| *** sdake_ has joined #openstack-nova | 00:06 | |
| *** sdake has quit IRC | 00:10 | |
| gmann | jogo: Please have a look on functional tests merge patches - https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:merge_sample_tests,n,z | 00:10 |
|---|---|---|
| *** Glass_saga has joined #openstack-nova | 00:23 | |
| Glass_saga | Can anyone review this? Thanks in advance. https://bugs.launchpad.net/python-novaclient/+bug/1248517 | 00:24 |
| openstack | Launchpad bug 1248517 in python-novaclient "Set admin_pass not possible through CLI" [Wishlist,In progress] - Assigned to Masaki Matsushita (mmasaki) | 00:24 |
| *** markvoelker has joined #openstack-nova | 00:25 | |
| *** markvoelker has quit IRC | 00:30 | |
| *** tonytan4ever has joined #openstack-nova | 00:31 | |
| *** sdake has joined #openstack-nova | 00:32 | |
| *** sdake_ has quit IRC | 00:35 | |
| *** igordcard has quit IRC | 00:36 | |
| *** mitz has joined #openstack-nova | 00:47 | |
| *** otter768 has joined #openstack-nova | 00:48 | |
| *** pixelbeat_ has quit IRC | 00:50 | |
| *** yingjun has joined #openstack-nova | 00:52 | |
| *** otter768 has quit IRC | 00:52 | |
| *** sdake_ has joined #openstack-nova | 01:06 | |
| *** dims__ has quit IRC | 01:06 | |
| *** jamielennox|away is now known as jamielennox | 01:07 | |
| *** sdake__ has joined #openstack-nova | 01:09 | |
| *** sdake has quit IRC | 01:09 | |
| *** sdake__ is now known as sdake | 01:09 | |
| *** mriedem has quit IRC | 01:11 | |
| *** sdake_ has quit IRC | 01:13 | |
| *** dave-mccowan has quit IRC | 01:24 | |
| *** aix has quit IRC | 01:26 | |
| *** eliqiao1 has quit IRC | 01:34 | |
| *** eliqiao has joined #openstack-nova | 01:37 | |
| *** eliqiao has left #openstack-nova | 01:39 | |
| *** heyongli has joined #openstack-nova | 01:40 | |
| heyongli | this patch broken the pci, https://review.openstack.org/#/c/188486/, this can be avoid by merge this fix:https://review.openstack.org/#/c/131321/, | 01:42 |
| heyongli | so, we might should review that fix, to avoid such conner bugs of pci. | 01:43 |
| *** sdake_ has joined #openstack-nova | 01:43 | |
| *** annashen has joined #openstack-nova | 01:45 | |
| heyongli | wait, let me check that log -:0 | 01:45 |
| *** dave-mccowan has joined #openstack-nova | 01:46 | |
| *** sdake has quit IRC | 01:46 | |
| *** sdake has joined #openstack-nova | 01:54 | |
| *** sdake_ has quit IRC | 01:58 | |
| *** dims_ has joined #openstack-nova | 01:58 | |
| *** jwcroppe_ has joined #openstack-nova | 02:03 | |
| *** Kennan2 has joined #openstack-nova | 02:03 | |
| *** dims_ has quit IRC | 02:03 | |
| *** Kennan has quit IRC | 02:04 | |
| *** baoli has joined #openstack-nova | 02:05 | |
| *** jwcroppe has quit IRC | 02:05 | |
| *** baoli has quit IRC | 02:09 | |
| openstackgerrit | heijlong proposed openstack/nova: Merge extended-volume functional tests of v2 and v2.1 https://review.openstack.org/173216 | 02:12 |
| *** park_heijlong has joined #openstack-nova | 02:12 | |
| *** markvoelker has joined #openstack-nova | 02:14 | |
| *** markvoelker has quit IRC | 02:19 | |
| *** tonytan4ever has quit IRC | 02:22 | |
| *** bzhao has joined #openstack-nova | 02:22 | |
| *** sdake_ has joined #openstack-nova | 02:25 | |
| *** bzhao has quit IRC | 02:29 | |
| *** sdake has quit IRC | 02:29 | |
| *** Glaupre has quit IRC | 02:32 | |
| *** davidlenwell has quit IRC | 02:33 | |
| *** sdake has joined #openstack-nova | 02:44 | |
| openstackgerrit | wangxiyuan proposed openstack/nova: Add security group check when boot with port https://review.openstack.org/188993 | 02:46 |
| *** sdake_ has quit IRC | 02:48 | |
| *** otter768 has joined #openstack-nova | 02:48 | |
| *** otter768 has quit IRC | 02:53 | |
| *** patrickeast has quit IRC | 02:58 | |
| *** Jeffrey4l has joined #openstack-nova | 02:58 | |
| *** zhenguo has joined #openstack-nova | 03:03 | |
| *** dims_ has joined #openstack-nova | 03:03 | |
| *** aswadr has joined #openstack-nova | 03:04 | |
| *** dims_ has quit IRC | 03:09 | |
| *** annashen has quit IRC | 03:12 | |
| *** liusheng has joined #openstack-nova | 03:13 | |
| *** heyongli has quit IRC | 03:13 | |
| *** heyongli has joined #openstack-nova | 03:14 | |
| *** vilobhmm has joined #openstack-nova | 03:16 | |
| *** annashen has joined #openstack-nova | 03:18 | |
| openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova: Add and use raise_feature_not_supported() https://review.openstack.org/163697 | 03:18 |
| openstackgerrit | Ken'ichi Ohmichi proposed openstack/nova: Add a hacking rule for consistent HTTP501 message https://review.openstack.org/163698 | 03:18 |
| openstackgerrit | Kevin Benton proposed openstack/nova: Set default event status to completed https://review.openstack.org/189176 | 03:22 |
| *** park_heijlong has quit IRC | 03:23 | |
| *** mdbooth has quit IRC | 03:27 | |
| *** pkholkin has quit IRC | 03:27 | |
| *** toure has quit IRC | 03:27 | |
| *** tanee has quit IRC | 03:27 | |
| *** heyongli has quit IRC | 03:27 | |
| *** josecastroleon has quit IRC | 03:27 | |
| *** bradjones|away has quit IRC | 03:27 | |
| *** winston-d has quit IRC | 03:27 | |
| *** mnestratov has quit IRC | 03:27 | |
| *** mestery_afk has quit IRC | 03:27 | |
| *** asselin has quit IRC | 03:27 | |
| *** aluria has quit IRC | 03:27 | |
| *** mikal has quit IRC | 03:27 | |
| *** ondergetekende has quit IRC | 03:27 | |
| *** toure has joined #openstack-nova | 03:27 | |
| *** pkholkin has joined #openstack-nova | 03:27 | |
| *** mdbooth has joined #openstack-nova | 03:27 | |
| *** tanee has joined #openstack-nova | 03:28 | |
| *** mikal has joined #openstack-nova | 03:28 | |
| *** winston-d has joined #openstack-nova | 03:28 | |
| *** aluria has joined #openstack-nova | 03:28 | |
| *** ondergetekende has joined #openstack-nova | 03:28 | |
| *** josecastroleon has joined #openstack-nova | 03:28 | |
| *** mnestratov|3 has joined #openstack-nova | 03:28 | |
| *** bradjones has joined #openstack-nova | 03:28 | |
| *** bradjones has quit IRC | 03:28 | |
| *** bradjones has joined #openstack-nova | 03:28 | |
| *** asselin has joined #openstack-nova | 03:28 | |
| *** mestery has joined #openstack-nova | 03:28 | |
| *** annashen has quit IRC | 03:31 | |
| *** annashen has joined #openstack-nova | 03:32 | |
| *** annashen has quit IRC | 03:37 | |
| *** annashen has joined #openstack-nova | 03:38 | |
| *** annashen has quit IRC | 03:43 | |
| *** coolsvap|afk is now known as coolsvap | 03:44 | |
| *** deepthi has joined #openstack-nova | 03:45 | |
| *** yingjun has quit IRC | 03:50 | |
| *** annashen has joined #openstack-nova | 03:54 | |
| *** davidlenwell has joined #openstack-nova | 03:55 | |
| *** dave-mcc_ has joined #openstack-nova | 03:55 | |
| openstackgerrit | Rui Chen proposed openstack/nova: Refactor network API 'get_instance_nw_info' https://review.openstack.org/146036 | 03:56 |
| *** dave-mccowan has quit IRC | 03:57 | |
| *** dave-mcc_ has quit IRC | 03:59 | |
| *** armax has joined #openstack-nova | 03:59 | |
| *** armax has quit IRC | 04:04 | |
| *** jang has joined #openstack-nova | 04:13 | |
| *** jwcroppe has joined #openstack-nova | 04:18 | |
| *** jwcroppe_ has quit IRC | 04:21 | |
| *** annashen has quit IRC | 04:23 | |
| *** annashen has joined #openstack-nova | 04:23 | |
| *** annashen has quit IRC | 04:25 | |
| *** annashen has joined #openstack-nova | 04:25 | |
| *** annashen has quit IRC | 04:26 | |
| *** irenab has quit IRC | 04:26 | |
| *** annashen has joined #openstack-nova | 04:26 | |
| *** annashen has quit IRC | 04:27 | |
| *** garyk has quit IRC | 04:29 | |
| *** armax has joined #openstack-nova | 04:30 | |
| *** unicell1 has joined #openstack-nova | 04:31 | |
| *** lpetrut has joined #openstack-nova | 04:32 | |
| *** achanda has joined #openstack-nova | 04:33 | |
| *** unicell has quit IRC | 04:33 | |
| *** annashen has joined #openstack-nova | 04:33 | |
| *** armax has quit IRC | 04:35 | |
| *** otter768 has joined #openstack-nova | 04:49 | |
| *** otter768 has quit IRC | 04:54 | |
| *** achanda has quit IRC | 04:55 | |
| *** Maike has joined #openstack-nova | 05:02 | |
| *** nihilifer_ has joined #openstack-nova | 05:09 | |
| *** annashen has quit IRC | 05:09 | |
| *** achanda has joined #openstack-nova | 05:13 | |
| *** belmoreira has joined #openstack-nova | 05:20 | |
| *** achanda_ has joined #openstack-nova | 05:22 | |
| *** achanda has quit IRC | 05:23 | |
| *** irenab has joined #openstack-nova | 05:25 | |
| openstackgerrit | Andrey Pavlov proposed openstack/nova: template for disabe_terminate and shutdown_terminate flags https://review.openstack.org/189112 | 05:25 |
| *** eliqiao has joined #openstack-nova | 05:27 | |
| *** achanda_ has quit IRC | 05:28 | |
| *** oomichi has joined #openstack-nova | 05:28 | |
| *** vilobhmm has quit IRC | 05:32 | |
| *** iamjarvo has quit IRC | 05:34 | |
| *** lsmola has joined #openstack-nova | 05:36 | |
| *** apuimedo has joined #openstack-nova | 05:37 | |
| *** vilobhmm has joined #openstack-nova | 05:39 | |
| *** yingjun has joined #openstack-nova | 05:40 | |
| *** lpetrut has left #openstack-nova | 05:45 | |
| *** Kennan2 has quit IRC | 05:47 | |
| *** arnaud____ has joined #openstack-nova | 05:47 | |
| *** Kennan has joined #openstack-nova | 05:48 | |
| *** markvoelker has joined #openstack-nova | 05:52 | |
| *** sudipto has joined #openstack-nova | 05:54 | |
| *** markvoelker has quit IRC | 05:56 | |
| *** jcoufal has joined #openstack-nova | 05:57 | |
| *** pradipta has joined #openstack-nova | 05:58 | |
| *** destiny has joined #openstack-nova | 06:02 | |
| *** tteggel has quit IRC | 06:04 | |
| *** belmoreira has quit IRC | 06:06 | |
| *** arnaud____ has quit IRC | 06:13 | |
| *** josecastroleon has quit IRC | 06:19 | |
| *** achanda has joined #openstack-nova | 06:19 | |
| *** arnaud____ has joined #openstack-nova | 06:20 | |
| *** moshele has joined #openstack-nova | 06:36 | |
| *** dulek has joined #openstack-nova | 06:39 | |
| openstackgerrit | Rui Chen proposed openstack/nova: Fix resource leaking when consume_from_instance raise exception https://review.openstack.org/182165 | 06:40 |
| openstackgerrit | shihanzhang proposed openstack/nova: Optimize validate_networks https://review.openstack.org/175742 | 06:40 |
| *** tobe has joined #openstack-nova | 06:45 | |
| *** mordred has quit IRC | 06:46 | |
| *** arnaud____ has quit IRC | 06:47 | |
| *** abhishekk has joined #openstack-nova | 06:49 | |
| openstackgerrit | Kevin Benton proposed openstack/nova: Set default event status to completed https://review.openstack.org/189176 | 06:50 |
| *** heyongli has joined #openstack-nova | 06:50 | |
| *** otter768 has joined #openstack-nova | 06:50 | |
| openstackgerrit | Yongli He proposed openstack/nova: resource tracker style pci resource management https://review.openstack.org/131321 | 06:51 |
| *** lennyb has joined #openstack-nova | 06:51 | |
| *** garyk has joined #openstack-nova | 06:51 | |
| *** ajayaa has joined #openstack-nova | 06:51 | |
| *** otter768 has quit IRC | 06:55 | |
| *** sahid has joined #openstack-nova | 06:57 | |
| *** vilobhmm has quit IRC | 07:02 | |
| *** achanda has quit IRC | 07:02 | |
| *** lpetrut has joined #openstack-nova | 07:03 | |
| *** tojuvone has joined #openstack-nova | 07:03 | |
| *** achanda has joined #openstack-nova | 07:04 | |
| *** arnaud____ has joined #openstack-nova | 07:05 | |
| garyk | BobBall: you around? Ci is down. | 07:09 |
| *** masayukig has quit IRC | 07:12 | |
| *** kiwik-chenrui has quit IRC | 07:14 | |
| *** markus_z has joined #openstack-nova | 07:16 | |
| *** masayukig has joined #openstack-nova | 07:17 | |
| *** achanda has quit IRC | 07:17 | |
| *** RuiChen has joined #openstack-nova | 07:19 | |
| *** lpetrut1 has joined #openstack-nova | 07:28 | |
| *** arnaud____ has quit IRC | 07:28 | |
| rgerganov | alex_xu: hi | 07:28 |
| rgerganov | alex_xu: any chance we can get some core attention on https://review.openstack.org/#/c/185844/ ? | 07:28 |
| *** lpetrut has quit IRC | 07:31 | |
| *** gilliard has joined #openstack-nova | 07:34 | |
| openstackgerrit | garyk proposed openstack/nova: VMware: convert driver to use nova.objects.ImageMeta https://review.openstack.org/189085 | 07:37 |
| *** markvoelker has joined #openstack-nova | 07:40 | |
| *** matrohon has joined #openstack-nova | 07:40 | |
| *** lpetrut1 has quit IRC | 07:43 | |
| *** markvoelker has quit IRC | 07:45 | |
| *** dguitarbite has joined #openstack-nova | 07:46 | |
| *** jlanoux has joined #openstack-nova | 07:48 | |
| *** lpetrut has joined #openstack-nova | 07:48 | |
| *** ian_brown has quit IRC | 07:50 | |
| *** ian_brown_ has joined #openstack-nova | 07:50 | |
| *** romainh has joined #openstack-nova | 07:52 | |
| *** jistr has joined #openstack-nova | 07:52 | |
| gilliard | oomichi: are you still there? | 07:56 |
| gilliard | oomichi: I have a question about https://review.openstack.org/#/c/163698/9 | 07:56 |
| *** ian_brown_ has quit IRC | 07:57 | |
| *** jlanoux has quit IRC | 07:57 | |
| *** lpetrut1 has joined #openstack-nova | 07:57 | |
| *** jlanoux has joined #openstack-nova | 07:58 | |
| *** pbandzi has joined #openstack-nova | 07:58 | |
| *** lpetrut has quit IRC | 08:00 | |
| *** belmoreira has joined #openstack-nova | 08:00 | |
| *** dguitarbite has quit IRC | 08:01 | |
| *** ZZelle has joined #openstack-nova | 08:01 | |
| *** pixelbeat_ has joined #openstack-nova | 08:01 | |
| *** matrohon has quit IRC | 08:05 | |
| gilliard | oomichi: OK I'll just comment on the patch... | 08:05 |
| *** lucasagomes has joined #openstack-nova | 08:05 | |
| *** davidlenwell has quit IRC | 08:06 | |
| *** afazekas_mtg has joined #openstack-nova | 08:06 | |
| *** davidlenwell has joined #openstack-nova | 08:09 | |
| *** annashen has joined #openstack-nova | 08:10 | |
| *** matrohon has joined #openstack-nova | 08:11 | |
| *** claudiub has joined #openstack-nova | 08:11 | |
| *** claudiub_ has joined #openstack-nova | 08:13 | |
| *** derekh has joined #openstack-nova | 08:14 | |
| *** achanda has joined #openstack-nova | 08:14 | |
| *** yingjun has quit IRC | 08:14 | |
| *** annashen has quit IRC | 08:14 | |
| *** claudiub has quit IRC | 08:15 | |
| *** ndipanov has joined #openstack-nova | 08:15 | |
| *** josecastroleon has joined #openstack-nova | 08:18 | |
| *** yingjun has joined #openstack-nova | 08:20 | |
| *** danpb has joined #openstack-nova | 08:22 | |
| openstackgerrit | Alex Xu proposed openstack/nova-specs: Remove v3 from nova code tree https://review.openstack.org/189218 | 08:22 |
| *** exploreshaifali has joined #openstack-nova | 08:23 | |
| *** davidlenwell has quit IRC | 08:23 | |
| *** lxsli has joined #openstack-nova | 08:24 | |
| BobBall | thanks garyk - investigating | 08:25 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: Split instance resize operation into 2 phases https://review.openstack.org/184495 | 08:27 |
| *** shihanzhang has joined #openstack-nova | 08:27 | |
| BobBall | garyk: Seems that it's working but it's taking forever to create new nodes so only a very few jobs run at any time and there is a huge backlog... | 08:27 |
| garyk | BobBall: thanks | 08:27 |
| BobBall | garyk: Your change got commented on late last night :) | 08:27 |
| BobBall | I'll fix it today | 08:28 |
| garyk | danpb: you around? i have some questions reqgarding the image_meta props | 08:28 |
| garyk | thanks! | 08:28 |
| *** dims_ has joined #openstack-nova | 08:28 | |
| openstackgerrit | Mh Raies proposed openstack/nova: Enable tenant/user specific instance filtering https://review.openstack.org/139025 | 08:28 |
| *** hanzhang has quit IRC | 08:28 | |
| *** shz has joined #openstack-nova | 08:29 | |
| *** achanda has quit IRC | 08:29 | |
| *** davidlenwell has joined #openstack-nova | 08:30 | |
| *** achanda has joined #openstack-nova | 08:30 | |
| *** yassine_ has joined #openstack-nova | 08:30 | |
| *** lpetrut1 has quit IRC | 08:31 | |
| *** yassine_ has quit IRC | 08:31 | |
| *** yassine_ has joined #openstack-nova | 08:32 | |
| *** shihanzhang has quit IRC | 08:33 | |
| *** dims_ has quit IRC | 08:33 | |
| *** dguitarbite has joined #openstack-nova | 08:34 | |
| *** achanda has quit IRC | 08:35 | |
| *** liusheng has quit IRC | 08:36 | |
| *** liusheng has joined #openstack-nova | 08:36 | |
| *** aix has joined #openstack-nova | 08:41 | |
| garyk | danpb: i have a few issue with the image meta which maybe you can help clarify. in the vmware driver there is a disk_type and a disk_adapter. the possible adapter types are: https://github.com/openstack/nova/blob/master/nova/virt/vmwareapi/constants.py#L50. In the image meta this is https://github.com/openstack/nova/blob/master/nova/objects/image_meta.py#L288 (which I do not think is correct). wnated to hear what you have to say | 08:45 |
