroll | mgagne: right, but there's no nodename there | 17:49 |
| mgagne | jroll: I sent an email in operators about the details but | 17:49 |
| dansmith | mriedem: it tests that, given a system upgraded to newton, that we can land that code | 17:49 |
| sdague | you make a fork | 17:49 |
| dansmith | mriedem: er, run | 17:49 |
| dansmith | sdague: um, what? | 17:49 |
| mgagne | jroll: in fact, with get_all_host_states looping on all hypervisors (not hosts), you end up with all instances being loaded for each hypervisor | 17:49 |
| sdague | dansmith: if we land this grenade patch, we don't ever test newton upgrades without cell0 | 17:50 |
| *** tongli has quit IRC | 17:50 | |
| mgagne | jroll: so 2000 hypervisors/instances, this means 2000^2 loaded in memory | 17:50 |
| dansmith | sdague: the grenade patch is already in, this is a devstack patch | 17:50 |
| alaski | mriedem: that checks that cell0 works in Newton. But we can't add a cell migration in grenade because it doesn't run new devstack and breaks on the db not being setup | 17:50 |
| jroll | mgagne: lol, yeah I think you're right | 17:50 |
| mgagne | jroll: this line will do just fine instead: inst_list = objects.InstanceList.get_by_host_and_node(context, host_name, compute.hypervisor_hostname) | 17:50 |
| *** tongli has joined #openstack-nova | 17:50 | |
| sdague | dansmith: I am staring at grenade patches | 17:50 |
| dansmith | sorry, I'm confusing the cell0 thing with the last one | 17:50 |
| sdague | https://review.openstack.org/#/c/361436/5 | 17:50 |
| mriedem | alaski: oh i see, we have a circular dependency :) | 17:50 |
| jroll | mgagne: maybe, yeah | 17:51 |
| alaski | mriedem: yeah | 17:51 |
| mgagne | tested and it works just fine, well no errors | 17:51 |
| *** dimtruck is now known as zz_dimtruck | 17:51 | |
| alaski | sdague: so right now grenade runs simple_cell_setup, which tells Nova that a cell0 db exists. However it doesn't actually exist yet, so Nova blows up when trying to use it | 17:51 |
| alaski | which is what my latest patch does | 17:51 |
| mgagne | jroll: but anyway, with your patch, it won't hit that line anymore | 17:51 |
| sdague | can we hangout, I fee like I'm playing who's on first? :) | 17:52 |
| mriedem | can simple_cell_setup detect that the cell0 db doesn't exist? | 17:52 |
| mriedem | yeah let's get on a call | 17:52 |
| alaski | sdague: so my Nova patch is blocked on having the db added in grenade, or reverting the cell migration in grenade. If we revert we can't ask users to migrate to cellsv2 in N because it's not at all tested | 17:52 |
| dansmith | aye | 17:52 |
| alaski | alright, one sec. I'm in a coffee shop atm | 17:53 |
| sdague | https://hangouts.google.com/call/vhef6o64srfalp63dfqizgfyuqe | 17:53 |
| *** ducttape_ has quit IRC | 17:53 | |
| *** zenoway has quit IRC | 17:53 | |
| *** ihrachys has quit IRC | 17:54 | |
| *** akshai has quit IRC | 17:57 | |
| *** tongli has quit IRC | 17:58 | |
| *** shewless has quit IRC | 17:58 | |
| *** tongli has joined #openstack-nova | 17:58 | |
| *** tbachman has joined #openstack-nova | 17:58 | |
| cdent | dansmith: I guess burning a generation when we do an allocation update would be wrong. I ask because that's the simplest way to reuse existing code. Updating to the same value if the value matches is what we said on the call, right? (just confirming) | 17:59 |
| *** tongli_ has joined #openstack-nova | 18:00 | |
| cmartin | Hey does anybody know if there is a situation where the destination for an instance can be a volume but the source isn't? | 18:01 |
| *** zz_dimtruck is now known as dimtruck | 18:01 | |
| *** diogogmt has quit IRC | 18:02 | |
| *** ducttape_ has joined #openstack-nova | 18:02 | |
| *** tongli has quit IRC | 18:03 | |
| *** rmcall_ has joined #openstack-nova | 18:03 | |
| *** rmcall has quit IRC | 18:05 | |
| *** rmcall_ is now known as rmcall | 18:05 | |
| *** thumpba has joined #openstack-nova | 18:05 | |
| *** akshai has joined #openstack-nova | 18:06 | |
| *** sudipto has joined #openstack-nova | 18:06 | |
| *** Kevin_Zheng has quit IRC | 18:06 | |
| dansmith | sdague: https://review.openstack.org/#/c/361436/5/projects/60_nova/upgrade.sh | 18:06 |
| *** su_zhang has quit IRC | 18:07 | |
| *** salv-orl_ has joined #openstack-nova | 18:08 | |
| *** salv-orlando has quit IRC | 18:10 | |
| *** tongli has joined #openstack-nova | 18:11 | |
| *** Kevin_Zheng has joined #openstack-nova | 18:11 | |
| *** lpetrut has joined #openstack-nova | 18:12 | |
| *** xenogear has joined #openstack-nova | 18:12 | |
| *** stewie925 has quit IRC | 18:13 | |
| *** cmartin has quit IRC | 18:13 | |
| *** syjulian has quit IRC | 18:13 | |
| *** syjulian_ is now known as syjulian | 18:13 | |
| *** jinli has quit IRC | 18:13 | |
| *** MikeG451 has quit IRC | 18:14 | |
| *** tongli_ has quit IRC | 18:14 | |
| *** xenogear has quit IRC | 18:17 | |
| *** tonytan4ever has joined #openstack-nova | 18:18 | |
| *** syjulian has quit IRC | 18:18 | |
| *** ducttape_ has quit IRC | 18:19 | |
| *** ducttape_ has joined #openstack-nova | 18:20 | |
| *** prateek has quit IRC | 18:20 | |
| nic | I'm wondering if I've hit a bug or a PEBCAK error with the API | 18:20 |
| nic | Use case: I'm trying to GET /servers/detail with a list of UUIDs, so that I don't have to a) make 250 API calls to get the 250 VMs I need, and b) I don't have to get all 30,000 VMs and filter out the 250 I want | 18:22 |
| nic | The API seems to support it: https://github.com/openstack/nova/blob/master/nova/db/sqlalchemy/api.py#L2369-L2374 | 18:22 |
| nic | And if I set a breakpoint in the debugger and change the filter to be a list, it works | 18:22 |
| nic | But the wsgi only appears to support setting strings as search_opts | 18:22 |
| nic | How does one pass a list in as a search query? | 18:23 |
| nic | The novaclient explicitly coerces all search arguments into strings, so I'm guessing there's a trick to it? | 18:24 |
| mriedem | nic: use --debug with novaclient to see how it forms the query | 18:25 |
| nic | I already have — it's smashing my list down into a string, which then fails to match | 18:25 |
| mriedem | nic: you can't filter on uuid https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/servers.py#L1102 | 18:26 |
| mriedem | from what i can tell | 18:26 |
| nic | Sure you can | 18:26 |
| *** Sukhdev has joined #openstack-nova | 18:27 | |
| *** hemanthm is now known as hemanthm|lunch | 18:27 | |
| nic | Admins can search on whatever they like | 18:27 |
| mriedem | oh right, i was looking at non-admin | 18:28 |
| nic | And again, if you set a breakpoint and un-stringify the search opt back into a list, it does the right thing | 18:28 |
| *** prateek has joined #openstack-nova | 18:29 | |
| *** amotoki has quit IRC | 18:31 | |
| nic | Do I have to do something like that horrible trick in PHP with the magic names on query strings? | 18:32 |
| nic | I mean, whatever works, but eww... | 18:33 |
| *** gszasz has quit IRC | 18:33 | |
| *** TravT_ has joined #openstack-nova | 18:33 | |
| mriedem | what does the encoded uuid query string look like? | 18:33 |
| mriedem | does it have commas in it or something? | 18:33 |
| mriedem | in sqla db api if we have a uuid filter and there is a comma in the list, we could just split that | 18:34 |
| sdague | johnthetubaguy: your comment on https://review.openstack.org/#/c/363061/7 made me lol :) | 18:35 |
| sdague | because we had just unwound that | 18:35 |
| *** MikeG451 has joined #openstack-nova | 18:35 | |
| *** sudipto has quit IRC | 18:35 | |
| mriedem | sdague: ok, in lieu of another etherpad for cellsv2, plan is create a variable in devstack, mirrored in grenade, to setup cell0 and run simple_cell_setup - and then we'll use that in one of the grenade jobs and the placement api job | 18:35 |
