| mriedem | melwitt: so that nvoaclient functional gate hook thing is our problem | 00:00 |
|---|---|---|
| melwitt | mriedem: do we need to do something/change something to handle whatever mtreinish mentioned before? | 00:01 |
| mriedem | yeah | 00:01 |
| mtreinish | melwitt: you just need to update the path in the gate hook | 00:01 |
| mriedem | something in ehre https://github.com/openstack/python-novaclient/blob/master/novaclient/tests/functional/hooks/post_test_hook.sh | 00:01 |
| melwitt | oh, I was confused because it seems to not always fail? | 00:02 |
| mtreinish | melwitt: like that: https://review.openstack.org/#/c/212864/3/jenkins/scripts/run-tox.sh | 00:02 |
| mriedem | i've got it | 00:02 |
| mtreinish | melwitt: yeah I was pretty sure the gate hooks didn't have errexit set | 00:02 |
| mtreinish | so I'm not sure why it would fail there | 00:02 |
| mtreinish | because it should fail to generate the html output and just move on | 00:02 |
| mtreinish | but whatever, it's a simple fix | 00:03 |
| mriedem | doing it locally | 00:03 |
| mriedem | adding the other projects to the bug | 00:03 |
| *** emagana has quit IRC | 00:03 | |
| *** takedakn1 has joined #openstack-nova | 00:03 | |
| mtreinish | melwitt: oh, you know it's only on the latest nodepool image, so whenever that started getting used is probably when things started failing | 00:04 |
| *** takedakn has quit IRC | 00:07 | |
| *** takedakn1 is now known as takedakn | 00:07 | |
| openstackgerrit | Matt Riedemann proposed openstack/python-novaclient: Update path to subunit2html in post_test_hook https://review.openstack.org/219924 | 00:07 |
| melwitt | mtreinish: ahhh | 00:07 |
| mriedem | mtreinish: ^ like that? | 00:08 |
| mtreinish | mriedem: yep | 00:08 |
| mriedem | mtreinish: shouldn't novaclient require os-testr for that? | 00:09 |
| mriedem | and does it need to be the absolute path? | 00:09 |
| mriedem | couldn't we just require os-testr and call subunit2html? | 00:09 |
| mtreinish | mriedem: no, nodepool is baking it into the image inside of a venv | 00:09 |
| mtreinish | mriedem: I mean you could (you kinda already are by pulling tempest-lib it's in the reqs list there) | 00:09 |
| mtreinish | and just call it directly | 00:09 |
| mriedem | ok | 00:09 |
| mriedem | well if this passes we'll just call it good | 00:10 |
| *** takedakn1 has joined #openstack-nova | 00:10 | |
| mtreinish | mriedem: yeah that's got fast approve written all over it | 00:10 |
| *** takedakn has quit IRC | 00:11 | |
| *** takedakn1 is now known as takedakn | 00:11 | |
| *** rushil_ has quit IRC | 00:13 | |
| *** rushil has joined #openstack-nova | 00:14 | |
| *** tobasco_ has quit IRC | 00:15 | |
| *** tobasco has joined #openstack-nova | 00:16 | |
| *** promulo__ has joined #openstack-nova | 00:18 | |
| *** goodygum has quit IRC | 00:19 | |
| *** wanghua has joined #openstack-nova | 00:20 | |
| *** mtanino_ has quit IRC | 00:21 | |
| *** jcoufal has joined #openstack-nova | 00:21 | |
| *** wanghua has quit IRC | 00:21 | |
| *** goodygum has joined #openstack-nova | 00:21 | |
| *** thorst has joined #openstack-nova | 00:21 | |
| *** wanghua has joined #openstack-nova | 00:21 | |
| *** promulo has quit IRC | 00:22 | |
| *** whenry_ has joined #openstack-nova | 00:22 | |
| *** thorst has quit IRC | 00:23 | |
| *** whenry_ has quit IRC | 00:27 | |
| *** achanda has joined #openstack-nova | 00:28 | |
| *** sachin_ has quit IRC | 00:28 | |
| *** sachin_ has joined #openstack-nova | 00:29 | |
| *** moshele has quit IRC | 00:31 | |
| mriedem | dhellmann: so i plan on doing a novaclient 2.27.1 release tonight to unblock manilaclient, if i request that through the releases repo, i'm assuming it won't actually happen until tomorrrow? | 00:33 |
| mriedem | dhellmann: is there anything preventing me from just doing the release myself? | 00:33 |
| *** Sukhdev_ has joined #openstack-nova | 00:37 | |
| *** takedakn1 has joined #openstack-nova | 00:40 | |
| *** takedakn has quit IRC | 00:44 | |
| *** IanGovett has quit IRC | 00:45 | |
| *** browne has quit IRC | 00:46 | |
| *** armax has quit IRC | 00:46 | |
| *** takedakn1 has quit IRC | 00:46 | |
| *** armax has joined #openstack-nova | 00:47 | |
| *** achanda has quit IRC | 00:51 | |
| *** achanda has joined #openstack-nova | 00:52 | |
| *** achanda has quit IRC | 00:52 | |
| *** electrocucaracha has quit IRC | 00:54 | |
| *** shoutm_ has joined #openstack-nova | 00:56 | |
| *** shoutm has quit IRC | 00:56 | |
| *** browne has joined #openstack-nova | 00:58 | |
| *** sachin_ has quit IRC | 00:59 | |
| *** sachin_ has joined #openstack-nova | 01:00 | |
| *** rushil has quit IRC | 01:02 | |
| *** zzzeek has joined #openstack-nova | 01:03 | |
| *** pixelb has quit IRC | 01:04 | |
| *** ducnc has joined #openstack-nova | 01:07 | |
| *** tjones has joined #openstack-nova | 01:11 | |
| *** zul has quit IRC | 01:14 | |
| *** tri2sing has quit IRC | 01:16 | |
| *** zul has joined #openstack-nova | 01:17 | |
| *** ijw__ has quit IRC | 01:20 | |
| *** mdrabe has joined #openstack-nova | 01:25 | |
| mriedem | sc68cal: HenryG: could one of you take a look at this and confirm? https://review.openstack.org/#/c/192043/ | 01:28 |
| mriedem | about the neutron security group API | 01:28 |
| sc68cal | mriedem: looking | 01:29 |
| *** tjones has quit IRC | 01:30 | |
| *** smatzek has joined #openstack-nova | 01:33 | |
| *** vilobhmm_11 has quit IRC | 01:36 | |
| sc68cal | mriedem: commented on the review | 01:39 |
| *** mcarden has quit IRC | 01:40 | |
| *** mcarden has joined #openstack-nova | 01:40 | |
| *** jcoufal has quit IRC | 01:41 | |
| *** achanda has joined #openstack-nova | 01:44 | |
| *** sachin_ has quit IRC | 01:45 | |
| *** boris-42 has quit IRC | 01:50 | |
| *** jcoufal has joined #openstack-nova | 01:52 | |
| mriedem | sc68cal: thanks | 01:52 |
| mriedem | mtreinish: melwitt: classic, i have to squash those two novaclient fixes to unblock novaclient | 02:05 |
| mriedem | http://logs.openstack.org/24/219924/1/check/gate-novaclient-dsvm-functional/bb93877/console.html#_2015-09-03_01_07_27_335 | 02:05 |
| *** baoli has quit IRC | 02:06 | |
| *** baoli has joined #openstack-nova | 02:06 | |
| *** SlickNik has joined #openstack-nova | 02:07 | |
| *** baoli has quit IRC | 02:07 | |
| *** baoli has joined #openstack-nova | 02:08 | |
| *** tjones has joined #openstack-nova | 02:08 | |
| SlickNik | hi folks — did something recently merge that disallows spaces in nova instance names? | 02:09 |
| *** zzzeek has quit IRC | 02:09 | |
| openstackgerrit | Matt Riedemann proposed openstack/python-novaclient: Update path to subunit2html in post_test_hook https://review.openstack.org/219835 | 02:09 |
| SlickNik | We're now getting this error when trying to do so: | 02:10 |
| mriedem | SlickNik: might be another symptom of https://github.com/openstack/nova/commit/4a18f7d3bafcdbede48500aac389e0a770b8e6a8 | 02:10 |
| SlickNik | Returning 400 to user: Invalid input for field/attribute name. Value: TEST_2015-09-02 21:39:34.721645_config. u'TEST_2015-09-02 21:39:34.721645_config' does not match '^[a-zA-Z0-9-._]*$' | 02:10 |
| mriedem | SlickNik: document it here https://bugs.launchpad.net/nova/+bug/1491511 | 02:11 |
| openstack | Launchpad bug 1491511 in OpenStack Compute (nova) "Behavior change with latest nova paste config" [Undecided,New] | 02:11 |
| mriedem | we're going to be talking about that in the nova meeting tomorrow morning | 02:11 |
| *** lyanchih has joined #openstack-nova | 02:12 | |
| *** armax has quit IRC | 02:12 | |
| SlickNik | will add the info to the bug. Thanks mriedem. | 02:13 |
| mriedem | np | 02:13 |
| mriedem | basically that change makes all v2.0 requests go to v2.1 | 02:13 |
| mriedem | which is apparently not completely backwards compatible when it comes to schema validation | 02:13 |
| SlickNik | ah, got it. | 02:14 |
| jroll | wasn't that fully intentional? | 02:14 |
| jroll | or at least known? | 02:14 |
| mriedem | jroll: well, i think you'd usually want people to opt-in | 02:15 |
| mriedem | rather than just break them :) | 02:15 |
| *** jcoufal has quit IRC | 02:16 | |
| jroll | mriedem: overrated :) | 02:17 |
| *** smatzek has quit IRC | 02:17 | |
| *** unicell has quit IRC | 02:18 | |
| mriedem | dansmith: sdague: melwitt: mikal: alaski: oomichi: both of these were +2/+W, i had to squash - needed to unblock novaclient (and manilaclient after we release this) https://review.openstack.org/#/c/219835/ | 02:18 |
| mriedem | BobBall: i've removed xenproject ci from https://review.openstack.org/#/admin/groups/511,members b/c it seems to have gone off the rails, -1ing changes everywhere and i'm getting 404s trying to reach the job logs | 02:22 |
| mriedem | johnthetubaguy: ^ | 02:22 |
| *** Kennan has quit IRC | 02:29 | |
| *** Kennan has joined #openstack-nova | 02:30 | |
| *** Sukhdev_ has quit IRC | 02:32 | |
| *** haomaiwang has quit IRC | 02:36 | |
| *** haomaiwang has joined #openstack-nova | 02:38 | |
| *** unicell has joined #openstack-nova | 02:41 | |
| oomichi | alex_xu, gmann: can you take a look at https://review.openstack.org/#/c/219580/ ? | 02:43 |
| *** kiran-r has joined #openstack-nova | 02:43 | |
| gmann | oomichi: looks good. Thanks | 02:44 |
| oomichi | gmann: thanks | 02:45 |
| gmann | oomichi: np | 02:45 |
| oomichi | sdague: can you also take a look at https://review.openstack.org/#/c/219580/ ? | 02:46 |
| *** humble_ has joined #openstack-nova | 02:47 | |
| *** armax has joined #openstack-nova | 02:49 | |
| *** wanghua has quit IRC | 02:50 | |
| *** hparekh has quit IRC | 02:52 | |
| *** tojuvone has joined #openstack-nova | 02:52 | |
| *** hakimo_ has quit IRC | 02:54 | |
| *** haomaiwang has quit IRC | 02:57 | |
| *** hakimo has joined #openstack-nova | 02:57 | |
| *** venkat_p has joined #openstack-nova | 02:58 | |
| *** kiran-r has quit IRC | 02:58 | |
| *** haomaiwa_ has joined #openstack-nova | 03:00 | |
| *** mriedem has quit IRC | 03:00 | |
| *** xiaohui has quit IRC | 03:02 | |
| *** xiaohui has joined #openstack-nova | 03:02 | |
| *** wanghua has joined #openstack-nova | 03:04 | |
| *** bkopilov has quit IRC | 03:05 | |
| *** sachin has joined #openstack-nova | 03:06 | |
| *** rushil has joined #openstack-nova | 03:07 | |
| *** humble_ has quit IRC | 03:08 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/216968 | 03:09 |
| *** sachin has quit IRC | 03:10 | |
| *** rushil has quit IRC | 03:17 | |
| *** shoutm_ has quit IRC | 03:20 | |
| *** bswartz has joined #openstack-nova | 03:21 | |
| *** BadCub has quit IRC | 03:27 | |
| *** BadCub has joined #openstack-nova | 03:27 | |
| *** vishwanathj has quit IRC | 03:28 | |
| *** achanda has quit IRC | 03:28 | |
| *** apmelton has quit IRC | 03:29 | |
| *** davidlenwell has quit IRC | 03:29 | |
| *** erhudy1 has quit IRC | 03:29 | |
| *** d34dh0r53 has quit IRC | 03:30 | |
| *** jroll has quit IRC | 03:30 | |
| *** lyanchih_ has joined #openstack-nova | 03:30 | |
| *** bswartz has quit IRC | 03:30 | |
| *** sabari has quit IRC | 03:30 | |
| *** tcammann has quit IRC | 03:30 | |
| *** lyanchih has quit IRC | 03:31 | |
| *** lyanchih_ is now known as lyanchih | 03:31 | |
| *** d34dh0r53 has joined #openstack-nova | 03:31 | |
| *** sabari has joined #openstack-nova | 03:31 | |
| *** davidlenwell has joined #openstack-nova | 03:32 | |
| *** apmelton has joined #openstack-nova | 03:32 | |
| *** tcammann has joined #openstack-nova | 03:32 | |
| *** jroll has joined #openstack-nova | 03:33 | |
| *** adam_g has quit IRC | 03:33 | |
| *** bswartz has joined #openstack-nova | 03:34 | |
| *** adam_g has joined #openstack-nova | 03:35 | |
| *** adam_g has quit IRC | 03:35 | |
| *** adam_g has joined #openstack-nova | 03:35 | |
| *** tpeoples has joined #openstack-nova | 03:35 | |
| *** dims has joined #openstack-nova | 03:35 | |
| *** tjones has quit IRC | 03:36 | |
| *** tjones has joined #openstack-nova | 03:36 | |
| *** shoutm has joined #openstack-nova | 03:41 | |
| *** deepthi has joined #openstack-nova | 03:42 | |
| *** shaohe_feng has joined #openstack-nova | 03:44 | |
| *** tpeoples has quit IRC | 03:46 | |
| *** jecarey has quit IRC | 03:46 | |
| openstackgerrit | Tony Breeds proposed openstack/nova: Fix mistake in UT:test_detach_unattached_volume https://review.openstack.org/214920 | 03:49 |
| openstackgerrit | lyanchih proposed openstack/nova: Creating instance fail when inject ssh key in cells mode https://review.openstack.org/197514 | 03:54 |
| openstackgerrit | Michael Still proposed openstack/nova: Handle nova-compute failure during a soft reboot https://review.openstack.org/219980 | 03:54 |
| openstackgerrit | Michael Still proposed openstack/nova: Have one list of soft reboot task_states https://review.openstack.org/219981 | 03:54 |
| openstackgerrit | lyanchih proposed openstack/nova: NUMATopologyFilter raise exception and not continue filter next node https://review.openstack.org/200630 | 03:55 |
| *** sachin has joined #openstack-nova | 03:55 | |
| openstackgerrit | lyanchih proposed openstack/nova: nova secgroup-list-rules shows empty table with neutron https://review.openstack.org/203022 | 03:56 |
| openstackgerrit | lyanchih proposed openstack/nova: libvirt - Flavor's quota instance resource not observed for non-ephemeral storage https://review.openstack.org/201019 | 03:57 |
| *** boris-42 has joined #openstack-nova | 03:59 | |
| *** dims has quit IRC | 04:00 | |
| *** links has joined #openstack-nova | 04:02 | |
| *** tjones has quit IRC | 04:02 | |
| *** rajesht has joined #openstack-nova | 04:03 | |
| *** sachin has quit IRC | 04:06 | |
| *** Sukhdev_ has joined #openstack-nova | 04:06 | |
| gmann | oomichi: alex_xu : edleafe : we planning to remove this too in L right?- https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/__init__.py#L168 | 04:07 |
| oomichi | gmann: maybe it is not possible in L because of grenede test | 04:08 |
| oomichi | gmann: s/grenade/grenade/ | 04:08 |
| *** sachin has joined #openstack-nova | 04:08 | |
| *** Sree has joined #openstack-nova | 04:09 | |
| *** Sree has quit IRC | 04:09 | |
| oomichi | gmann: Kilo's api-paste.ini contains APIRouterV3.factory. so if removing it, the test will fail. | 04:10 |
| *** Sree has joined #openstack-nova | 04:10 | |
| gmann | oomichi: ahh, right even kilo functional tests need those | 04:10 |
| gmann | oomichi: Thanks :) | 04:10 |
| oomichi | gmann: :) | 04:11 |
| *** Sree has quit IRC | 04:11 | |
| *** baoli has quit IRC | 04:14 | |
| *** shaohe_feng has quit IRC | 04:14 | |
| *** Sree has joined #openstack-nova | 04:14 | |
| *** sachin has quit IRC | 04:14 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/216968 | 04:17 |
| *** mdrabe has quit IRC | 04:19 | |
| *** rhedlind has joined #openstack-nova | 04:22 | |