| *** Kevin_Zheng has quit IRC | 08:47 | |
| *** Kevin_Zheng has joined #openstack-nova | 08:48 | |
| *** achanda has joined #openstack-nova | 08:49 | |
| *** lpetrut has joined #openstack-nova | 08:49 | |
| *** eglynn has joined #openstack-nova | 08:50 | |
| *** otter768 has joined #openstack-nova | 08:51 | |
| *** prashant-up has joined #openstack-nova | 08:51 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: introduce method to attach device https://review.openstack.org/188763 | 08:53 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: introduce method detach_device to Guest object https://review.openstack.org/188765 | 08:53 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: remove _get_disk_xml to use get_disk from Guest https://review.openstack.org/188764 | 08:53 |
| *** claudiub_ has quit IRC | 08:54 | |
| *** lpetrut has quit IRC | 08:54 | |
| *** leopoldj has joined #openstack-nova | 08:54 | |
| *** achanda has quit IRC | 08:56 | |
| *** otter768 has quit IRC | 08:56 | |
| *** dulek has quit IRC | 08:56 | |
| *** achanda has joined #openstack-nova | 08:57 | |
| *** danpb has left #openstack-nova | 08:58 | |
| *** danpb has joined #openstack-nova | 08:58 | |
| *** dulek has joined #openstack-nova | 08:59 | |
| *** ajayaa has quit IRC | 09:00 | |
| *** ociuhandu has quit IRC | 09:01 | |
| *** achanda has quit IRC | 09:03 | |
| *** prashant-up has quit IRC | 09:05 | |
| openstackgerrit | garyk proposed openstack/nova: VMware: convert driver to use nova.objects.ImageMeta https://review.openstack.org/189085 | 09:05 |
| openstackgerrit | John Garbutt proposed openstack/nova-specs: Add liberty priorities https://review.openstack.org/187272 | 09:05 |
| garyk | danpb: not sure if you saw the message above. but i think that for the vmware driver we should be using 'hw_disk_bus' instead of ''hw_scsi_model' | 09:08 |
| *** lpetrut has joined #openstack-nova | 09:09 | |
| *** varya has joined #openstack-nova | 09:12 | |
| johnthetubaguy | superdan: there is an interesting spec if you get a moment, it relates to some of your evacuate work I think: https://review.openstack.org/#/c/183033/1/specs/liberty/approved/optional-shared-storage-flag-in-evacuate-api.rst,cm | 09:12 |
| danpb | garyk: in which part ? | 09:12 |
| danpb | oh, you mean as a replacement for vmware_adaptertype ? | 09:13 |
| garyk | danpb: yes, correct | 09:13 |
| garyk | i started a pacth on https://review.openstack.org/189086 but wanted to chat with you first. | 09:13 |
| openstackgerrit | Andrey Pavlov proposed openstack/nova: template for disabe_terminate and shutdown_terminate flags https://review.openstack.org/189112 | 09:14 |
| *** prashup has joined #openstack-nova | 09:15 | |
| openstackgerrit | Kevin Benton proposed openstack/nova: Set default event status to completed https://review.openstack.org/189176 | 09:15 |
| danpb | garyk: no, that doesn't really work - i'll reply on the review rather than here | 09:16 |
| *** ajayaa has joined #openstack-nova | 09:16 | |
| *** haypo has joined #openstack-nova | 09:18 | |
| garyk | danok, thanks. | 09:19 |
| *** Kevin_Zheng has quit IRC | 09:21 | |
| *** Kevin_Zheng has joined #openstack-nova | 09:21 | |
| *** Nisha has joined #openstack-nova | 09:23 | |
| *** dulek has quit IRC | 09:23 | |
| kashyap | sahid: Not sure if you've noticed this, seems like you're already working on this area ("libvirt-set-admin-password -- https://bugs.launchpad.net/nova/+bug/1462239 | 09:23 |
| openstack | Launchpad bug 1462239 in OpenStack Compute (nova) "libvirt: thrown NotImplementedError when running "nova root-password"" [Undecided,New] - Assigned to javeme (javaloveme) | 09:23 |
| sahid | kashyap: nice i did not notived this | 09:24 |
| kashyap | sahid: Yeah, he might be duplicating the work, so I added that note with a pointer to your in-progress work. | 09:24 |
| sahid | kashyap: thank yyou i also have a spec in progress | 09:25 |
| sahid | bw if you want to review it https://review.openstack.org/#/c/187509/ | 09:25 |
| sahid | :) | 09:25 |
| kashyap | Thanks, will read. I briefly tested it w/ QMP command and w/ virsh's `qemu-agent-command`. | 09:26 |
| *** dulek has joined #openstack-nova | 09:28 | |
| *** lpetrut1 has joined #openstack-nova | 09:29 | |
| *** dulek has quit IRC | 09:29 | |
| *** markvoelker has joined #openstack-nova | 09:29 | |
| *** dulek has joined #openstack-nova | 09:30 | |
| *** lpetrut has quit IRC | 09:31 | |
| *** achanda has joined #openstack-nova | 09:31 | |
| openstackgerrit | Matthew Booth proposed openstack/nova: Implement compare-and-swap for instance update https://review.openstack.org/141115 | 09:32 |
| *** yingjun has quit IRC | 09:33 | |
| *** alexpilotti has joined #openstack-nova | 09:33 | |
| *** achanda has quit IRC | 09:33 | |
| *** markvoelker has quit IRC | 09:34 | |
| *** achanda has joined #openstack-nova | 09:34 | |
| *** claudiub has joined #openstack-nova | 09:35 | |
| *** gszasz has joined #openstack-nova | 09:35 | |
| claudiub | johnthetubaguy: hello | 09:37 |
| johnthetubaguy | claudiub: hi | 09:38 |
| claudiub | johnthetubaguy: so, I've read the nova-liberty-summit-action-items you've created on etherpad, and there are a couple of things | 09:39 |
| sahid | garyk: i rplied on your comments https://review.openstack.org/#/c/188763/ - not sure if that will clearly respond of them, please let me know | 09:40 |
| claudiub | johnthetubaguy: 1st of all, the hyper-v cluster spec is here: https://review.openstack.org/188374 | 09:40 |
| *** haigang has joined #openstack-nova | 09:40 | |
| claudiub | johnthetubaguy: secondly, will the instance live-resize bp still be a thing? | 09:41 |
| garyk | sahid: thanks! i have changed my score. i was just not clear | 09:41 |
| claudiub | johnthetubaguy: because I saw it as an item, adding a devref for that | 09:41 |
| johnthetubaguy | claudiub: in its current form, no, we need to work out how it would look across multiple hypervisors | 09:41 |
| claudiub | johnthetubaguy: ps: i will have to update tomorrow the hyper-v cluster spec. | 09:42 |
| johnthetubaguy | claudiub: so I want to refine our scope document to better reflect whats going on with that debate, so we can revisit the concept | 09:42 |
| claudiub | johnthetubaguy: well, if there is anything I can help on that, let me know. | 09:43 |
| openstackgerrit | John Garbutt proposed openstack/nova-specs: Remove disable_terminate https://review.openstack.org/183023 | 09:44 |
| johnthetubaguy | claudiub: thanks, will need a good review when its ready, but right now focusing on spec reviews, given the looming deadlines there | 09:44 |
| *** kbyrne has quit IRC | 09:45 | |
| openstackgerrit | garyk proposed openstack/nova: VMware: convert driver to use nova.objects.ImageMeta https://review.openstack.org/189085 | 09:45 |
| sahid | garyk: cool :) | 09:46 |
| eliqiao | johnthetubaguy: hi john, can you help update spec on https://review.openstack.org/#/c/163274/ https://review.openstack.org/#/c/163275/ ? | 09:46 |
| johnthetubaguy | eliqiao: hello, let me take a look at your comments, and try answer the questions there | 09:47 |
| eliqiao | johnthetubaguy: okay, I believe they are only nics.. | 09:48 |
| *** kbyrne has joined #openstack-nova | 09:48 | |
| johnthetubaguy | eliqiao: for this one: https://review.openstack.org/#/c/163274/9/specs/liberty/approved/validate-service-state-before-deleting.rst,cm | 09:48 |
| johnthetubaguy | eliqiao: what I mean is, requests that come to the API before your new microversion will get the existing behaviour, right? | 09:49 |
| *** claudiub has quit IRC | 09:51 | |
| *** achanda_ has joined #openstack-nova | 09:51 | |
| eliqiao | johnthetubaguy: for API, do you mean compute_api ? then I get you. | 09:51 |
| *** claudiub has joined #openstack-nova | 09:51 | |
| johnthetubaguy | eliqiao: I mean the public REST API | 09:52 |
| johnthetubaguy | eliqiao: your change should only apply to REST requests with a version header that are equal or higher to when you introduced the change | 09:52 |
| eliqiao | johnthetubaguy: ah, I get you ;) I will add it more explicitly | 09:52 |
| johnthetubaguy | eliqiao: cool, thanks | 09:53 |
| eliqiao | johnthetubaguy: you are nice, I think that is a stupid question. thanks again. | 09:53 |
| johnthetubaguy | eliqiao: yeah, its more that its not totally clear from the current wording (actually it sounds a bit the other way around) | 09:53 |
| johnthetubaguy | eliqiao: its a very good question, no problems at all | 09:53 |
| *** pbandzi has quit IRC | 09:53 | |
| *** achanda has quit IRC | 09:53 | |
| *** pbandzi has joined #openstack-nova | 09:55 | |
| *** yassine_ has quit IRC | 09:55 | |
| *** prashup has quit IRC | 09:56 | |
| mnestratov|3 | danpb: Hi | 09:58 |
| mnestratov|3 | I have a question about HVSpec object versioning. If I change the number of reported hv_types, should I bump up HVSpec version? | 09:58 |
| *** mnestratov|3 is now known as mnestratov | 09:58 | |
| *** L33 is now known as lyarwood | 09:59 | |
| *** sdake has quit IRC | 10:00 | |
| *** varya_ has joined #openstack-nova | 10:02 | |
| *** varya has quit IRC | 10:03 | |
| *** boris-42 has joined #openstack-nova | 10:04 | |
| *** claudiub has quit IRC | 10:06 | |
| *** tteggel has joined #openstack-nova | 10:08 | |
| *** openstackgerrit has quit IRC | 10:09 | |
| *** dims_ has joined #openstack-nova | 10:09 | |
| *** openstackgerrit has joined #openstack-nova | 10:09 | |
| johnthetubaguy | danpb: ndipanov: if either of you have a moment, I think this spec would benefit from you having a look at it: https://review.openstack.org/#/c/183057/3/specs/liberty/approved/nova-api-hypervsor-cpu-info.rst,cm | 10:09 |
| ndipanov | johnthetubaguy, hmmm, I thought that was settled :) I guess not | 10:10 |
| openstackgerrit | Alvaro Lopez Garcia proposed openstack/nova: scheduler: weight nodes based on their cached images https://review.openstack.org/173304 | 10:10 |
| openstackgerrit | Alvaro Lopez Garcia proposed openstack/nova: Add new image cache monitor in compute nodes https://review.openstack.org/173303 | 10:10 |
| openstackgerrit | Alvaro Lopez Garcia proposed openstack/nova: vmware: add image_cache_manager property https://review.openstack.org/187602 | 10:10 |
| openstackgerrit | Alvaro Lopez Garcia proposed openstack/nova: compute monitors: set update delta from base class https://review.openstack.org/173987 | 10:10 |
| johnthetubaguy | ndipanov: did we write down what we agreed somewhere? | 10:11 |
| ndipanov | oh no this is something else | 10:11 |
| ndipanov | so | 10:11 |
| ndipanov | I was under the impression that we want to make this an object | 10:12 |
| ndipanov | actually there is a comment in the code for that | 10:12 |
| *** tteggel has quit IRC | 10:12 | |
| ndipanov | not really related to exposing it in the API | 10:12 |
| openstackgerrit | Rui Chen proposed openstack/nova-specs: CreateVM supports subnet specified https://review.openstack.org/169638 | 10:12 |
| ndipanov | how does this work with microversions johnthetubaguy ? | 10:13 |
| ndipanov | if it's business as usual - add a microversion and that's it - why not just call this a bugfix? | 10:13 |
| *** tteggel has joined #openstack-nova | 10:13 | |
| johnthetubaguy | ndipanov: all API changes need a spec, for the moment | 10:14 |
| ndipanov | I see | 10:14 |
| ndipanov | that is actually probably a good idea | 10:14 |
| johnthetubaguy | ndipanov: when we didn't require that, we kept screwing up the API in ways that we can't fix | 10:14 |
| ndipanov | yep completely agree | 10:14 |
| ndipanov | well | 10:15 |
| johnthetubaguy | I am hoping we will find some simple patterns, eventually, where its just a bug fix | 10:15 |
| johnthetubaguy | but right now, we haven't seen those emerge, I guess | 10:15 |
| ndipanov | so the thing is - it would be really good to objectify this as part of this changer | 10:15 |
| *** Nisha has quit IRC | 10:15 | |
| johnthetubaguy | 500-> existing error code is likely to be one | 10:15 |
| johnthetubaguy | ndipanov: for the above change, I think it just gets added in a specific microversion, which removes the old method at the same time, which works OK (although none of that is stated in the spec, I added a comment about that). | 10:16 |
| johnthetubaguy | ndipanov: good point about the object in the code though, ideally its close to the API we come up with | 10:17 |
| ndipanov | johnthetubaguy, https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L4998 | 10:17 |
| johnthetubaguy | I mean, there are good reasons they might be totally different, but it would be nice not to miss stuff | 10:17 |
| johnthetubaguy | the reason why to covert it, is so it means something in the API, and isn't driver specific, and we can share scheduler logic, etc, etc | 10:18 |
| johnthetubaguy | but anyways | 10:18 |
| ndipanov | and then make sure that there is only one way to convert that object to a dict that is used in the API | 10:18 |
| ndipanov | yeah so making it an object would mean it's not driver specific anymore | 10:18 |
| johnthetubaguy | ndipanov: +1 and I like that a lot | 10:19 |
| ndipanov | (well unless people stick more json fields on the cpu_info object) :D | 10:19 |
| johnthetubaguy | ah, true | 10:19 |
| *** claudiub has joined #openstack-nova | 10:19 | |
| ndipanov | but let's hope we catch taht in review | 10:19 |
| ndipanov | I can comment | 10:19 |
| ndipanov | I feel bad telling ppl | 10:19 |
| ndipanov | you need to do this other work we should have done before you can do your stuff | 10:19 |
| ndipanov | but in this case it's kind of a cleanup anyway | 10:19 |
| ndipanov | so.. | 10:20 |
| ndipanov | might as well do it right | 10:20 |
| johnthetubaguy | ndipanov: yeah, its hard, we keep doing that a lot right now, because we have so much technical debt | 10:20 |
| *** park_heijlong has joined #openstack-nova | 10:20 | |
| johnthetubaguy | ndipanov: this one is interesting though, I am tempted to let it land with all the things as optional, so drivers slowly push their data out the API in the "official" format | 10:20 |
| johnthetubaguy | ... its not clean, but its lets "queued" behind lots of work | 10:21 |
| johnthetubaguy | duno | 10:21 |
| *** prashup has joined #openstack-nova | 10:21 | |
| danpb | johnthetubaguy: ndipanov we already have an object for modelling CPU model info | 10:21 |
| ndipanov | orly | 10:21 |
| danpb | we didn't convert the host resources reported data to use the object yet for the sake of back compatibility | 10:21 |
| johnthetubaguy | ndipanov: so I am hoping to write down the work that people are queued behind, so they can offer to un-block themselves, even if its some random backlog spec | 10:22 |
| *** kaisers1 has quit IRC | 10:22 | |
| danpb | as & when we object-ify the host resources data itself, then I'd expect the cpu info data to be objectified at the same time | 10:22 |
| danpb | so we can deal with back-compat support in one go | 10:22 |
| ndipanov | danpb, where is the object for this | 10:23 |
| ndipanov | but fwiw there is a patch up for objectifying RT | 10:23 |
| ndipanov | so at that point everything should be using the compute node object | 10:23 |
| *** kaisers has quit IRC | 10:23 | |
| danpb | nova/objects/vcpu_model.py | 10:23 |
| ndipanov | johnthetubaguy, ^^ | 10:23 |
| danpb | it was designed so that we use the same object model for representing host and guest CPU model config | 10:23 |
| *** kaisers has joined #openstack-nova | 10:24 | |
| johnthetubaguy | not sure thats quite the same as the host CPU model though, but if its good for both, thats cool | 10:24 |
| ndipanov | so in that case - I think this is pretty clear then | 10:24 |
| ndipanov | was just about to say | 10:24 |
| *** kaisers1 has joined #openstack-nova | 10:24 | |
| johnthetubaguy | ndipanov: danpb: if you can comment on the spec to add these details, that would be awesome | 10:24 |
| danpb | libvirt internally uses the same data structure to represent host & guest CPU models | 10:25 |
| danpb | and this was designed to follow the same structure, so I don't forsee any problems | 10:25 |
| *** oomichi has quit IRC | 10:25 | |
| danpb | just a couple of the attributes will not be set in the host model (eg the "match" and "policy" attributes only make sense for the guest) | 10:26 |
| openstackgerrit | Mh Raies proposed openstack/python-novaclient: Adding missing nova read only CLI test https://review.openstack.org/171148 | 10:26 |
| *** RuiChen has quit IRC | 10:26 | |
| *** RuiChen has joined #openstack-nova | 10:27 | |
| *** flwang1 has joined #openstack-nova | 10:27 | |
| *** claudiub has quit IRC | 10:27 | |
| *** achanda_ has quit IRC | 10:28 | |