| mriedem | dansmith: alaski: ^ correct? | 18:36 |
| *** TravT has quit IRC | 18:36 | |
| dansmith | yeah, and I guess we just do those in laski's patches | 18:36 |
| mriedem | yup | 18:36 |
| sdague | mriedem: that seems like what we just decide | 18:36 |
| sdague | decided | 18:36 |
| *** jaypipes has joined #openstack-nova | 18:36 | |
| mriedem | ok, word up | 18:36 |
| mriedem | that gives us something to do until his jayness shows up | 18:36 |
| dansmith | alaski: so since those are your patches do you actually want to just do that? if not, I will be glad to | 18:36 |
| dansmith | I'm going to need a lot of beer in barcelona | 18:37 |
| alaski | mriedem: yep | 18:37 |
| mriedem | jaypipes: oh his jayness is here | 18:37 |
| dansmith | and by a lot, I mean like.. two | 18:37 |
| mriedem | dansmith is going to require exactly 6.5 oz of beer | 18:37 |
| alaski | dansmith: if you have experience doing that it might be easier for you, but I'm happy to look for something to copy | 18:38 |
| jaypipes | mriedem: please don't call me that. I did my best to get back to an Internet location after dinner and family things. | 18:38 |
| dansmith | alaski: okay | 18:38 |
| mriedem | jaypipes: i'm not trying to dump on you, we do have two things that need your attention though when you can | 18:38 |
| alaski | dansmith: I will provide you with beverages of your choice, or some plain food. I don't actually know what to expect there foodwise | 18:39 |
| mriedem | cdent put those in the ML thread | 18:39 |
| *** bvanhav has quit IRC | 18:39 | |
| *** moshele has joined #openstack-nova | 18:40 | |
| *** bvanhav has joined #openstack-nova | 18:40 | |
| *** xenogear has joined #openstack-nova | 18:40 | |
| jaypipes | mriedem: I just read them and am reading the etherpad. | 18:41 |
| mriedem | the etherpad is probably mostly distraction at this point | 18:41 |
| *** cmartin has joined #openstack-nova | 18:41 | |
| jaypipes | mriedem: ok... so you want me to focus on cdent's allocation SQL patch, yes? | 18:42 |
| mriedem | correc | 18:42 |
| mriedem | and write the client side RT change to populate allocations in the placement API | 18:43 |
| cdent | jaypipes: I'm just finishing a tweak to that, which is probably still wrong but more like what we'll eventually want | 18:43 |
| mriedem | if you haven't already done so | 18:43 |
| jaypipes | k | 18:43 |
| cdent | such that the tests may still be useful | 18:43 |
| openstackgerrit | Ludovic Beliveau proposed openstack/nova: PCI: Fix pci_passthrough_whitelist with fully qualified PF PCI address https://review.openstack.org/363884 | 18:43 |
| *** tongli has quit IRC | 18:43 | |
| jaypipes | mriedem: I suppose I should write it as a quick hack (the client side RT stuff)? | 18:43 |
| *** Sukhdev has quit IRC | 18:43 | |
| *** jinli has joined #openstack-nova | 18:43 | |
| *** tongli has joined #openstack-nova | 18:44 | |
| openstackgerrit | Paulo Matias proposed openstack/nova: Fill destination check data with VNC/SPICE listen addresses https://review.openstack.org/358599 | 18:44 |
| jaypipes | mriedem: since my attempts to do things properly have mostly been rejected or failed outright. | 18:44 |
| mriedem | jaypipes: in the interest of time i think we want bare bones functionality | 18:44 |
| jaypipes | I gathered that, yes. | 18:45 |
| mriedem | we can optimize later | 18:45 |
| mriedem | sdague: dansmith: alaski: posted the cells v2 plan to the dev list also, for general awareness | 18:45 |
| dansmith | mriedem: okay, got the devstack bits up for them to look at and then will do the grenade part | 18:46 |
| *** tongli_ has joined #openstack-nova | 18:46 | |
| *** kfarr has joined #openstack-nova | 18:47 | |
| *** stewie925 has joined #openstack-nova | 18:47 | |
| alaski | mriedem: cool | 18:47 |
| *** diogogmt has joined #openstack-nova | 18:48 | |
| *** tongli has quit IRC | 18:48 | |
| openstackgerrit | Sean Dague proposed openstack/nova: Add basic logging to placement api https://review.openstack.org/363130 | 18:48 |
| *** thumpba has quit IRC | 18:49 | |
| *** thumpba has joined #openstack-nova | 18:49 | |
| *** syjulian has joined #openstack-nova | 18:51 | |
| mriedem | dansmith: ok, comments in the devstack changes. not -1, but not +1 yet either | 18:52 |
| *** david-lyle_ has joined #openstack-nova | 18:52 | |
| *** david-lyle_ has quit IRC | 18:52 | |
| melwitt | sdague: https://review.openstack.org/363295 is ready with RABBIT_HOST moved as well | 18:53 |
| *** rmcall has quit IRC | 18:54 | |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP Add set_all and delete_all to AllocationList https://review.openstack.org/363209 | 18:54 |
| cdent | jaypipes, dansmith ^ there's a new version | 18:54 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Pull from cell0 and build_requests for instance list https://review.openstack.org/356138 | 18:54 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Fix service version lookup during delete https://review.openstack.org/363888 | 18:54 |
| *** akshai has quit IRC | 18:55 | |
| cdent | that one changes create_all to set_all so that existing allocations for the rp, consumer pair are (sort of) updated instead of being added to (there's no unique constraint on those two columns?). It also checks the resource provider generation | 18:55 |
| cdent | there's no test, as I thought best to pass this on to jay now, on the generation handling | 18:55 |
| *** akshai has joined #openstack-nova | 18:55 | |
| cdent | it uses the current generation and updates the generation to the current if it can find the current. This is to avoid changing the generation for non-inventory things. So we know if things have changed out from under us. | 18:56 |
| openstackgerrit | Andrew Laski proposed openstack/nova: DNM logs for debugging https://review.openstack.org/362442 | 18:56 |
| cdent | Now that I've commited it, I think that's wrong. We should burn a generation/ | 18:56 |
| *** sheel has quit IRC | 18:56 | |
| cdent | but I'll let jaypipes give it a read through | 18:56 |
| cdent | i'm going to fix the follow up patch so it is using the right method | 18:57 |
| *** ijw has joined #openstack-nova | 18:57 | |
| *** diogogmt has quit IRC | 18:57 | |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP Add support for allocations in placement API https://review.openstack.org/339874 | 18:57 |
| *** moshele has quit IRC | 18:59 | |
| *** ildikov has joined #openstack-nova | 19:01 | |
| dansmith | alaski: okay so what is the deal with this: https://review.openstack.org/#/c/361436/5/projects/60_nova/upgrade.sh | 19:05 |
| dansmith | alaski: do we still need to move that ordering? and/or should I just move it to before the online part? | 19:05 |
| sdague | melwitt: oh, gdi, I know why the RABBIT thing wasn't moved | 19:05 |
| sdague | because of SERVICE_HOST calculation | 19:06 |
| sdague | ok, I suck, if you go back to patch 1 I'll approve | 19:06 |
| alaski | dansmith: that's so that db sync gets the cell0 db | 19:06 |
| dansmith | hah | 19:06 |
| dansmith | alaski: but not db_sync, just the online_data_migrations part right? | 19:06 |
| alaski | dansmith: no, the actual db sync command. it fans out to cell0 now | 19:07 |
| dansmith | what the... I thought I asked you that in the call | 19:07 |
| dansmith | lemme go look | 19:07 |
| alaski | https://review.openstack.org/#/c/274885/ | 19:07 |
| melwitt | sdague: no worries. I'm not so familiar with these, so thanks for telling the right way | 19:07 |
| alaski | dansmith: yeah, that was my initial faulty thumbsup. | 19:08 |
| dansmith | alaski: okay, heh | 19:08 |
| *** raunak has quit IRC | 19:09 | |