| *** vishwanathj has joined #openstack-nova | 04:22 | |
| *** sachin has joined #openstack-nova | 04:31 | |
| *** lyanchih has quit IRC | 04:32 | |
| *** garyk has quit IRC | 04:35 | |
| *** sachin has quit IRC | 04:35 | |
| *** tri2sing has joined #openstack-nova | 04:37 | |
| *** tpeoples has joined #openstack-nova | 04:37 | |
| *** shoutm has quit IRC | 04:38 | |
| *** tri2sing has quit IRC | 04:39 | |
| *** hparekh has joined #openstack-nova | 04:40 | |
| *** moshele has joined #openstack-nova | 04:43 | |
| *** tri2sing has joined #openstack-nova | 04:45 | |
| *** shoutm has joined #openstack-nova | 04:47 | |
| *** moshele has quit IRC | 04:50 | |
| *** dannywilson has joined #openstack-nova | 04:52 | |
| *** ducnc is now known as ducnguyen | 04:55 | |
| *** tpeoples has quit IRC | 04:55 | |
| *** bkopilov has joined #openstack-nova | 04:57 | |
| *** tri2sing has quit IRC | 04:58 | |
| *** dims has joined #openstack-nova | 04:59 | |
| *** unicell1 has joined #openstack-nova | 05:00 | |
| *** unicell has quit IRC | 05:00 | |
| *** lyanchih has joined #openstack-nova | 05:02 | |
| *** tri2sing has joined #openstack-nova | 05:05 | |
| *** vivekd has joined #openstack-nova | 05:05 | |
| *** dims has quit IRC | 05:06 | |
| *** vilobhmm_11 has joined #openstack-nova | 05:08 | |
| *** shaohe_feng has joined #openstack-nova | 05:08 | |
| *** kiran-r has joined #openstack-nova | 05:14 | |
| *** vilobhmm_11 has quit IRC | 05:14 | |
| *** deepakcs has joined #openstack-nova | 05:18 | |
| *** lyanchih has quit IRC | 05:23 | |
| *** Kennan2 has joined #openstack-nova | 05:24 | |
| openstackgerrit | Michael Still proposed openstack/nova: Adds instance_uuid index for instance_system_metadata https://review.openstack.org/208277 | 05:24 |
| *** Kennan has quit IRC | 05:25 | |
| *** derekh has joined #openstack-nova | 05:26 | |
| *** angdraug has joined #openstack-nova | 05:26 | |
| *** lpetrut has joined #openstack-nova | 05:29 | |
| *** shaohe_feng has quit IRC | 05:30 | |
| *** asettle is now known as asettle-afk | 05:33 | |
| *** achanda has joined #openstack-nova | 05:34 | |
| *** asettle-afk is now known as asettle | 05:36 | |
| *** shoutm has quit IRC | 05:36 | |
| *** shoutm has joined #openstack-nova | 05:37 | |
| *** achanda has quit IRC | 05:40 | |
| *** dannywilson has quit IRC | 05:41 | |
| *** asettle is now known as asettle-afk | 05:42 | |
| *** ishant has joined #openstack-nova | 05:43 | |
| *** ducnguyen has quit IRC | 05:44 | |
| *** kiran-r has quit IRC | 05:44 | |
| *** moshele has joined #openstack-nova | 05:49 | |
| *** VikasC has joined #openstack-nova | 05:49 | |
| *** sahid has joined #openstack-nova | 05:49 | |
| alex_xu | oomichi: +1, thanks | 05:54 |
| alex_xu | gmann: yea, as oomichi said, we can't remove it in L | 05:54 |
| openstackgerrit | Bartosz Górski proposed openstack/python-novaclient: Adds missing internationalization for help message https://review.openstack.org/219768 | 05:54 |
| oomichi | alex_xu: much thanks :) | 05:55 |
| *** unmeshg has joined #openstack-nova | 05:55 | |
| alex_xu | oomichi: np :) | 05:56 |
| oomichi | alex_xu: btw we need to update microversion-spec on api-wg. | 05:56 |
| oomichi | alex_xu: https://review.openstack.org/#/c/196918/ is one of good input for that | 05:56 |
| oomichi | alex_xu: but it needs merger of https://review.openstack.org/#/c/181393/ before that | 05:57 |
| oomichi | alex_xu: can you help it also? | 05:57 |
| alex_xu | oomichi: actually I already updated the spec on api-wg https://review.openstack.org/187112 :) | 05:57 |
| *** shoutm has quit IRC | 05:58 | |
| *** shoutm has joined #openstack-nova | 05:58 | |
| oomichi | alex_xu: oops, I missed that. ok, will see it. | 05:58 |
| alex_xu | oomichi: thanks :) | 05:59 |
| unmeshg | Hi, facing an error "Unknown protocol 'rbd'" upon attaching a Ceph backed volume (versions: libvirt 1.2.9-9, qemu-kvm 2.3) | 05:59 |
| unmeshg | has anyone seen this? is it a known issue with the libvirt/qemu version ? | 06:00 |
| unmeshg | thanks | 06:00 |
| *** achanda has joined #openstack-nova | 06:00 | |
| *** lyanchih has joined #openstack-nova | 06:01 | |
| oomichi | alex_xu: can we separate it to several patches for each topics like Versioning, Client Interaction and Version Discovery? | 06:01 |
| *** nkrinner has joined #openstack-nova | 06:02 | |
| oomichi | alex_xu: via our experience, it was very diffult to get a consensus only for the header name. | 06:02 |
| alex_xu | oomichi: emm....sounds good idea | 06:02 |
| oomichi | alex_xu: it is enough that the first patch contains the overview only and the other part is TBD like the other doc in api-wg spec. | 06:03 |
| *** cfriesen has quit IRC | 06:04 | |
| alex_xu | oomichi: looks like client interaction and version discovery touch the header both | 06:04 |
| alex_xu | oomichi: but just a reference in version discovery | 06:05 |
| *** Sukhdev_ has quit IRC | 06:06 | |
| *** garyk has joined #openstack-nova | 06:07 | |
| alex_xu | oomichi: it's worth a try, I will take more look at this idea more. thanks for teach me some smart way of separated patch :) | 06:09 |
| oomichi | alex_xu: thanks in advance. I really hope we can get a consensus about microversions thing between projects :) | 06:10 |
| alex_xu | oomichi: yea, me too | 06:10 |
| *** vivekd has quit IRC | 06:12 | |
| *** baoli has joined #openstack-nova | 06:14 | |
| *** baoli has quit IRC | 06:19 | |
| *** achanda has quit IRC | 06:19 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Imported Translations from Transifex https://review.openstack.org/212367 | 06:24 |
| *** vilobhmm_11 has joined #openstack-nova | 06:25 | |
| *** shoutm_ has joined #openstack-nova | 06:25 | |
| *** shoutm has quit IRC | 06:26 | |
| *** mpaolino has joined #openstack-nova | 06:31 | |
| *** kaisers1 has quit IRC | 06:34 | |
| *** lpetrut has quit IRC | 06:37 | |
| *** ftersin has joined #openstack-nova | 06:40 | |
| *** abhishekk has joined #openstack-nova | 06:44 | |
| *** kiran-r has joined #openstack-nova | 06:44 | |
| *** tri2sing has quit IRC | 06:45 | |
| *** achanda has joined #openstack-nova | 06:48 | |
| *** vivekd has joined #openstack-nova | 06:50 | |
| *** vilobhmm_11 has quit IRC | 06:54 | |
| *** alex_klimov has joined #openstack-nova | 06:55 | |
| openstackgerrit | Hans Lindgren proposed openstack/nova: Remove and deprecate conductor provider_fw_rule_get_all() https://review.openstack.org/216241 | 06:56 |
| openstackgerrit | Hans Lindgren proposed openstack/nova: Remove unused provider firewall rules functionality in nova https://review.openstack.org/184027 | 06:56 |
| openstackgerrit | Hans Lindgren proposed openstack/nova: Remove provider_fw_rules tables and associated db api methods https://review.openstack.org/216242 | 06:56 |
| *** stefan_amann has joined #openstack-nova | 06:57 | |
| *** armax has quit IRC | 06:58 | |
| *** kiran-r has quit IRC | 06:59 | |
| openstackgerrit | Hans Lindgren proposed openstack/nova: Drop unused conductor manager vol_usage_update() mock https://review.openstack.org/218652 | 06:59 |
| *** vilobhmm_11 has joined #openstack-nova | 07:02 | |
| *** ildikov has joined #openstack-nova | 07:05 | |
| *** browne has quit IRC | 07:07 | |
| *** Murali has joined #openstack-nova | 07:07 | |
| *** rdopiera has joined #openstack-nova | 07:08 | |
| *** browne has joined #openstack-nova | 07:09 | |
| *** browne has quit IRC | 07:09 | |
| openstackgerrit | Michael Still proposed openstack/nova: [libvirt] Move cleanup of imported files to imagebackend https://review.openstack.org/220018 | 07:10 |
| *** xgsa has quit IRC | 07:10 | |
| *** husanu has joined #openstack-nova | 07:13 | |
| *** Marga_ has joined #openstack-nova | 07:15 | |
| *** Marga_ has quit IRC | 07:15 | |
| *** Marga_ has joined #openstack-nova | 07:16 | |
| *** xgsa has joined #openstack-nova | 07:18 | |
| *** husanu has quit IRC | 07:18 | |
| *** y_sawai has joined #openstack-nova | 07:18 | |
| *** husanu has joined #openstack-nova | 07:20 | |
| *** jichen has joined #openstack-nova | 07:21 | |
| *** matrohon has joined #openstack-nova | 07:28 | |
| *** obondarev has joined #openstack-nova | 07:30 | |
| openstackgerrit | Michael Still proposed openstack/nova: Add notes explaining vmware's suds usage https://review.openstack.org/220023 | 07:30 |
| oomichi | mikal: hi, can you take a look at https://review.openstack.org/#/c/219570/ and https://review.openstack.org/#/c/219580/ ? | 07:31 |
| mikal | Sure | 07:32 |
| mikal | oomichi: did you see the api v2.1 question on openstack-operators by the way? | 07:33 |
| *** afazekas__ has joined #openstack-nova | 07:33 | |
| oomichi | mikal: ah, no. I didn't see it because I don't receive openstack-operators | 07:34 |
| mikal | I shall copy you in then | 07:34 |
| *** mpaolino has quit IRC | 07:35 | |
| oomichi | mikal: thanks, I see it: http://lists.openstack.org/pipermail/openstack-operators/2015-September/008045.html | 07:35 |
| *** mpaolino has joined #openstack-nova | 07:36 | |
| *** jlanoux has joined #openstack-nova | 07:36 | |
| *** derekh has quit IRC | 07:36 | |
| oomichi | mikal: ah, I see. the reason is v2.1 api was disabled as the default config value in kilo. | 07:37 |
| *** gszasz has joined #openstack-nova | 07:37 | |
| mikal | Yep | 07:38 |
| oomichi | mikal: and devstack made it enabled | 07:38 |
| oomichi | mikal: we have already fixed it by changing it enabled as the default value on Liverty | 07:38 |
| *** rotbeard has joined #openstack-nova | 07:41 | |
| *** ociuhandu has quit IRC | 07:41 | |
| oomichi | mikal: https://review.openstack.org/#/c/202724 | 07:41 |
| oomichi | mikal: the above fixed it | 07:41 |
| * mikal ducks out for a bit | 07:43 | |
| *** vilobhmm_11 has quit IRC | 07:43 | |
| *** y_sawai has quit IRC | 07:45 | |
| *** lpetrut has joined #openstack-nova | 07:47 | |
| *** angdraug has quit IRC | 07:47 | |
| *** lyanchih has quit IRC | 07:48 | |
| oomichi | mikal: I sent a mail on openstack-operators | 07:50 |
| oomichi | mikal: thanks for reviews, anyway :) | 07:51 |
| *** Marga_ has quit IRC | 07:52 | |
| *** ociuhandu has joined #openstack-nova | 07:52 | |
| *** browne has joined #openstack-nova | 07:53 | |
| *** Marga_ has joined #openstack-nova | 07:56 | |
| *** lyanchih has joined #openstack-nova | 07:57 | |
| *** Sree_ has joined #openstack-nova | 08:00 | |
| *** Sree_ is now known as Guest82844 | 08:01 | |
| *** links has quit IRC | 08:02 | |
| *** y_sawai has joined #openstack-nova | 08:03 | |
| *** Sree has quit IRC | 08:03 | |
| *** ndipanov has joined #openstack-nova | 08:06 | |
| oomichi | johnthetubaguy: hi, morning | 08:06 |
| oomichi | johnthetubaguy: can you take a look at https://review.openstack.org/#/c/219570/ and https://review.openstack.org/#/c/219580/ ? | 08:06 |
| *** vivekd has quit IRC | 08:07 | |
| mkoderer | oomichi: do you know who can add a new version tag on python-novaclient? | 08:07 |
| *** y_sawai has quit IRC | 08:08 | |
| lyanchih | Hi, recently I try to implement wish list. It is talking about migrate instance to different tenant. It needs to write blueprint + spec. But I’m first time to write that. I had notice liberty will be freeze soon, should I write rst in mitaka for nova-specs? | 08:08 |
| mkoderer | we need a new version since https://bugs.launchpad.net/nova/+bug/1491325 breaks the Manila gate | 08:08 |
| openstack | Launchpad bug 1491325 in python-novaclient "nova api v2.1 does not allow to use autodetection of volume device path" [High,Fix committed] - Assigned to Matt Riedemann (mriedem) | 08:08 |
| oomichi | mkoderer: unfortunately, I cannot because of non nova-drivers. | 08:09 |
| *** unmeshg has quit IRC | 08:10 | |
| *** y_sawai has joined #openstack-nova | 08:10 | |
| mkoderer | oomichi: do you know a nova-driver in our time zone? :) | 08:10 |
| *** y_sawai has quit IRC | 08:10 | |
| oomichi | mkoderer: mikal can do that | 08:10 |
| mkoderer | mikal: hi there :) | 08:10 |
| oomichi | mikal: ^^^ can you help it ? | 08:10 |
| *** y_sawai has joined #openstack-nova | 08:11 | |
| mkoderer | vponomaryov: ^^ | 08:11 |
| *** yassine_ has joined #openstack-nova | 08:13 | |
| *** romainh has joined #openstack-nova | 08:13 | |
| *** y_sawai has quit IRC | 08:15 | |
| *** alexschm has joined #openstack-nova | 08:17 | |
| *** lucasagomes has joined #openstack-nova | 08:18 | |
| openstackgerrit | garyk proposed openstack/nova: Logging: add in a comment indicating why we need the suds setting https://review.openstack.org/220032 | 08:21 |
| garyk | johnthetubaguy: ^ | 08:22 |
| *** danpb has joined #openstack-nova | 08:22 | |
| oomichi | garyk: that seems conflict with mikal's patch: https://review.openstack.org/#/c/220023/ | 08:25 |
| gmann | johnthetubaguy: mikal : mriedem: Please have a look, has one +2 - https://review.openstack.org/#/c/198944/ | 08:26 |
| *** y_sawai has joined #openstack-nova | 08:27 | |
| garyk | oomichi: thanks! | 08:29 |
| *** vivekd has joined #openstack-nova | 08:33 | |
| *** browne has quit IRC | 08:35 | |
| *** xgsa has quit IRC | 08:36 | |
| *** achanda has quit IRC | 08:36 | |
| *** jistr has joined #openstack-nova | 08:37 | |
| openstackgerrit | Hans Lindgren proposed openstack/nova: Remove and deprecate conductor compute_node_create() https://review.openstack.org/217132 | 08:38 |
| openstackgerrit | Hans Lindgren proposed openstack/nova: Add version 3.0 of conductor RPC interface https://review.openstack.org/168857 | 08:38 |
| openstackgerrit | Hans Lindgren proposed openstack/nova: Remove lazy-loading property compute_task_api from compute api https://review.openstack.org/168859 | 08:38 |
| openstackgerrit | Hans Lindgren proposed openstack/nova: Remove conductor 2.x RPC API https://review.openstack.org/168858 | 08:38 |
| openstackgerrit | Rajesh Tailor proposed openstack/nova: Fix quota-update when nova-compute startup finish https://review.openstack.org/170118 | 08:42 |
| *** doude_ has quit IRC | 08:44 | |
| *** y_sawai has quit IRC | 08:46 | |
| *** vivekd has quit IRC | 08:47 | |
| *** y_sawai has joined #openstack-nova | 08:49 | |
| *** xgsa has joined #openstack-nova | 08:52 | |
| *** asettle-afk is now known as asettle | 08:53 | |
| *** y_sawai has quit IRC | 08:54 | |
| *** vivekd has joined #openstack-nova | 08:55 | |
| *** derekh has joined #openstack-nova | 08:55 | |
| BobBall | Can someone please re-add the XenProject CI to https://review.openstack.org/#/admin/groups/511,members? There was a temporary breakage last night which prompted mriedem to remove it, however that breakage has resolved itself. | 08:59 |
| johnthetubaguy | oomichi: I added your patch into here: https://etherpad.openstack.org/p/liberty-nova-priorities-tracking | 08:59 |
| BobBall | johnthetubaguy: ^^? | 08:59 |
| BobBall | Ah - you are lurking anyway :D | 08:59 |
| johnthetubaguy | just got back | 08:59 |