| *** arnaud____ has joined #openstack-nova | 10:29 | |
| johnthetubaguy | danpb: we should work out how to document some of these concepts, eventually, but we have bigger issues in those docs first (like removing the stuff thats just incorrect!) | 10:30 |
| ndipanov | johnthetubaguy, so a question | 10:31 |
| johnthetubaguy | ndipanov: sure | 10:31 |
| ndipanov | do we keep python novaclient only on 'latest' or do we support all microversions? | 10:31 |
| kashyap | sahid: Added two small comments. Hope that's reasonable. | 10:31 |
| johnthetubaguy | ndipanov: not sure we have properly answered that, I think we want it supporting all the latest features via the CLI | 10:32 |
| johnthetubaguy | ndipanov: I think the python lib side is still a bit TBC | 10:32 |
| johnthetubaguy | ndipanov: so we need all the features to work with servers that support older versions, via the CLI, thats a hard requirement, as far as I see it | 10:33 |
| johnthetubaguy | ndipanov: I suspect that will define what the python API looks like | 10:33 |
| *** arnaud____ has quit IRC | 10:33 | |
| ndipanov | ok so in this case we can't just remove support we have to keep both in the client | 10:33 |
| johnthetubaguy | ndipanov: and in the API, it probably stays longer in the client than the API | 10:34 |
| *** pradipta has quit IRC | 10:34 | |
| johnthetubaguy | ndipanov: its not "free", but it gives us more ability to move forward while staying interoperable (ish) | 10:34 |
| *** lpetrut1 has quit IRC | 10:35 | |
| johnthetubaguy | ndipanov: honestly, I think we will only start to see the cost of microversions towards the end of M | 10:35 |
| *** achanda has joined #openstack-nova | 10:35 | |
| *** dulek_ has joined #openstack-nova | 10:37 | |
| *** dulek has quit IRC | 10:37 | |
| *** park_heijlong has quit IRC | 10:37 | |
| johnthetubaguy | RuiChen: hi, just wondering about a few little updates on the subnet_id spec | 10:38 |
| *** dulek___ has joined #openstack-nova | 10:39 | |
| *** dulek_ has quit IRC | 10:39 | |
| *** dulek___ has quit IRC | 10:40 | |
| *** dulek has joined #openstack-nova | 10:41 | |
| *** deepthi has quit IRC | 10:41 | |
| ndipanov | johnthetubaguy, I like how sdague rationalized what we do in his blog where he says we are not a service we are an oss codebase and we decided we want to be interoperable | 10:42 |
| *** achanda has quit IRC | 10:42 | |
| johnthetubaguy | lascii: do you have your flavor_class stuff up for review yet? https://review.openstack.org/#/c/182345/3/specs/backlog/approved/flavor-resize-restriction.rst,cm | 10:42 |
| ndipanov | so then it will have a cost of course | 10:42 |
| johnthetubaguy | ndipanov: +1 I am really pushing we are better at explaining the *why* behind what we decided, so we remember why we are taking that cost! | 10:43 |
| johnthetubaguy | and so we don't end up doing something thats costly, but doesn't actually achieve the original goal | 10:43 |
| *** achanda has joined #openstack-nova | 10:43 | |
| *** dulek has quit IRC | 10:43 | |
| johnthetubaguy | (for the process stuff, and the code!) | 10:43 |
| ndipanov | so I was thinking about this a bit over the weekend and I think this really aligns with several other themes like cells v2 | 10:44 |
| ndipanov | where we actually kinda have 2 nova dbs that decouple things even further etc | 10:44 |
| *** achanda has quit IRC | 10:45 | |
| *** ajayaa has quit IRC | 10:46 | |
| ndipanov | so definitely makes sense for where we are going... but it's a long road :) | 10:46 |
| *** achanda has joined #openstack-nova | 10:46 | |
| johnthetubaguy | ndipanov: +1 | 10:46 |
| mnestratov | Good day all | 10:47 |
| johnthetubaguy | ndipanov: I am wanting to get that clearer in our docs, we have quite a strong vision, but mostly we don't communicate it well, even to each other | 10:47 |
| mnestratov | I have a question | 10:47 |
| *** coolsvap is now known as coolsvap|afk | 10:47 | |
| ndipanov | johnthetubaguy, also a lot of this stuff is a set of features that only very serious deployers will care about, which I think has been voiced as a concern a lot lately... | 10:49 |
| ndipanov | that we are moving away from something "the little shop" can get their head around | 10:49 |
| ndipanov | maybe not | 10:49 |
| ndipanov | that's what I was thinking anyway | 10:49 |
| mnestratov | A question about HVSpec object versioning. If I change the number of reported hv_types, should I bump up HVSpec object version or just update its hash? | 10:50 |
| johnthetubaguy | ndipanov: thats a valid point, it needs to be easy for the little folks too | 10:52 |
| *** achanda has quit IRC | 10:52 | |
| *** otter768 has joined #openstack-nova | 10:52 | |
| johnthetubaguy | ndipanov: but we mostly don't describe anything to anyone, which I think is the issue | 10:52 |
| ndipanov | right :) | 10:52 |
| johnthetubaguy | ndipanov: little folks will want to add servers, cells v2 makes that easier in many ways, but we don't tell anyone that, because that would be... useful? | 10:53 |
| johnthetubaguy | mnestratov: sounds like you might want to bump the version, because there is a downgrade issue that needs fixing somehow | 10:53 |
| johnthetubaguy | mnestratov: but I am not the expert in that area, hoping they will be able to answer your question | 10:54 |
| johnthetubaguy | mnestratov: AFAIK we added the hash change to make it more obvious when the version likely needs an update | 10:54 |
| ndipanov | yes I see what you are saying - having a high level overview of the mission will definitely help folks get involved properly even as users alone | 10:55 |
| mnestratov | johnthetubaguy: ok thank you. I think since don't change interface I don't need to change the version either. | 10:56 |
| *** otter768 has quit IRC | 10:57 | |
| johnthetubaguy | ndipanov: so I am going to try take on that task this release, mostly as I hope thats easier to do along side the other PTL stuff compared to coding, but lets see, maybe the opposite is true! | 10:57 |
| johnthetubaguy | mnestratov: so you change the type of a field, basically, so I suspect it might need a bump, but I am sorry, I just don't know the details there | 10:58 |
| *** afazekas_mtg is now known as afazekas | 10:58 | |
| openstackgerrit | garyk proposed openstack/nova: VMware: enable a resize of instance with no root disk https://review.openstack.org/173802 | 10:58 |
| *** ajayaa has joined #openstack-nova | 10:58 | |
| *** ociuhandu has joined #openstack-nova | 10:59 | |
| ndipanov | well I think coding is important for your own sanity :) but yeah having a clear mission and scope will help us when we need to decide which of the features companies/users want to see in Nova actually make sense and also prioritize them | 10:59 |
| mnestratov | johnthetubaguy: anyway thanks | 10:59 |
| johnthetubaguy | ndipanov: +1 to all of that | 11:02 |
| johnthetubaguy | thinking of sanity... its time for lunch I think | 11:02 |
| openstackgerrit | Yusuke Ide proposed openstack/python-novaclient: Add help message for secgroup-add/del-default-rule https://review.openstack.org/189263 | 11:05 |
| *** wuhg has joined #openstack-nova | 11:08 | |
| *** deepthi has joined #openstack-nova | 11:09 | |
| varya_ | Hi all. I have a question. as part of creating a new tenant can we perform some custom actions in nova or neutron. For example creating a default network/subnet for every tenant as soon as the tenant is created? Sorry if this is not the right forum to ask this question. | 11:10 |
| *** varya_ has quit IRC | 11:12 | |
| *** ferest has joined #openstack-nova | 11:14 | |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: WIP: docs: Add bug handling description to developer reference https://review.openstack.org/187571 | 11:15 |
| *** ajayaa has quit IRC | 11:18 | |
| *** ferest has quit IRC | 11:19 | |
| *** fesp has joined #openstack-nova | 11:20 | |
| sdague | ndipanov: glad you liked the post :) | 11:22 |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova-specs: Remove disable_terminate https://review.openstack.org/183023 | 11:22 |
| mkoderer | sdague: ping | 11:22 |
| sdague | mkoderer: pong | 11:22 |
| *** rushiagr_away is now known as rushiagr | 11:22 | |
| mkoderer | sdague: I am bit confused about https://review.openstack.org/#/c/188687/1 | 11:23 |
| mkoderer | sdague: since I was involed in this... was the actual nova bug fixed already? | 11:23 |
| sdague | so, it's being unskipped because cirros supports it? Yeh, honestly i'm not sure | 11:26 |
| sdague | though it's more of a -qa question | 11:26 |
| *** tobe has quit IRC | 11:26 | |
| sdague | perhaps dkranz has some background on it | 11:26 |
| mkoderer | sdague: yeah your right I just realized that it's not really releated to nova at all | 11:27 |
| Kevin_Zheng | Hi, could anyone kindly approve this tiny fix? already alot +1s :-) https://review.openstack.org/#/c/184662/ | 11:28 |
| *** markvoelker has joined #openstack-nova | 11:30 | |
| *** yassine_ has joined #openstack-nova | 11:34 | |
| *** markvoelker has quit IRC | 11:35 | |
| *** claudiub has joined #openstack-nova | 11:35 | |
| *** rpodolyaka has joined #openstack-nova | 11:36 | |
| *** ajayaa has joined #openstack-nova | 11:37 | |
| *** tlbr has joined #openstack-nova | 11:37 | |
| *** haigang has quit IRC | 11:41 | |
| *** haigang has joined #openstack-nova | 11:42 | |
| *** lpetrut has joined #openstack-nova | 11:43 | |
| *** Jeffrey4l has quit IRC | 11:44 | |
| *** lucasagomes is now known as lucas-hungry | 11:44 | |
| *** deepthi has quit IRC | 11:45 | |
| *** claudiub has quit IRC | 11:45 | |
| *** alex_klimov has joined #openstack-nova | 11:46 | |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: libvirt: make default_device_names DRY-er https://review.openstack.org/180636 | 11:49 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: virt: add get_device_name_for_instance to the base driver class https://review.openstack.org/180637 | 11:49 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: libvirt: implement get_device_name_for_instance https://review.openstack.org/180638 | 11:49 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: virt: Move building the block_device_info dict into a method https://review.openstack.org/180635 | 11:49 |
| *** tojuvone has quit IRC | 11:56 | |
| *** exploreshaifali has quit IRC | 11:58 | |
| openstackgerrit | Andrey Kurilin proposed openstack/nova-specs: Nova API Microversions support in NovaClient https://review.openstack.org/188410 | 11:59 |
| *** Jeffrey4l has joined #openstack-nova | 12:00 | |
| openstackgerrit | Timofey Durakov proposed openstack/nova: Split instance resize operation into 2 phases https://review.openstack.org/184495 | 12:04 |
| *** prashup has quit IRC | 12:06 | |
| *** ajayaa has quit IRC | 12:07 | |
| *** zhenguo has quit IRC | 12:07 | |
| *** tdurakov has joined #openstack-nova | 12:09 | |
| *** annegentle has joined #openstack-nova | 12:09 | |
| *** raildo has joined #openstack-nova | 12:11 | |
| *** ferest has joined #openstack-nova | 12:15 | |
| *** haigang has quit IRC | 12:16 | |
| *** haigang has joined #openstack-nova | 12:16 | |
| openstackgerrit | Maxim Nestratov proposed openstack/nova: libvirt: rename parallels driver to virtuozzo https://review.openstack.org/184311 | 12:16 |
| *** zz_jgrimm is now known as jgrimm | 12:17 | |
| *** haigang has quit IRC | 12:18 | |
| *** artom has joined #openstack-nova | 12:18 | |
| *** aysyanne has joined #openstack-nova | 12:19 | |
| *** ferest has quit IRC | 12:20 | |
| *** sudipto has quit IRC | 12:24 | |
| openstackgerrit | Andrey Kurilin proposed openstack/nova-specs: Raname header X-OpenStack-Nova-API-Version https://review.openstack.org/189278 | 12:24 |
| *** fesp is now known as flaper87 | 12:25 | |
| *** flaper87 has quit IRC | 12:25 | |
| *** flaper87 has joined #openstack-nova | 12:25 | |
| *** ivasev has joined #openstack-nova | 12:28 | |
| *** nihilifer_ has quit IRC | 12:28 | |
| openstackgerrit | garyk proposed openstack/nova: Objects: update missing adapter types https://review.openstack.org/189086 | 12:29 |
| *** pece has joined #openstack-nova | 12:29 | |
| openstackgerrit | Rodolfo Alonso proposed openstack/nova: Added new scheduler filter: AggregateTypeExtraSpecsAffinityFilter https://review.openstack.org/189279 | 12:30 |
| garyk | danpb: ^ | 12:30 |
| *** rfolco has joined #openstack-nova | 12:31 | |
| *** VW has joined #openstack-nova | 12:35 | |
| *** lpetrut has quit IRC | 12:37 | |
| *** lucas-hungry is now known as lucasagomes | 12:38 | |
| sahid | kashyap: thanks for your review, i replied to one of your point | 12:38 |
| sahid | please let me know if that makes sense for you | 12:38 |
| sahid | https://review.openstack.org/#/c/187509/ | 12:38 |
| kashyap | sahid: Yep, it does. | 12:40 |
| kashyap | I hesitated briefly whether to write that point (a) or not. Thanks for correcting. | 12:40 |
| sahid | kashyap: np at all thank for reviewing | 12:40 |
| *** lpetrut has joined #openstack-nova | 12:41 | |
| *** dulek has joined #openstack-nova | 12:42 | |
| *** marun has joined #openstack-nova | 12:44 | |
| *** ociuhandu has quit IRC | 12:49 | |
| *** thiagop has quit IRC | 12:49 | |
| *** raildo has quit IRC | 12:50 | |
| *** gabriel-bezerra has quit IRC | 12:50 | |
| *** tellesnobrega has quit IRC | 12:50 | |
| *** rushiagr is now known as rushiagr_away | 12:50 | |
| *** flwang1 has quit IRC | 12:51 | |
| *** ericksonsantos has quit IRC | 12:51 | |
| *** achanda has joined #openstack-nova | 12:52 | |
| *** ociuhandu has joined #openstack-nova | 12:52 | |
| *** dave-mccowan has joined #openstack-nova | 12:53 | |
| *** otter768 has joined #openstack-nova | 12:53 | |
| *** baoli has joined #openstack-nova | 12:54 | |
| *** pilgrimstack has joined #openstack-nova | 12:55 | |
| mnestratov | sahid, danpb, vladik: could you please, as you are libvirt subteam members, consider code corresponding to the following blueprints as "ready for review"? | 12:56 |
| mnestratov | https://blueprints.launchpad.net/nova/+spec/virtuozzo-instance-resize-support | 12:56 |
| mnestratov | https://blueprints.launchpad.net/nova/+spec/ploop-snapshot-support | 12:56 |
| mnestratov | https://blueprints.launchpad.net/nova/+spec/virtuozzo-container-boot-from-volume | 12:56 |
| *** otter768 has quit IRC | 12:57 | |
| *** yassine__ has joined #openstack-nova | 12:58 | |
| *** Maike_ has joined #openstack-nova | 12:58 | |
| *** yassine_ has quit IRC | 13:01 | |
| openstackgerrit | Eugeniya Kudryashova proposed openstack/python-novaclient: Raise exception when API does't support microversions https://review.openstack.org/188816 | 13:01 |
| sahid | mnestratov: it seems any of these bleuprints has been approved ? | 13:01 |
| *** Maike has quit IRC | 13:01 | |
| sahid | did you get opportunity to talk about them during nova meeting or mailing liest? | 13:02 |
| *** irenab has quit IRC | 13:04 | |
| mnestratov | sahid: only during summit design session | 13:04 |
| sahid | from my point of view the first one do not need spec | 13:05 |
| *** ericksonsantos has joined #openstack-nova | 13:06 | |
| *** tellesnobrega has joined #openstack-nova | 13:06 | |
| *** raildo has joined #openstack-nova | 13:06 | |
| sahid | perhaps you could try to talk about it during meeting to make it approved | 13:06 |
| *** dane-fichter has joined #openstack-nova | 13:06 | |
| sahid | then that will help core to review it | 13:06 |
| mnestratov | sahid: why specs? the spec for parallels/vz driver was approved for the previous release cycle | 13:07 |
| *** heyongli has quit IRC | 13:07 | |
| sahid | mnestratov: hum... i wanted to say the bp does not need spec yes | 13:07 |
| mnestratov | these changes are internal for libvirt/parallels driver- they are expanding feature coverage we lack for the moment | 13:08 |
| openstackgerrit | garyk proposed openstack/nova: Objects: update missing adapter types https://review.openstack.org/189086 | 13:08 |
| *** heyongli has joined #openstack-nova | 13:08 | |
| sahid | from my point of view, but others people on the community can have questions | 13:08 |
| *** ferest has joined #openstack-nova | 13:08 | |
| *** gabriel-bezerra has joined #openstack-nova | 13:08 | |
| sahid | anyway you should to make them approved if you want to expect coresto reviews it | 13:08 |
| *** VW has quit IRC | 13:09 | |
| mnestratov | ok | 13:09 |
| *** dprince has joined #openstack-nova | 13:09 | |
| mnestratov | to make them approved should I add them to the next nova meeting agenda? | 13:09 |
| sahid | mnestratov: yes that should be to way to follow | 13:10 |
| sahid | johnthetubaguy: ^ can you confirm me? | 13:10 |
| *** ferest_ has joined #openstack-nova | 13:11 | |
| *** ferest has quit IRC | 13:13 | |