| mriedem | placement job results are in on the inventory change fwiw http://logs.openstack.org/61/363061/7/check/gate-tempest-dsvm-neutron-placement-full-ubuntu-xenial-nv/3445b46/ | 19:11 |
| * bauzas back up | 19:12 | |
| bauzas | dansmith: sdague: mriedem: anything I could help ? | 19:13 |
| mriedem | review https://review.openstack.org/#/c/363061/ | 19:13 |
| *** links has quit IRC | 19:13 | |
| mriedem | bauzas: and https://review.openstack.org/#/c/356138/ | 19:13 |
| dansmith | bauzas: thank you very much for writing those unit tests | 19:14 |
| bauzas | mriedem: roger, roger. | 19:14 |
| dansmith | mriedem: http://logs.openstack.org/61/363061/7/check/gate-tempest-dsvm-neutron-placement-full-ubuntu-xenial-nv/3445b46/logs/screen-n-cpu.txt.gz?#_2016-08-31_18_30_26_132 | 19:14 |
| mriedem | yup | 19:14 |
| melwitt | sdague: done | 19:14 |
| sdague | melwitt: yep, thanks, +A | 19:14 |
| bauzas | dansmith: heh, you should rather thank the TZs | 19:15 |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP Add support for allocations in placement API https://review.openstack.org/339874 | 19:15 |
| dansmith | the one time TZs helped | 19:15 |
| melwitt | sdague: thanks! | 19:15 |
| cdent | sdague: do you have any pending placement devstack changes? | 19:16 |
| cdent | there's a rather critical thing missing from the apache conf template | 19:16 |
| sdague | cdent: I do not | 19:16 |
| sdague | cdent: which bit is that? | 19:17 |
| cdent | CustomLog | 19:17 |
| cdent | I'll add it | 19:17 |
| mriedem | so, something i'm confused about with the inventorys put | 19:17 |
| sdague | cdent: ok | 19:17 |
| cdent | and let you know when it is there | 19:17 |
| sdague | it was all working locally | 19:17 |
| jaypipes | cdent: so do you want me to push a revision that does all this in a single SQL transaction? | 19:17 |
| mriedem | cdent: dansmith: we pass a body with 2 keys https://review.openstack.org/#/c/363061/7/nova/scheduler/client/report.py@228 - but the gabbit doesn't pass an inventories list https://github.com/openstack/nova/blob/master/nova/tests/functional/api/openstack/placement/gabbits/inventory.yaml#L72 | 19:18 |
| cdent | jaypipes: yes that would be great. What's there right now is chaos. it "works" but does so in a very crufty way | 19:18 |
| *** fragatin_ has joined #openstack-nova | 19:18 | |
| mriedem | oh i guess we have a gabbit for that here https://github.com/openstack/nova/blob/master/nova/tests/functional/api/openstack/placement/gabbits/inventory.yaml#L254 | 19:18 |
| sdague | cdent: https://dl.dropboxusercontent.com/1/view/q3r20hjyw1fdrxh/Apps/Shutter/screenshot_324.png | 19:18 |
| dansmith | mriedem: it does, down further | 19:18 |
| cdent | mriedem: there are two ways to PUT inventory (accident of history) | 19:18 |
| mriedem | yeah | 19:18 |
| mriedem | see that now | 19:18 |
| cdent | one is to inventories, the other is to inventories/{resource_class} | 19:18 |
| mriedem | ah ok | 19:19 |
| mriedem | $LAST_URL was not clear to me | 19:19 |
| cdent | sdague: yeah, that's because of what you added, right? but it isn't apache doing it's own logging, which is something that should still happen, isn't it? | 19:19 |
| *** salv-orl_ has quit IRC | 19:19 | |
| cdent | mriedem: yeah, I resisted adding that feature for a _long_ time | 19:20 |
| *** coolsvap has quit IRC | 19:22 | |
| bauzas | cdent: mriedem: I looked at the gabbits to see what was possible to do with the inventories, that's cool | 19:22 |
| *** fragatina has quit IRC | 19:22 | |
| bauzas | I'm just used to look other ways, but this one looks good | 19:22 |
| *** rfolco has quit IRC | 19:22 | |
| cdent | bauzas: I'm glad you like it. It does take a bit to get used to, but usually ends up being relatively readable | 19:22 |
| bauzas | cdent: just one note tho | 19:23 |
| cdent | sure | 19:23 |
| *** daemontool has quit IRC | 19:23 | |
| *** rfolco has joined #openstack-nova | 19:23 | |
| bauzas | cdent: any idea why you're only verifying those https://github.com/openstack/nova/blob/master/nova/tests/functional/api/openstack/placement/gabbits/inventory.yaml#L266-L270 | 19:24 |
| bauzas | ? | 19:24 |
| cdent | you mean instead of everything in the response? | 19:24 |
| bauzas | cdent: I thought we were returning more than those 2 fields | 19:24 |
| bauzas | yeah | 19:24 |
| cdent | because the idea is to confirm that the api is behaving as expected not to test that serializers work | 19:24 |
| bauzas | I wanted to have kind of a exhaustive answer and I needed to look at code rather | 19:24 |
| cdent | if every single test tested the entire response body it would get very noisy | 19:25 |
| cdent | each test has a particular purpose, despite other things happening | 19:25 |
| bauzas | cdent: okay, but I haven't found a full answer in that gabbit file | 19:25 |
| *** akshai has quit IRC | 19:25 | |
| *** jvgrant_ has quit IRC | 19:25 | |
| cdent | bauzas: I hadn't really written the gabbits a full spec of the api because we have http://specs.openstack.org/openstack/nova-specs/specs/newton/approved/generic-resource-pools.html | 19:26 |
| bauzas | cdent: I can understand gabbit is not for doc | 19:26 |
| cdent | gabbits certainly _can_ be that, though | 19:26 |
| cdent | and it would be easy to do | 19:26 |
| dansmith | cdent: that spec is not the right place to keep the canonical version, IMHO | 19:26 |
| dansmith | it's too easy for it to be wrong | 19:26 |
| cdent | I have a friend who insists that gabbi should be used canonical documentation | 19:26 |
| bauzas | cdent: but I just want to make sure we have *somewhere* to look when we want to look at the apidocs | 19:26 |
| cdent | and it would work well for that | 19:26 |
| cdent | bauzas: we can add more tests at any time | 19:27 |
| cdent | _we_ | 19:27 |
| bauzas | cdent: well, I'm not sure gabbi is the answer to this, I would ideally like to see some formal doc I could refer to know what I can get when I call PUT /inventories, that's it :) | 19:27 |
| *** akshai has joined #openstack-nova | 19:28 | |
| *** jvgrant has joined #openstack-nova | 19:28 | |
| bauzas | cdent: and I agree with dansmith, a spec is not for getting the right documentation | 19:29 |
| *** claudiub has quit IRC | 19:29 | |
| bauzas | cdent: at least because the implementation can evolve while the spec is mostly stale once accepted | 19:29 |
| *** hemanthm|lunch is now known as hemanthm | 19:29 | |
| *** Swami has joined #openstack-nova | 19:30 | |
| cdent | docs is on the todo list, on the spec, so those things are on the list, but not enough people have had the free time to be involved (until now) | 19:30 |
| cdent | so we can probalby get started on that soon | 19:30 |
| alaski | melwitt: couple of questions on https://review.openstack.org/#/c/326906/10 | 19:31 |
| *** akshai has quit IRC | 19:31 | |
| bauzas | cdent: sure, I'd definitly love some kind of autodocumentation :) | 19:31 |
| * melwitt looks | 19:31 | |
| bauzas | just for this, Pecan/WSME <3 | 19:31 |
| *** aswadr_ has quit IRC | 19:31 | |
| * bauzas has to kill the troll | 19:32 | |
| *** akshai has joined #openstack-nova | 19:32 | |
| *** sdake_ has joined #openstack-nova | 19:32 | |
| cdent | jaypipes: I responded to your comments on https://review.openstack.org/#/c/363209/3 just you're aware of how/why it got that way. No disputes. | 19:32 |
| *** sdake has quit IRC | 19:33 | |
| *** bvanhav has quit IRC | 19:33 | |
| jaypipes | cdent: ya, not to worry, thx | 19:34 |
| *** sdake has joined #openstack-nova | 19:35 | |
| dansmith | cdent: jaypipes' burn-a-generation logic makes sense to me and should handle the case we talked about this morning I think | 19:35 |