| johnthetubaguy | BobBall: how do I find your user name, its not finding you for some reason | 09:00 |
| johnthetubaguy | ah, there we go | 09:01 |
| BobBall | Great | 09:01 |
| BobBall | It's XenProject-CI I think though? | 09:01 |
| *** Guest82844 has quit IRC | 09:01 | |
| BobBall | Just needed a new image rebuilding - which was rebuilt overnight! | 09:01 |
| BobBall | Gotta love the self-healing CI | 09:01 |
| *** Sree has joined #openstack-nova | 09:02 | |
| oomichi | johnthetubaguy: ok, I will do that. | 09:06 |
| johnthetubaguy | oomichi: working through liberty-3 tagging right now, so best to put it there so I can get back to it | 09:07 |
| bauzas | BobBall: per the above, I guess I can ask for a xen: recheck ? :) | 09:09 |
| BobBall | Yes bauzas | 09:09 |
| bauzas | BobBall: coolness | 09:10 |
| BobBall | I guess I should look into fixing the way the CI works... There is one dynamic dependency which is what caused last nights breakage... and it's installed post-build, so we can upload logs. | 09:10 |
| BobBall | If it fails, no logs :D | 09:10 |
| bauzas | BobBall: fasten your seatbelt, I'm gonna hit your CI with 14 changes | 09:11 |
| * BobBall blacklists bauzas | 09:11 | |
| bauzas | heh | 09:11 |
| BobBall | uhhh I mean great! | 09:11 |
| BobBall | hang on | 09:12 |
| BobBall | I'll do it for you bauzas | 09:12 |
| bauzas | nah, that's okay | 09:12 |
| BobBall | I'll recheck all changes that failed last night - easy enough to ... or is it. | 09:12 |
| * BobBall wonders if there is a way to query gerrit for all changes a particular account negative voted on | 09:12 | |
| bauzas | BobBall: my Ctrl-C/Ctrl-V skills are ood | 09:12 |
| bauzas | BobBall: yup | 09:12 |
| *** kse has joined #openstack-nova | 09:12 | |
| BobBall | I assume you meant good rather than odd :D | 09:12 |
| bauzas | BobBall: label:Verified<=-1,<user> | 09:13 |
| BobBall | Curious - there aren't as many as I would expect | 09:13 |
| bauzas | BobBall: https://review.openstack.org/#/q/status:open+project:openstack/nova+label:Verified%253C%253D-1%252CXenProject-CI,n,z | 09:14 |
| BobBall | Don't hit us for 14 changes - looks like only a few failed xenproject CI for you? | 09:14 |
| BobBall | Anyway - doesn't matter | 09:14 |
| BobBall | it'll churn until they are done | 09:14 |
| *** alex_klimov has quit IRC | 09:15 | |
| *** alex_klimov has joined #openstack-nova | 09:15 | |
| BobBall | drat - seems that only the web interface supports that type of query... Can't add the username to the query over SSH | 09:16 |
| *** Marga_ has quit IRC | 09:17 | |
| bauzas | BobBall: I personnally prefer to use the REST API for calling Gerrit programatically | 09:17 |
| *** Marga_ has joined #openstack-nova | 09:17 | |
| bauzas | and that ^should be possible thru the HTTP calls | 09:18 |
| BobBall | Might look into that at some point. Found an easier way using scoreboard | 09:18 |
| openstackgerrit | John Garbutt proposed openstack/nova: objects: Hook migration object into Instance https://review.openstack.org/216733 | 09:18 |
| *** MarcF has quit IRC | 09:19 | |
| *** eglynn has joined #openstack-nova | 09:19 | |
| lyanchih | sahid: Hi, are you busy? I want to discuss with you about your last comment in https://review.openstack.org/#/c/200630/5/nova/virt/hardware.py | 09:19 |
| *** jamespd has quit IRC | 09:22 | |
| *** angdraug has joined #openstack-nova | 09:22 | |
| *** Marga_ has quit IRC | 09:24 | |
| BobBall | bauzas: All requeued. | 09:24 |
| *** Marga_ has joined #openstack-nova | 09:24 | |
| sahid | lyanchih: what is your question? | 09:25 |
| *** belmoreira has joined #openstack-nova | 09:26 | |
| *** jamespd has joined #openstack-nova | 09:28 | |
| *** Marga_ has quit IRC | 09:29 | |
| bauzas | BobBall: ack | 09:30 |
| bauzas | thanks | 09:30 |
| *** MarcF has joined #openstack-nova | 09:31 | |
| *** rmart04 has joined #openstack-nova | 09:37 | |
| lyanchih | sahid: Do you though we should verify whether host had adaptive pagesize at begin of filter schdule? | 09:38 |
| johnthetubaguy | danpb: could you rebase this chain please? would be great to attempt to get this in for liberty, if we can: https://review.openstack.org/#/c/189323 | 09:39 |
| rmart04 | Afternoon, could anyone be able to give me an idea of the flow of processes involved in terminating an ephemeral disk that is backed by ceph rbd backend? | 09:41 |
| *** husanu has quit IRC | 09:41 | |
| *** marzif has joined #openstack-nova | 09:42 | |
| *** Marga_ has joined #openstack-nova | 09:42 | |
| *** apuimedo has joined #openstack-nova | 09:42 | |
| sahid | lyanchih: you have two ways to fix that, catch exception MemoryPageSizeNotSupported in filter or early verify in filter if host can handle pagesize requested | 09:42 |
| sahid | i mean if host support pagesize requested | 09:43 |
| sahid | lyanchih: first solution is reasonable in my point of view | 09:43 |
| *** husanu has joined #openstack-nova | 09:43 | |
| danpb | johnthetubaguy: yeah doing that this morning | 09:45 |
| lyanchih | But the exception should’t been handle in filter for correct architecture. | 09:45 |
| johnthetubaguy | danpb: thank you, I think you are +Wed all the way through that now | 09:45 |
| danpb | just never ending merge conflicts | 09:46 |
| lyanchih | It will interrupt other cells match. | 09:47 |
| johnthetubaguy | ndipanov: I dropped some -2s on a few patches, so we can limit what we are waiting for, from a brief look, I think they could get moved into the bug fix, and I can drop those -2s, assuming there are no extra log messages (didn't look closely enough) | 09:47 |
| ndipanov | johnthetubaguy, looking | 09:48 |
| johnthetubaguy | ndipanov: thank you | 09:48 |
| *** y_sawai has joined #openstack-nova | 09:50 | |
| sahid | danpb: i would like to update realtime's spec according my comment on https://review.openstack.org/#/c/219692/ | 09:50 |
| *** lalitd has quit IRC | 09:51 | |
| ndipanov | johnthetubaguy, there is one string change in those 3 patches you -2ed that is not too important | 09:51 |
| johnthetubaguy | ndipanov: cool, could we just make it a debug log so it doesn't count? | 09:51 |
| *** y_sawai_ has joined #openstack-nova | 09:51 | |
| ndipanov | yes | 09:51 |
| johnthetubaguy | ndipanov: sweet | 09:52 |
| ndipanov | was just going to suggest that | 09:52 |
| sahid | danpb: please let me know when you have a moment if that makes sense for you too | 09:52 |
| ndipanov | will do it today | 09:52 |
| ndipanov | I see dansmith complained about ordering of those patches | 09:52 |
| ndipanov | so might have to change more with those actually | 09:52 |
| johnthetubaguy | ndipanov: cool, ping me when its updated with a bug or something, and I can drop the -2 | 09:52 |
| ndipanov | (though I don't fully agree with the arguments) | 09:52 |
| ndipanov | johnthetubaguy, awesome will do | 09:53 |
| *** y_sawai has quit IRC | 09:54 | |
| lyanchih | sahid: Oh!! I finally know what is your concern. You are concern about we should handle exception in filter file instead of in hardware file, aren’t you? | 09:56 |
| *** y_sawai_ has quit IRC | 09:56 | |
| johnthetubaguy | bauzas: I have a question for you about this one: https://review.openstack.org/#/c/188487/33/nova/objects/request_spec.py,cm | 09:57 |
| * bauzas looking | 09:57 | |
| *** jlanoux has quit IRC | 09:58 | |
| ndipanov | johnthetubaguy, so is this OK to go provided I get another +2 | 09:59 |
| bauzas | johnthetubaguy: well, not exactly, see the primitive keys and mem | 09:59 |
| ndipanov | I think jay just missed that one by mistake | 09:59 |
| bauzas | johnthetubaguy: I actually moved that logic from where it was previously | 09:59 |
| johnthetubaguy | ndipanov: which one is that? | 10:00 |
| ndipanov | lol I forgot to paste | 10:00 |
| ndipanov | https://review.openstack.org/#/c/218385/ | 10:00 |
| ndipanov | johnthetubaguy, ^ | 10:01 |
| johnthetubaguy | ndipanov: ah, yeah, I am trying to get the stuff thats critical for liberty-3 into the gate first, will try come back to that | 10:01 |
| johnthetubaguy | more worried about this one right this second: https://review.openstack.org/#/c/216733 | 10:01 |
| *** IanGovett has joined #openstack-nova | 10:02 | |
| ndipanov | johnthetubaguy, iyeah that one is really important | 10:02 |
| *** jlanoux has joined #openstack-nova | 10:02 | |
| *** Marga_ has quit IRC | 10:02 | |
| johnthetubaguy | yeah, give me a nudge if thats not in by tomorrow | 10:02 |
| johnthetubaguy | thats the second one | 10:02 |
| *** Marga_ has joined #openstack-nova | 10:02 | |
| sahid | lyanchih: hum well i think it could be better to catch ti in hardware.py#_numa_fit_instance_cell | 10:02 |
| johnthetubaguy | oops, I mean the bug one | 10:02 |
| ndipanov | johnthetubaguy, when do we actually want to cut rc1? | 10:02 |
| johnthetubaguy | ndipanov: on tuesday just gone, ideally | 10:03 |
| *** pixelb has joined #openstack-nova | 10:03 | |
| ndipanov | that may be difficult | 10:03 |
| ndipanov | :) | 10:03 |
| johnthetubaguy | ndipanov: has to be sometime today, at the latest | 10:03 |
| lyanchih | sahid: Oh, no!! I misunderstand again…..Sorry…:( | 10:04 |
| johnthetubaguy | ndipanov: I am being slap dash on the definition of today on purpose | 10:04 |
| *** MarcF has quit IRC | 10:04 | |
| sahid | lyanchih: wait let me check | 10:05 |
| bauzas | johnthetubaguy: so, reading again your comment, I think there is a difference between what's coming from the DB (the sahid method you gave) and what was primitived by to_primitive() | 10:05 |
| ndipanov | johnthetubaguy, yeah of course | 10:05 |
| bauzas | johnthetubaguy: like I commented, it's because we're flattening the objects when passing over the wire | 10:06 |
| bauzas | johnthetubaguy: which is quite an hybrid dict very different from what we have as the original legacy dict | 10:07 |
| johnthetubaguy | bauzas: yeah, I am just finding it hard to convince myself its the correct magic combination | 10:07 |
| bauzas | johnthetubaguy: the main problem is that obj_base.obj_to_primitive() doesn't actually provide the legacy primitive | 10:08 |
| bauzas | johnthetubaguy: hence that stuff | 10:08 |
| *** marzif has quit IRC | 10:08 | |
| johnthetubaguy | bauzas: for object fields, I guess? | 10:08 |
| *** xgsa has quit IRC | 10:08 | |
| johnthetubaguy | its not recursive or something? | 10:08 |
| sahid | lyanchih: ok so in method numa_fit_instance_to_host you should to add try..exept to around method _numa_fit_instance_cell - so if PageSizenotSupported is raised you stop process and return None | 10:09 |
| bauzas | johnthetubaguy: it's taking the fields and providing them as keys for the new dict | 10:09 |
| sahid | means if a topology is requested and we are not able to fit it, so filter will pass that host | 10:09 |
| bauzas | johnthetubaguy: if the primitive was not using the same field names, it's not good | 10:09 |
| *** karimb has joined #openstack-nova | 10:09 | |
| bauzas | johnthetubaguy: or if we hydrate an object differently from its primitive, it will return a totally different dict | 10:10 |
| bauzas | johnthetubaguy: I basically provided the same logic (and comment) as in https://github.com/openstack/nova/blob/master/nova/virt/hardware.py#L1200-L1222 | 10:10 |
| johnthetubaguy | bauzas: agreed, I think, just getting my head around where we are at, and where we are going | 10:10 |
| bauzas | johnthetubaguy: all of that is planned to be removed once we'll pass a ReqSpec object over the wire | 10:10 |
| johnthetubaguy | bauzas: honestly, I can't see us getting any of this merged at this point :( | 10:11 |
| bauzas | johnthetubaguy: jay was basically agreeing the whole series (but the last one), but I lost his voice since we discovered a last minute bug that needed some rebase | 10:12 |
| *** kbyrne has quit IRC | 10:12 | |
| johnthetubaguy | bauzas: yeah, its just the gate is a 12 hour queue, and we need to tag today :( | 10:12 |
| bauzas | yeah I understand that :( | 10:13 |
| bauzas | it's your call I mean | 10:13 |
| *** sc68cal has quit IRC | 10:13 | |
| bauzas | johnthetubaguy: are you planning some FFE discussion for https://wiki.openstack.org/wiki/Meetings/Nova#Agenda_for_next_meeting ? | 10:13 |
| johnthetubaguy | bauzas: yeah we could | 10:14 |
| johnthetubaguy | bauzas: although we don't totally get to decide | 10:14 |
| lyanchih | sahid: Oh…But we shouldn’t return None at there if exception was raised. Progess still need to check other cells. Therefore I can’t understand benefit to put catach block there. | 10:15 |
| bauzas | johnthetubaguy: MHO is that since it's a pretty big refactoring, I'm fine with considering to be merging either for Liberty or Mitaka as soon as it opens | 10:15 |
| *** kbyrne has joined #openstack-nova | 10:15 | |
| bauzas | johnthetubaguy: ideally, we should have tried to handle the RPC API change for Liberty, but that's not doable | 10:15 |
| bauzas | because it's the last bit in the chain | 10:16 |
| bauzas | johnthetubaguy: so I'd love to see some discussion around FFE | 10:16 |
| oomichi | johnthetubaguy: done, thanks in advance :) | 10:16 |
| johnthetubaguy | bauzas: so I am struggling with the end user benefit vs the risk of the size of the chain | 10:16 |
| bauzas | johnthetubaguy: and if some changes can land there, I'm fine | 10:16 |
| johnthetubaguy | bauzas: we have the RPC verison bump as an FFE | 10:16 |
| *** MarcF has joined #openstack-nova | 10:16 | |
| bauzas | johnthetubaguy: nope, I honestly don't want to have the RPC bump as a FEE, that's pretty huge | 10:17 |
| johnthetubaguy | bauzas: there isn't an alternative really, has to be post the other changes, and if we don't do it, we get loads of legacy we can't drop for another release | 10:17 |
| bauzas | johnthetubaguy: oh maybe I misunderstood | 10:18 |
| johnthetubaguy | would get a deadline of this time next week, for merge | 10:18 |
| bauzas | johnthetubaguy: you mean the usual Liberty RPC version bump, not the one I was doing ? | 10:18 |
| *** Sree_ has joined #openstack-nova | 10:18 | |
| johnthetubaguy | bauzas: its quite small at this point, bump instance and bump conductor RPC API is all I think | 10:19 |
| johnthetubaguy | bauzas: ah, yes, sorry, the general one | 10:19 |
| *** Sree_ is now known as Guest53061 | 10:19 | |
| bauzas | johnthetubaguy: oh ok, my bad, was talking about https://review.openstack.org/#/c/199205/27 | 10:19 |
| sahid | lyanchih: if a host cell does not support pagesize there are not need to continue the process (means all cells do not support pagesize) | 10:19 |
| *** Guest53061 has quit IRC | 10:19 | |
| bauzas | johnthetubaguy: in particular https://review.openstack.org/#/c/199205/27/nova/scheduler/manager.py,cm | 10:19 |
| johnthetubaguy | bauzas: ah, yeah, not that one | 10:19 |
| *** vivekd has quit IRC | 10:19 | |