| *** prashup has joined #openstack-nova | 13:13 | |
| mnestratov | sahid: can you please take a look at https://review.openstack.org/#/c/186858/ - it is approved | 13:14 |
| mnestratov | sahid: sorry - thois one https://review.openstack.org/#/c/149253/ | 13:14 |
| sahid | mnestratov: sure i will do, just would like to finish something before. | 13:15 |
| mnestratov | sahid: thank you | 13:16 |
| *** markvoelker has joined #openstack-nova | 13:20 | |
| johnthetubaguy | sahid: mnestratov: yeah put spec-less BPs on the meeting agenda where it mentioned about approving spec-less blueprints | 13:20 |
| *** achanda has quit IRC | 13:20 | |
| *** ferest_ has quit IRC | 13:20 | |
| *** achanda has joined #openstack-nova | 13:21 | |
| johnthetubaguy | mnestratov: honestly, that looks a lot like a bug fix... rather than a blueprint, did you want a blueprint for some reason? | 13:21 |
| openstackgerrit | Andrey Kurilin proposed openstack/nova-specs: Nova API Microversions support in NovaClient https://review.openstack.org/188410 | 13:21 |
| johnthetubaguy | mnestratov: ah, so I see what its a blueprint now, my bad | 13:22 |
| lxsli | edleafe: up yet? | 13:22 |
| *** varya has joined #openstack-nova | 13:23 | |
| *** markvoelker has quit IRC | 13:25 | |
| *** abhishekk has quit IRC | 13:26 | |
| *** achanda has quit IRC | 13:26 | |
| *** superdan is now known as dansmith | 13:26 | |
| *** thangp has joined #openstack-nova | 13:28 | |
| *** Maike has joined #openstack-nova | 13:28 | |
| *** rook has quit IRC | 13:29 | |
| moshele | johnthetubaguy: hi | 13:30 |
| *** Maike_ has quit IRC | 13:30 | |
| *** afazekas has quit IRC | 13:31 | |
| openstackgerrit | Baodong (Robert) Li proposed openstack/nova-specs: Support live migration with macvtap SR-IOV https://review.openstack.org/136077 | 13:31 |
| *** eharney has quit IRC | 13:32 | |
| *** thangp has quit IRC | 13:32 | |
| *** wverdugo500 has joined #openstack-nova | 13:32 | |
| *** wverdugo500 has left #openstack-nova | 13:32 | |
| *** thorst has joined #openstack-nova | 13:33 | |
| *** rook has joined #openstack-nova | 13:35 | |
| *** dboik has quit IRC | 13:36 | |
| mnestratov | johnthetubaguy: added to the next meeting agenda. thanks | 13:36 |
| *** thangp has joined #openstack-nova | 13:37 | |
| johnthetubaguy | moshele: hi, I am around, fire away | 13:37 |
| edleafe | lxsli: yep | 13:37 |
| *** marun has quit IRC | 13:37 | |
| lxsli | edleafe: goood morning | 13:37 |
| *** mriedem has joined #openstack-nova | 13:38 | |
| lxsli | edleafe: has anything been written up about the (eventual) plans to create a scheduler library? | 13:38 |
| *** vladikr has joined #openstack-nova | 13:38 | |
| *** igordcard_ has joined #openstack-nova | 13:38 | |
| edleafe | lxsli: not that I'm aware of | 13:38 |
| moshele | johnthetubaguy: I add a comment to https://review.openstack.org/#/c/169638/ regarding the general direction on the nova api | 13:38 |
| edleafe | I do remember talking about it in Paris when the idea of jaypipes Resource Objects was being discussed | 13:39 |
| lxsli | edleafe: OK, I'll put it near the end of my todo list. It would be nice if a newcomer to Nova could get a reasonable sense of what's going on by reading specs rather than spending a year talking to people | 13:39 |
| edleafe | lxsli: the problem with a separate scheduler service was what you were talking about on Friday | 13:40 |
| lxsli | edleafe: yes I remember being very confused while that discussion was going on! | 13:40 |
| edleafe | lxsli: namely how to pass the class info and logic between the scheduled service and the scheduler | 13:40 |
| *** mtanino has quit IRC | 13:40 | |
| johnthetubaguy | moshele: vnic_type is quite different, subnet is something that got missed out of the existing API, vnic_type is a new concept | 13:40 |
| lxsli | edleafe: yep, there is another answer but I'm not sure people will ike it | 13:41 |
| edleafe | lxsli: yes, it's unfortunate that side discussions broke out | 13:41 |
| lxsli | *like | 13:41 |
| * edleafe is waiting to hear it... | 13:41 | |
| *** artom has quit IRC | 13:41 | |
| johnthetubaguy | moshele: if vnic_type mapped to some "type" owned by neutron, that looks just like the Cinder volume_type, I would consider including it, although its possible the other driver will disagree with me on that | 13:41 |
| lxsli | edleafe: we could provide a resource provision and matching language, adequate to support all reasonable requirements | 13:41 |
| *** mpaolino has joined #openstack-nova | 13:42 | |
| *** mpaolino has quit IRC | 13:42 | |
| lxsli | edleafe: that way, services can take care of translating their requests into that language and the scheduler is independent | 13:42 |
| *** amotoki has joined #openstack-nova | 13:42 | |
| lxsli | I see no real problem with the scheduler-lib solution though (maybe I just need to think harder) | 13:43 |
| edleafe | lxsli: there are two words in that sentence that jump out at me: 'adequate' and 'reasonable' | 13:43 |
| lxsli | :D | 13:43 |
| moshele | johnthetubaguy: vnic_type tells nova if it need to allocate pci device so it affect nova code flow for launch a vm | 13:43 |
| edleafe | lxsli: you know that we'll be forever finding new requirements that no one thought was reasonable at first | 13:44 |
| *** dpaterson has joined #openstack-nova | 13:44 | |
| *** burt has joined #openstack-nova | 13:44 | |
| lxsli | I'm using those in a fairly academic sense, "adequate" -> "fully supporting" and "reasonable" -> "not more than NP-hard" | 13:44 |
| edleafe | lxsli: ah, ok | 13:44 |
| *** htruta has joined #openstack-nova | 13:44 | |
| johnthetubaguy | moshele: its way too implementation specific for a concept in the end user facing Nova API, thats my big issue with it | 13:45 |
| edleafe | lxsli: now it's just the 'fully' part that is glaring | 13:45 |
| *** xyang1 has joined #openstack-nova | 13:46 | |
| johnthetubaguy | moshele: so the route cause of the issue here is that I need to write up the devref documentation on our architectural direction, and we should actually remove the API we are considering removing, rather than leaving them in limbo | 13:46 |
| lxsli | edleafe: I think you can do an awful lot with relations like "M to N relations between the sets A and B which meet these constraints" for example | 13:46 |
| *** mpaolino has joined #openstack-nova | 13:47 | |
| lxsli | EG NUMA | 13:47 |
| edleafe | lxsli: I think you would need at least a POC to serve as an example | 13:47 |
| *** amotoki has quit IRC | 13:47 | |
| edleafe | lxsli: so that others can see what a resource would look like | 13:47 |
| moshele | johnthetubaguy: I have also different proposal like using https://github.com/openstack/python-openstackclient api to build all the required elements in neutron/cinder and then using nova api just to attach specific element like port | 13:47 |
| johnthetubaguy | moshele: I plan to get this all straightened out by the end of Liberty, but honestly, thats probably a little optimistic, I think these API direction things are most likely to get pushed to the next summit, at least to some extent | 13:47 |
| edleafe | lxsli: as well as a request and claim | 13:48 |
| dansmith | dims_: how do you feel about my abbreviating o.vo like this? https://review.openstack.org/#/c/187719/3//COMMIT_MSG,cm | 13:48 |
| dansmith | dims_: I'm pretty consistent about it and thus I think it's fine (and helps with wrapping and such).. do you have an opinion/ | 13:48 |
| *** sdake has joined #openstack-nova | 13:48 | |
| edleafe | dansmith: In general, I'd spell it out once, and abbreviate everywhere else | 13:48 |
| johnthetubaguy | moshele: we will need REST API support for this kind of thing | 13:49 |
| dansmith | edleafe: once per patch you mean? I probably only ever say it once per patch anyway, on average | 13:49 |
| johnthetubaguy | moshele: I am still tempted to -2 the other spec, but I like how it simplifies the API significantly, in a way our users probably want | 13:49 |
| dansmith | edleafe: I feel like I've got a pretty good amount of context going on "o.vo" but maybe that's all in my head | 13:49 |
| lxsli | edleafe: yes I played with this a little while ago http://potassco.sourceforge.net/ | 13:50 |
| dims_ | dansmith: replied on the review. am good | 13:50 |
| edleafe | dansmith: yes, for exactly that reason: it's clear to you, but won't be to someone who might not be living and breathing that stuff | 13:50 |
| dansmith | dims_: ah, cool, thanks | 13:50 |
| dims_ | edleafe: if it were a spec, sure. this is ok | 13:50 |
| lxsli | dansmith: if this is a straw poll, +1 to spelling it out, sorry | 13:51 |
| lxsli | :ab o.vo oslo_versionedobjects | 13:51 |
| dansmith | it's not about the typing really | 13:51 |
| *** Maike has quit IRC | 13:51 | |
| lxsli | the first time I saw one of your reviews with o.vo I had to come here and ask | 13:52 |
| dims_ | lxsli: we've been using o.vo for a while now as a short hand for the project, so i really don't mind | 13:52 |
| lxsli | OK just giving a not-core perspective :) | 13:52 |
| dansmith | lxsli: what if it's defined in the blueprint/spec? (it might actually be, I'd have to check) | 13:52 |
| dansmith | lxsli: what does core have to do with it? | 13:52 |
| * dims_ tries to find my core hat and does not find any :) | 13:53 | |
| lxsli | dansmith: well, it's not impossible to work out but if the typing isn't an issue, just do it? | 13:53 |
| lxsli | dansmith: just a proxy for participation level | 13:53 |
| moshele | johnthetubaguy: every additional attribute will simplifies the api, but I think the question here is what are the guideline for adding new attribute for the api | 13:53 |
| *** atuvenie has joined #openstack-nova | 13:53 | |
| dansmith | lxsli: the reason I like the abbreviation is that it wraps better in comments and commit messages and is easier to read in general I think | 13:53 |
| *** amotoki has joined #openstack-nova | 13:54 | |
| dims_ | dansmith: let's add it to the o.vo README :) https://github.com/openstack/oslo.versionedobjects/ | 13:54 |
| lxsli | surely if you're reading the o.vo README, you can guess what it means? | 13:55 |
| dansmith | dims_: that way it should be google-able | 13:55 |
| dims_ | dansmith: +1 | 13:55 |
| lxsli | Showing results for ovo.com | 13:55 |
| dims_ | lxsli: it was a joke :) | 13:55 |
| lxsli | >.< sorry | 13:56 |
| dims_ | :) | 13:56 |
| *** lascii is now known as alaski | 13:56 | |
| lxsli | dansmith: anyway I'm not telling you what to do, just saying it confused me briefly one time ^^ | 13:56 |
| dansmith | lxsli: what if there was at least one grep'able reference in the nova tree? | 13:57 |
| *** dboik has joined #openstack-nova | 13:57 | |
| dansmith | like somewhere I said "#NOTE(danms): Inherit from oslo.versionedobjects (o.vo) for the base class" | 13:58 |
| lxsli | that helps I guess | 13:58 |
| dansmith | nova/objects/base.py: # Temporary until we inherit from o.vo.base.VersionedObject | 13:58 |
| dansmith | nova/objects/base.py: # NOTE(danms): This is nova-specific, so don't copy this to o.vo | 13:58 |
| dansmith | those are already there, FWIW | 13:58 |
| *** rushiagr_away is now known as rushiagr | 13:58 | |
| dansmith | I don't want to do something that increases obscurity, I just much prefer the abbreviation, both in conversation and text | 13:58 |
| *** mpaolino has quit IRC | 13:59 | |
| dansmith | johnthetubaguy: alaski: thoughts on using the abbreviation? | 13:59 |
| dims_ | dansmith: +1 to add it to a few places to help with discovery | 13:59 |
| alaski | dansmith: is there a patch in question? | 14:00 |
| dansmith | alaski: there are several, | 14:00 |
| *** sdake_ has joined #openstack-nova | 14:00 | |
| dansmith | alaski: in general, just the question of using o.vo in place of oslo.versionedobjects in things like commit messages | 14:00 |
| *** VW has joined #openstack-nova | 14:00 | |
| dansmith | (and comments) | 14:00 |
| lxsli | you could build the bikeshed as "oslo.vo"? | 14:01 |
| *** jaypipes has joined #openstack-nova | 14:01 | |
| dansmith | lxsli: I could, although we've been using o.vo for quite some time now, so another shorthand makes it worse I think | 14:01 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: convert driver to use nova.objects.ImageMeta https://review.openstack.org/188119 | 14:01 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: make instance compulsory in blockinfo APIs https://review.openstack.org/188118 | 14:01 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: objects: add os_require_quiesce image property https://review.openstack.org/188787 | 14:01 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: virt: convert hardware module to use nova.objects.ImageMeta https://review.openstack.org/188789 | 14:01 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: compute: only use non_inheritable_image_properties if snapshotting https://review.openstack.org/187250 | 14:01 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: objects: fix parsing of NUMA cpu/mem properties https://review.openstack.org/188788 | 14:01 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: compute: remove get_image_metadata method https://review.openstack.org/187251 | 14:01 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: xenapi: convert driver to use nova.objects.ImageMeta https://review.openstack.org/188790 | 14:01 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: hyperv: convert driver to use nova.objects.ImageMeta https://review.openstack.org/189302 | 14:01 |
| lxsli | OK, I'm going to let you build it how you want now :) | 14:01 |
| johnthetubaguy | dansmith: code and commits are write once read many, which leads me towards avoiding those abbreviations, for clarity | 14:02 |
| dansmith | alrighty | 14:02 |
| alaski | dansmith: in comments I think I would prefer a full name, but it's not a hard preference from me | 14:02 |
| alaski | in commit messages I would say use once then abbreviate | 14:02 |
| dansmith | dims_ and I will just use it as our secret handshake then | 14:02 |
| dims_ | :) | 14:03 |
| *** eharney has joined #openstack-nova | 14:03 | |
| *** sdake has quit IRC | 14:04 | |
| alaski | I think the context is important. If you have a comment above a line like "something = ovo.Foo" then using o.vo should be fine | 14:04 |
| dansmith | yeah, the code isn't the question, since it's an import and searchable I think | 14:04 |
| alaski | yeah, I more meant the comment above it | 14:05 |
| dansmith | ah | 14:05 |
| *** BadCub-Hiding is now known as BadCub | 14:05 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: introduce method to get domain XML https://review.openstack.org/189307 | 14:05 |
| BobBall | In the object-world is it possible for a nova period task to pull an object from the DB before it's fully populated, and then that object is in memory and might cause an issue later on? https://bugs.launchpad.net/nova/+bug/1463039 is a potential race we're hitting in the XS CI. Just wondering if my first thought (that the periodic task happens to kick in before the object is fully populated) might be the issue | 14:06 |
| openstack | Launchpad bug 1463039 in OpenStack Compute (nova) "BandwidthUsage object has no attribute 'instance_uuid'" [Undecided,New] | 14:06 |
| dims_ | haypo: welcome back! Victor, managed to get around 30% of unit tests working in py3 with 5 reviews so far (https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:python3-porting,n,z) | 14:06 |
| dansmith | BobBall: no more so that before when we were pulling dicts, I think | 14:08 |
| haypo | dims_: oh, i didn't know that you was interested to port nova to python 3 | 14:08 |
| dims_ | haypo: i had some time last week :) | 14:08 |
| BobBall | Shame... Was worth a try! :) | 14:08 |
| mdbooth | jaypipes: Instance compare-and-swap is good to go: https://review.openstack.org/#/c/141115/ | 14:08 |
| mdbooth | It would be good to knock that on the head | 14:09 |
| mdbooth | It had a +1 from zzzeek until I made a trivial update, btw | 14:09 |
| haypo | dims_: 30% is a lot, good job | 14:09 |
| openstackgerrit | Petrut Lucian proposed openstack/nova: Hyper-V: Fix missing WMI namespace issue on Windows 2008 R2 https://review.openstack.org/189308 | 14:09 |
| haypo | dims_: i was in holiday during 5 days, and now i'm trying to read all emails... i didn't have time to check my pending patches | 14:10 |
| dims_ | haypo: i'll let you catch up :) no worries | 14:10 |
| openstackgerrit | Dan Smith proposed openstack/nova: Fix an objects layering violation in compute/api https://review.openstack.org/188110 | 14:10 |
| openstackgerrit | Dan Smith proposed openstack/nova: Replace metaclass registry with explicit opt-in registry from oslo https://review.openstack.org/188111 | 14:10 |