| cdent | bauzas: this is another useful gabbi trick: https://gist.github.com/cdent/a9590764fbc7402d450fa36df14f35e0 | 19:35 |
| cdent | yeah, after I went to all the trouble of making it not inc I changed my mind | 19:35 |
| mriedem | dansmith: comments in the code part of https://review.openstack.org/#/c/363061/ | 19:35 |
| mriedem | before going on to tests | 19:35 |
| *** sdake_ has quit IRC | 19:36 | |
| dansmith | mriedem: okay, will start on those | 19:38 |
| mriedem | is devstack/lib/nova sourced before this runs? https://review.openstack.org/#/c/362310/3/projects/60_nova/from-mitaka/upgrade-nova | 19:38 |
| *** riccardo has joined #openstack-nova | 19:40 | |
| mriedem | oh i see it's in grenade/projects/60_nova/upgrade.sh | 19:40 |
| mriedem | source $TARGET_DEVSTACK_DIR/lib/nova | 19:40 |
| riccardo | hi, what is now used instead of nova.openstack.common.importutils ? | 19:40 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Pull from cell0 and build_requests for instance list https://review.openstack.org/356138 | 19:41 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Delete BuildRequest regardless of service_version https://review.openstack.org/363911 | 19:41 |
| mriedem | riccardo: oslo something | 19:41 |
| mriedem | riccardo: https://github.com/openstack/oslo.utils/blob/master/oslo_utils/importutils.py | 19:41 |
| riccardo | mriedem, is there a list with the mapping? | 19:41 |
| riccardo | mriedem, thx | 19:41 |
| mriedem | riccardo: http://codesearch.openstack.org/ | 19:41 |
| *** TravT_ has quit IRC | 19:41 | |
| *** moshele has joined #openstack-nova | 19:42 | |
| riccardo | mriedem, cool, thx | 19:43 |
| melwitt | alaski: commented on https://review.openstack.org/#/c/363888/ and https://review.openstack.org/#/c/356138/16 (an update was pushed while I was reviewing) | 19:43 |
| *** rfolco has quit IRC | 19:43 | |
| alaski | melwitt: I do that on purpose (not really) | 19:44 |
| melwitt | heh :) | 19:44 |
| *** rfolco has joined #openstack-nova | 19:44 | |
| *** fragatin_ has quit IRC | 19:45 | |
| *** fragatina has joined #openstack-nova | 19:45 | |
| *** akshai has quit IRC | 19:49 | |
| alaski | dansmith: you led me astray with http://git.openstack.org/cgit/openstack/nova/tree/nova/service.py#n282 . That should be osapi_compute | 19:51 |
| melwitt | alaski: I think that might be a typo in the comment. I did a quick check in the services table of my devstack installation and saw nova-osapi_compute and I find it elsewhere in the code | 19:51 |
| mriedem | easter egg | 19:51 |
| alaski | melwitt: yeah, I just commented in the review that you were right. yet again... sigh | 19:52 |
| mriedem | # A list of APIs to enable by default (list value) #enabled_apis = osapi_compute,metadata | 19:52 |
| *** raunak has joined #openstack-nova | 19:52 | |
| dansmith | alaski: sorry, I suck | 19:52 |
| melwitt | alaski: it's what I do | 19:52 |
| melwitt | :P | 19:53 |
| alaski | dansmith: well, you did mention that that was probably my issue with the service version lookup which helped. so I guess we'll call it even | 19:53 |
| dansmith | dayum | 19:53 |
| dansmith | alaski: heh, alright | 19:53 |
| *** praneshp has quit IRC | 19:56 | |
| jaypipes | dansmith, cdent: should have the all-in-one SQL patch pushed within a couple hours... | 19:56 |
| dansmith | jaypipes: okay | 19:57 |
| *** code-R has joined #openstack-nova | 19:57 | |
| cdent | thanks jaypipes | 19:57 |
| cdent | If I'm still awake I'll check in | 19:57 |
| dansmith | the question is.. will it be a *mondo* SQL patch? | 19:57 |
| cdent | I believe mondo is the order of the day | 19:57 |
| * dansmith nods | 19:57 | |
| *** su_zhang has joined #openstack-nova | 19:58 | |
| cdent | sdague: turns out the logging thing was a misconfiguration on my side (I had turned off the default "site" so the default access log was not being used). Since placement runs under /placement, not a virtualhost, it wants to use that for access logging | 19:59 |
| *** moshele has quit IRC | 20:00 | |
| *** sean-k-mooney is now known as sean-k-mooneyAFK | 20:01 | |
| cdent | dansmith, sdague is there pending stuff in the placement realm that needs review? | 20:01 |
| dansmith | cdent: I'm waiting for mriedem to finish with my inventory tests, but that is still un-+2d | 20:01 |
| mriedem | dansmith: just did | 20:01 |
| mriedem | missing a test | 20:01 |
| *** scottynomad has quit IRC | 20:01 | |
| dansmith | cdent: soon I will push a new set with an extra test | 20:01 |
| * cdent gazes avidly | 20:02 | |
| sdague | cdent: the apache access logs should probably go to another place all together | 20:03 |
| sdague | http://logs.openstack.org/99/363699/4/check/gate-tempest-dsvm-full-ubuntu-xenial/424d857/logs/apache/ - note the keystone_access file | 20:03 |
| dansmith | what's the missing test? that the generation is bumped? | 20:03 |
| *** tonytan4ever has quit IRC | 20:03 | |
| mriedem | dansmith: really just that we don't call update_inventory from update_resource_stats if the uuid isn't in the cache | 20:03 |
| mriedem | the updated generatoin thing was more of a question | 20:04 |
| dansmith | mriedem: okay | 20:04 |
| *** rfolco has quit IRC | 20:04 | |
| cdent | I can't decide how I feel about the phrasing "Placement API is more sensitive to this than other APIs". | 20:04 |
| sdague | cdent: well, it's true :) | 20:04 |
| mriedem | the placement API listens a lot of emo in the basement | 20:05 |
| mriedem | the cure is on full rotatoin | 20:05 |
| cdent | elliot smith | 20:05 |
| *** TravT has joined #openstack-nova | 20:05 | |
| *** rfolco has joined #openstack-nova | 20:05 | |
| cdent | sdague: the flip side is that sending an incorrect or absent content-type is a bug? | 20:05 |
| *** TravT_ has joined #openstack-nova | 20:05 | |
| cdent | so it is _correctly_ more sensitive, thus my indecision on phrasing | 20:06 |
| jaypipes | dansmith: no, just a Brawndo SQL statement. | 20:06 |
| cdent | Brawndo! \o/ | 20:06 |
| dansmith | jaypipes: it's what plants crave | 20:06 |
| jaypipes | indeed. | 20:06 |
| mriedem | i had to look that up | 20:06 |
| sdague | cdent: sure, just most of the rest of the API are ok with sloppy content types, because they only know / can respond in application/json | 20:06 |
| *** liverpooler has quit IRC | 20:07 | |
| sdague | which means you can use requests.post(data=""), for instance. But this requires we use the json= field | 20:07 |
| cdent | in which case you're sending bad http, which is icky and wrong and should be shunned | 20:08 |
| mriedem | alaski: dansmith: want me to update this? https://review.openstack.org/#/c/361436/6/projects/60_nova/upgrade.sh | 20:08 |
| bauzas | mriedem: dansmith: yup, I commented in PS3 that I missed to update the existing test for checking that we call the inventory | 20:08 |
| sdague | cdent: that's all fine and good, it makes more effort for all client writers who may not be expecting that level of subtlety. Like the original client patch. | 20:09 |
| *** code-R_ has joined #openstack-nova | 20:09 | |
| bauzas | dansmith: you want me updating given mriedem's comments ? | 20:09 |
| sdague | anyway, I think it's pointless to argue about this right now | 20:09 |
| cdent | sdague: I'm not arguing, I'm making conversation | 20:10 |
| dansmith | bauzas: I'm doign ti now | 20:10 |
| bauzas | k | 20:10 |
| bauzas | leaving it for you :p | 20:10 |
| sdague | it is the behavior that's there, I left a note because it was code I had to change to get to work | 20:10 |
| *** TravT has quit IRC | 20:10 | |
| alaski | mriedem: it would be a bit before I get around to it, so fine with me | 20:10 |
| cdent | sdague: totally fair | 20:10 |