| bauzas | johnthetubaguy: oh okay, then yeah it's totally fine to bump the RPC versions as usual | 10:20 |
| bauzas | johnthetubaguy: I should just rebase on top of them for https://review.openstack.org/#/c/199205/27/nova/scheduler/manager.py,cm | 10:20 |
| johnthetubaguy | bauzas: so I am thinking I should -2 your series, but we can get it back up the day mitaka opens? | 10:20 |
| bauzas | johnthetubaguy: provided we have a FFE discussion, that's understandable | 10:20 |
| bauzas | to see what we can still consider to merge | 10:21 |
| bauzas | (again, not speaking to merge https://review.openstack.org/#/c/199205/27/nova/scheduler/manager.py,cm that one is definitely -2) | 10:21 |
| johnthetubaguy | bauzas: what is the end user benifit of merging it, or release critical to merge? | 10:21 |
| bauzas | johnthetubaguy: that's not an enduser thing, rather a developer change, ie. mostly https://review.openstack.org/#/c/197912/29 | 10:22 |
| *** Sree has quit IRC | 10:22 | |
| *** asettle has quit IRC | 10:22 | |
| bauzas | johnthetubaguy: TBH, I'm not that attached to Liberty and I totally get your point | 10:22 |
| bauzas | johnthetubaguy: I just want to be sure we don't land such things : | 10:23 |
| bauzas | (in the intervalà) | 10:23 |
| bauzas | (sorry finding the change) | 10:23 |
| bauzas | that I found it => https://review.openstack.org/#/c/194059/ | 10:23 |
| bauzas | johnthetubaguy: ^ | 10:23 |
| BobBall | johnthetubaguy: vGPU. What's the preferred way to specify that we want a particular "thing" these days? I believe there has been some discussion around not using extra_specs for this recently? | 10:24 |
| bauzas | johnthetubaguy: I mean, we can fully defer the series to Mitaka, I'm okay with that | 10:24 |
| johnthetubaguy | BobBall: haven't really moved that discussion too far forward yet, only agreed it needs fixing | 10:24 |
| bauzas | johnthetubaguy: I just want to make sure that during the RC period and before https://review.openstack.org/#/c/197912/29 merges, we prevent such things | 10:24 |
| lxsli | john is very sad today :( | 10:25 |
| BobBall | OK, so if we were to put a spec in for K then we should start assuming to use extra_specs and see where the discussion goes? | 10:25 |
| bauzas | johnthetubaguy: you see my concern ? | 10:25 |
| johnthetubaguy | lxsli: yeah, I hate delaying things | 10:25 |
| bauzas | johnthetubaguy: https://review.openstack.org/#/c/194059/2/nova/compute/api.py,cm L980 is just a terrible thing if getting merged | 10:25 |
| johnthetubaguy | bauzas: we should block terrible things from merging | 10:26 |
| alex_xu | johnthetubaguy: oomichi are you around?, should we have a talk about https://review.openstack.org/#/c/217727/ | 10:26 |
| johnthetubaguy | bauzas: thats always true right? | 10:26 |
| bauzas | johnthetubaguy: yep :) | 10:26 |
| bauzas | johnthetubaguy: I just wanted to explain why I'm pretty concerned by merging the Spec series - because it automatically prevents that | 10:27 |
| bauzas | johnthetubaguy: but that's fair :) | 10:27 |
| bauzas | johnthetubaguy: you can throw the axe | 10:27 |
| *** ildikov has quit IRC | 10:28 | |
| johnthetubaguy | bauzas: I mean I know its important, hence its a priority, but I don't think we delay any bug fixes, or features too much by merging this in a month rather than today | 10:28 |
| bauzas | johnthetubaguy: yeah, you're totally right | 10:28 |
| bauzas | johnthetubaguy: just cut the wood | 10:28 |
| oomichi | alex_xu: sorry, I need to go home now. the time is already over I promissed with my family | 10:29 |
| bauzas | johnthetubaguy: thanks for your support btw. :) | 10:29 |
| alex_xu | oomichi: ok, no problem | 10:29 |
| *** oomichi is now known as oomichi_away | 10:29 | |
| johnthetubaguy | bauzas: so lets just be sure to get this up on the day mitaka opens, and merge it pre summit, I think thats a good way to go here | 10:29 |
| bauzas | johnthetubaguy: agreed | 10:30 |
| bauzas | johnthetubaguy: you're right, the balance risks vs. benefits is too high | 10:30 |
| johnthetubaguy | bauzas: the example is ndipanov had a thing thats a bug fix, so we just called it a bug fix, and trying to merge the feature like bit before we cut, that patch we might give a FFE to, if its the only one waiting | 10:30 |
| bauzas | johnthetubaguy: yeah to that | 10:31 |
| bauzas | johnthetubaguy: if we want to properly 'do' things, there is no reason to rush | 10:31 |
| *** links has joined #openstack-nova | 10:31 | |
| johnthetubaguy | rushing is very rarely the correct answer | 10:32 |
| johnthetubaguy | its a useful tool, but its a very blunt one | 10:32 |
| bauzas | :) | 10:33 |
| *** xgsa has joined #openstack-nova | 10:34 | |
| bauzas | johnthetubaguy: FWIW, -W'd the first change of the series until either you throw -2 or Mitaka opens | 10:36 |
| openstackgerrit | George Peristerakis proposed openstack/python-novaclient: Search the list of servers with unicode text. https://review.openstack.org/200649 | 10:36 |
| johnthetubaguy | bauzas: do you want the blueprint moved to mitaka or partially complete for liberty, I am thinking the later (but that means updating the blueprint references on the patches) | 10:37 |
| bauzas | johnthetubaguy: Partially Complete is better because now the community has to deal with a ReqSpec object | 10:38 |
| bauzas | johnthetubaguy: which means dependencies etc. | 10:38 |
| johnthetubaguy | bauzas: yeah, thats what I am thinking, cool | 10:38 |
| bauzas | johnthetubaguy: so I guess I should open another BP for Mitaka and point to another spec to fast-approve, right? | 10:39 |
| lyanchih | sahid: How did you know all other host cells do not support pagesize if a instance cell can’t find match pagesize with one of host cell? | 10:39 |
| johnthetubaguy | bauzas: yeah | 10:39 |
| bauzas | johnthetubaguy: and then update the whole series to match with that new ref? | 10:39 |
| lyanchih | sahid: Sorry, I still can’t understood. | 10:39 |
| bauzas | johnthetubaguy: okay, gotcha, will do that today since mitaka-specs is open IIRC | 10:39 |
| bauzas | johnthetubaguy: thanks for your help | 10:40 |
| *** smatzek has joined #openstack-nova | 10:40 | |
| johnthetubaguy | bauzas: no problems | 10:41 |
| *** claudiub has joined #openstack-nova | 10:44 | |
| *** Kennan2 is now known as Kennan_on_vacati | 10:44 | |
| johnthetubaguy | danpb: so having said all that earlier, the image_meta thing is now about the only remaining thing that is blocking us tagging liberty-3 (assuming ndipanov patch can get an FFE, or merges in time), just looking through yours trying to work out the best path forward... | 10:45 |
| *** Kennan_on_vacati is now known as Kennan_Vacation | 10:45 | |
| *** vivekd has joined #openstack-nova | 10:45 | |
| *** otter768_ has quit IRC | 10:45 | |
| * alex_xu back to public vacation.... | 10:45 | |
| *** aix has quit IRC | 10:46 | |
| *** marzif has joined #openstack-nova | 10:47 | |
| *** ZZelle has quit IRC | 10:48 | |
| vponomaryov | Guys from core team who is able to tag a novaclient? it is required for unblock whole Manila project. See bug #1491325 | 10:48 |
| openstack | bug 1491325 in python-novaclient "nova api v2.1 does not allow to use autodetection of volume device path" [High,Fix committed] https://launchpad.net/bugs/1491325 - Assigned to Matt Riedemann (mriedem) | 10:48 |
| *** ZZelle has joined #openstack-nova | 10:48 | |
| *** jcoufal has joined #openstack-nova | 10:51 | |
| openstackgerrit | garyk proposed openstack/nova: VMware: add method for getting hosts attached to datastore https://review.openstack.org/220085 | 10:52 |
| johnthetubaguy | danpb: I am thinking this one blocks the possibility of a FFE: https://review.openstack.org/#/c/189685/27 | 10:53 |
| garyk | johnthetubaguy: why should that be blocked? | 10:54 |
| *** deepakcs has quit IRC | 10:54 | |
| garyk | drivers out of tree should adapt or die | 10:54 |
| garyk | if the drivers has decent unit tests then they will find the issues as soon as this lands and they will fix it | 10:55 |
| johnthetubaguy | garyk: thats not the bit that worries me, its more the size of it and risk of that creating a bug we don't find in regular testing | 10:55 |
| garyk | that is what we do in nuetron | 10:55 |
| garyk | we still have another month to test. i am really not that concerned about it | 10:56 |
| garyk | it has been in review for months and i have tested this many times | 10:56 |
| johnthetubaguy | if we had proper coverage for our use of image properties, I would be fine with it | 10:58 |
| johnthetubaguy | at this point its delaying the tagging of liberty-3 | 10:58 |
| johnthetubaguy | and has been in merge conflict since monday, its now were the time bound thing applies, so we leave time for bug fixes | 10:59 |
| garyk | johnthetubaguy: that sounds reasonable then | 11:00 |
| johnthetubaguy | we can consider an exception for this one later | 11:01 |
| openstackgerrit | Michael Still proposed openstack/nova: [libvirt] Move cleanup of imported files to imagebackend https://review.openstack.org/220018 | 11:02 |
| johnthetubaguy | just not sure its a strong case, given its lack of user impact, and slightly higher than average risk | 11:02 |
| danpb | johnthetubaguy: i'm not sure what you mean by that ? | 11:03 |
| *** y_sawai has joined #openstack-nova | 11:03 | |
| *** angdraug has quit IRC | 11:03 | |
| johnthetubaguy | danpb: so its an important change, and I want it merged, just at this point it seems like its medium risk, but low immediate user benefit, so merging this in a month might be the best trade off here | 11:04 |
| danpb | you mean merging it in M-1 ? | 11:04 |
| johnthetubaguy | danpb: yeah, pre summit | 11:04 |
| *** y_sawai has quit IRC | 11:05 | |
| danpb | johnthetubaguy: are you referring just to the compute manager one, or the whole series ? | 11:05 |
| danpb | johnthetubaguy: i'd like to at least get the ironic + hyperv conversions merged | 11:05 |
| danpb | so that all in-tree virt drivers are using ImageMeta object | 11:05 |
| *** y_sawai has joined #openstack-nova | 11:05 | |
| johnthetubaguy | danpb: so at this point, I am thinking the whole series, but open to ideas | 11:06 |
| danpb | after those two, i'm fine if the rest wait another cycle | 11:06 |
| johnthetubaguy | danpb: cool, that sounds doable to me | 11:06 |
| danpb | i'm just waiting for unit tests to complete locally before psting a refresh | 11:06 |
| johnthetubaguy | danpb: cool | 11:06 |
| danpb | i'll put a temporary -2 on the ones to hold | 11:07 |
| johnthetubaguy | danpb: OK, so I will -2 the back end, and FFE the front end, does that work? | 11:07 |
| johnthetubaguy | no need for the -2 on the front ones, we can assume a FFE on those, assuming no string changes are in there | 11:07 |
| danpb | i stuck -2 on this one https://review.openstack.org/#/c/189685/ | 11:08 |
| *** xgsa has quit IRC | 11:08 | |
| danpb | that lets the ironci, hyperv & config drive ones merge and blocks the other ones | 11:08 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: ironic: convert driver to use nova.objects.ImageMeta https://review.openstack.org/189362 | 11:09 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: objects: convert config drive to use ImageMeta object https://review.openstack.org/189323 | 11:09 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: hyperv: convert driver to use nova.objects.ImageMeta https://review.openstack.org/189302 | 11:09 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: objects: add virtual 'image_meta' property to Instance object https://review.openstack.org/190145 | 11:09 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: libvirt: convert to use instance.image_meta property https://review.openstack.org/190146 | 11:09 |
| openstackgerrit | Daniel Berrange proposed openstack/nova: compute: convert manager to use nova.objects.ImageMeta https://review.openstack.org/189685 | 11:09 |
| johnthetubaguy | danpb: sounds good to me, thanks | 11:09 |
| ndipanov | johnthetubaguy, https://review.openstack.org/#/c/216733/7 failed the gate :( | 11:09 |
| danpb | wow, my pending nova patch queue is empty besides this series - been a long time since i had near zero pending reviews | 11:09 |
| johnthetubaguy | ndipanov: so I will get that an FFE, but its worth a recheck I guess | 11:10 |
| *** y_sawai has quit IRC | 11:10 | |
| ndipanov | I did it already | 11:11 |
| *** apuimedo is now known as apuimedo|lunch | 11:11 | |
| *** kse has quit IRC | 11:11 | |
| johnthetubaguy | ndipanov: sweet, thanks | 11:11 |
| *** Marga_ has quit IRC | 11:12 | |
| *** tojuvone has quit IRC | 11:14 | |
| *** vivekd has quit IRC | 11:14 | |
| *** aix has joined #openstack-nova | 11:15 | |
| *** xgsa has joined #openstack-nova | 11:15 | |
| alex_xu | bauzas: are you around? | 11:17 |
| bauzas | alex_xu: should be refueling my stomach but still there | 11:17 |
| *** shoutm has joined #openstack-nova | 11:17 | |
| *** vivekd has joined #openstack-nova | 11:18 | |
| *** shoutm_ has quit IRC | 11:18 | |
| *** y_sawai has joined #openstack-nova | 11:18 | |
| alex_xu | bauzas: just want to ask, whether we allow out-of-tree filters, and do we allow user extend the accepted value of scheduler-hints | 11:18 |
| bauzas | alex_xu: speaking about the API hints then ? | 11:19 |
| *** ishant has quit IRC | 11:19 | |
| *** lyanchih has quit IRC | 11:19 | |
| bauzas | alex_xu: because we just relaxed the hints in the Spec obj | 11:19 |
| alex_xu | bauzas: yea | 11:19 |
| *** xgsa has quit IRC | 11:20 | |
| alex_xu | bauzas: so for now, we still expect the API accept user defined scheduler-hints? | 11:20 |
| bauzas | alex_xu: so yup, we do easily allow use of out-of-tree filters, see eg. https://github.com/stackforge/blazar-nova/blob/master/blazarnova/scheduler/filters/blazar_filter.py | 11:20 |
| bauzas | alex_xu: in that ^ filter, it defines a new hint | 11:20 |
| johnthetubaguy | alex_xu: did you see this bug: https://bugs.launchpad.net/nova/+bug/1491511 and this one: https://bugs.launchpad.net/python-novaclient/+bug/1491579 will chat about those in the nova meeting later | 11:20 |
| openstack | Launchpad bug 1491511 in OpenStack Compute (nova) "Behavior change with latest nova paste config" [Undecided,Confirmed] | 11:20 |
| openstack | Launchpad bug 1491579 in python-novaclient "against all sanity, nova needs to work around broken public clouds" [Critical,Confirmed] | 11:20 |
| alex_xu | bauzas: :) | 11:21 |
| bauzas | johnthetubaguy: I have ideas for both of them | 11:21 |
| sdague | johnthetubaguy: I'm working on the novaclient one right now | 11:21 |
| bauzas | johnthetubaguy: but I leave the meeting run | 11:21 |
| alex_xu | johnthetubaguy: not yet, will check now | 11:21 |
| johnthetubaguy | sdague: awesome | 11:21 |