| openstackgerrit | Dan Smith proposed openstack/nova: Remove assertRemotes() from objects tests https://review.openstack.org/188109 | 14:10 |
| openstackgerrit | Dan Smith proposed openstack/nova: Convert test objects to new field formats https://review.openstack.org/187719 | 14:10 |
| openstackgerrit | Dan Smith proposed openstack/nova: Begin the transition to an explicit object registry https://review.openstack.org/187718 | 14:10 |
| openstackgerrit | Dan Smith proposed openstack/nova: Use fields from oslo.versionedobjects https://review.openstack.org/187720 | 14:10 |
| openstackgerrit | Dan Smith proposed openstack/nova: Make readonly field tests use exception from o.vo https://review.openstack.org/188112 | 14:10 |
| dansmith | doh | 14:10 |
| haypo | dims_: i saw that the py34 gate of nova was broken because of a setup.py file of a project expected the locale encoding was UTF-8, the issue was worked around in tox.ini | 14:11 |
| openstackgerrit | Dan Smith proposed openstack/nova: Make readonly field tests use exception from oslo.versionedobjects https://review.openstack.org/188112 | 14:11 |
| haypo | dims_: if the gate is stable, I would like to make it voting, that's the next step of the spec | 14:11 |
| dims_ | haypo: y and then one more thing broke and was fixed | 14:11 |
| *** logan2 has quit IRC | 14:12 | |
| *** achanda has joined #openstack-nova | 14:12 | |
| openstackgerrit | Petrut Lucian proposed openstack/nova: Hyper-V: Fix missing WMI namespace issue on Windows 2008 R2 https://review.openstack.org/189308 | 14:12 |
| *** logan2 has joined #openstack-nova | 14:13 | |
| dims_ | haypo: for the record it was a jsonschema release - https://github.com/Julian/jsonschema/issues/228 | 14:14 |
| haypo | dims_: no :) i was rfc3986, i proposed a real fix and it was merged a few minutes later! https://github.com/sigmavirus24/rfc3986/pull/14 | 14:15 |
| *** artom has joined #openstack-nova | 14:15 | |
| haypo | dims_: oh, this morning i saw a failure with functools32 yes | 14:15 |
| dims_ | haypo: y that was the first one, and on friday we hit the jsonschema with bad wheel | 14:15 |
| dims_ | biab | 14:16 |
| haypo | dims_: so what do you think of making the gate voting? | 14:16 |
| *** READ10 has joined #openstack-nova | 14:16 | |
| *** jecarey has joined #openstack-nova | 14:17 | |
| *** achanda has quit IRC | 14:18 | |
| *** pbandzi has quit IRC | 14:18 | |
| *** varya has quit IRC | 14:20 | |
| dansmith | johnthetubaguy: threw some comments on that evacuate spec.. I don't think there is any overlap with mine, but I pointed out another big issue I think | 14:21 |
| *** agireud has joined #openstack-nova | 14:21 | |
| *** IanGovett has joined #openstack-nova | 14:23 | |
| *** prashant__ has joined #openstack-nova | 14:23 | |
| *** prashup has quit IRC | 14:24 | |
| *** _gryf has joined #openstack-nova | 14:29 | |
| *** annegentle has quit IRC | 14:29 | |
| dansmith | ndipanov: did you seem my question here? https://review.openstack.org/#/c/183354/ | 14:29 |
| _gryf | dansmith, hi | 14:30 |
| dansmith | ndipanov: I think that the filter/comment in RT addresses the concern enough to proceed, but if not, please let me know what else you think needs doing | 14:30 |
| dansmith | _gryf: hi | 14:30 |
| openstackgerrit | Jean-Daniel Bonnetot proposed openstack/nova-specs: Workflow to manage unauthorised resizing https://review.openstack.org/182345 | 14:30 |
| ndipanov | dansmith, looking | 14:30 |
| _gryf | dansmith, is this a problem, the implementation will stall for two weeks? | 14:31 |
| _gryf | dansmith, I'm on the vacation right now | 14:31 |
| *** unicell has joined #openstack-nova | 14:31 | |
| *** unicell1 has quit IRC | 14:31 | |
| dansmith | _gryf: no, it's fine.. I need to get back to look at your patches again, I've just had lots of other things to look at | 14:31 |
| *** dguitarbite has quit IRC | 14:32 | |
| _gryf | dansmith, rihgt, nevertheless i didn't push the implementation using microversion yet, since i have some issues with the tests | 14:32 |
| dansmith | _gryf: okay | 14:32 |
| *** Maike has joined #openstack-nova | 14:32 | |
| openstackgerrit | Rodolfo Alonso proposed openstack/nova: Added new scheduler filter: AggregateTypeExtraSpecsAffinityFilter https://review.openstack.org/189279 | 14:33 |
| mriedem | jaypipes: you were +2 on this on friday, want to take a look again now that i'm +2? https://review.openstack.org/#/c/177084/ | 14:33 |
| ndipanov | dansmith, I remember this - well filtering them out is fine, as it will make it exactly as bad as it is now... | 14:34 |
| jaypipes | mriedem: done. thank you sir :) | 14:35 |
| openstackgerrit | Jean-Daniel Bonnetot proposed openstack/nova-specs: Workflow to manage unauthorised resizing https://review.openstack.org/182345 | 14:35 |
| *** electrocucaracha has joined #openstack-nova | 14:35 | |
| ndipanov | I kinda feel that this kind of stuff ties in very tightly with the convo we started on the RequestSpec bp | 14:35 |
| dansmith | ndipanov: right.. "just as sucky" was what I was going for | 14:35 |
| *** afazekas has joined #openstack-nova | 14:36 | |
| dansmith | ndipanov: in that the only goal of that patch is to make the movement of the guest visible through the api.. RT enhancements are a different thing | 14:36 |
| ndipanov | so I feel really bad saying that this is A-OK, when I gave bauzas a hard time ... but I will remove the -1 as the direct concern was addressed | 14:36 |
| *** tjones1 has joined #openstack-nova | 14:37 | |
| dansmith | ndipanov: okay, not sure I get what the concern is since it doesn't break what is there and the goal is orthogonal to any request or RT stuff | 14:37 |
| ndipanov | hence removing -1 | 14:37 |
| dansmith | I mean.. whatever lingering guilt you may have, but that's fine :) | 14:38 |
| *** jamielennox is now known as jamielennox|away | 14:38 | |
| *** jgrimm is now known as zz_jgrimm | 14:38 | |
| *** lpetrut has quit IRC | 14:38 | |
| *** lpetrut has joined #openstack-nova | 14:39 | |
| *** Maike has quit IRC | 14:39 | |
| sahid | markus_z: hello, please before to unassign me from a bug don't hesistate to ping me on IRC i'm usually always connected | 14:40 |
| sahid | the bug needs to be addressed prior to this spec https://review.openstack.org/#/c/165838/ | 14:40 |
| sahid | i'm currently working on it | 14:40 |
| *** cbader has joined #openstack-nova | 14:41 | |
| *** armax has joined #openstack-nova | 14:42 | |
| lxsli | sbauza.beers += 1 | 14:42 |
| ndipanov | dansmith, https://review.openstack.org/#/c/180637/ - I also replied to you there so check it out when yolu get a chance | 14:43 |
| *** yingjun has joined #openstack-nova | 14:43 | |
| markus_z | sahid: OK, sorry, I try to go through some of the bugs and look what's up. I didn't want to offend you in any way with this. | 14:43 |
| dansmith | lxsli: careful, cfriesen actually paid me a beer in vancouver | 14:43 |
| dansmith | lxsli: so people might start looking to collect on those.. | 14:43 |
| markus_z | sahid: Good to hear that you are working on this :) | 14:43 |
| lxsli | dansmith: yeah but the beer was free right? | 14:43 |
| sahid | markus_z: np at all really | 14:43 |
| dansmith | lxsli: it was free to me, but the one he brought was externally-sourced and funded | 14:44 |
| *** cfriesen has joined #openstack-nova | 14:44 | |
| lxsli | dansmith: it'll never catch on | 14:44 |
| dansmith | heh | 14:44 |
| dansmith | ndipanov: hmm, so boot doesn't call that reserve method like attach does? | 14:45 |
| jaypipes | I promised a case of beer to the Neutron dev who implemented "get me a network" API. | 14:45 |
| ndipanov | dansmith, nope | 14:45 |
| ndipanov | how can it - it has to ask the scheduler first | 14:45 |
| ndipanov | where do I put this | 14:45 |
| dansmith | ah, right | 14:45 |
| dansmith | ndipanov: well, tbh, the confusion I see the most is from attach, because on boot nothing is used, so you can't easily have a conflict unless you double up, right? | 14:46 |
| dansmith | ndipanov: so it seems like improving the experience for attach is still worthwhile, | 14:46 |
| dansmith | as people tend to attach and detach /dev/sdb over and over and hit the problem | 14:46 |
| *** pilgrimstack has quit IRC | 14:47 | |
| ndipanov | dansmith, yeah it's quite tricky thought - this will very much be visible from the API to a lot of people | 14:49 |
| danpb | arrrgh, having a file nova/tests/unit/virt/.#test_configdrive.py saved by emacs breaks python unit test enumeration ! | 14:50 |
| dansmith | yeah, I'm sure it will, but it's not good now.. | 14:50 |
| ndipanov | if we choose to reject requests with a device name period. | 14:50 |
| * danpb just wasted half an hour figuring that out | 14:50 | |
| dansmith | ndipanov: right now, they're using the API wrong and we're lying to them | 14:51 |
| jaypipes | :waq | 14:51 |
| jaypipes | guh... | 14:51 |
| *** annegentle has joined #openstack-nova | 14:52 | |
| dansmith | ndipanov: so making it break explicitly or just go ignored both seem like better options than what we have | 14:52 |
| *** tjones1 has quit IRC | 14:52 | |
| ndipanov | dansmith, so completely ignoring it and deciding all of them from scratch is what you would rather see | 14:53 |
| ndipanov | I agree actually | 14:53 |
| dansmith | ndipanov: well, if we're on libvirt, I think so yeah | 14:54 |
| ndipanov | I actually have a BP I will propose that will tie into this | 14:54 |
| edleafe | jaypipes: it's Monday | 14:54 |
| *** otter768 has joined #openstack-nova | 14:54 | |
| ndipanov | OK that's one more easy patch in libvirt | 14:54 |
| *** yassine_ has joined #openstack-nova | 14:54 | |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: WIP: docs: Add bug handling description to developer reference https://review.openstack.org/187571 | 14:54 |
| ndipanov | plus I will sit down and document all of that junk this week | 14:54 |
| *** sahid has quit IRC | 14:54 | |
| ndipanov | cool | 14:54 |
| dansmith | ndipanov: just re-reading what you have, | 14:55 |
| *** sahid has joined #openstack-nova | 14:55 | |
| dansmith | ndipanov: your libvirt implementation ignores suggested_device_name, right? | 14:55 |
| ndipanov | latest patch does not | 14:55 |
| ndipanov | the one you commented on does | 14:55 |
| dansmith | ah, that's what I'm looking at | 14:55 |
| ndipanov | but boot does not as well which is what needs to be fixed too | 14:55 |
| dansmith | right | 14:55 |
| ndipanov | but that is real easy | 14:56 |
| ndipanov | ok will do that now | 14:56 |
| jaypipes | edleafe, bauzas, ndipanov, lxsli: the resource object models patch series is back on track: https://review.openstack.org/#/c/128992 | 14:56 |
| dansmith | ndipanov: cool, sounds like a plan, thanks | 14:56 |
| *** achanda has joined #openstack-nova | 14:56 | |
| edleafe | jaypipes: \o/ | 14:56 |
| edleafe | jaypipes: I'll go over them a little later | 14:56 |
| jaypipes | edleafe: yeah, worked last night to get it back on track. nicely, the work from sahid on cleaning up the NUMA stuff made the patch more readable. | 14:57 |
| *** yassine__ has quit IRC | 14:57 | |
| openstackgerrit | Daniel Berrange proposed openstack/nova: objects: convert config drive to use ImageMeta object https://review.openstack.org/189323 | 14:58 |
| ndipanov | jaypipes, awesome - looking forward to that one | 14:58 |
| *** otter768 has quit IRC | 14:58 | |
| *** achanda has quit IRC | 15:02 | |
| openstackgerrit | Andrea Rosa proposed openstack/nova: WIP: adding a volume detach_force call https://review.openstack.org/184537 | 15:02 |
| *** zzzeek has joined #openstack-nova | 15:02 | |
| *** tonytan4ever has joined #openstack-nova | 15:04 | |
| *** IanGovett has quit IRC | 15:06 | |
| *** belmoreira has quit IRC | 15:06 | |
| *** jcoufal has quit IRC | 15:08 | |
| *** baoli has quit IRC | 15:08 | |
| *** markvoelker has joined #openstack-nova | 15:08 | |
| *** baoli_ has joined #openstack-nova | 15:10 | |
| *** andrearosa has quit IRC | 15:12 | |
| *** markvoelker has quit IRC | 15:13 | |
| BobBall | How/what does the translation from 'uuid' to 'instance_uuid' in BandwidthUsage? BandwidthUsage.create takes a uuid which seems to be set directly in the DB but _from_db_object seems to expect an instance_uuid field. It does seem to work sometimes though (I don't always hit https://bugs.launchpad.net/nova/+bug/1463039) | 15:16 |
| openstack | Launchpad bug 1463039 in OpenStack Compute (nova) "BandwidthUsage object has no attribute 'instance_uuid'" [Undecided,New] | 15:16 |
| *** dprince has quit IRC | 15:17 | |
| *** andrearosa has joined #openstack-nova | 15:19 | |
| openstackgerrit | Alexandre Levine proposed openstack/nova-specs: Added a spec for disable-termination and shutdown-termination flags https://review.openstack.org/189334 | 15:20 |
| *** irenab has joined #openstack-nova | 15:21 | |
| openstackgerrit | Andrey Pavlov proposed openstack/nova: template for disabe_terminate and shutdown_terminate flags https://review.openstack.org/189112 | 15:21 |
| *** cbader has quit IRC | 15:21 | |
| *** cbader has joined #openstack-nova | 15:21 | |
| alaski | BobBall: afaik it doesn't happen. I have code to do it in a review of mine https://review.openstack.org/#/c/172524/3/nova/objects/bandwidth_usage.py,cm | 15:23 |
| *** Mike_D_laptop has joined #openstack-nova | 15:26 | |
| openstackgerrit | Andrey Pavlov proposed openstack/nova: template for disabe_terminate and shutdown_terminate flags https://review.openstack.org/189112 | 15:27 |
| openstackgerrit | Alexandre Levine proposed openstack/nova-specs: Added a spec for disable-termination and shutdown-termination flags https://review.openstack.org/189334 | 15:28 |
| BobBall | alaski: Yay | 15:28 |
| BobBall | alaski: That might be the cause of my bug then :) | 15:28 |
| * BobBall looks a little closer | 15:28 | |
| *** Mike_D_laptop has quit IRC | 15:28 | |
| *** markus_z has quit IRC | 15:29 | |
| *** ijuwang has quit IRC | 15:30 | |
| alaski | BobBall: looks like it. I'll refresh my patch today | 15:30 |
| *** Mike_D_laptop has joined #openstack-nova | 15:30 | |
| *** tjones1 has joined #openstack-nova | 15:31 | |
| *** leopoldj has quit IRC | 15:31 | |
| BobBall | Great - thanks! | 15:32 |
| BobBall | Let me know if you want me to pull out the fix into a different review or something | 15:32 |
| *** electrocucaracha has quit IRC | 15:34 | |
| *** raies has joined #openstack-nova | 15:34 | |
| raies | lxsli: hi | 15:35 |
| *** electrocucaracha has joined #openstack-nova | 15:35 | |
| *** electrocucaracha has quit IRC | 15:35 | |
| *** Jeffrey4l has quit IRC | 15:36 | |
| *** aix has quit IRC | 15:36 | |
| lxsli | raies: hiya | 15:36 |
| raies | lxsli: I want to discuss regarding patch https://review.openstack.org/#/c/139025/ | 15:37 |
| *** agireud has quit IRC | 15:37 | |
| raies | lxsli: lets discuss it point to point | 15:37 |
| lxsli | thought so :) | 15:37 |
| *** mtanino has joined #openstack-nova | 15:38 | |
| *** Sukhdev has joined #openstack-nova | 15:39 | |
| lxsli | Want to start? | 15:40 |
| raies | lxsli: https://review.openstack.org/#/c/139025/29/nova/api/openstack/compute/plugins/v3/servers.py #L374 | 15:40 |
| lxsli | B64_REGEX? | 15:41 |
| raies | lxsli: what way I should make it sorted | 15:41 |
| lxsli | oh, wrong file | 15:41 |
| raies | lxsli: I didn't get B64_REGEX | 15:41 |
| *** unicell has quit IRC | 15:41 | |
| lxsli | my mistake, I see it now | 15:41 |
| *** sdake_ has quit IRC | 15:42 | |
| lxsli | so garyk has a point, OpenStack style is to prefer line continuation with () rather than \ | 15:42 |
| lxsli | or if you're talking about my comment on L382, it should be possible to not nest four ifs together | 15:42 |
| raies | lxsli: okk thats good, I will amend it in next patch | 15:43 |
| raies | yes I am talking about yours | 15:43 |
| lxsli | I refactored your code in patch 28, I hoped you'd fix that rather than completely reverting | 15:43 |
| *** tsekiyama has joined #openstack-nova | 15:44 | |
| raies | lxsli: ok I will check it once again, actually I thought you do not have comment on my portion of code | 15:44 |
| *** Nisha has joined #openstack-nova | 15:44 | |
| raies | lxsli: but its ok I will check and amend it | 15:44 |
| raies | lxsli: one more thing in same file | 15:45 |
| lxsli | I had a lot of comments on PS#26 and thought that rather than tear your patch apart, it would be more constructive to show how I thought it could be done better | 15:45 |
| raies | whether I should remove L#331 onward comments | 15:45 |
| raies | ?? | 15:45 |
| lxsli | mm | 15:45 |