| bauzas | folks, just lemme know if you need some implementation/PS for tomorrow's EU morning, I can work on it | 20:10 |
| mriedem | alaski: ok i'll do it | 20:10 |
| *** moshele has joined #openstack-nova | 20:11 | |
| *** eharney has quit IRC | 20:11 | |
| *** code-R has quit IRC | 20:12 | |
| sdague | bauzas: did you end up working on any unit test patches for the safe_connect decorator? | 20:12 |
| bauzas | sdague: nope sorry | 20:12 |
| sdague | before I go and do that, I just want to make sure I'm not duplicating folks | 20:13 |
| bauzas | sdague: sorry, I had to go to my garage this afternoon :/ | 20:13 |
| sdague | bauzas: no worries | 20:13 |
| sdague | just checking in | 20:13 |
| openstackgerrit | Dan Smith proposed openstack/nova: Report compute node inventories through placement https://review.openstack.org/363061 | 20:14 |
| mriedem | what happens in the garage? | 20:14 |
| sdague | mriedem: on https://review.openstack.org/#/c/363061/7/nova/scheduler/client/report.py - there will always be a result | 20:14 |
| *** diogogmt has joined #openstack-nova | 20:14 | |
| mriedem | sdague: before you dive into unit tests, i think this is ready https://review.openstack.org/#/q/topic:setup_cell0_before_migrations | 20:15 |
| mriedem | sdague: yeah i talked myself out of that one | 20:15 |
| sdague | mriedem: ok, can you boil down your -1 then? | 20:16 |
| sdague | missing unit test for last function? | 20:16 |
| cdent | dansmith: I left a simple question on patchset 7 as you were pushing 8: not very relevant, but it's there if you care | 20:16 |
| mriedem | sdague: yeah, the unit test | 20:16 |
| mriedem | dan is updating it | 20:16 |
| dansmith | bauzas: sdague: Fixed mriedem's thing | 20:16 |
| *** tongli_ has quit IRC | 20:16 | |
| *** abalutoiu_ has joined #openstack-nova | 20:17 | |
| sdague | it's sometimes hard to read your reviews with what you actually care about, because they get a little meandering with convincing yourself back and forth :) | 20:17 |
| dansmith | mriedem: already pushed the fix with the unit test and other corretions | 20:17 |
| dansmith | sdague: +1000 | 20:17 |
| sdague | I think it's a midwest thing | 20:17 |
| mriedem | oh come on | 20:17 |
| dansmith | hah | 20:17 |
| cdent | that explains why I find it natural | 20:17 |
| dansmith | cdent: ot | 20:18 |
| mriedem | dansmith: did you see cdent's comment about just updating the generation in the 409 case? | 20:18 |
| mriedem | rather than rebuilding the payload? not sure it matters since we're not like calling out to a db or anything | 20:18 |
| mriedem | shouldn't hurt | 20:18 |
| dansmith | cdent: it's not really necessary to regenerate all the data, but it's technically more correct because the point of a generation conflict is that we're supposed to merge our views, | 20:19 |
| cdent | but where would things have changed? | 20:19 |
| dansmith | cdent: so later if we were merging in inventories we don't control there or something, we'd need to kinda re-do the whole dance to make sure we were up-to-date | 20:19 |
| cdent | ah okay | 20:19 |
| dansmith | cdent: nothing in this particular path, | 20:19 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Fix service version lookups https://review.openstack.org/363888 | 20:19 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Pull from cell0 and build_requests for instance list https://review.openstack.org/356138 | 20:19 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Delete BuildRequest regardless of service_version https://review.openstack.org/363911 | 20:19 |
| cdent | the "later" was the part I wasn't thinking of | 20:19 |
| dansmith | but it could be different in the future, and this encodes the intent I think | 20:19 |
| *** TravT_ is now known as TravT | 20:20 | |
| cdent | ✔ | 20:20 |
| mriedem | +W on https://review.openstack.org/#/c/363061/ dontchaknow | 20:21 |
| openstackgerrit | Andrew Laski proposed openstack/nova: DNM logs for debugging https://review.openstack.org/362442 | 20:21 |
| sdague | mriedem: ok, the devstack and grenade patches I think are all good now that we are feature flagged | 20:22 |
| sdague | +2 all around | 20:22 |
| *** rstarmer has quit IRC | 20:22 | |
| dansmith | fuckin. A. | 20:22 |
| *** dimtruck is now known as zz_dimtruck | 20:23 | |
| mriedem | sdague: cool, i'll start the d-g and p-c changes | 20:23 |
| sdague | mriedem: sounds great | 20:23 |
| *** mingyu has joined #openstack-nova | 20:23 | |
| *** diablo_rojo has joined #openstack-nova | 20:23 | |
| sdague | mriedem: logging change - http://logs.openstack.org/30/363130/5/check/gate-tempest-dsvm-neutron-placement-full-ubuntu-xenial-nv/8cb677d/logs/apache/placement-api.txt.gz | 20:25 |
| sdague | https://review.openstack.org/#/c/363130/ | 20:25 |
| cdent | I'm going to split for a while and will check back in later to see how it is going with jaypipes. If there is something I can attend to in my AM tomorrow please send me some email, or if you leave it as a message to me here, I ought to find it. thanks. | 20:26 |
| mriedem | heh, 99% config option dump | 20:26 |
| dansmith | sdague: no request logging in there/. | 20:26 |
| sdague | dansmith: yes there is | 20:26 |
| mriedem | dansmith: very very bottom | 20:26 |
| mriedem | http://logs.openstack.org/30/363130/5/check/gate-tempest-dsvm-neutron-placement-full-ubuntu-xenial-nv/8cb677d/logs/apache/placement-api.txt.gz#_2016-08-31_19_14_25_457 | 20:26 |
| mriedem | Starting request: 127.0.0.1 "GET /resource_providers/31c3a568-f605-47f7-a9be-b72ce49f51bf" status: 200 len: 0 __call__ /opt/stack/new/nova/nova/api/openstack/placement/handler.py:155 | 20:26 |
| dansmith | sdague: but only like two things.. is this not configured for placement? | 20:26 |
| dansmith | from the computes | 20:26 |
| dansmith | oh, this doesn't have my patch I guess/ | 20:26 |
| sdague | dansmith: it is | 20:26 |
| mriedem | right | 20:26 |
| mriedem | only the RP creation | 20:27 |
| dansmith | gotch | 20:27 |
| sdague | this is before your patch | 20:27 |
| mriedem | which is a 1 time thing in single node devstack | 20:27 |
| dansmith | will be many more after mine | 20:27 |
| *** akshai has joined #openstack-nova | 20:27 | |
| sdague | your patch hits it like every 2 minutes | 20:27 |
| sdague | dansmith: yes | 20:27 |
| *** cdent has quit IRC | 20:27 | |
| mriedem | this series on down is also ready https://review.openstack.org/#/c/363771/ | 20:27 |
| dansmith | cools | 20:27 |
| sdague | mriedem: ok, hitting that | 20:28 |
| *** mingyu has quit IRC | 20:28 | |
| mriedem | +2 on the logging change | 20:28 |
| riccardo | mriedem, what about from nova.api.openstack import xmlutil | 20:28 |
| riccardo | ? | 20:28 |
| riccardo | mriedem, i cannot find an oslo substitute | 20:29 |
| mriedem | riccardo: umm, which release are you looking at? | 20:29 |
| mriedem | xml was dropped from the api years ago | 20:29 |
| riccardo | mitaka | 20:29 |
| mriedem | riccardo: that was dropped in kilo | 20:30 |
| mriedem | so idk | 20:30 |
| riccardo | mriedem, btw, what IRC client are you using? I am on HexChat but get no notifications. | 20:30 |
| mriedem | pidgin | 20:30 |
| sdague | riccardo: xmlutil? | 20:31 |
| mriedem | you have to configure pidgin to get notified | 20:31 |
| *** moshele has quit IRC | 20:31 | |
| riccardo | ok | 20:31 |
| mriedem | riccardo: are you upgrading from like juno to mitaka or something? | 20:31 |
| mriedem | and updating patches | 20:31 |
| riccardo | yes, i am porting an selfmade extention from juno to mitaka | 20:32 |
| riccardo | sdague, yes :-) | 20:32 |
| mriedem | riccardo: holy rap | 20:32 |
| mriedem | *crap | 20:32 |
| mriedem | well, good luck | 20:32 |
| mriedem | edleafe: you can drop https://review.openstack.org/#/c/363125/ | 20:32 |