| johnthetubaguy | bauzas: comments on the bug welcome | 11:21 |
| *** eharney has joined #openstack-nova | 11:21 | |
| johnthetubaguy | alex_xu: you should enjoy your holiday, but wanted to give you a heads up | 11:22 |
| bauzas | johnthetubaguy: I need to eat but will do | 11:22 |
| alex_xu | johnthetubaguy: it's fine, actually I plan back to work few hours | 11:22 |
| bswartz | hey guys, the manila gate is still broken until the latest commit for novaclient (eaf1e56b210c94e7d26ca99b5931f3688cef984c) gets a tag | 11:23 |
| bswartz | is there anyone here who can push a tag to novaclient? | 11:23 |
| *** y_sawai has quit IRC | 11:23 | |
| johnthetubaguy | sdague: technically speaking, on the rackspace side, we have repose (a proxy) that does rate limiting and auth, sat in front of all of the API, much to my annoyance | 11:23 |
| openstackgerrit | Alex Xu proposed openstack/nova: api: allow any scheduler hints https://review.openstack.org/217727 | 11:23 |
| johnthetubaguy | bswartz: do you have the bug that we fixed? | 11:23 |
| alex_xu | johnthetubaguy: anyway I updated a version for allow scheduler hints ^ | 11:24 |
| johnthetubaguy | oh wait, you gave me the commit | 11:24 |
| bswartz | johnthetubaguy: the bug affects us -- but without a tag we aren't picking it up | 11:24 |
| bauzas | alex_xu: thanks | 11:24 |
| alex_xu | bauzas: np, if you like, hope you can take a look at that patch also | 11:25 |
| *** takedakn has joined #openstack-nova | 11:25 | |
| bauzas | alex_xu: will do, it's pretty critical | 11:25 |
| sdague | johnthetubaguy: right, however, the blocking of valid API urls gets me really concerned | 11:25 |
| sdague | because wtf are we doing if we have basic things like version info in our API, and it gets blocked by service providers. | 11:26 |
| *** sayali has quit IRC | 11:26 | |
| *** alex_klimov has quit IRC | 11:26 | |
| johnthetubaguy | sdague: if you pass a valid token, we will return the version, I believe | 11:26 |
| sdague | nope | 11:27 |
| bauzas | sdague: MHO is that RAX is maybe caring about not exposing that info explicitely | 11:27 |
| bauzas | sdague: maybe having a content negociation implicitely is okay | 11:27 |
| *** pixelb has quit IRC | 11:27 | |
| bauzas | sdague: meaning that instead of answering to a request with a specific version number, we should find some way to hide the negociation | 11:27 |
| *** venkat_p has quit IRC | 11:27 | |
| johnthetubaguy | sdague: hmm, so I don't think the block was intentional on our side at least, thats just broken then :( | 11:28 |
| johnthetubaguy | sdague: I think this is something we need to get to the defcore folks as required | 11:28 |
| sdague | bauzas: I am seriously not going to play a guessing game with a black box that has blocked parts of our code that were designed for a reason | 11:28 |
| bauzas | johnthetubaguy: from an operator POV, that's quite understandable that you want to block the version exposure of your webservices | 11:28 |
| bauzas | because that's vector #1 for security flaws | 11:29 |
| *** MarcF has quit IRC | 11:29 | |
| sdague | bauzas: we're talking about API version. How the f* can you write against an API if the API hides it's version from you. | 11:29 |
| openstackgerrit | Vipin Balachandran proposed openstack/nova: Detach and terminate conn if Cinder attach fails https://review.openstack.org/186742 | 11:30 |
| bauzas | sdague: yeah I understood | 11:30 |
| bauzas | sdague: I'm just thinking about a possible handshake | 11:30 |
| sdague | bauzas: which they are blocking | 11:30 |
| johnthetubaguy | bswartz: so I need to get some food, but I will reach out to some folks to see where we are at with client releases | 11:30 |
| sdague | we're trying to do that handshake | 11:30 |
| johnthetubaguy | bauzas: thats not reasonable | 11:31 |
| johnthetubaguy | bauzas: for the reasons sdague was saying really | 11:31 |
| * johnthetubaguy really needs to go get some food | 11:31 | |
| bauzas | +1 | 11:31 |
| sdague | bauzas: so, right now, all you're going to do is piss me off if you keep suggestiong that it's acceptable to block the version GET request | 11:31 |
| sdague | because it's not | 11:31 |
| bauzas | sdague: okay, lemme rephrase | 11:31 |
| sdague | it's a fundamental part of openstack api infrastructure since long before I got here | 11:32 |
| bauzas | sdague: yeah to that | 11:32 |
| sdague | anyway, I'm going to go back to working on the work around code | 11:32 |
| *** Marga_ has joined #openstack-nova | 11:35 | |
| *** ociuhandu has quit IRC | 11:36 | |
| *** takedakn has quit IRC | 11:36 | |
| *** takedakn has joined #openstack-nova | 11:36 | |
| *** thorst has joined #openstack-nova | 11:37 | |
| *** ildikov has joined #openstack-nova | 11:39 | |
| *** y_sawai has joined #openstack-nova | 11:39 | |
| *** takedakn1 has joined #openstack-nova | 11:41 | |
| *** y_sawai has quit IRC | 11:43 | |
| *** MarcF has joined #openstack-nova | 11:43 | |
| *** y_sawai has joined #openstack-nova | 11:43 | |
| *** pixelb has joined #openstack-nova | 11:44 | |
| alex_xu | sdague: https://bugs.launchpad.net/python-novaclient/+bug/1491579 is because we need token when execute version-list? | 11:44 |
| openstack | Launchpad bug 1491579 in python-novaclient "against all sanity, nova needs to work around broken public clouds" [Critical,Confirmed] | 11:44 |
| *** takedakn has quit IRC | 11:44 | |
| *** takedakn1 is now known as takedakn | 11:44 | |
| sdague | alex_xu: nope | 11:45 |
| sdague | we've gotten the token already | 11:45 |
| sdague | it's not even getting to nova | 11:46 |
| sdague | at least as far as we could tell | 11:46 |
| *** lucasagomes is now known as lucas-hungry | 11:46 | |
| openstackgerrit | sulochan-acharya proposed openstack/nova: xapi: cleanup volume sr on live migration rollback https://review.openstack.org/217105 | 11:46 |
| alex_xu | sdague: strange, my local version works, but let me update latest code | 11:46 |
| sdague | alex_xu: against rax? | 11:47 |
| *** y_sawai_ has joined #openstack-nova | 11:47 | |
| alex_xu | sdague: no, just upstream code | 11:47 |
| sdague | right, this all works with upstream code | 11:47 |
| sdague | the issue is things service providers had done in their deploys | 11:47 |
| alex_xu | sdague: strange... | 11:48 |
| *** y_sawai has quit IRC | 11:48 | |
| *** bkopilov has quit IRC | 11:51 | |
| *** y_sawai_ has quit IRC | 11:51 | |
| *** shaohe_feng has joined #openstack-nova | 11:52 | |
| *** takedakn has quit IRC | 11:52 | |
| *** Sree has joined #openstack-nova | 11:52 | |
| *** ociuhandu has joined #openstack-nova | 11:55 | |
| *** diogogmt has joined #openstack-nova | 11:56 | |
| openstackgerrit | George Peristerakis proposed openstack/python-novaclient: Search the list of servers with unicode text. https://review.openstack.org/200649 | 11:57 |
| *** alexpilotti_ has joined #openstack-nova | 11:58 | |
| *** Marga_ has quit IRC | 12:01 | |
| sdague | hmmmmm... I wonder if there is a slightly simpler bit here | 12:03 |
| sdague | johnthetubaguy: any chance you can get me rax credentials for testing a thing? | 12:03 |
| *** alex_klimov has joined #openstack-nova | 12:05 | |
| *** hparekh has quit IRC | 12:07 | |
| *** shaohe_feng has quit IRC | 12:11 | |
| *** shaohe_feng has joined #openstack-nova | 12:11 | |
| *** bfic has joined #openstack-nova | 12:11 | |
| sdague | ah, it looks like there is a simpler fix, maybe | 12:12 |
| *** marzif has quit IRC | 12:14 | |
| *** marzif has joined #openstack-nova | 12:15 | |
| *** raildo-afk is now known as raildo | 12:18 | |
| *** deepthi has quit IRC | 12:19 | |
| johnthetubaguy | sdague: honestly not sure how to do that safely, but let me know if I can test something on my box | 12:22 |
| sdague | well, I tested a fix with hp cloud | 12:22 |
| sdague | that's the only place I have credentials, and this worked there | 12:22 |
| sdague | let me push the patch and just see if you can do a nova list on rax after that | 12:23 |
| sdague | I'm just working on a commit message | 12:23 |
| johnthetubaguy | sdague: cool, happy to try that | 12:24 |
| *** xyang1 has joined #openstack-nova | 12:24 | |
| sdague | also, in general, it would be useful if rax gave some of us free accounts again :P | 12:25 |
| *** aysyd has joined #openstack-nova | 12:25 | |
| sdague | to actually try stuff like this. | 12:25 |
| johnthetubaguy | sdague: yeah, I know mikal was pushing on that, not sure if he has got further with that than I managed before | 12:25 |
| johnthetubaguy | sdague: let me drop him an email before I forget | 12:26 |
| *** diogogmt has quit IRC | 12:27 | |
| *** VikasC has quit IRC | 12:28 | |
| alexpilotti_ | danpb: hi, small semi-nit question: may I ask why did you remove a constant and hardcoded the string literal here 'hw_machine_type'? https://review.openstack.org/#/c/189302/25/nova/virt/hyperv/vmops.py | 12:28 |
| *** ttrumm has joined #openstack-nova | 12:28 | |
| *** shaohe_feng has quit IRC | 12:28 | |
| openstackgerrit | Sean Dague proposed openstack/python-novaclient: Don't assume oscomputeversions is correctly deployed https://review.openstack.org/220111 | 12:28 |
| danpb | alexpilotti_: it is a pointless constant | 12:28 |
| *** cfarquhar has joined #openstack-nova | 12:28 | |
| *** cfarquhar has joined #openstack-nova | 12:28 | |
| danpb | and it obscures what image property we're accessing | 12:29 |
| sdague | johnthetubaguy: ok, so if you can pull https://review.openstack.org/220111 install into a venv, and do "nova list" with your rax account, that would be great. | 12:29 |
| sdague | I have a bunch of bits of a more extensive fix as well, however this might be good enough | 12:30 |
| *** lucas-hungry is now known as lucasagomes | 12:30 | |
| *** diogogmt has joined #openstack-nova | 12:31 | |
| openstackgerrit | Alexis Lee proposed openstack/nova: WIP Unify on _schedule_instances https://review.openstack.org/219761 | 12:31 |
| alexpilotti_ | danpb: hmmm ok, wonder where the fine line between “pointless” and “non pointless” constants resides then :-) | 12:31 |
| alexpilotti_ | danpb: the question was mostly to see if there are some style guidelines around constants that we should follow | 12:32 |
| danpb | this piece of code is essentially doing an object attribute access | 12:32 |
| *** burgerk has joined #openstack-nova | 12:33 | |
| *** baoli has joined #openstack-nova | 12:33 | |
| danpb | we don't need to define constants for the attributes, as they're already validated by the object and it just obscures what attribute you're accessing | 12:33 |
| *** stefan_amann has quit IRC | 12:34 | |
| alexpilotti_ | ok, so it’s just moving avay from the dict “get” to a more object based syntax, if I got it right | 12:35 |
| *** stefan_amann has joined #openstack-nova | 12:35 | |
| johnthetubaguy | sdague: pulling that down | 12:35 |
| alexpilotti_ | something like this wouldnt work? image_prop_vm = image_meta.properties.hw_machine_type or default_vm_gen | 12:35 |
| alexpilotti_ | didnt check if the object raises if a nonexistent property is being accessed | 12:37 |
| *** baoli_ has joined #openstack-nova | 12:37 | |
| danpb | alexpilotti_: no, objects will raise an attribute error if hw_machine_type isn't set | 12:38 |
| *** baoli has quit IRC | 12:38 | |
| *** diogogmt has quit IRC | 12:40 | |
| *** dims has joined #openstack-nova | 12:41 | |
| *** sayali has joined #openstack-nova | 12:42 | |
| *** y_sawai has joined #openstack-nova | 12:43 | |
| sdague | johnthetubaguy: how's it looking? | 12:44 |
| openstackgerrit | garyk proposed openstack/nova: VMware: ensure volume is accessible by instance https://review.openstack.org/218639 | 12:45 |
| *** diogogmt has joined #openstack-nova | 12:46 | |
| *** shoutm_ has joined #openstack-nova | 12:47 | |
| *** shoutm has quit IRC | 12:48 | |
| *** y_sawai has quit IRC | 12:48 | |
| *** jcoufal has quit IRC | 12:49 | |
| *** dguitarbite has quit IRC | 12:51 | |
| *** sayali has quit IRC | 12:52 | |
| *** dguitarbite has joined #openstack-nova | 12:52 | |
| sdague | johnthetubaguy: ok, well I'm stepping away from the computer for a bit. Please let me know how that goes. Will also talk about it in the meeting in an hour | 12:52 |
| johnthetubaguy | sdague: ack | 12:53 |
| johnthetubaguy | sdague: my environment is giving me problems, working through those | 12:53 |
| *** lbeliveau has joined #openstack-nova | 12:54 | |
| *** chutwig has joined #openstack-nova | 12:55 | |
| *** sayali has joined #openstack-nova | 12:56 | |
| *** jcoufal has joined #openstack-nova | 13:02 | |
| *** vladikr has joined #openstack-nova | 13:08 | |
| openstackgerrit | Rajesh Tailor proposed openstack/nova: Fix order of arguments in assertEqual https://review.openstack.org/220134 | 13:11 |
| *** vladikr has quit IRC | 13:12 | |
| *** diogogmt_ has joined #openstack-nova | 13:12 | |
| bswartz | johnthetubaguy: any chance we can get a tag for novaclient pushed this morning? | 13:12 |
| *** vladikr has joined #openstack-nova | 13:12 | |
| *** diogogmt has quit IRC | 13:13 | |
| *** diogogmt_ is now known as diogogmt | 13:13 | |
| *** pixelbeat has joined #openstack-nova | 13:13 | |
| *** edmondsw has joined #openstack-nova | 13:13 | |
| *** pixelb has quit IRC | 13:13 | |
| *** dustins has joined #openstack-nova | 13:15 | |
| *** ivasev has joined #openstack-nova | 13:16 | |
| *** sayali has quit IRC | 13:16 | |
| openstackgerrit | Pranali Deore proposed openstack/nova: Fix order of arguments in assertEqual https://review.openstack.org/218817 | 13:16 |
| johnthetubaguy | bswartz: I think so, just trying to see if we have fixed some other things also, to save us having to tag tomorrow as well | 13:16 |
| *** rlrossit has joined #openstack-nova | 13:18 | |
| *** ctrath has joined #openstack-nova | 13:19 | |
| *** xgsa has joined #openstack-nova | 13:19 | |
| *** sayali has joined #openstack-nova | 13:22 | |
| ndipanov | dansmith, https://review.openstack.org/#/c/214713/ see my comment when you get the chance | 13:22 |
| bswartz | johnthetubaguy: ty | 13:22 |
| *** sayali has quit IRC | 13:24 | |
| alex_xu | johnthetubaguy: sdague ...strange, I didn't found any behavior change which describe in https://bugs.launchpad.net/nova/+bug/1491511 | 13:24 |
| openstack | Launchpad bug 1491511 in OpenStack Compute (nova) "Behavior change with latest nova paste config" [Undecided,Confirmed] | 13:24 |
| johnthetubaguy | sdague: that fix is working for me | 13:25 |
| openstackgerrit | Pranali Deore proposed openstack/nova: Fix order of arguments in assertEqual https://review.openstack.org/219265 | 13:25 |
| alex_xu | delete floating ip will return fault if pass an ip address in legacy v2 also | 13:25 |
| alex_xu | and we didn't have any related change recently | 13:25 |
| johnthetubaguy | sdague: oh dear, digging, I think my environment is confusing me | 13:25 |