| johnthetubaguy | lxsli: did we write that down somewhere, and preferring () over \? I thought in Nova we were saying something like, if the machine doesn't check it, we don't have an opinion. (i am excluding discussions on names, etc, and stuff thats not machine checkable) | 15:45 |
| *** agireud has joined #openstack-nova | 15:46 | |
| garyk | lxsli: my preference is the () | 15:46 |
| lxsli | johnthetubaguy: I think it may have been checked for a bit but maybe it isn't now? | 15:46 |
| *** matrohon has quit IRC | 15:46 | |
| danpb | i believe we've got a policy of using () based on the number of -1's i've had for using \ | 15:47 |
| lxsli | garyk: honestly I'm fine with either, but our style seems more like () and I do like consistency | 15:47 |
| lxsli | raies: I think the TODO still needs to be done | 15:47 |
| *** sdake has joined #openstack-nova | 15:47 | |
| danpb | johnthetubaguy: in general, I always push for an addition to the nova/hacking/checks.py file for anything code style related | 15:47 |
| garyk | my two cents is that whever possible we should use () and if there is no way around then to use \ | 15:47 |
| johnthetubaguy | right local consistency wins here, just doesn't seem worth delaying a patch for, unless we get the machine to check for that, at least I thoughts that what our policy was here | 15:47 |
| *** pece has quit IRC | 15:48 | |
| johnthetubaguy | garyk: can we add a machine check for that? | 15:48 |
| lxsli | raies: you could do it, as you're bumping the microversion anyway now, and remove the TODO - it's trivial | 15:48 |
| danpb | but sometimes we have to be pragmatic because writing a machine check for somethings is a total PITA | 15:48 |
| garyk | johnthetubaguy: i do not think so. there are so many lines of the db code that has the \ | 15:48 |
| *** electrocucaracha has joined #openstack-nova | 15:48 | |
| *** sdake_ has joined #openstack-nova | 15:48 | |
| lxsli | raies: you can drop lines 335-344 I think but I'd keep 346-348 | 15:49 |
| *** afazekas has quit IRC | 15:49 | |
| lxsli | raies: I'll put some extra comments | 15:50 |
| raies | lxsli: okk | 15:50 |
| *** dane-fichter has quit IRC | 15:50 | |
| raies | lxsli: So according to you, your lines L#346 covers these ? | 15:50 |
| lxsli | raies: I think the comment at line 346 is still useful | 15:51 |
| raies | lxsli: let me check | 15:51 |
| johnthetubaguy | garyk: danpb: hmm, personally I think we should stop -1 ing for it, or machine check it somehow, as its going to just waste time for folks submitting code and reviewer time. I kinda thought we agreed that at some point (wonders if jogo wrote that down somewhere), either way, we should probably right down what we choose | 15:51 |
| johnthetubaguy | anyways, not going to dig into that right now, more important things need documenting | 15:52 |
| *** sdake has quit IRC | 15:52 | |
| *** gszasz has quit IRC | 15:53 | |
| *** yingjun has quit IRC | 15:53 | |
| raies | lxsli: r u talking about commented lines at 346 ? | 15:53 |
| raies | lxsli: I do not see any comment near 346 | 15:53 |
| lxsli | yes | 15:53 |
| raies | which patch you ae talking about | 15:53 |
| raies | okk | 15:54 |
| lxsli | https://review.openstack.org/#/c/139025/29/nova/api/openstack/compute/plugins/v3/servers.py | 15:54 |
| lxsli | # If all tenants is passed with 0 or false as the value | 15:54 |
| raies | okk | 15:54 |
| raies | lxsli: Please confirm, I think TODO is done in our code. what you say ? | 15:55 |
| raies | del search_opts['all_tenants'] | 15:56 |
| *** sdake_ is now known as sdake | 15:56 | |
| mriedem | alexpilotti: looks like hyper-v ci is sad on stable/juno http://64.119.130.115/178446/2/Hyper-V_logs/create-environment-c2-r2-u06.openstack.tld.log.gz | 15:56 |
| raies | I think this removes TODO. what you say I think we should exclude this TODO | 15:56 |
| raies | lxsli: ^^ | 15:57 |
| *** wuhg has quit IRC | 15:57 | |
| alexpilotti | ociuhandu: ^ | 15:57 |
| ociuhandu | mriedem: looking now into this | 15:57 |
| lxsli | raies: just reading carefully | 15:58 |
| *** moshele has quit IRC | 15:58 | |
| *** jlanoux has quit IRC | 15:59 | |
| *** devananda has joined #openstack-nova | 15:59 | |
| *** varya has joined #openstack-nova | 16:01 | |
| *** garyk has quit IRC | 16:01 | |
| lxsli | raies: so the TODO is to automatically set all_tenants if tenant_id is passed | 16:01 |
| lxsli | raies: unless the context project id matches the search tenant_id, because then you don't need it | 16:02 |
| *** tonytan4ever has quit IRC | 16:02 | |
| raies | lxsli: yes it seems to me same | 16:02 |
| *** rfolco has quit IRC | 16:02 | |
| *** tonytan4ever has joined #openstack-nova | 16:02 | |
| lxsli | raies: whereas if I read it right, your code allows anyone to query any tenant without checking they have all_tenants? | 16:02 |
| raies | lxsli: in my code one tenant can't se other's | 16:04 |
| *** yassine_ has quit IRC | 16:04 | |
| raies | for this purpose get_per_tenant policy is implemented in api.py | 16:04 |
| *** romainh has left #openstack-nova | 16:04 | |
| lxsli | ok let me check that | 16:04 |
| *** rushiagr is now known as rushiagr_away | 16:04 | |
| *** BobBall is now known as BobBall_AWOL | 16:04 | |
| *** whenry has quit IRC | 16:05 | |
| *** annegentle has quit IRC | 16:06 | |
| raies | lxsli: https://review.openstack.org/#/c/139025/29/nova/compute/api.py L#1951 | 16:07 |
| *** tonytan4ever has quit IRC | 16:07 | |
| *** IanGovett has joined #openstack-nova | 16:07 | |
| lxsli | we want a user from tenant A to be able to query tenant B if they have permission get_all_tenants; or if they have get_per_tenant for tenant B, is that correct? | 16:07 |
| raies | lxsli: let me clarify | 16:08 |
| *** amotoki has quit IRC | 16:08 | |
| *** amotoki has joined #openstack-nova | 16:08 | |
| raies | only a user in admin tenant can see tenant B | 16:08 |
| raies | 2. a user in Tenant A can't see Tenant B until unless Tenant A is admin | 16:09 |
| raies | 3. nova list ==> gives list in current context | 16:09 |
| *** jistr has quit IRC | 16:10 | |
| raies | 4. nova list --tenant <tenant id> ==> gives all servers for all users in tenant with tenant_id | 16:10 |
| *** sahid has quit IRC | 16:11 | |
| raies | 5. nova list --all-tenants --tenant <tenant_id> gives all servers for all-tenants in current context | 16:11 |
| raies | as all-tenants are entertained first | 16:11 |
| *** patrickeast has joined #openstack-nova | 16:11 | |
| raies | 6. nova list --tenant <tenant-id> --user <user id> ==> gives all servers for user/tenant combination | 16:12 |
| *** apuimedo has quit IRC | 16:12 | |
| lxsli | this would be helpful in the commit message :) | 16:12 |
| raies | 7. nova list --all-users ==> all servers for all users in current context | 16:13 |
| *** tonytan4ever has joined #openstack-nova | 16:13 | |
| raies | lxsli: now any confusion ? | 16:13 |
| lxsli | let me process that | 16:13 |
| lxsli | what's get_per_tenant for then? | 16:13 |
| raies | lxsli: one user in Tenant A should be isolated fron another user in Tenant B until unless it is Admin Tenant | 16:14 |
| *** yamahata has joined #openstack-nova | 16:15 | |
| *** rfolco has joined #openstack-nova | 16:15 | |
| raies | get_per_tenant ===> policy to isolate one tenant from other | 16:15 |
| lxsli | might be clearer as "get_any_user_in_tenant" | 16:16 |
| *** achanda has joined #openstack-nova | 16:16 | |
| raies | lxsli: No it should not be at user level. It should be at teant level | 16:16 |
| raies | as we are only isolation tenants | 16:17 |
| raies | lxsli: users under tenant will automatically be isolated | 16:17 |
| *** hemna_ is now known as hemna | 16:17 | |
| raies | lxsli: You can run above 7 scenarios and get result | 16:17 |
| raies | in a tenant any user can see other user in same tenant | 16:18 |
| raies | but isolation outer boundary should be Tenant only | 16:18 |
| lxsli | so "one user in Tenant A should be isolated fron another user in Tenant B unless it is Admin Tenant" seems to contradict "in a tenant any user can see other user in same tenant" | 16:19 |
| lxsli | Oh wait, I see | 16:19 |
| raies | Tenant A===>(user1, user2, user3) : any user in tenant A can see each other | 16:20 |
| lxsli | So can you phrase what having get_per_tenant lets you do? | 16:20 |
| raies | Tenant B===>(user4, user5, user6) : any user in tenant B can see each other | 16:20 |
| raies | But user 1 can't see user 4 | 16:20 |
| raies | ok I can do this. | 16:21 |
| raies | Actually this is a bit tricky, I had to find hard how to implement above scenarios | 16:22 |
| raies | lxsli: implementation is simple but letting it known to someone is a bit difficult. | 16:22 |
| raies | So, I think https://review.openstack.org/#/c/139025/29/nova/api/openstack/compute/plugins/v3/servers.py is clear | 16:22 |
| raies | ?? | 16:22 |
| raies | again any thing else on https://review.openstack.org/#/c/139025/29/nova/compute/api.py ?? | 16:23 |
| lxsli | I can't see get_per_tenant mentioned at all in v3/servers.py | 16:24 |
| raies | servers.py calls get_all of api.py | 16:24 |
| lxsli | it seems that it allows you to query by tenant_id | 16:24 |
| raies | yes that why I have mentioned it -- You can query like ===> nova list --tenant <tenant-id> | 16:25 |
| lxsli | but, it seems to me that if you have get_all_tenants, you should implicitly have get_per_tenant and vice versa. is that true or can you explain why not? | 16:25 |
| lxsli | in other words, you're admin or not | 16:26 |
| *** nic has joined #openstack-nova | 16:26 | |
| raies | if you have get_all_tenant | 16:26 |
| raies | lxsli: see servers.py | 16:27 |
| raies | you will go at L#365 | 16:27 |
| *** tonytan4ever has quit IRC | 16:27 | |
| raies | lxsli: after that u have only all_tenants in search_opts | 16:27 |
| raies | lxsli: all_tenants is getting deleted in L#386 | 16:28 |
| raies | lxsli: now get_all of api.py | 16:28 |
| raies | lxsli: get_per_tenant will not be executed | 16:29 |
| raies | it means it is going in all_tenants for current context | 16:29 |
| *** unicell has joined #openstack-nova | 16:29 | |
| raies | lxsli: whether it is admin context or non admin context | 16:29 |
| raies | lxsli: are you getting it | 16:29 |
| lxsli | I see that they are both checked | 16:30 |
| raies | lxsli: Thus implementation is correct | 16:30 |
| lxsli | I don't see what benefit having an extra permission provides | 16:30 |
| raies | lxsli: No they both are not checked when passed all-tenants | 16:30 |
| raies | please check again | 16:30 |
| raies | lxsli: if all-tenants in servers.py l#365 | 16:31 |
| lxsli | raies: please listen for a moment | 16:31 |
| lxsli | raies: the implementation is not at issue right now | 16:31 |
| raies | ok | 16:31 |
| lxsli | raies: what benefit does adding get_per_tenant provide? really? | 16:31 |
| *** dprince has joined #openstack-nova | 16:31 | |
| *** lpetrut has quit IRC | 16:31 | |
| lxsli | raies: if you can see all tenants, you should be able to query them individually and vice versa | 16:32 |
| raies | isolation one tenant from other. If one tenant tries to see other Policy does not allow error should come | 16:32 |
| lxsli | users by default are isolated across tenants | 16:33 |
| raies | lxsli: Again get_all_tenant and get_per_tenant are not called simultneously | 16:33 |
| raies | if all-tenant is in search_opts | 16:33 |
| raies | the get_all_tenant will be executed | 16:33 |
| lxsli | the old solution was to require all_tenants if tenant_id is passed; this is a bit clunky from a user POV so the TODO said to implicitly set all_tenants if tenant_id is passed and check that permission | 16:33 |
| raies | and if not passed then else part at line L#379 will be executed | 16:34 |
| *** diogogmt has joined #openstack-nova | 16:34 | |
| diogogmt | any technical reason why this validation is being applied when a volume is detached from an instance? https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/contrib/volumes.py#L412 | 16:34 |
| raies | lxsli: in this case whats happening let me tell you once again | 16:35 |
| raies | lxsli: 1. case1. -all-tenants ==> | 16:35 |
| lxsli | actually lets take a break | 16:35 |
| raies | search_opts will have all_tenants | 16:36 |
| lxsli | I need to go home and I think some space to process would help | 16:36 |
| *** annashen has joined #openstack-nova | 16:37 | |
| *** davideagnello has joined #openstack-nova | 16:37 | |
| raies | lxsli: if we have all-tenants then it mean we should give output all servers from all users in all tenants (in case of admin tenant) or we should give output all servers from alll users in CURRENT tenants (In case of non admin) | 16:38 |
| *** _gryf has quit IRC | 16:38 | |
| raies | thinking this please see | 16:38 |
| kashyap | mriedem: When you get a sec, maybe you're planning to resubmit this by retrying the virt_dom.destroy() call? - https://review.openstack.org/#/c/181781/ | 16:38 |
| raies | all-tenants ==> get_all_tenants permision will executed | 16:39 |
| *** ijw has joined #openstack-nova | 16:39 | |
| mriedem | kashyap: i still have to process danpb's comment | 16:39 |
| lxsli | what I'd recommend right now is putting some of this in the commit message and trying to refactor the code to make it clearer that all the cases are covered | 16:39 |
| mriedem | sitting in my email | 16:39 |
| raies | if not all-tenants ==> per tenant check is required | 16:39 |
| raies | lxsli: hmm good approach | 16:39 |
| kashyap | mriedem: I ask because it's still sitting there in the elastic-recheck page as one of the top bugs. | 16:39 |
| kashyap | mriedem: Thanks. | 16:40 |
| raies | I will mention all above usecases to verify this | 16:40 |
| *** ijw has quit IRC | 16:40 | |
| danpb | mriedem: which comment ? | 16:40 |
| mriedem | danpb: the one i haven't parsed yet | 16:40 |
| kashyap | :-) | 16:40 |
| mriedem | on a call, sorry | 16:40 |
| kashyap | danpb: https://review.openstack.org/#/c/181781/2/nova/virt/libvirt/driver.py | 16:40 |
| *** varya has quit IRC | 16:40 | |
| kashyap | That one | 16:40 |
| lxsli | raies: it's quite confusing having a mutable dict passed between two places, it might help to try to clarify the responsibilities of server.py vs api.py | 16:40 |
| *** ijw has joined #openstack-nova | 16:40 | |
| *** achanda has quit IRC | 16:41 | |
| raies | lxsli: I will try to capture codes from patch#26 but still if there would be any gap I will chat you over that | 16:41 |
| danpb | oh right | 16:41 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: objects: convert config drive to use ImageMeta object https://review.openstack.org/189323 | 16:41 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: hyperv: convert driver to use nova.objects.ImageMeta https://review.openstack.org/189302 | 16:41 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: ironic: convert driver to use nova.objects.ImageMeta https://review.openstack.org/189362 | 16:41 |
| lxsli | raies: sure thing, chat again later | 16:41 |
| kashyap | danpb: I was checking here, because I see people mention it on other IRC channels, and have triaged atleast 6 bugs related to that. | 16:41 |
| raies | lxsli: in which file you are talking about ? | 16:42 |
| raies | mutabledict ? | 16:42 |
| lxsli | mutable => a thing which can be changed | 16:42 |
| lxsli | specifically, the search_opts | 16:42 |
| raies | lxsli: I know that | 16:42 |
| raies | I am asking you where you are pointing it in code | 16:43 |
| lxsli | because it keeps changing, you have to keep track of its current state in your head as you read the code, which is a nightmare | 16:43 |
| *** belmoreira has joined #openstack-nova | 16:44 | |
| raies | ok leave it all, I will upload a next patch which will sort out major problem and confusions | 16:44 |
| *** rushiagr_away is now known as rushiagr | 16:44 | |
| lxsli | raies: ok, I'll have a look when it goes up. Bye for now o/ | 16:44 |
| *** lxsli is now known as lxsli_away | 16:44 | |
| *** achanda has joined #openstack-nova | 16:44 | |
| raies | lxsli: okk thanks bye | 16:44 |
| *** raies has left #openstack-nova | 16:44 | |
| mriedem | kashyap: got a few things on my plate today that take priority, but i should be able to fit that in later this afternoon | 16:44 |
| *** bnemec has quit IRC | 16:45 | |
| kashyap | mriedem: Sure, didn't mean imply to attend to it immediately. | 16:45 |
| mriedem | kashyap: fwiw we hit it in our internal CI also but not in a resize case | 16:45 |
| *** mdrabe has joined #openstack-nova | 16:45 | |
| mriedem | which reminds me, i got libvirt debug logging enabled on friday to track that down internally.... | 16:45 |
| *** bnemec has joined #openstack-nova | 16:46 | |
| mriedem | oh awesome, now that i turned on libvirtd debug logging the problem isn't happening anymore internally :) | 16:46 |