| riccardo | hahaha | 20:32 |
| *** claudiub has joined #openstack-nova | 20:32 | |
| sdague | riccardo: we deleted that 20 months ago - https://review.openstack.org/#/c/139650/ | 20:33 |
| mriedem | riccardo: btw, extensions are dead in newton | 20:33 |
| mriedem | http://specs.openstack.org/openstack/nova-specs/specs/newton/approved/api-no-more-extensions.html | 20:33 |
| edleafe | mriedem: yeah, was planning on it | 20:33 |
| riccardo | sdague, ok, lemme see for what it is used. | 20:33 |
| openstackgerrit | Slawek Kaplonski proposed openstack/nova: Fix race condition bug during live_snapshot https://review.openstack.org/363926 | 20:33 |
| *** Jeffrey4l__ has joined #openstack-nova | 20:34 | |
| riccardo | mriedem, really? so how should i do an extention? I am working on a project of the university of zurich | 20:34 |
| mriedem | riccardo: if you haven't noticed, a group of us are working frantically to land some features in the next <24 hours | 20:34 |
| *** Jeffrey4l_ has quit IRC | 20:35 | |
| mriedem | so helping with a juno API extension port to mitaka is not high priority right now | 20:35 |
| mriedem | riccardo: i'd suggest reading that spec | 20:35 |
| riccardo | mriedem, i see. i didn't know. | 20:35 |
| riccardo | sorry | 20:35 |
| *** whenry has quit IRC | 20:35 | |
| mriedem | np, just read the spec | 20:35 |
| riccardo | which one? | 20:35 |
| mriedem | http://specs.openstack.org/openstack/nova-specs/specs/newton/approved/api-no-more-extensions.html | 20:36 |
| riccardo | thx | 20:36 |
| mriedem | if you're going to have forked extensions, you'll need to carry a revert of the patch in that blueprint that kills extensions | 20:36 |
| mriedem | or upstream your extension | 20:36 |
| *** rlrossit has quit IRC | 20:37 | |
| mriedem | sdague: just fyi, since we have 3 grenade jobs in the nova check queue, but only gate-grenade-dsvm-neutron-ubuntu-trusty runs with neutron, that's the one i'm going to turn on for the cellsv2 stuff | 20:40 |
| mriedem | multinode and the other are n-net | 20:40 |
| mriedem | honestly we should probably just drop gate-grenade-dsvm-ubuntu-trusty since the grenade multinode job runs n-net | 20:40 |
| mriedem | so single-node + neutron + cellsv2, and multinode + nova-network and no cells v2 | 20:41 |
| sdague | mriedem: sure | 20:41 |
| *** ansiwen has quit IRC | 20:41 | |
| sdague | well, honestly, we probably want to turn our multinode job into the neutron one | 20:41 |
| sdague | there is going to need to be a rebalance of more jobs default to neutron post release | 20:42 |
| *** riccardo has quit IRC | 20:42 | |
| clarkb | sdague: we should be able to do something like neutron everywhere except one job on nova ya? | 20:44 |
| mriedem | sdague: yeah, i actually thought we were running neutron multinode grenade already | 20:44 |
| mriedem | oh that's right we run gate-tempest-dsvm-neutron-multinode-full | 20:45 |
| mriedem | which isn't grenade | 20:45 |
| mriedem | and gate-tempest-dsvm-multinode-live-migration which is nova-net | 20:45 |
| *** praneshp has joined #openstack-nova | 20:46 | |
| sdague | mriedem: there is a neutron version | 20:46 |
| mriedem | yeah i know | 20:46 |
| sdague | it's just run only on neutron I think | 20:46 |
| mriedem | anyway, don't need to sort that out right now | 20:47 |
| sdague | clarkb: yeh, basically we should make neutron the default service, and make explicit nova-net jobs when we need them | 20:47 |
| sdague | clarkb: I did a little reorg on that 2 weeks ago to get started | 20:47 |
| sdague | clarkb: https://review.openstack.org/#/c/352557/ | 20:48 |
| clarkb | cool | 20:48 |
| *** ducttape_ has quit IRC | 20:49 | |
| *** ducttape_ has joined #openstack-nova | 20:49 | |
| mriedem | weird, DEVSTACK_GATE_CELLS isn't used in d-g anymore | 20:51 |
| *** pwnall has quit IRC | 20:54 | |
| *** eharney has joined #openstack-nova | 20:55 | |
| *** rtheis has quit IRC | 20:56 | |
| mriedem | with DEVSTACK_LOCAL_CONFIG i don't have to make any d-g variables, so that's good | 20:57 |
| clarkb | ya we are trying to move more and more towards only having d-g variabels if d-g itself needs to act on them beyond passthrough to localrc | 20:58 |
| *** vladikr has quit IRC | 20:58 | |
| *** rfolco has quit IRC | 21:00 | |
| *** baoli_ has quit IRC | 21:02 | |
| *** baoli has joined #openstack-nova | 21:03 | |
| *** salv-orlando has joined #openstack-nova | 21:03 | |
| *** ansiwen has joined #openstack-nova | 21:03 | |
| *** tonytan4ever has joined #openstack-nova | 21:04 | |
| *** kfarr has quit IRC | 21:08 | |
| mriedem | sdague: dansmith: alaski: melwitt: ok here are the job changes for running cells v2 with cell0 https://review.openstack.org/#/c/363937/ | 21:08 |
| *** tonytan4ever has quit IRC | 21:09 | |
| dansmith | mriedem: cool | 21:09 |
| *** ayoung has quit IRC | 21:10 | |
| *** priteau has quit IRC | 21:11 | |
| alaski | mriedem: excellent | 21:11 |
| alaski | dansmith: looking at https://review.openstack.org/#/c/355602/9 we're still missing one bit. the db sync needs to run after the simple_cell_setup | 21:12 |
| *** code-R_ has quit IRC | 21:13 | |
| *** riccardo has joined #openstack-nova | 21:13 | |
| mriedem | gah | 21:13 |
| mriedem | those changes probably should have been squashed | 21:14 |
| mriedem | alaski: you should probably -1 if there is an issue | 21:14 |
| dansmith | alaski: hmm, but we have to run it late | 21:14 |
| mriedem | i was just about to ping someone to look at those | 21:14 |
| alaski | dansmith: yeah, I think we need to run db_sync again there | 21:15 |
| dansmith | ugh | 21:15 |
| dansmith | alaski: well, you gonna edit it? | 21:15 |
| mriedem | can simple_cell_setup run db_sync? | 21:15 |
| dansmith | mriedem: you might not want it to | 21:15 |
| mriedem | so we don't have to run db sync 50 times? | 21:15 |
| alaski | mriedem: we discussed doing it and decided against | 21:16 |
| alaski | dansmith: sure, I'll update | 21:16 |
| mriedem | we should double check that the release notes are clear about this | 21:16 |
| mriedem | and the order it has to happen in | 21:16 |
| mriedem | release notes and/or whatever docs we're going to have for this | 21:16 |
| dansmith | alaski: what if that command just syncs the cell0 database? | 21:18 |
| dansmith | alaski: that would be okay with me I think, and thus either ordering would work | 21:18 |
| alaski | dansmith: okay. I assumed that's what it would do if it synced | 21:18 |
| *** gouthamr has quit IRC | 21:18 | |
| alaski | so you'd prefer to update simple_cell_setup? | 21:19 |
| alaski | rather, have me update :) | 21:19 |
| sdague | dansmith: you want to bring us home with the logging - https://review.openstack.org/#/c/363130/ ? | 21:19 |
| dansmith | alaski: I think if I previously argued against the sync, I was talking about it affecting the main db | 21:19 |
| alaski | dansmith: okay | 21:19 |
| dansmith | alaski: definitely don't want it doing that, but.. I dunno, seems okay to do the cell0 one | 21:19 |
| dansmith | alaski: as long as the regular sync still hits cell0 as well, so it's not like the only way to sync cell0 is with setup | 21:20 |
| alaski | dansmith: roger | 21:20 |
| *** ansiwen has quit IRC | 21:21 | |
| dansmith | alaski: sorry if I'm the one that prevented us from doing that | 21:21 |
| openstackgerrit | Mathieu Gagné proposed openstack/nova: Make scheduler less greedy when loading instances for tracking https://review.openstack.org/363944 | 21:21 |
| alaski | no worries, we'll get this sorted eventually | 21:21 |
| mriedem | if by eventually you mean the next 2 hours :) | 21:22 |
| *** mdrabe has quit IRC | 21:24 | |
| *** mdrabe has joined #openstack-nova | 21:25 | |
| *** sdague has quit IRC | 21:28 | |