| *** sayali has joined #openstack-nova | 13:27 | |
| alaski | johnthetubaguy: sdague there's been some discussion on this internally and I'm pushing towards a resolution | 13:28 |
| alaski | our Repose proxy has a regex of allowable URLs it will pass through, which requires a tenant-id | 13:28 |
| alaski | which is of course incorrect | 13:29 |
| alaski | modifying the regex to allow the call then causes an error with Nova because of some interaction with filling out the context with auth info | 13:29 |
| alaski | so we're chasing that now | 13:29 |
| johnthetubaguy | alaski: yes, that came up a little while back, its incorrectly doing some RBAC | 13:29 |
| *** amotoki has joined #openstack-nova | 13:29 | |
| johnthetubaguy | alaski: current just testing sdague's workaround, and my environment is messing my about, but I think I get a 400 error with this in place | 13:30 |
| mordred | sdague: testing your workaround on the other clouds | 13:31 |
| johnthetubaguy | alaski: thanks for pushing on the proper fix internally, thats would be good to fix | 13:31 |
| sdague | mordred: thanks | 13:31 |
| *** VW_ has joined #openstack-nova | 13:32 | |
| sdague | well, I've got the beginnings of a much more extensive work around if we end up thinking we need that | 13:32 |
| sdague | johnthetubaguy: so... what's the verdict? I see conflicting statements of it working and then not working | 13:33 |
| alaski | a question that this raises is how to do quotas/rate-limiting on this call. every other call has a tenant so it can be scoped that way, this is just an open call | 13:34 |
| alaski | limiting by IP has been suggested, which I don't like | 13:34 |
| dansmith | ndipanov: honestly, I don't get it | 13:34 |
| *** jecarey has joined #openstack-nova | 13:34 | |
| johnthetubaguy | sdague: honestly, I am just confused, I am now seeing it succeed with the broken version :S | 13:35 |
| mordred | it fixes auro. runabove still hangs. unitedstack continues to work (as expected) | 13:35 |
| mordred | johnthetubaguy: oh good! | 13:35 |
| mordred | which - rax? | 13:35 |
| johnthetubaguy | yeah, I think I just screwed up my venv though... | 13:35 |
| dansmith | ndipanov: one of the patches is 40 line.. squashing that into the 100 line patch makes a 140 line patch. that's not "way more difficult to review" IMHO | 13:35 |
| dansmith | ndipanov: if you're worried, you could pull out nearly 40 lines in the conductor bits that actually pass the information, but I don't think it's necessary | 13:36 |
| mordred | I still get 401 from rax with sean's fix. 2.26 works for me. 2.27 401 | 13:36 |
| sdague | alaski: so, honestly, this url always returns the same until you upgrade | 13:36 |
| sdague | it should just be cached hard by an upper layer if you are concerned | 13:36 |
| mordred | johnthetubaguy: fwiw, that's against DFW - I have not checked other rax regions | 13:36 |
| dansmith | ndipanov: I can't recall when we've ever done an rpc change like this, and I don't think that we need to do it here... | 13:36 |
| ndipanov | dansmith, well what worries me is that I'll do the factor out a submethod and changes in the same PS | 13:37 |
| johnthetubaguy | mordred: I was against IAD, but yeah, I saw 2.27 fail, then 2.26 work, now I am seeing everything work, somehow | 13:37 |
| ndipanov | but if you're happy to review it like that - I'm cool | 13:37 |
| *** ijuwang has joined #openstack-nova | 13:37 | |
| mordred | johnthetubaguy: can I have your computer? | 13:37 |
| mordred | :) | 13:37 |
| johnthetubaguy | :) | 13:37 |
| alaski | sdague: I personally am not concerned, but those who manage repose are. caching is a reasonable answer though | 13:37 |
| openstackgerrit | Rajesh Tailor proposed openstack/nova: Fix order of arguments in assertEqual https://review.openstack.org/220142 | 13:37 |
| sdague | johnthetubaguy: well, run "nova --debug list" and make sure that the GET /v2/ call is getting made | 13:37 |
| dansmith | ndipanov: I'm not sure what you mean.. neither of these patches factor out a submethod, do they? | 13:38 |
| sdague | or get someone to give me a free rax account and I can poke myself :) | 13:38 |
| johnthetubaguy | sdague: its sending a null API key, I got that messed up somehow | 13:38 |
| sdague | johnthetubaguy: wait, so if you don't set the API key, it all works? | 13:39 |
| mordred | sdague: I'm only testing list - not version-list btw - any benefit in testing version list? | 13:39 |
| ndipanov | dansmith, they do | 13:39 |
| sdague | mordred: the change doesn't impact that call | 13:39 |
| mordred | kk | 13:39 |
| ndipanov | dansmith, oh wait | 13:39 |
| mordred | I thought so | 13:39 |
| ndipanov | that merged | 13:39 |
| ndipanov | hmmm | 13:39 |
| mordred | just making sure | 13:39 |
| sdague | yep, np | 13:39 |
| openstackgerrit | garyk proposed openstack/nova: VMware: ensure volume is accessible by instance https://review.openstack.org/218639 | 13:39 |
| openstackgerrit | Chuck Carmack proposed openstack/nova: Remove the blacklisted nova-cells shelve tests https://review.openstack.org/219778 | 13:40 |
| mordred | sdague: also fixes dreamhost | 13:40 |
| ndipanov | dansmith, never mind | 13:40 |
| sdague | mordred: so are we down to all the clouds that fail version-list are the only ones that fail now? | 13:40 |
| *** links has quit IRC | 13:40 | |
| mordred | no | 13:41 |
| *** lyanchih has joined #openstack-nova | 13:41 | |
| *** burgerk has quit IRC | 13:41 | |
| johnthetubaguy | sdague: sorry, too many variables here, something is messed up | 13:41 |
| mordred | sdague: the patch fixes auro, dreamhost and hp | 13:42 |
| mordred | sdague: the patch does not fix runabove or ra | 13:42 |
| mordred | rax | 13:42 |
| mordred | sdague: rax throws the 401, runabove hangs | 13:42 |
| sdague | right, but those fail nova version-list, right? | 13:42 |
| *** ccarmack has joined #openstack-nova | 13:42 | |
| mordred | oh! yes. sorry, I misread my original notes | 13:42 |
| mordred | you are correct | 13:42 |
| mordred | this fixes everyone who had not broken version-list | 13:43 |
| mordred | and the ones who have fail this in the same way they fail version-list | 13:43 |
| openstackgerrit | Rajesh Tailor proposed openstack/nova: Fix order of arguments in assertEqual https://review.openstack.org/220143 | 13:43 |
| *** shaohe_feng has joined #openstack-nova | 13:43 | |
| mordred | sdague: so I think it's a win for the new code | 13:44 |
| *** y_sawai has joined #openstack-nova | 13:44 | |
| *** moshele has quit IRC | 13:44 | |
| johnthetubaguy | sdague: yeah, I am seeing the 401 with your change now I have my more sorted | 13:45 |
| johnthetubaguy | sdague: I think its most likely the repose issue that alaski was mentioning | 13:45 |
| *** atuvenie has joined #openstack-nova | 13:46 | |
| sdague | ok, so we should definitely land this fix then, it makes it better for the majority | 13:46 |
| johnthetubaguy | sdague: +1 that | 13:46 |
| dansmith | if we get a reliable 401 | 13:46 |
| sdague | then we should decide what level of additional work around we want | 13:46 |
| dansmith | can we just treat that as 2.0 and barf a warning? | 13:46 |
| sdague | dansmith: yeh, I was thinking that | 13:46 |
| johnthetubaguy | sdague: should we wait for this to merge before tagging again? | 13:46 |
| sdague | johnthetubaguy: well, I'd like to | 13:47 |
| sdague | the gate is 10 hours deep right now | 13:47 |
| sdague | so that's a decision to sort out | 13:47 |
| johnthetubaguy | sdague: yeah | 13:47 |
| *** bfic has left #openstack-nova | 13:47 | |
| johnthetubaguy | bswartz was talking about manila gate having issues | 13:48 |
| sdague | yeh | 13:48 |
| sdague | but that's also about an API compat issue | 13:48 |
| sdague | which we're going to discuss in the meeting | 13:48 |
| *** y_sawai has quit IRC | 13:49 | |
| *** eharney has quit IRC | 13:49 | |
| sdague | and, honestly, I was reading the entire device path yesterday after I -2ed that patch, and I'm more sympathetic to the change | 13:49 |
| sdague | because we basically treat device as a required element which might be None all the way through our code | 13:49 |
| bswartz | manila gate is broken until the nova client fix goes to pypi | 13:49 |
| sdague | but, we should discuss | 13:49 |
| bswartz | or that server fix comes back | 13:50 |
| *** lyanchih has quit IRC | 13:50 | |
| *** ttrumm has quit IRC | 13:50 | |
| dansmith | what's the server "fix"? I have strong opinions about this path.. | 13:50 |
| sdague | allow device to be None | 13:50 |
| bswartz | dansmith: https://review.openstack.org/#/c/219696/ | 13:51 |
| *** zzzeek has joined #openstack-nova | 13:51 | |
| bswartz | dansmith: it got double -2'd | 13:51 |
| sdague | https://review.openstack.org/#/c/219696/ | 13:51 |
| mordred | johnthetubaguy, sdague: infra is pinned on an old version of novaclient until a new thing comes out with this fix _and_ with something to deal with the 401. HOWEVER, that's workable at the moment - so I vote both for waiting and not-waiting | 13:51 |
| sdague | mordred: so, we clearly want a release this week that gets us through the 401 bit as well | 13:52 |
| mordred | sdague: cool | 13:52 |
| dims | sdague: want https://review.openstack.org/#/c/219696/ back? | 13:52 |
| sdague | dims: yeh, I just restored | 13:52 |
| sdague | so it can gen test results | 13:52 |
| dims | ah k | 13:52 |
| johnthetubaguy | yeah, need a release this week | 13:53 |
| *** alexpilotti_ is now known as alexpilotti | 13:53 | |
| johnthetubaguy | when is the tricky part | 13:53 |
| bauzas | sdague: so I thought about the problem we discussed and I think we should treat those requests as v2.0 | 13:53 |
| dansmith | sdague: bswartz: so that's the thing we wish we never had in our API | 13:53 |
| dansmith | sdague: bswartz: and the thing we discussed deprecating | 13:54 |
| bauzas | sdague: sorry if you misunderstood my opinion about not providing APIversions | 13:54 |
| bswartz | dansmith: I don't disagree | 13:54 |
| sdague | dansmith: yes | 13:54 |
| dansmith | sdague: bswartz: not passing it from the novaclient unless it's provided makes it easier to do that in the future I think | 13:54 |
| sdague | however, python-novaclient isn't the only client that passes it | 13:54 |
| sdague | the ruby fog code does as well | 13:54 |
| ndipanov | dansmith, 9 files changed, 258 insertions(+), 48 deletions(-) | 13:54 |
| bswartz | dansmith: our problem is that the server side change makes it impossible for old clients to talk to new servers unless they specify the thing you want to deprecate | 13:54 |
| ndipanov | not too bad I guess | 13:54 |
| *** jroll has quit IRC | 13:54 | |
| *** jroll has joined #openstack-nova | 13:54 | |
| dansmith | which maybe isn't an argument not to accept null as long as we do, I dunno | 13:54 |
| sdague | which makes me think we break a lot more people if we fail there | 13:54 |
| bswartz | and 219696 was supposed to undo that breakage | 13:55 |
| dansmith | I don't think that passing None here makes sense | 13:55 |
| dansmith | from the semantic pov of the api | 13:56 |
| bswartz | nevertheless, old clients do pass None | 13:56 |
| dansmith | but if we're deciding to do the thing that breaks fewer people.. | 13:56 |
| sdague | yeh, if it was just python-novaclient, I'd be fine fixing it there | 13:56 |
| sdague | but I did go off and read the fog code | 13:56 |
| *** sayali has quit IRC | 13:57 | |
| sdague | https://github.com/fog/fog/blob/23467f391cbc9d8d3262da33cec533f8b8e80cee/lib/fog/openstack/requests/compute/attach_volume.rb#L9 | 13:57 |
| *** vivekd has quit IRC | 13:57 | |
| sdague | it's always passed | 13:57 |
| dansmith | I'm not clicking on that because I value my eyes and my dignity | 13:57 |
| sdague | the code looks a ton better than python-novaclient :) | 13:57 |
| mordred | I will confirm that | 13:57 |
| dansmith | but it's ruby so it's actually terrible | 13:58 |
| mordred | dansmith: well, it rewrites itself inside of a closure that implements a dsl - but that's normal, right? | 13:58 |
| *** tpeoples has joined #openstack-nova | 13:58 | |
| dansmith | that ^ | 13:58 |
| *** shaohe_feng1 has joined #openstack-nova | 13:59 | |
| *** mriedem has joined #openstack-nova | 13:59 | |
| sdague | no, it just mixes in ever method individually | 13:59 |
| mordred | sdague: btw - I never realized that I should have such a strong opinion on the difference between "pass value of None" and "don't pass value" until this year | 13:59 |
| *** shaohe_feng has quit IRC | 13:59 | |
| mriedem | howdy | 13:59 |
| dims | hey mriedem | 13:59 |
| sdague | mordred: yeh | 13:59 |
| *** sayali has joined #openstack-nova | 13:59 | |
| dims | so consensus then? "pass value of None" and "don't pass value" are the same? | 14:00 |
| dims | treat them as the same? | 14:00 |
| dansmith | no | 14:00 |
| *** MarcF has quit IRC | 14:00 | |
| dansmith | they are NOT the same | 14:00 |
| mriedem | meeting? | 14:00 |
| dansmith | we're agreed to accept *either* I think | 14:00 |
| mordred | sdague: the neutronclient api has very strong and explicit semantics on this front - so we have a custom decorator to implement dealing with it on the python side | 14:00 |
| sdague | dims: no, they aren't the same. However, in this case, we're deciding based on the "break less users" to be pragmatic | 14:00 |
| mordred | the decorator hurts my eye | 14:01 |
| sdague | mriedem: lets do it | 14:01 |
| dims | ack, thats what i meant | 14:01 |
| dims | treat them as the same by accepting either one | 14:01 |
| johnthetubaguy | mikal tjones cburgess jgrimm adrian_otto funzo mjturek jcookekhugen irina_pov krtaylor danpb alexpilotti flip214 jaypipes garyk edleafe dims moshele anteaya Nisha sileht claudiub lxsli neiljerram markus_z swamireddy alevine tonyb andreykurilin ndipanov sc68cal akuriata artom jlvillal mnestratov kashyap aloga rgeragnov bauzas xyang tpatil med_ nic scottda nagyz dannywilson belliott dguitarbite sdague its nova meeting time | 14:01 |
| sdague | yeh, because, device is turned into None later anyway | 14:01 |
| sdague | https://github.com/openstack/nova/blob/b3f07e8c1fa45e9ae2835ed028a092436ae2ad10/nova/compute/api.py#L3057 | 14:01 |
| dims | right | 14:01 |
| sdague | so it's 2 layers of transit before it was coerced into that | 14:02 |
| *** bnemec has quit IRC | 14:02 | |
| dims | sdague: oh, on the novaclient we change "auto" to None as well :) | 14:02 |
| *** bnemec has joined #openstack-nova | 14:02 | |
| sdague | the novaclient does tons of dumb stuff | 14:03 |
| sdague | we should fix those | 14:03 |
| *** MarcF has joined #openstack-nova | 14:03 | |
| mriedem | BobBall: is xenproject ci back in order? | 14:03 |
| *** burgerk has joined #openstack-nova | 14:03 | |