| kashyap | :-) | 16:48 |
| kashyap | mriedem: Yeah, this looks like one of those crazy bugs to find the root cause | 16:48 |
| mriedem | we are running tempest on a smaller flavor, only 4 VCPUs | 16:49 |
| *** garyk has joined #openstack-nova | 16:49 | |
| jaypipes | cfriesen: please take a look at the resource objects implementation here: https://review.openstack.org/#/c/128992 | 16:51 |
| jaypipes | cfriesen, lxsli_away: so a "resource" is neither an amount nor a usage. A "resource" is a class of quantifiable things that may be requested by a user. An "amount" is the abstract quantity of that resource. A "usage" is an abstract representation of capacity and used amounts of the resource. | 16:52 |
| *** exploreshaifali has joined #openstack-nova | 16:53 | |
| jaypipes | cfriesen, lxsli_away: furthermore, you are conflating a resource "provider" -- e.g. a compute node -- with the usage records of individual resources that the provider provides. | 16:53 |
| *** alex_klimov has quit IRC | 16:55 | |
| *** otter768 has joined #openstack-nova | 16:55 | |
| *** markvoelker has joined #openstack-nova | 16:57 | |
| *** otter768 has quit IRC | 16:59 | |
| *** lucasagomes is now known as lucas-dinner | 17:01 | |
| *** derekh has quit IRC | 17:01 | |
| *** markvoelker has quit IRC | 17:02 | |
| *** vilobhmm has joined #openstack-nova | 17:03 | |
| *** pixelbeat_ has quit IRC | 17:03 | |
| *** amotoki has quit IRC | 17:04 | |
| *** atuvenie has quit IRC | 17:04 | |
| *** jodah has joined #openstack-nova | 17:06 | |
| *** annegentle has joined #openstack-nova | 17:06 | |
| *** dims has joined #openstack-nova | 17:07 | |
| *** vipul has joined #openstack-nova | 17:08 | |
| jaypipes | eliqiao: you planning to rebase https://review.openstack.org/#/c/156112/ soon? :) | 17:08 |
| *** lpetrut has joined #openstack-nova | 17:09 | |
| *** dims_ has quit IRC | 17:10 | |
| openstackgerrit | Alvaro Lopez Garcia proposed openstack/nova-specs: WIP: Propose to support spot instances https://review.openstack.org/104883 | 17:10 |
| *** ssurana has joined #openstack-nova | 17:10 | |
| *** bradjones has quit IRC | 17:10 | |
| ericksonsantos | johnthetubaguy, ping about: https://etherpad.openstack.org/p/YVR-nova-liberty-summit-action-items | 17:11 |
| *** bradjones has joined #openstack-nova | 17:11 | |
| *** bradjones has quit IRC | 17:11 | |
| *** bradjones has joined #openstack-nova | 17:11 | |
| *** annegentle has quit IRC | 17:12 | |
| *** agireud has quit IRC | 17:14 | |
| *** erw has left #openstack-nova | 17:16 | |
| *** IanGovett has quit IRC | 17:16 | |
| *** IanGovett has joined #openstack-nova | 17:18 | |
| *** annegentle has joined #openstack-nova | 17:20 | |
| *** annashen has quit IRC | 17:21 | |
| *** lpetrut has left #openstack-nova | 17:21 | |
| *** sudipto has joined #openstack-nova | 17:22 | |
| *** annashen has joined #openstack-nova | 17:22 | |
| *** Marga_ has joined #openstack-nova | 17:23 | |
| *** atuvenie has joined #openstack-nova | 17:26 | |
| *** Marga_ has quit IRC | 17:27 | |
| *** baoli_ has quit IRC | 17:27 | |
| *** Marga_ has joined #openstack-nova | 17:27 | |
| *** baoli has joined #openstack-nova | 17:29 | |
| *** dsanders has joined #openstack-nova | 17:30 | |
| *** eglynn has quit IRC | 17:30 | |
| *** lsmola has quit IRC | 17:32 | |
| *** rushiagr is now known as rushiagr_away | 17:33 | |
| openstackgerrit | Dave Walker proposed openstack/nova: Removed a non-conditional 'if' statement https://review.openstack.org/189377 | 17:34 |
| *** Sukhdev has quit IRC | 17:39 | |
| *** Sukhdev has joined #openstack-nova | 17:40 | |
| mriedem | jaypipes: i have a mind to just rebase that myself | 17:40 |
| jaypipes | mriedem: feel free... I'm eager to review it. | 17:41 |
| jaypipes | sdague: best line from that blog post: "But the moment you start thinking about what the client code would look like to talk to that, you want to throw up a little bit." | 17:46 |
| *** electrocucaracha has quit IRC | 17:47 | |
| *** annashen has quit IRC | 17:48 | |
| *** gzhai has quit IRC | 17:51 | |
| *** aswadr has quit IRC | 17:51 | |
| *** eliqiao has quit IRC | 17:51 | |
| openstackgerrit | Dave Walker proposed openstack/nova: Removed a non-conditional 'if' statement https://review.openstack.org/189377 | 17:51 |
| *** xiaohui has quit IRC | 17:52 | |
| *** xiaohui has joined #openstack-nova | 17:52 | |
| *** zz_jgrimm is now known as jgrimm | 17:53 | |
| *** gzhai has joined #openstack-nova | 17:53 | |
| *** ZZelle_ has joined #openstack-nova | 17:54 | |
| sdague | jaypipes: heh, thanks :) | 17:55 |
| *** annegentle has quit IRC | 17:55 | |
| *** annegentle has joined #openstack-nova | 17:56 | |
| *** marun has joined #openstack-nova | 17:56 | |
| *** eliqiao has joined #openstack-nova | 17:57 | |
| *** ijw_ has joined #openstack-nova | 17:58 | |
| *** achanda_ has joined #openstack-nova | 17:59 | |
| *** dontalton has joined #openstack-nova | 18:00 | |
| *** ijw has quit IRC | 18:01 | |
| *** achanda has quit IRC | 18:02 | |
| mriedem | danpb: what do you think about just calling back into _destroy() again if we hit EBUSY here? https://review.openstack.org/#/c/181781/2/nova/virt/libvirt/driver.py | 18:02 |
| *** sdake has quit IRC | 18:02 | |
| *** sdake has joined #openstack-nova | 18:02 | |
| *** exploreshaifali has quit IRC | 18:04 | |
| danpb | mriedem: we should probably use a loop instead, otherwise the stack could get very deep | 18:04 |
| mriedem | danpb: well i was thinking about killing it at 3 | 18:05 |
| ericksonsantos | Hi! :) | 18:06 |
| ericksonsantos | Does someone know how to use keystoneclient from Nova? | 18:06 |
| ericksonsantos | I want to call keystoneclient.projects.get(...), but I don't know the best way to instantiate keystoneclient. | 18:07 |
| mriedem | danpb: like this http://paste.openstack.org/show/274985/ | 18:07 |
| danpb | mriedem: so that'll result it the instance going into ERROR state on the 4th loop ? | 18:09 |
| *** openstackgerrit has quit IRC | 18:09 | |
| mriedem | it would result in whatever happens today when we hit that | 18:09 |
| *** openstackgerrit has joined #openstack-nova | 18:09 | |
| danpb | heh | 18:09 |
| mriedem | danpb: which is yeah, ERROR state | 18:11 |
| danpb | thats ok then | 18:11 |
| mriedem | ok, i'll write up a test and see how this looks | 18:11 |
| mriedem | and recheck it a few times to see if i can recreate that bug | 18:11 |
| lifeless | cfriesen: hi | 18:15 |
| lifeless | cfriesen: on this ChunkedFile thing | 18:15 |
| lifeless | cfriesen: you're using the filesystem backend for glance yeah ? | 18:16 |
| *** danpb has quit IRC | 18:16 | |
| *** whenry has joined #openstack-nova | 18:16 | |
| lifeless | cfriesen: or rather, let me put it another way - how does nova's process end up with a ChunkedFile ? | 18:16 |
| *** rushiagr_away is now known as rushiagr | 18:18 | |
| *** browne has joined #openstack-nova | 18:19 | |
| *** gary-smith has joined #openstack-nova | 18:24 | |
| *** agireud has joined #openstack-nova | 18:24 | |
| *** pixelbeat_ has joined #openstack-nova | 18:29 | |
| *** IanGovett has quit IRC | 18:30 | |
| *** achanda_ has quit IRC | 18:32 | |
| *** agireud_ has joined #openstack-nova | 18:33 | |
| *** agireud has quit IRC | 18:34 | |
| *** agireud_ is now known as agireud | 18:34 | |
| *** eliqiao has quit IRC | 18:37 | |
| *** Marga_ has quit IRC | 18:40 | |
| *** rushiagr is now known as rushiagr_away | 18:40 | |
| *** Sukhdev has quit IRC | 18:41 | |
| *** achanda has joined #openstack-nova | 18:41 | |
| *** eglynn has joined #openstack-nova | 18:41 | |
| *** eliqiao has joined #openstack-nova | 18:41 | |
| *** achanda has quit IRC | 18:46 | |
| *** markvoelker has joined #openstack-nova | 18:46 | |
| *** Marga_ has joined #openstack-nova | 18:47 | |
| *** electrocucaracha has joined #openstack-nova | 18:48 | |
| *** annashen has joined #openstack-nova | 18:49 | |
| *** agireud has quit IRC | 18:50 | |
| *** markvoelker has quit IRC | 18:52 | |
| *** agireud has joined #openstack-nova | 18:53 | |
| openstackgerrit | Lee Yarwood proposed openstack/nova: Multipath commands with error messages in stdout fail to parse https://review.openstack.org/172660 | 18:54 |
| *** annashen has quit IRC | 18:54 | |
| openstackgerrit | Andrew Laski proposed openstack/nova: Return bandwidth usage after updating https://review.openstack.org/172524 | 18:55 |
| openstackgerrit | Andrew Laski proposed openstack/nova: Map uuid db field to instance_uuid in BandwidthUsage object https://review.openstack.org/189397 | 18:55 |
| *** sudipto has quit IRC | 18:55 | |
| *** otter768 has joined #openstack-nova | 18:56 | |
| vilobhmm | jaypipes : qq | 18:56 |
| *** electrocucaracha has quit IRC | 18:57 | |
| *** baoli has quit IRC | 18:59 | |
| *** otter768 has quit IRC | 19:01 | |
| *** annashen has joined #openstack-nova | 19:01 | |
| *** ijw_ has quit IRC | 19:01 | |
| *** ijw has joined #openstack-nova | 19:02 | |
| *** electrocucaracha has joined #openstack-nova | 19:03 | |
| *** eglynn has quit IRC | 19:06 | |
| *** annegentle has quit IRC | 19:06 | |
| *** annegentle has joined #openstack-nova | 19:06 | |
| *** eglynn has joined #openstack-nova | 19:13 | |
| *** FL1SK_ has joined #openstack-nova | 19:14 | |
| *** FL1SK has quit IRC | 19:15 | |
| *** igordcard has joined #openstack-nova | 19:15 | |
| *** BadCub is now known as BadCub-Hiding | 19:20 | |
| *** annashen has quit IRC | 19:28 | |
| vilobhmm | bauzas, jaypipes : can we say the same thing about computefilter for stable/icehouse "that its never executed"…(i guess no) say https://github.com/openstack/nova/blob/stable/icehouse/nova/scheduler/filter_scheduler.py#L317 -> https://github.com/openstack/nova/blob/stable/icehouse/nova/scheduler/filter_scheduler.py#L358 -> https://github.com/openstack/nova/blob/stable/icehouse/nova/scheduler/host_manager.py#L443 | 19:38 |
| *** tjones1 has quit IRC | 19:38 | |
| *** electrocucaracha has quit IRC | 19:38 | |
| *** electrocucaracha has joined #openstack-nova | 19:38 | |
| *** dprince has quit IRC | 19:39 | |
| vilobhmm | my bad since this executes before any filter https://github.com/openstack/nova/blob/stable/icehouse/nova/scheduler/driver.py#L111-L117 it should be true…can you please confirm that | 19:41 |
| *** ociuhandu has quit IRC | 19:41 | |
| openstackgerrit | Ed Leafe proposed openstack/nova-specs: Add liberty priorities https://review.openstack.org/187272 | 19:43 |
| *** Nisha has quit IRC | 19:44 | |
| *** FL1SK_ is now known as FL1SK | 19:48 | |
| openstackgerrit | Andrey Pavlov proposed openstack/nova: template for disabe_terminate and shutdown_terminate flags https://review.openstack.org/189112 | 19:51 |
| *** eharney has quit IRC | 19:54 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: libvirt: handle code=38 + sigkill (ebusy) in destroy() https://review.openstack.org/181781 | 19:55 |
| *** tjones1 has joined #openstack-nova | 19:55 | |
| *** alex_klimov has joined #openstack-nova | 19:56 | |
| mriedem | kashyap: ^ happy birthday | 19:56 |
| *** Marga_ has quit IRC | 19:57 | |
| *** dave-mcc_ has joined #openstack-nova | 19:59 | |
| *** dprince has joined #openstack-nova | 20:01 | |
| *** dave-mccowan has quit IRC | 20:02 | |
| *** dulek has quit IRC | 20:05 | |
| cfriesen | does anyone know why "allow_resize_to_same_host" defaults to False? | 20:10 |
| cfriesen | lifeless: you around? missed you earlier. I think what's happening is that nova calls out to glance-api, which interally creates the ChunkedFile object and returns a URL that nova can read from to get the image data. Nova starts reading from that URL, but then hits an exception and then stops. | 20:12 |
| *** afazekas has joined #openstack-nova | 20:13 | |
| cfriesen | lifeless: so then glance-api has this ChunkedFile object sitting there and it's just waiting to be read from but nova won't ever read from it anymore. | 20:13 |
| lifeless | cfriesen: hi | 20:14 |
| lifeless | cfriesen: right, did you see my reply on the thread | 20:14 |
| lifeless | cfriesen: I'm trying to unpick all the way back up | 20:14 |
| cfriesen | lifeless: checking | 20:14 |
| lifeless | cfriesen: if the http connection is closed the api container (whatever it is) will cleanup | 20:14 |
| lifeless | unless thats buggy | 20:15 |
| lifeless | we'll find the bug and the world will be better | 20:15 |
| cfriesen | lifeless: so you're saying that when nova leaves image.glance.GlanceImageService.download() then image_chunks goes out of scope and that should result in the http connection closing, which should get propagated down (in glance-api) down into the ChunkedFile object being destroyed? | 20:18 |
| lifeless | yes | 20:19 |
| lifeless | tcp can signal close at any point | 20:19 |
| lifeless | in your environment we don't need to worry about firewallsmessing with tcp state either | 20:19 |
| cfriesen | will that result in self.fp being being closed at the time ChunkedFile is destroyed? or does ChunkedFile need a destructor that closes the file? | 20:20 |
| *** Mike_D_laptop has quit IRC | 20:21 | |
| *** dprince has quit IRC | 20:22 | |
| *** exploreshaifali has joined #openstack-nova | 20:23 | |
| jaypipes | cfriesen: have you done a lsof to see who owns the file descriptors you see as leaking? | 20:26 |
| cfriesen | jaypipes: glance-api | 20:26 |
| jaypipes | cfriesen: k. thx, looking something right quick. | 20:26 |
| cfriesen | jaypipes: pretty sure it's the glance_store._drivers.filesystem.ChunkedFile class | 20:27 |
| jaypipes | cfriesen: you using v1 or v2 Glance API? | 20:28 |
| mriedem | dansmith: i think you wanted to see this happen https://review.openstack.org/#/c/183711/ | 20:28 |
| *** annashen has joined #openstack-nova | 20:28 | |
| *** eharney has joined #openstack-nova | 20:28 | |
| jaypipes | cfriesen: pretty sure it's not the ChunkedFile in filesystem driver. | 20:29 |
| jaypipes | cfriesen: more likely to be the utils.cooperative_iter(): https://github.com/openstack/glance/blob/master/glance/api/v1/images.py#L499 | 20:29 |
| dansmith | mriedem: well, I wanted to see us fail on <5.0, but yes this is better than nothing | 20:30 |
| lifeless | jaypipes: that also uses a clean for loop | 20:30 |
| lifeless | jaypipes: so should be triggering refcount-zero cleanups on exiting scope | 20:30 |
| *** prashant__ has quit IRC | 20:31 | |
| cfriesen | jaypipes: not sure which version actually. easiest way to check? | 20:31 |
| jaypipes | cfriesen: keystone service_catalog, check the version in the image API endpoint URI. | 20:31 |
| jaypipes | sorry, keystone catalog | 20:31 |
| lifeless | cfriesen: how is glance-api deployed? mod_wsgi? gunicorn? ... ? | 20:31 |
| jaypipes | lifeless: I don't see https://github.com/openstack/glance/blob/master/glance/common/utils.py#L151 close()'ing the file descriptor it gets passed... | 20:32 |
| lifeless | jaypipes: its not passed an fd, but an iterable | 20:33 |
| lifeless | jaypipes: the ChunkedFile iterable has a try:finally:close: | 20:33 |
| lifeless | jaypipes: when the iterator created from the iterable goes out of the for loop its refcount hits zero and the generator gets an exception thrown from whatever yield it was on, which will hit the finally. | 20:33 |
| jaypipes | lifeless: it's passed an fd, not an iterable. | 20:33 |
| *** annashen has quit IRC | 20:34 | |
| lifeless | jaypipes: cooperative_iter? | 20:34 |
| jaypipes | CooperativeReader. | 20:34 |
| lifeless | jaypipes: its passed the result of _get_from_store()[0] | 20:34 |
| jaypipes | its constructor. | 20:34 |
| lifeless | jaypipes: sorry, I was looking at cooperative_iter, not cooperative_read | 20:34 |
| *** markvoelker has joined #openstack-nova | 20:35 | |
| cfriesen | jaypipes: no version in the url: http://192.168.204.2:9292 | 20:35 |
| jaypipes | lifeless: hmm, question for you... | 20:35 |
| *** baoli has joined #openstack-nova | 20:35 | |
| jaypipes | lifeless: what if ChunkedFile was never iterated through? ChunkedFile.close() would never be called then? | 20:36 |
| lifeless | jaypipes: yes | 20:36 |
| jaypipes | lifeless: so do we think that's the issue? | 20:36 |
| lifeless | jaypipes: thats a clear bug, it might not be this one. | 20:36 |
| lifeless | it should open at the start of __iter__ IMO | 20:37 |
| jaypipes | agreed., | 20:37 |
| jaypipes | man, this code is a blast from the past :) | 20:37 |
| *** baoli has quit IRC | 20:37 | |
| jaypipes | I wrote a lot of this shit about 4 years ago. | 20:37 |
| *** vladikr has quit IRC | 20:38 | |
| *** baoli has joined #openstack-nova | 20:38 | |
| cburgess | What jaypipes is saying is that its all his fault. | 20:38 |
| cfriesen | lol | 20:38 |
| jaypipes | yup. | 20:38 |