| *** praneshp has quit IRC | 21:28 | |
| *** abalutoiu_ has quit IRC | 21:28 | |
| bauzas | folks, bailing out for tonight, but given tomorrow is the last day before our FF, ping me if you want me to help you about some review needed or some implementation change \o | 21:29 |
| *** tblakes has quit IRC | 21:29 | |
| mriedem | bauzas: thanks | 21:30 |
| dansmith | o/ | 21:30 |
| *** zz_dimtruck is now known as dimtruck | 21:31 | |
| *** pwnall has joined #openstack-nova | 21:32 | |
| mriedem | melwitt: is your only thing on this the commit message https://review.openstack.org/#/c/363888/ ? | 21:32 |
| mriedem | b/c otherwise it looks good to me | 21:32 |
| melwitt | mriedem: yeah | 21:32 |
| mriedem | then let's fix it inline and +W | 21:32 |
| melwitt | k | 21:32 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Fix service version lookups https://review.openstack.org/363888 | 21:32 |
| *** zengine has quit IRC | 21:35 | |
| *** kevinz has joined #openstack-nova | 21:38 | |
| *** thorst has quit IRC | 21:38 | |
| mikal | mriedem: did you see that service account patch I emailed about? | 21:38 |
| *** pwnall has quit IRC | 21:38 | |
| mikal | mriedem: is there any point in trying to get that to somewhere near ready to merge? Or do you want to hold off for newton? | 21:38 |
| mriedem | mikal: yeah, we've been pushing all day on placement service and cellsv2 so i dont' think we deal with the vd2 service account in newton | 21:39 |
| *** pwnall has joined #openstack-nova | 21:39 | |
| mikal | mriedem: that's fine, I'll keep pottering. Its actually a relatively small change, but does require that a bunch of the vendordata flags move into a config group, which causes some churn. | 21:40 |
| mikal | mriedem: its also untested as I haven't gotten that far yet | 21:40 |
| mriedem | right, with 24 hours to FF i'd like to hold off | 21:40 |
| mikal | mriedem: that's fair | 21:41 |
| *** scsnow has quit IRC | 21:41 | |
| *** thumpba has quit IRC | 21:41 | |
| mikal | mriedem: we can talk about if not working with keystone middleware is a "bug" next week I suppose | 21:41 |
| mriedem | alright | 21:42 |
| *** kevinz has quit IRC | 21:42 | |
| mriedem | alaski: on https://review.openstack.org/#/c/363911/ it says "Because an instance list will attempt to pull instances from the build_requests table" - isn't that only true after the next change in the series? | 21:44 |
| *** praneshp has joined #openstack-nova | 21:44 | |
| *** Hosam has quit IRC | 21:44 | |
| alaski | mriedem: yes | 21:44 |
| alaski | would a different wording be better | 21:44 |
| mriedem | so should those be re-ordered? | 21:44 |
| alaski | that will introduce a potential bug | 21:45 |
| mriedem | squash them? | 21:45 |
| alaski | I can, but the next change is somewhat large already | 21:46 |
| *** dimtruck is now known as zz_dimtruck | 21:46 | |
| openstackgerrit | Andrew Laski proposed openstack/nova: Run cell0 db migrations during nova-manage simple_cell_setup https://review.openstack.org/363956 | 21:47 |
| alaski | dansmith: ^ | 21:47 |
| *** cleong has quit IRC | 21:48 | |
| *** diogogmt has quit IRC | 21:48 | |
| *** zz_dimtruck is now known as dimtruck | 21:52 | |
| *** pwnall1 has joined #openstack-nova | 21:53 | |
| *** pwnall has quit IRC | 21:53 | |
| *** chrichip has quit IRC | 21:54 | |
| openstackgerrit | Andrew Laski proposed openstack/nova: Delete BuildRequest regardless of service_version https://review.openstack.org/363911 | 21:54 |
| *** chrichip has joined #openstack-nova | 21:55 | |
| alaski | mriedem: updated the comment | 21:56 |
| mriedem | alaski: ok, one small issue in the test in https://review.openstack.org/#/c/363956/ | 21:57 |
| mriedem | otherwise +2 | 21:57 |
| *** anush__ has quit IRC | 21:58 | |
| alaski | mriedem: hmm, I've been using assert_not_called in other places. | 21:59 |
| alaski | but I'll update | 21:59 |
| mriedem | oh, i'm confusing the lib with the py version | 22:00 |
| mriedem | yeah https://github.com/testing-cabal/mock/blob/master/mock/mock.py#L890 | 22:00 |
| mriedem | don't udpate it | 22:00 |
| alaski | okay | 22:00 |
| *** cdent has joined #openstack-nova | 22:00 | |
| mriedem | +2 now | 22:01 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Run cell0 db migrations during nova-manage simple_cell_setup https://review.openstack.org/363956 | 22:01 |
| alaski | gah | 22:01 |
| mriedem | ha | 22:01 |
| alaski | just pushed for the bp in commit message | 22:01 |
| mriedem | ok +2 again | 22:01 |
| mriedem | the devstack change should probably depends on this right? | 22:01 |
| alaski | \o/ | 22:01 |
| alaski | I updated that | 22:01 |
| mriedem | this one https://review.openstack.org/#/c/355602/ | 22:01 |
| mriedem | ok | 22:01 |
| mriedem | now i can pester mtreinish | 22:02 |
| cdent | jaypipes: making good headway? If you get it to an almost state where the next steps are clear I can pick it up in the morning. | 22:02 |
| alaski | I'm out for a bit now. will try to check back in later | 22:02 |
| mriedem | mtreinish: can you take a gander at these? https://review.openstack.org/#/q/topic:cell0_db+project:openstack-dev/devstack | 22:03 |
| edleafe | cdent: jaypipes: I can take a whack at it this evening, too, if needed | 22:03 |
| *** ansiwen has joined #openstack-nova | 22:05 | |
| cdent | thanks edleafe. I'm going to bed soon, will cruise gerrit in the morning | 22:05 |
| cdent | 'night | 22:05 |
| *** cdent has quit IRC | 22:05 | |
| *** marst has quit IRC | 22:07 | |
| edleafe | 'night | 22:07 |
| *** mingyu has joined #openstack-nova | 22:12 | |
| mriedem | my stomach is eating itself so i'm going to go home, will be back online later | 22:12 |
| *** mriedem has quit IRC | 22:12 | |
| *** akshai has quit IRC | 22:12 | |
| openstackgerrit | Jay Pipes proposed openstack/nova: WIP create_all and delete_all for AllocationList https://review.openstack.org/363209 | 22:13 |
| jaypipes | edleafe: ^^ | 22:13 |
| jaypipes | edleafe: missing unit tests but it should be good to have a whack at. | 22:14 |
| edleafe | jaypipes: ok, I'll whack away :) | 22:14 |
| jaypipes | cheers | 22:14 |
| edleafe | jaypipes: enjoy what's left of your evening | 22:14 |
| *** xyang1 has quit IRC | 22:15 | |
| *** mingyu has quit IRC | 22:16 | |
| *** eharney has quit IRC | 22:17 | |
| *** praneshp has quit IRC | 22:23 | |
| *** lpetrut has quit IRC | 22:26 | |
| *** Sukhdev has joined #openstack-nova | 22:28 | |
| *** akshai has joined #openstack-nova | 22:32 | |
| *** su_zhang has quit IRC | 22:32 | |
| openstackgerrit | melanie witt proposed openstack/nova: Move cell message queue switching and add caching https://review.openstack.org/326906 | 22:33 |
| *** ducttape_ has quit IRC | 22:33 | |
| *** mdrabe has quit IRC | 22:33 | |
| *** Sukhdev has quit IRC | 22:34 | |
| *** akshai has quit IRC | 22:34 | |
| jaypipes | edleafe: I'm heading to bed. No promises on any of that code. Likely won't work but it should be self-explanatory and if not, accompanied by extensive code comments. | 22:35 |
| jaypipes | edleafe: ty very much for looking at it. | 22:36 |
| *** sbezverk_ has left #openstack-nova | 22:36 | |
| dansmith | jaypipes: thanks for doing that | 22:37 |
| edleafe | jaypipes: it looks sane after a quick review. Gonna dig a little deeper now | 22:39 |
| *** hemna is now known as hemnafk | 22:40 | |
| jaypipes | dansmith: yeah, I'll check in bright and early (it's almost midnight here) and work on the client (RT) side of things tomorrow. | 22:40 |
| *** efried has quit IRC | 22:40 | |
| dansmith | jaypipes: right on, I'll see what you got done when I get up and push on it then | 22:41 |
| *** akshai has joined #openstack-nova | 22:42 | |
| *** riccardo has quit IRC | 22:45 | |