| *** rushil has joined #openstack-nova | 14:04 | |
| *** e0ne has joined #openstack-nova | 14:04 | |
| garyk | sdague: would it be possible that you please take alook at https://review.openstack.org/212786 | 14:06 |
| garyk | sdague: i guess i should have done that on another channel but not sure that there is a devstack one | 14:07 |
| *** apuimedo|lunch has quit IRC | 14:08 | |
| *** thangp has joined #openstack-nova | 14:08 | |
| *** stefan_amann has quit IRC | 14:08 | |
| *** rushil has quit IRC | 14:09 | |
| BobBall | yes mriedem | 14:09 |
| BobBall | It autofixed | 14:09 |
| mriedem | BobBall: ok i'll add back to nova-ci in a bit | 14:09 |
| BobBall | I think johnthetubaguy already did | 14:09 |
| *** cfriesen has joined #openstack-nova | 14:10 | |
| sdague | garyk: that's what -qa is for | 14:10 |
| garyk | sdague: thanks | 14:11 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: RT: Migration resource tracking uses migration context https://review.openstack.org/218938 | 14:11 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: compute: migrate/resize paths properly handle stashed numa_topology https://review.openstack.org/218500 | 14:11 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: rebuild: RPC sends additional args and claims are done https://review.openstack.org/200485 | 14:11 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: Claims: Make sure move claims create a migration context records https://review.openstack.org/218385 | 14:11 |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: objects: Hook migration object into Instance https://review.openstack.org/216733 | 14:11 |
| mriedem | rlrossit: you should take a pass through this https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/liberty-bump-object-and-rpcapi-versions,n,z | 14:12 |
| *** apuimedo|lunch has joined #openstack-nova | 14:12 | |
| rlrossit | mriedem: yeah I'm already looking at them | 14:12 |
| *** stefan_amann has joined #openstack-nova | 14:12 | |
| *** electrocucaracha has joined #openstack-nova | 14:13 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Update from global requirements https://review.openstack.org/220157 | 14:13 |
| mriedem | dansmith: sdague: ^ | 14:13 |
| dansmith | mriedem: remember how I said I'd like to do that? | 14:14 |
| mriedem | i can abandon | 14:14 |
| dansmith | no | 14:14 |
| openstackgerrit | Chuck Carmack proposed openstack/nova: Remove the blacklisted nova-cells shelve tests https://review.openstack.org/220158 | 14:16 |
| *** shoutm_ has quit IRC | 14:18 | |
| *** dave-mccowan has quit IRC | 14:18 | |
| *** MarcF has quit IRC | 14:19 | |
| *** eharney has joined #openstack-nova | 14:19 | |
| *** Murali has quit IRC | 14:21 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: Allow to use autodetection of volume device path https://review.openstack.org/219696 | 14:22 |
| *** yamahata has joined #openstack-nova | 14:22 | |
| mriedem | bswartz: so i guess i won't do a novaclient 2.27.1 release since https://review.openstack.org/#/c/219696/ will fix the manila blocker | 14:24 |
| bswartz | mriedem: that's fine | 14:25 |
| *** atuvenie has quit IRC | 14:28 | |
| *** tonytan4ever has joined #openstack-nova | 14:28 | |
| *** rushil_ has joined #openstack-nova | 14:28 | |
| *** Piet has quit IRC | 14:30 | |
| mriedem | sdague: i have a fix for the unit test failure in https://review.openstack.org/#/c/220111/ if you want me to push that up | 14:30 |
| *** MarcF has joined #openstack-nova | 14:30 | |
| *** browne has joined #openstack-nova | 14:31 | |
| *** dave-mccowan has joined #openstack-nova | 14:31 | |
| sdague | mriedem: yes please | 14:31 |
| mriedem | ok | 14:31 |
| *** rods has joined #openstack-nova | 14:31 | |
| *** bkopilov has joined #openstack-nova | 14:32 | |
| electrocucaracha | Hey there, this is my first meeting I'm not sure if you have an agenda, I would like to talk about https://review.openstack.org/#/c/215414/ | 14:32 |
| *** devlaps has quit IRC | 14:33 | |
| edleafe | electrocucaracha: the nova meeting is going on in #openstack-meeting channel | 14:33 |
| openstackgerrit | Matt Riedemann proposed openstack/python-novaclient: Don't assume oscomputeversions is correctly deployed https://review.openstack.org/220111 | 14:33 |
| electrocucaracha | thankd edleafe | 14:33 |
| *** eharney has quit IRC | 14:34 | |
| mriedem | electrocucaracha: it's open discussion right now so post your hting | 14:34 |
| *** mtanino has joined #openstack-nova | 14:35 | |
| electrocucaracha | ok, maybe I have update my calendar | 14:36 |
| *** whenry_ has joined #openstack-nova | 14:36 | |
| *** chutwig has quit IRC | 14:36 | |
| * alex_xu out of power...go to take a rest | 14:38 | |
| * mriedem replaces alex_xu's batteries | 14:39 | |
| *** sayali has quit IRC | 14:40 | |
| mriedem | danpb: do you know if qemu-nbd -d is aysnc? | 14:40 |
| mriedem | *async | 14:40 |
| danpb | mriedem: not sure off hand | 14:40 |
| *** Marga_ has joined #openstack-nova | 14:43 | |
| *** whenry_ has quit IRC | 14:43 | |
| *** pilgrimstack has joined #openstack-nova | 14:44 | |
| pilgrimstack | hi all | 14:44 |
| *** romainh has left #openstack-nova | 14:44 | |
| *** ftersin has left #openstack-nova | 14:44 | |
| *** y_sawai has joined #openstack-nova | 14:45 | |
| pilgrimstack | I'm from RunAbove | 14:46 |
| *** diogogmt_ has joined #openstack-nova | 14:46 | |
| pilgrimstack | how can I help for https://bugs.launchpad.net/python-novaclient/+bug/1491579 | 14:46 |
| openstack | Launchpad bug 1491579 in python-novaclient "against all sanity, nova needs to work around broken public clouds" [Critical,In progress] - Assigned to Matt Riedemann (mriedem) | 14:46 |
| pilgrimstack | ? | 14:46 |
| mriedem | johnthetubaguy: danpb: updated https://wiki.openstack.org/wiki/Meetings/Libvirt#Libvirt_Meetings | 14:47 |
| *** diogogmt has quit IRC | 14:47 | |
| *** diogogmt_ is now known as diogogmt | 14:47 | |
| mriedem | pilgrimstack: verify https://review.openstack.org/#/c/220111/ ? | 14:47 |
| bauzas | mordred: ^ | 14:47 |
| sdague | mordred: can you give me a "nova --debug list" dump from rax again? | 14:47 |
| *** atuvenie has joined #openstack-nova | 14:47 | |
| sdague | pilgrimstack: runabove appears to be hanging on a GET of /v2/ | 14:47 |
| sdague | also a GET of / | 14:47 |
| mordred | sdague: with or without your patch? | 14:47 |
| bauzas | pilgrimstack: so as said in the bugfile, novaclient doesn't work with runabove for 2.27 | 14:48 |
| sdague | mordred: ... with | 14:48 |
| ndipanov | johnthetubaguy, https://review.openstack.org/#/c/200485/ as per your suggestions | 14:48 |
| sdague | mordred: just need some more details to build the 401 work around | 14:48 |
| bauzas | pilgrimstack: could you please create a virtualenv and deploy the above change to test it? | 14:48 |
| bauzas | pilgrimstack: using your creds | 14:48 |
| *** eharney has joined #openstack-nova | 14:48 | |
| mordred | sdague: http://paste.openstack.org/show/444358/ | 14:49 |
| sdague | pilgrimstack: yeh, the test case is that "nova list" actually returns in a reasonable amount of time | 14:49 |
| pilgrimstack | doing | 14:49 |
| sdague | pilgrimstack: a separate, but related bug, is that "nova version-list" hangs against runabove | 14:49 |
| *** wanghua has quit IRC | 14:50 | |
| *** y_sawai has quit IRC | 14:50 | |
| pilgrimstack | https://review.openstack.org/#/c/220111/ ? | 14:51 |
| sdague | pilgrimstack: yes | 14:52 |
| *** mdrabe has joined #openstack-nova | 14:53 | |
| *** abhishekk has left #openstack-nova | 14:53 | |
| BobBall | Should configdrive include the private IP in the openstack/latest/meta_data.json file? I can only see the private ipv4 listed in the ec2 metadata my instances which seems ... unusual :) | 14:53 |
| *** mtanino has quit IRC | 14:53 | |
| *** y_sawai has joined #openstack-nova | 14:54 | |
| *** gilliard has left #openstack-nova | 14:54 | |
| *** tri2sing has joined #openstack-nova | 14:55 | |
| dansmith | sdague: wanna +W the requirements thing? https://review.openstack.org/#/c/220157/1 | 14:55 |
| sdague | dansmith: sure | 14:56 |
| sdague | also, we should just fast approve requirements in the future | 14:56 |
| dansmith | oh, is that a thing? | 14:56 |
| sdague | especially if they come from the bot | 14:56 |
| dansmith | this one being from mriedem it probably deserves scrutiny | 14:56 |
| sdague | well, it looks almost like the bot one did | 14:57 |
| dansmith | he used the tool | 14:57 |
| sdague | yeh, everything in that is sane | 14:57 |
| dansmith | so the bot one is in gate again, | 14:57 |
| dansmith | I wonder if zuul will be fine queuing this up behind it when the net will be zero? | 14:57 |
| sdague | yeh, it might | 14:58 |
| sdague | we'll see | 14:58 |
| dansmith | so in the future, | 14:58 |
| dansmith | I could have depends-on'd this sync, | 14:58 |
| dansmith | but the code would still be waiting to merge | 14:59 |
| *** VW_ has quit IRC | 14:59 | |
| *** y_sawai has quit IRC | 14:59 | |
| dansmith | is it kosher to just do one add in a patch and let the requirements job keep us honest? | 14:59 |
| *** unicell has joined #openstack-nova | 14:59 | |
| *** tri2sing has quit IRC | 14:59 | |
| *** erhudy has joined #openstack-nova | 14:59 | |
| *** VW_ has joined #openstack-nova | 14:59 | |
| *** husanu has quit IRC | 14:59 | |
| dansmith | actually, hmm | 14:59 |
| dansmith | is there no requirements job anymore? | 14:59 |
| *** emagana has joined #openstack-nova | 15:01 | |
| *** sachin has joined #openstack-nova | 15:01 | |
| *** husanu has joined #openstack-nova | 15:01 | |
| *** unicell1 has quit IRC | 15:01 | |
| mriedem | dansmith: it's fine to add the requirement you need in your change outside of the sync patch | 15:02 |
| *** jistr is now known as jistr|call | 15:02 | |
| mriedem | there is a reqs job http://logs.openstack.org/68/216968/16/check/gate-nova-requirements/1456462/ | 15:03 |
| *** armax has joined #openstack-nova | 15:03 | |
| *** VW_ has quit IRC | 15:03 | |
| *** VW_ has joined #openstack-nova | 15:03 | |
| *** rushil_ has quit IRC | 15:04 | |
| dansmith | mriedem: okay, I don't see that in jenkins reports | 15:04 |
| dansmith | does it only comment if it fails? | 15:04 |
| *** whenry has joined #openstack-nova | 15:04 | |
| *** rotbeard has quit IRC | 15:04 | |
| mriedem | dansmith: i've only ever noticed it when it fails i guess | 15:04 |
| *** Marga_ has quit IRC | 15:05 | |
| *** sayali has joined #openstack-nova | 15:05 | |
| dansmith | okay | 15:06 |
| dansmith | well, not putting that req in my patch was my bad | 15:06 |
| dansmith | I just didn't know that was legit | 15:06 |
| openstackgerrit | jichenjc proposed openstack/nova: API: remove unused parameter https://review.openstack.org/218932 | 15:06 |
| mriedem | usually only see it when adding a new dependency i guess | 15:07 |
| dansmith | well, if the bot was less stupid, depends-on is the right way I guess | 15:07 |
| *** markvoelker has joined #openstack-nova | 15:08 | |
| *** rmart04 has quit IRC | 15:09 | |
| *** sayali has quit IRC | 15:10 | |
| *** husanu has quit IRC | 15:10 | |
| *** husanu has joined #openstack-nova | 15:12 | |
| *** MarcF has quit IRC | 15:13 | |
| *** chutwig has joined #openstack-nova | 15:14 | |
| pilgrimstack | mordred: sdague: $ nova --version | 15:15 |
| pilgrimstack | 2.27.1 | 15:15 |
| pilgrimstack | nova list working | 15:15 |
| pilgrimstack | nova version-list too | 15:16 |
| *** alexpilotti_ has joined #openstack-nova | 15:16 | |
| mordred | pilgrimstack: oh yeah? that's with sean's patch? let me try again on my side | 15:16 |
| *** alexpilotti has quit IRC | 15:17 | |
| *** alexpilotti_ is now known as alexpilotti | 15:17 | |
| openstackgerrit | Sean Dague proposed openstack/python-novaclient: workaround for RAX repose configuration https://review.openstack.org/220192 | 15:17 |
| sdague | mordred: if you, or a rax person could test ^^^ | 15:17 |
| sdague | I think that addresses the 401 correctly | 15:17 |
| mordred | pilgrimstack: it's still hanging for me - you're hitting the public endpoint and not an admin endpoint, yeah? | 15:17 |
| pilgrimstack | public yes | 15:18 |
| mordred | hrm | 15:18 |
| mordred | sdague: patch fixes rax | 15:19 |
| pilgrimstack | how can I verify I'm using the right patch ? | 15:19 |
| *** tri2sing has joined #openstack-nova | 15:19 | |
| sdague | mordred: great, if you could +1 that review with that info, would be appreciated | 15:19 |
| sdague | pilgrimstack: "nova --debug list" pastebin the output | 15:19 |
| sdague | that might give us enough info | 15:20 |
| mordred | sdague, pilgrimstack: aha! it worked | 15:20 |
| mordred | I think I had a weird config - started a clean shell and retested | 15:20 |
| mordred | so - with both of sean's patches applied, all the clouds work again | 15:20 |
| *** shaohe_feng1 has quit IRC | 15:21 | |
| *** mriedem has quit IRC | 15:21 | |
| *** annegentle has joined #openstack-nova | 15:21 | |
| sdague | mordred: so out of curiosity, could you figure out if runabove works with 2.27 raw then? maybe their hangs were just your weird config. | 15:22 |
| mordred | sdague: yes. one sec | 15:22 |
| *** tri2sing has quit IRC | 15:22 | |
| *** sayali has joined #openstack-nova | 15:22 | |
| *** tri2sing has joined #openstack-nova | 15:24 | |
| mordred | sdague: with raw 2.27.0 I get this: | 15:24 |
| mordred | ConnectionRefused: Unable to establish connection to http://compute.sbg-1.runabove.io/v2/ | 15:24 |
| mordred | ERROR (ConnectionRefused): Unable to establish connection to http://compute.sbg-1.runabove.io/v2/ | 15:24 |
| *** whenry has quit IRC | 15:25 | |
| mordred | so it looks like the rax fix also fixed runabo e | 15:25 |
| sdague | hmmm, it shouldn't have | 15:25 |
| *** dims has quit IRC | 15:26 | |
| sdague | but I'll take it | 15:26 |
| Vek | +2'd the rax patch, and I'm trying to track down the repose team to make sure they have the note | 15:26 |
| pilgrimstack | mordred: sdague: http://paste.openstack.org/show/444378/ | 15:26 |
| *** hemnafk is now known as hemna | 15:28 | |
| *** MarcF has joined #openstack-nova | 15:28 | |
| *** yamahata has quit IRC | 15:28 | |
| *** matrohon has quit IRC | 15:29 | |
| *** mcarden has quit IRC | 15:29 | |
| *** thangp has quit IRC | 15:29 | |
| *** mcarden has joined #openstack-nova | 15:30 | |
| pilgrimstack | It's ok for you? | 15:30 |
| *** emagana has quit IRC | 15:30 | |
| sdague | Vek thanks | 15:31 |
| *** dannywilson has joined #openstack-nova | 15:31 | |
| *** sachin has quit IRC | 15:32 | |
| *** emagana has joined #openstack-nova | 15:32 | |
| *** sayali has quit IRC | 15:32 | |
| *** nkrinner has quit IRC | 15:34 | |
| *** stefan_amann has quit IRC | 15:35 | |
| Vek | sdague: Reported it. For reference, #repose here on Freenode or reposecore@rackspace.com. | 15:35 |