| cfriesen | jaypipes: does anything ever explicitly call ChunkedFile.close()? | 20:38 |
| jaypipes | of course, since glance_store was split out without the git history, there's no mention of me in the git blame... wooohooo! :) | 20:39 |
| jaypipes | cfriesen: no, only the __iter__'s finally: block. | 20:39 |
| jaypipes | cfriesen: which is why I mention to lifeless above that it may be that the ChunkedFile object is never iterated against, and that is the source of the leak. | 20:40 |
| *** markvoelker has quit IRC | 20:40 | |
| lifeless | jaypipes: it seems like a much narrower race to hit | 20:40 |
| cfriesen | jaypipes: in the success case we don't see the dangling refcounts | 20:40 |
| cfriesen | jaypipes: if I hack nova to keep reading on error the problem goes away | 20:41 |
| lifeless | jaypipes: and - since we have to run out of disk space on nova to trigger it AIUI | 20:41 |
| lifeless | jaypipes: that implies we've read at least one segment, so I think we can rule it out. | 20:41 |
| jaypipes | lifeless: no, not necessarily... | 20:41 |
| jaypipes | lifeless: one sec, grabbing a link.. | 20:41 |
| lifeless | k | 20:41 |
| mriedem | garyk: added some comments in this detach volume change https://review.openstack.org/#/c/169732/10/nova/virt/vmwareapi/driver.py - specifically the thing where the driver has to set a vm_state on the instance for volumeops to work is kind of sad | 20:42 |
| mriedem | garyk: can't volumeops get the power_state of the vm from vcenter and use that rather than the nova db values on the instance? | 20:43 |
| mriedem | adds tight coupling between the driver and volumeops which is prone to break at some point when someone doesn't know to set the vm_state before calling volumeops.detach_volume | 20:43 |
| *** annegentle has quit IRC | 20:44 | |
| lifeless | hmm the chunkedfile chunk size is a bit small | 20:45 |
| lifeless | for 10G connections you really want that up at say 4M | 20:45 |
| lifeless | shrug | 20:45 |
| lifeless | one bug at a time | 20:45 |
| jaypipes | lifeless, cfriesen: https://github.com/openstack/glance/blob/master/glance/image_cache/drivers/sqlite.py#L361-L378 | 20:46 |
| jaypipes | so, I think this may have to do with the image cache middleware and how it returns a response iterator: https://github.com/openstack/glance/blob/master/glance/api/middleware/cache.py#L306-L308 | 20:47 |
| lifeless | jaypipes: that bit of code is fine, its using with to do RAII | 20:47 |
| lifeless | jaypipes: the caller of open_for_read could leak potentially | 20:47 |
| jaypipes | lifeless: https://github.com/openstack/glance/blob/master/glance/image_cache/__init__.py#L228-L244 | 20:48 |
| jaypipes | lifeless: the idea behind the caching middleware is to write images to a cache *while they are being streamed to a client*. | 20:49 |
| *** apuimedo has joined #openstack-nova | 20:49 | |
| jaypipes | lifeless: which is why you see the writer and reader iterables happening at the same time there. | 20:50 |
| *** dave-mccowan has joined #openstack-nova | 20:50 | |
| *** Sukhdev has joined #openstack-nova | 20:50 | |
| lifeless | k | 20:50 |
| cfriesen | jaypipes: so if the client never finishes reading the file... | 20:50 |
| jaypipes | lifeless: I have a feeling that cache_tee_iter() needs to have some sort of finally: block | 20:50 |
| *** thangp has quit IRC | 20:50 | |
| *** dave-mcc_ has quit IRC | 20:50 | |
| jaypipes | lifeless: bingo. | 20:50 |
| *** achanda has joined #openstack-nova | 20:51 | |
| lifeless | you tested something ? | 20:51 |
| jaypipes | lifeless: no, I think you hit the nail on the head. | 20:51 |
| jaypipes | cfriesen: an easy way to test this: can you remove (or disable) the image cache middleware and re-run your tests? | 20:52 |
| openstackgerrit | Don Dugger proposed openstack/nova-specs: Remove flavor from scheduler request https://review.openstack.org/189444 | 20:52 |
| lifeless | get_from_cache is what triggers open_for_read | 20:52 |
| jaypipes | cfriesen: if the tests no longer trigger this leak, we've determined the cause is the caching middleware. | 20:53 |
| lifeless | glance/api/middleware/cache.py line 168 is a problem | 20:54 |
| *** Marga_ has joined #openstack-nova | 20:54 | |
| lifeless | it calls a generator directly | 20:54 |
| cfriesen | jaypipes: just comment out the "paste.filter_factory = glance.api.middleware.cache:CacheFilter.factory" line? | 20:54 |
| lifeless | potential problem I should say | 20:54 |
| lifeless | which will prime the generator and then its passed into method() | 20:55 |
| lifeless | but exceptions thrown here could keep the thing alive on the stack | 20:55 |
| lifeless | because its context is not self-managing like for loops | 20:55 |
| lifeless | just a possibility | 20:55 |
| *** belmoreira has quit IRC | 20:56 | |
| *** achanda has quit IRC | 20:56 | |
| jaypipes | lifeless: no, it doesn't call a generator directly. it returns the generator in the image_iterator element of the response (https://github.com/openstack/glance/blob/master/glance/api/middleware/cache.py#L214) | 20:56 |
| *** otter768 has joined #openstack-nova | 20:57 | |
| cfriesen | jaypipes: it'll take some time to set up the test...I'll let you know how it goes. | 20:57 |
| jaypipes | lifeless: nm, that's the incoming request, not te resp[onse. | 20:57 |
| jaypipes | lifeless: this is the process of the GET /images/{id} response: https://github.com/openstack/glance/blob/master/glance/api/middleware/cache.py#L306-L307 | 20:58 |
| *** Marga_ has quit IRC | 20:59 | |
| *** ivasev has quit IRC | 20:59 | |
| *** Marga_ has joined #openstack-nova | 20:59 | |
| *** raildo has quit IRC | 21:00 | |
| *** otter768 has quit IRC | 21:01 | |
| *** jdurgin has quit IRC | 21:03 | |
| *** annashen has joined #openstack-nova | 21:04 | |
| *** dave-mcc_ has joined #openstack-nova | 21:06 | |
| *** thorst has quit IRC | 21:07 | |
| lifeless | jaypipes: correctme if I'm wrong | 21:08 |
| jaypipes | you're wrong. | 21:08 |
| lifeless | jaypipes: but isn't caching middleware of file store a unuseful combination? | 21:08 |
| jaypipes | lifeless: it is indeed. | 21:08 |
| cfriesen | jaypipes: so easy workaround if this fixes things. | 21:08 |
| jaypipes | lifeless: but you're still wrong, because, well, it's better that way. | 21:08 |
| lifeless | jaypipes: speaking of | 21:08 |
| lifeless | jaypipes: how should we get this architecture spelunking thing going | 21:09 |
| lifeless | like | 21:09 |
| lifeless | can we block out a time slot, pick a project, pick some basic outcomes, whatever... | 21:09 |
| *** dave-mccowan has quit IRC | 21:10 | |
| jaypipes | lifeless: unfortunately, I am way over-committed already to things this cycle, including the API WG. I really don't think I'll be able to contribute much other than the occasional review or opinion :( | 21:10 |
| lifeless | ok | 21:10 |
| jaypipes | lifeless: sorry :( | 21:10 |
| lifeless | clear signalling is useful | 21:10 |
| lifeless | nothing to apologise for | 21:10 |
| *** annegentle has joined #openstack-nova | 21:11 | |
| *** burt has quit IRC | 21:12 | |
| *** afazekas has quit IRC | 21:13 | |
| *** jgrimm is now known as zz_jgrimm | 21:17 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/189461 | 21:19 |
| *** sdake_ has joined #openstack-nova | 21:19 | |
| *** mriedem is now known as mriedem_away | 21:20 | |
| *** eglynn has quit IRC | 21:22 | |
| *** jdurgin has joined #openstack-nova | 21:23 | |
| *** sdake has quit IRC | 21:23 | |
| *** artom has quit IRC | 21:25 | |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: virt: introduce libosinfo library to set hardware policy https://review.openstack.org/149625 | 21:27 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: use osinfo when configuring network model https://review.openstack.org/149627 | 21:27 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: adding libosinfo configuration https://review.openstack.org/149626 | 21:27 |
| openstackgerrit | Vladik Romanovsky proposed openstack/nova: libvirt: use osinfo when configuring the disk bus https://review.openstack.org/149628 | 21:27 |
| *** afazekas has joined #openstack-nova | 21:27 | |
| *** eglynn has joined #openstack-nova | 21:29 | |
| *** jdurgin has quit IRC | 21:30 | |
| *** ociuhandu has joined #openstack-nova | 21:32 | |
| *** marun has quit IRC | 21:33 | |
| *** Marga_ has quit IRC | 21:37 | |
| *** lucas-dinner has quit IRC | 21:37 | |
| cfriesen | jaypipes: I commented out "paste.filter_factory = glance.api.middleware.cache:CacheFilter.factory" and restarted glance-api. a failed instance creation following by deleting the image in question results in two separate glance-api processes holding open file descriptors for "/opt/cgcs/glance/images/5222b99f-6fc7-4bab-96d9-de82f7e07287 (deleted)" | 21:39 |
| jaypipes | cfriesen: on phone... gimme a few | 21:40 |
| *** exploreshaifali has quit IRC | 21:40 | |
| *** exploreshaifali has joined #openstack-nova | 21:41 | |
| *** EmilienM is now known as EmilienM|afk | 21:41 | |
| *** VW has quit IRC | 21:46 | |
| *** dboik has quit IRC | 21:46 | |
| *** BadCub-Hiding is now known as BadCub | 21:47 | |
| *** ijw has quit IRC | 21:47 | |
| *** ijw has joined #openstack-nova | 21:48 | |
| *** jdurgin has joined #openstack-nova | 21:48 | |
| *** dprince has joined #openstack-nova | 21:49 | |
| *** agireud has quit IRC | 21:49 | |
| *** lucas-dinner has joined #openstack-nova | 21:50 | |
| *** lucas-dinner has quit IRC | 21:50 | |
| *** achanda has joined #openstack-nova | 21:51 | |
| *** aysyanne has quit IRC | 21:53 | |
| openstackgerrit | Don Dugger proposed openstack/nova-specs: Remove flavor from scheduler request https://review.openstack.org/189444 | 21:53 |
| *** baoli has quit IRC | 21:55 | |
| *** achanda has quit IRC | 21:57 | |
| *** achanda has joined #openstack-nova | 21:59 | |
| *** eglynn has quit IRC | 22:02 | |
| *** baoli has joined #openstack-nova | 22:05 | |
| *** exploreshaifali has quit IRC | 22:05 | |
| *** achanda has quit IRC | 22:08 | |
| *** annegentle has quit IRC | 22:10 | |
| *** promulo has joined #openstack-nova | 22:14 | |
| *** annegentle has joined #openstack-nova | 22:15 | |
| *** promulo has quit IRC | 22:15 | |
| *** dave-mccowan has joined #openstack-nova | 22:15 | |
| *** VW has joined #openstack-nova | 22:16 | |
| *** afazekas has quit IRC | 22:17 | |
| *** dave-mcc_ has quit IRC | 22:18 | |
| *** baoli has quit IRC | 22:20 | |
| *** markvoelker has joined #openstack-nova | 22:21 | |
| *** markvoelker has quit IRC | 22:26 | |
| *** baoli has joined #openstack-nova | 22:27 | |
| *** diegows has joined #openstack-nova | 22:27 | |
| *** liusheng has quit IRC | 22:28 | |
| *** ijw has quit IRC | 22:28 | |
| *** liusheng has joined #openstack-nova | 22:28 | |
| *** ijw has joined #openstack-nova | 22:29 | |
| *** ijw has quit IRC | 22:30 | |
| *** penick has joined #openstack-nova | 22:31 | |
| *** VW_ has joined #openstack-nova | 22:38 | |
| *** VW has quit IRC | 22:38 | |
| flwang | dansmith: ping | 22:39 |
| dansmith | flwang: ack | 22:39 |
| flwang | dansmith: may i know what kind of actions should be listed for the action-list command? | 22:39 |
| flwang | is there a policy for that? | 22:40 |
| *** pilgrimstack has joined #openstack-nova | 22:40 | |
| flwang | i asked since I found the nova backup action can't be shown with the action-list command | 22:40 |
| dansmith | flwang: I don't think there is any restriction around adding actions | 22:40 |
| dansmith | flwang: ask alaski | 22:41 |
| *** nic has quit IRC | 22:41 | |
| *** baoli has quit IRC | 22:41 | |
| *** nic has joined #openstack-nova | 22:41 | |
| flwang | dansmith: ok, cool, thanks | 22:41 |
| *** diogogmt has quit IRC | 22:41 | |
| flwang | alaski: ^ | 22:42 |
| *** apuimedo has quit IRC | 22:42 | |
| mriedem_away | appears we have a new multinode job failure | 22:43 |
| openstackgerrit | Nicolas Simonds proposed openstack/nova: imagebackend: Split out resize_image logic from create_image https://review.openstack.org/187857 | 22:44 |
| openstackgerrit | Nicolas Simonds proposed openstack/nova: Fix/implement revert-resize for RBD-backed images https://review.openstack.org/187395 | 22:44 |
| mriedem_away | http://goo.gl/dpFqTV | 22:44 |
| openstackgerrit | Nicolas Simonds proposed openstack/nova: libvirt: Make the guest console configurable https://review.openstack.org/122523 | 22:44 |
| *** ijw has joined #openstack-nova | 22:44 | |
| *** pilgrimstack has quit IRC | 22:45 | |
| *** mdrabe has quit IRC | 22:48 | |
| *** baoli has joined #openstack-nova | 22:49 | |
| *** Sukhdev has quit IRC | 22:50 | |
| cfriesen | jaypipes: lifeless: I was mislead earlier by the fact that it takes some time for the resources to be cleaned up. With the image cache present it takes about 4 minutes for the file descriptor to be closed. With the image cache not present it takes about 90 seconds. In both cases it did eventually close in my testing today. | 22:52 |
| *** jecarey has quit IRC | 22:53 | |
| lifeless | cool | 22:55 |
| mriedem_away | multinode resize/migrate fail bug https://bugs.launchpad.net/nova/+bug/1463200 | 22:55 |
| openstack | Launchpad bug 1463200 in OpenStack Compute (nova) "check-tempest-dsvm-multinode-full fails due to "Failed to compute_task_migrate_server: No valid host was found"" [Undecided,New] | 22:55 |
| mriedem_away | looks like one of the compute host services is being reported as down intermittently | 22:55 |
| cfriesen | jaypipes: lifeless: so I don't know if that's expected behaviour of the GC, or whether we can do better with ensuring that the fd is closed when the http connection is closed by nova. | 22:56 |
| cfriesen | jaypipes: lifeless: gah! I may have forgotten to remove our workaround. one sec, will check and retest if necessary. | 22:57 |
| *** otter768 has joined #openstack-nova | 22:57 | |
| cfriesen | jaypipes: lifeless: phew. no workaround present, results are valid | 22:59 |
| *** VW_ has quit IRC | 23:00 | |
| *** ZZelle_ has quit IRC | 23:01 | |
| *** VW has joined #openstack-nova | 23:01 | |
| *** annegentle has quit IRC | 23:02 | |
| *** otter768 has quit IRC | 23:02 | |
| *** xyang1 has quit IRC | 23:03 | |
| *** baoli has quit IRC | 23:03 | |
| *** jwcroppe_ has joined #openstack-nova | 23:05 | |
| *** VW has quit IRC | 23:06 | |
| *** jwcroppe has quit IRC | 23:07 | |
| *** zzzeek has quit IRC | 23:11 | |
| *** atuvenie has quit IRC | 23:12 | |
| *** jwcroppe has joined #openstack-nova | 23:13 | |
| *** jwcroppe_ has quit IRC | 23:14 | |
| melwitt | mriedem_away: why is it the else block should be removed here? https://review.openstack.org/#/c/189377/2/nova/virt/libvirt/host.py,cm | 23:16 |
| mriedem_away | melwitt: the formatting is weird, but there are 2 else blocks, the one that's left in there replaces it | 23:17 |
| mriedem_away | melwitt: it would have been more clear if the outer else was removed and the rest was shifted left | 23:17 |
| melwitt | mriedem_away: oh, duh. I see it now, thanks | 23:18 |
| *** dprince has quit IRC | 23:18 | |
| mriedem_away | yeah it's weird | 23:18 |
| mriedem_away | i added a comment on that | 23:18 |
| *** Sukhdev has joined #openstack-nova | 23:19 | |
| *** EmilienM|afk is now known as EmilienM | 23:20 | |
| *** electrocucaracha has quit IRC | 23:21 | |
| *** annegentle has joined #openstack-nova | 23:22 | |
| tonyb | Isn't it just git diff being too "smart" | 23:22 |
| *** ijw has quit IRC | 23:23 | |
| *** dave-mcc_ has joined #openstack-nova | 23:23 | |
| *** Maike has joined #openstack-nova | 23:24 | |
| *** kmartin_ is now known as kmartin | 23:26 | |
| *** dave-mccowan has quit IRC | 23:26 | |
| *** flwang has quit IRC | 23:27 | |
| *** alex_klimov has quit IRC | 23:29 | |
| jaypipes | edleafe, lxsli_away: If I renamed Usage to Inventory, would you be cool with that? | 23:31 |
| *** hemna is now known as hemnafk | 23:32 | |
| *** zhenguo has joined #openstack-nova | 23:35 | |
| *** jaypipes has quit IRC | 23:35 | |
| *** thorst has joined #openstack-nova | 23:36 | |
| *** annegentle has quit IRC | 23:37 | |
| *** diogogmt has joined #openstack-nova | 23:38 | |
| *** flwang has joined #openstack-nova | 23:39 | |
| *** thorst has quit IRC | 23:39 | |
| openstackgerrit | Don Dugger proposed openstack/nova-specs: Remove flavor from scheduler request https://review.openstack.org/189444 | 23:39 |
| *** cfriesen has quit IRC | 23:42 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/189461 | 23:44 |
| *** dontalton has quit IRC | 23:45 | |
| *** annegentle has joined #openstack-nova | 23:51 | |
| *** Maike has quit IRC | 23:53 | |
| *** annegentle has quit IRC | 23:54 | |
| *** diogogmt has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!