| *** Sukhdev has joined #openstack-nova | 22:45 | |
| *** mriedem has joined #openstack-nova | 22:47 | |
| auggy | i have a gerrit question, the tests in this patch are fixed in the related change - https://review.openstack.org/#/c/357726 | 22:47 |
| auggy | what's the best way to link these two changes? | 22:48 |
| *** TravT has quit IRC | 22:49 | |
| auggy | i can also split up the work i did to update the tests so it applies to both patches, but that seems kind of silly | 22:50 |
| *** praneshp has joined #openstack-nova | 22:52 | |
| *** jaypipes has quit IRC | 22:55 | |
| *** Sukhdev has quit IRC | 22:55 | |
| mriedem | auggy: put them in a dependent series | 22:55 |
| mriedem | http://docs.openstack.org/infra/manual/developers.html#adding-a-dependency | 22:56 |
| mriedem | dansmith: alaski: the cellsv2 grenade and devstack changes are being tested here https://review.openstack.org/#/c/363971/ | 22:58 |
| mriedem | ^ is the same thing that will test that project-config change to those 2 jobs | 22:59 |
| mriedem | i'm probably just going to babysit changes tonight | 22:59 |
| mriedem | dansmith: maybe one that d-g test change passes grenade you want to move these in https://review.openstack.org/#/q/topic:setup_cell0_before_migrations | 23:00 |
| mriedem | unless there is some conflict of interest with you +Wing those | 23:00 |
| mriedem | *once | 23:00 |
| armax | mriedem: ping | 23:01 |
| kevinbenton | mriedem: yoo | 23:01 |
| mriedem | crap https://review.openstack.org/#/c/363956/ already hit an infra failure | 23:01 |
| mriedem | oh no | 23:01 |
| mriedem | neutron double whammy | 23:01 |
| melwitt | auggy: I don't know the details of those patches but generally the tests should be updated alongside each patch i.e. if implementation changes and test correspondingly need to be changed | 23:02 |
| *** ducttape_ has joined #openstack-nova | 23:02 | |
| armax | what happen? I open the channel and I see neutron double whammy | 23:02 |
| kevinbenton | mriedem: armax and i have poor coordination skills :) | 23:02 |
| armax | do I need to go away? | 23:02 |
| mriedem | "let's go pounce on mriedem in -nova" | 23:02 |
| *** gouthamr has joined #openstack-nova | 23:02 | |
| kevinbenton | armax: you want to ask or me? | 23:02 |
| armax | kevinbenton: you go right ahead | 23:03 |
| armax | I can listen | 23:03 |
| kevinbenton | mriedem: os-vif fix for a linux bridge race merged | 23:03 |
| mriedem | patch? | 23:03 |
| mriedem | *link | 23:03 |
| *** pradk has quit IRC | 23:03 | |
| kevinbenton | mriedem: https://review.openstack.org/#/c/361516/ | 23:03 |
| auggy | mriedem: kk thx | 23:03 |
| mriedem | which is not in 1.2.0 | 23:03 |
| mriedem | or stable/newton | 23:03 |
| kevinbenton | mriedem: oh wow, already stable/newton is cut | 23:04 |
| mriedem | yeah, yesterday | 23:04 |
| armax | mriedem: we discoverd it last Friday, dhellmann said we had time until Monday to catch last minute fixes | 23:04 |
| kevinbenton | mriedem: ok, so i'll propose cherry-pick | 23:04 |
| mriedem | yeah so let's backport and release as 1.2.1 | 23:04 |
| kevinbenton | mriedem: https://review.openstack.org/#/c/363975/ | 23:04 |
| kevinbenton | mriedem: so another release is ok then? | 23:04 |
| mriedem | apparently stable-maint-core isn't on os-vif stable branches | 23:04 |
| mriedem | kevinbenton: 1.2.1 for stable/newton should be ok i think | 23:05 |
| kevinbenton | mriedem: it's an unmanned ship! | 23:05 |
| auggy | melwitt: yeah the issue was I'd fix broken Mox tests for the first change and then have to fix them again for the second change | 23:05 |
| mriedem | kevinbenton: probably an oversight, i can push a project-config change for that | 23:05 |
| kevinbenton | mriedem: thanks | 23:05 |
| auggy | I'll do the series thing, but I can split it out if that ends up making more sense | 23:06 |
| *** sameo has quit IRC | 23:07 | |
| melwitt | auggy: each logical change should be separate and the test fixings are usually logically separate. if there's really a dependency there then it sounds like you need to swap the order of the commits | 23:07 |
| mriedem | gerrit acl for os-vif stable https://review.openstack.org/#/c/363978/ | 23:10 |
| *** mingyu has joined #openstack-nova | 23:12 | |
| *** jamielennox|away is now known as jamielennox | 23:14 | |
| *** akshai_ has joined #openstack-nova | 23:15 | |
| *** mingyu has quit IRC | 23:17 | |
| mriedem | kevinbenton: armax: ok +1, i can't +2 until that project-config change is merged though | 23:17 |
| mriedem | but once it's merged we can do a 1.2.1 release for stable/newton | 23:17 |
| armax | mriedem: ack | 23:17 |
| *** akshai has quit IRC | 23:19 | |
| *** krtaylor has quit IRC | 23:22 | |
| *** jamielennox is now known as jamielennox|away | 23:22 | |
| *** cdent has joined #openstack-nova | 23:22 | |
| cdent | edleafe: I left some comments on jay's stuff identifying a couple of issues that if you feel inclined ought to be relatively doable | 23:23 |
| *** salv-orlando has quit IRC | 23:25 | |
| *** pwnall1 has quit IRC | 23:26 | |
| *** jamielennox|away is now known as jamielennox | 23:28 | |
| *** cdent has quit IRC | 23:29 | |
| *** nic has quit IRC | 23:30 | |
| *** krtaylor has joined #openstack-nova | 23:34 | |
| auggy | melwitt: alright that makes sense | 23:34 |
| mriedem | dansmith: alaski: crap the simple_cell_setup change to run db sync for cell0 fails grenade as there is no cell0 db | 23:35 |
| mriedem | https://review.openstack.org/#/c/363956/2 | 23:35 |
| mriedem | http://logs.openstack.org/56/363956/2/check/gate-grenade-dsvm-ubuntu-trusty/1f9aa24/logs/grenade.sh.txt.gz#_2016-08-31_23_24_52_904 | 23:35 |
| mriedem | we should have probably realized that given the call today :) | 23:35 |
| edleafe | cdent: you should be asleep now. I've already found a few things; I'll look at your comments | 23:35 |
| *** chrichip has quit IRC | 23:36 | |
| *** jamielennox is now known as jamielennox|away | 23:36 | |
| alaski | mriedem: I thought we were only calling that if the conditional was True | 23:36 |
| alaski | oh, I guess it needs a depends-on | 23:37 |
| *** gongysh has joined #openstack-nova | 23:37 | |
| *** chrichip has joined #openstack-nova | 23:37 | |
| *** apuimedo has joined #openstack-nova | 23:37 | |
| *** Swami has quit IRC | 23:37 | |
| *** apuimedo has left #openstack-nova | 23:37 | |
| openstackgerrit | Andrew Laski proposed openstack/nova: Run cell0 db migrations during nova-manage simple_cell_setup https://review.openstack.org/363956 | 23:38 |
| alaski | added it ^ | 23:38 |
| *** oomichi has quit IRC | 23:39 | |
| *** oomichi has joined #openstack-nova | 23:39 | |
| *** hemanthm is now known as hemanthm|afk | 23:40 | |
| *** oomichi has quit IRC | 23:40 | |
| *** Sukhdev has joined #openstack-nova | 23:41 | |
| *** chrichip has quit IRC | 23:42 | |
| *** claudiub has quit IRC | 23:42 | |
| *** Sukhdev has quit IRC | 23:43 | |
| *** chrichip has joined #openstack-nova | 23:43 | |
| *** oomichi has joined #openstack-nova | 23:44 | |
| *** diogogmt has joined #openstack-nova | 23:46 | |
| edleafe | brb, grabbing some food | 23:47 |
| melwitt | alaski: I noticed https://review.openstack.org/#/c/363888/ already failed the functional job. I don't yet understand why. get_minimum_version now returns 15 instead of 0 but it looks like it should just fall back on what it was doing before the change | 23:47 |
| alaski | and in the functional tests it matters even less because the AllServicesCurrent fixture uncondionally returns 15, so even the old lookups were returning 15 | 23:52 |
| melwitt | hm | 23:54 |
| *** rama_y has quit IRC | 23:58 | |
| *** hieulq has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