| *** dims has joined #openstack-nova | 15:36 | |
| *** sayali has joined #openstack-nova | 15:36 | |
| *** mtanino has joined #openstack-nova | 15:36 | |
| *** y_sawai has joined #openstack-nova | 15:37 | |
| *** Sree has quit IRC | 15:37 | |
| dansmith | johnthetubaguy: I think you can drop your -2 from this now: https://review.openstack.org/#/c/200485/19 | 15:37 |
| *** Sree has joined #openstack-nova | 15:38 | |
| sdague | Vek: ok, cool, thanks | 15:38 |
| *** sachin has joined #openstack-nova | 15:38 | |
| *** y_sawai has quit IRC | 15:38 | |
| sdague | Vek: honestly, I don't want to dive all the way down the well, I just want you all to fix this thing :) | 15:38 |
| openstackgerrit | garyk proposed openstack/nova: VMware: ensure volume is accessible by instance https://review.openstack.org/218639 | 15:39 |
| *** amotoki has quit IRC | 15:40 | |
| sdague | johnthetubaguy: if you want to +A - https://review.openstack.org/#/c/220192 we could cut another novaclient and be back to 2.26 levels of working | 15:41 |
| *** jichen has quit IRC | 15:41 | |
| sdague | once all that merges | 15:41 |
| *** hhuang has quit IRC | 15:42 | |
| *** Sree has quit IRC | 15:42 | |
| *** markvoelker_ has joined #openstack-nova | 15:42 | |
| *** thangp has joined #openstack-nova | 15:42 | |
| *** ArchiFleKs has joined #openstack-nova | 15:43 | |
| *** ccarmack has quit IRC | 15:43 | |
| *** sachin has quit IRC | 15:43 | |
| dansmith | sdague: you want his +W or anyone's? | 15:44 |
| *** jistr|call is now known as jistr | 15:44 | |
| *** electrocucaracha has quit IRC | 15:45 | |
| *** markvoelker has quit IRC | 15:45 | |
| *** electrocucaracha has joined #openstack-nova | 15:45 | |
| openstackgerrit | Timofey Durakov proposed openstack/nova: Change auth_plugin for neutron_client https://review.openstack.org/220207 | 15:46 |
| bauzas | sdague: I guess runabove blocks the API call too | 15:46 |
| bauzas | pilgrimstack: ^ | 15:46 |
| bauzas | sdague: so your change is good | 15:46 |
| bauzas | sdague: because it becomes a v2.0 version | 15:46 |
| *** ccarmack has joined #openstack-nova | 15:46 | |
| *** ccarmack has quit IRC | 15:46 | |
| bauzas | pilgrimstack: do you know if you're blocking /v2/ ? | 15:46 |
| sdague | dansmith: anyone's | 15:46 |
| bauzas | pilgrimstack: by 'you', I mean runabove of course :) | 15:47 |
| *** garyk has quit IRC | 15:47 | |
| *** whenry has joined #openstack-nova | 15:47 | |
| sdague | bauzas: so I'm satisfied if pilgrimstack believes it now works | 15:47 |
| *** burt has quit IRC | 15:47 | |
| bauzas | sdague: yeah, returning a v2.0 compat version is good IMHO | 15:47 |
| *** gszasz has quit IRC | 15:47 | |
| *** ccarmack has joined #openstack-nova | 15:48 | |
| bauzas | sdague: because it means that since the operators are blocking the versions, they accept to use the minor version | 15:48 |
| dansmith | I don't think it means that | 15:48 |
| dansmith | I think it means we're very gracious in letting them continue to work at v2.0 levels for the time being, despite their silliness :) | 15:49 |
| sdague | right, we're allowing a work around now. | 15:49 |
| sdague | but we will be really explicit that this is required to be called openstack | 15:49 |
| bauzas | fair enough | 15:49 |
| *** yamahata has joined #openstack-nova | 15:50 | |
| sdague | just trying to not let our users get caught in the crossfire. | 15:50 |
| bauzas | that said, probably something that the operators should know | 15:50 |
| bauzas | sdague: agreed | 15:50 |
| sdague | right, and defcore is definitely a way to communicate that | 15:50 |
| bauzas | sure thing | 15:50 |
| *** mdrabe has quit IRC | 15:50 | |
| mtreinish | sdague: https://review.openstack.org/219873 is the first step towards that | 15:52 |
| mtreinish | I want to make sure that's sufficient to catch this | 15:52 |
| sdague | mtreinish: +A | 15:53 |
| *** cp16net has joined #openstack-nova | 15:54 | |
| sdague | yeh, I think we need one more test after that, but lets get that in | 15:54 |
| *** browne has quit IRC | 15:54 | |
| cp16net | i think i found an issue using the v2 api booting an instance with a block device mapping | 15:54 |
| cp16net | https://github.com/openstack/python-novaclient/commit/58cdcabf1c6580fbf04ae54ca846692ec2127d78 | 15:54 |
| ArchiFleKs | Hi, I'm not sure it the place to post this, i'm sorry if it is not | 15:55 |
| ArchiFleKs | did anyone experience this bug https://bugs.launchpad.net/nova/+bug/1272076. When creating a volume from image on instance boot,Nova is trying to mount the block storage every 3 second for | 15:55 |
| openstack | Launchpad bug 1272076 in Cinder "VolumeNotCreated - Instance failed, cinder too slow with Ceph" [Undecided,Confirmed] | 15:55 |
| cp16net | that patch was to fix v3 but causes regression when using v2 | 15:55 |
| ArchiFleKs | 60 times without checking if the volume is ready on Cinder, today I had to increase the block_device_allocate_retries = 60 and block_device_allocate_retries_interval = 3 to force Nova to wait longer. Shouldn't Nova check the status of the volume with cinder before trying to attach it to the instance ? (The volume is still in downloading state when Nova tries to attach it) or is there a reason why it | 15:55 |
| ArchiFleKs | does not wait ? FYI : when deploying with Heat there is no problem, maybe because Heat is waiting for every object of the stack to be ready. | 15:55 |
| pilgrimstack | bauzas: do you know if you're blocking /v2/ ? => what do you mean? | 15:55 |
| bauzas | pilgrimstack: so, the client calls nova to get a list of versions | 15:56 |
| *** rushil has joined #openstack-nova | 15:56 | |
| *** SourabhP has joined #openstack-nova | 15:56 | |
| dansmith | mtreinish: that test verifies that the result is not none, but also compares the response against a schema, is that right? | 15:57 |
| bauzas | pilgrimstack: the change you tested was basically making sure that if the call was returning a 401, we were using the minimum version instead of something negociated | 15:58 |
| mtreinish | dansmith: yeah it does a jsonschema validate on the response | 15:58 |
| dansmith | mtreinish: okay sweet | 15:58 |
| *** tjones has joined #openstack-nova | 15:58 | |
| dansmith | mtreinish: just wanted to make sure you couldn't pass that test with returning "go away" or something | 15:58 |
| *** bkopilov has quit IRC | 15:58 | |
| *** SourabhP_ has joined #openstack-nova | 15:58 | |
| mtreinish | dansmith: heh, yeah that would have been no good | 15:58 |
| mtreinish | dansmith: fwiw, it's in the client method: https://review.openstack.org/#/c/219873/7/tempest/services/compute/json/versions_client.py | 15:59 |
| mtreinish | L36 | 15:59 |
| dansmith | mtreinish: so this is still fake-able if I wanted to be a real dick, right? like I could just have my proxy return a static file with a fake response here that would fit the schema | 15:59 |
| pilgrimstack | bauzas: curl -g -i -X GET https://compute.bhs-1.runabove.io/v2/ give me a 401 | 15:59 |
| bauzas | sdague: ^ | 15:59 |
| dansmith | mtreinish: do we care that much? like, do we validate that your response looks like the release you say you're running, etc? from a defcore perspective I mean | 15:59 |
| *** afazekas__ has quit IRC | 15:59 | |
| bauzas | pilgrimstack: oh but you need to pass a token | 15:59 |
| sdague | right, you need the token, it will give you a 401 without it | 16:00 |
| mtreinish | dansmith: yeah you could do that | 16:00 |
| pilgrimstack | right, I'll not past my token here :p | 16:00 |
| *** tjones has quit IRC | 16:00 | |
| mtreinish | but tbh, it'd be easier to just upload a fake response to defcore | 16:00 |
| mtreinish | that's easy to fake too | 16:00 |
| sdague | mtreinish: that assumes mallice | 16:00 |
| sdague | honestly, I don't think there is malice here | 16:01 |
| sdague | it's just not realizing that this is important | 16:01 |
| *** SourabhP has quit IRC | 16:01 | |
| *** SourabhP_ is now known as SourabhP | 16:01 | |
| mtreinish | dansmith: right now tempest doesn't understand nova microversions, I think we'll likely expand that test when it does | 16:01 |
| mtreinish | sdague: sure, I was just playing the game | 16:01 |
| mtreinish | and if I wanted to be evil I'd just upload fake test results | 16:01 |
| sdague | pilgrimstack: ok, so if you are returning a 401 on that url, realize that will eventually cause you to fail defcore | 16:02 |
| pilgrimstack | bauzas: sdague: mybe it was not well express, I use a token and I have a 401 | 16:02 |
| sdague | pilgrimstack: ok, so then it acts like rax | 16:03 |
| pilgrimstack | realize that will eventually cause you to fail defcore => what do you mean ? | 16:03 |
| sdague | you will lose the ability to call the cloud openstack | 16:03 |
| Vek | sdague: story seems to be that it's a configuration issue; we're working on how best to correct the configuration. | 16:03 |
| sdague | Vek: ok, great, thanks | 16:03 |
| *** marzif has quit IRC | 16:04 | |
| sdague | mtreinish: so the test we need on top of the one oomichi_away added, is actually that one, in getting /v2 and that being a thing. | 16:04 |
| sdague | because it's supposed to be | 16:04 |
| mtreinish | sdague: oh, yeah. That's actually what I originally wrote it as | 16:05 |
| mtreinish | I didn't realize he changed it | 16:05 |
| *** jlanoux has quit IRC | 16:05 | |
| *** browne has joined #openstack-nova | 16:06 | |
| *** rajesht has quit IRC | 16:06 | |
| *** mdrabe has joined #openstack-nova | 16:06 | |
| mtreinish | sdague: that should be easy enough to add | 16:06 |
| sdague | yeh, he was doing the top level version list | 16:06 |
| sdague | which, is also required and good | 16:06 |
| *** jistr has quit IRC | 16:06 | |
| sdague | mtreinish: any idea when defcore picks up their next set of tests? | 16:07 |
| *** jaypipes has joined #openstack-nova | 16:07 | |
| mtreinish | sdague: can we assume the service catalog will always have the project id as the last piece? | 16:07 |
| mtreinish | sdague: hmm I forget when there next definition round is | 16:07 |
| mtreinish | sdague: but hogepodge is aware of the test at least, so I'm sure if we bug him enough he'll add it to the list | 16:08 |
| sdague | ok, I'll follow up with him directly | 16:08 |
| *** ijw_ has joined #openstack-nova | 16:08 | |
| mtreinish | sdague: fwiw, I think it's just a json file in a git repo so you can propose the addition | 16:08 |
| *** _diana_ has joined #openstack-nova | 16:09 | |
| *** bkopilov has joined #openstack-nova | 16:10 | |
| *** alex_klimov has quit IRC | 16:11 | |
| *** tjones has joined #openstack-nova | 16:11 | |
| *** _diana__ has joined #openstack-nova | 16:11 | |
| *** aix has quit IRC | 16:12 | |
| *** Piet has joined #openstack-nova | 16:12 | |
| *** VikasC has joined #openstack-nova | 16:12 | |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: claims: Remove compat code with instance dicts https://review.openstack.org/220215 | 16:12 |
| *** unicell has quit IRC | 16:14 | |
| *** _diana__ has quit IRC | 16:14 | |
| *** unicell has joined #openstack-nova | 16:14 | |
| *** _diana__ has joined #openstack-nova | 16:15 | |
| *** _diana__ has quit IRC | 16:15 | |
| *** MarcF has quit IRC | 16:17 | |
| *** karimb has quit IRC | 16:17 | |
| *** derekh has quit IRC | 16:19 | |
| *** ndipanov has quit IRC | 16:21 | |
| *** sahid has quit IRC | 16:21 | |
| *** _diana_ has quit IRC | 16:22 | |
| *** _diana_ has joined #openstack-nova | 16:22 | |
| *** atuvenie has quit IRC | 16:22 | |
| hogepodge | mtreinish: yup | 16:24 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: Change auth_plugin for neutron_client https://review.openstack.org/220207 | 16:24 |
| *** nic has joined #openstack-nova | 16:25 | |
| *** chutwig has quit IRC | 16:25 | |
| cp16net | anyone seen this error crop up lately? http://logs.openstack.org/72/219972/1/check/gate-trove-functional-dsvm-mysql/78fdb96/logs/screen-n-api.txt.gz#_2015-09-03_13_17_19_169 | 16:25 |
| *** _diana_ has quit IRC | 16:25 | |
| *** sayali has quit IRC | 16:26 | |
| *** ctrath has quit IRC | 16:27 | |
| *** VW_ has quit IRC | 16:28 | |
| *** VW_ has joined #openstack-nova | 16:28 | |
| *** MarcF has joined #openstack-nova | 16:29 | |
| *** chutwig has joined #openstack-nova | 16:29 | |
| *** eglynn has quit IRC | 16:31 | |
| *** VW_ has quit IRC | 16:33 | |
| *** ccarmack has quit IRC | 16:34 | |
| *** e0ne has quit IRC | 16:34 | |
| *** sayali has joined #openstack-nova | 16:35 | |
| *** mpaolino has quit IRC | 16:36 | |
| *** annegentle has quit IRC | 16:36 | |
| *** SourabhP has quit IRC | 16:36 | |
| *** annegentle has joined #openstack-nova | 16:37 | |
| *** lucasagomes has quit IRC | 16:39 | |
| *** ctrath has joined #openstack-nova | 16:40 | |
| *** afazekas__ has joined #openstack-nova | 16:40 | |
| *** VW_ has joined #openstack-nova | 16:41 | |
| *** ccarmack has joined #openstack-nova | 16:42 | |
| *** VW_ has quit IRC | 16:42 | |
| *** mriedem has joined #openstack-nova | 16:42 | |
| *** VW_ has joined #openstack-nova | 16:42 | |
| *** tjones has left #openstack-nova | 16:43 | |
| *** VW_ has quit IRC | 16:43 | |
| *** smatzek has quit IRC | 16:43 | |
| *** VW_ has joined #openstack-nova | 16:43 | |
| *** VW_ has quit IRC | 16:44 | |
| *** VW_ has joined #openstack-nova | 16:44 | |
| *** atuvenie has joined #openstack-nova | 16:45 | |
| openstackgerrit | Vladyslav Drok proposed openstack/nova: DO NOT MERGE - Triggering experimental pipeline https://review.openstack.org/220229 | 16:45 |
| *** _diana_ has joined #openstack-nova | 16:45 | |
| *** _diana_ has quit IRC | 16:46 | |
| mriedem | ooo large ops test failures http://logs.openstack.org/96/219696/4/check/gate-tempest-dsvm-large-ops/158f061/logs/screen-n-cpu-1.txt.gz?level=TRACE | 16:47 |
| mriedem | how exciting | 16:47 |
| *** SourabhP has joined #openstack-nova | 16:47 | |
| dansmith | no fails in conductor log, | 16:47 |
| mriedem | nope, rpc timeouts | 16:47 |
| dansmith | which probably means n actual timeout | 16:47 |
| dansmith | right, but usually the reason you get those timeouts is because conductor is gone, or failing | 16:48 |
| *** lpetrut has quit IRC | 16:49 | |
| *** VW_ has quit IRC | 16:49 | |
| *** garyk has joined #openstack-nova | 16:49 | |
| mriedem | yar | 16:49 |
| mriedem | http://logs.openstack.org/96/219696/4/check/gate-tempest-dsvm-large-ops/158f061/logs/screen-n-net.txt.gz#_2015-09-03_15_11_10_664 | 16:49 |
| mriedem | so i can fix the var sub in there, but that is only logged in certain cases i think | 16:50 |
| mriedem | like i think there is literally a note in the code saying 'we shouldn't get here but if we do, do this' | 16:50 |
| mriedem | https://github.com/openstack/nova/blob/master/nova/network/manager.py#L1068 | 16:52 |
| *** VW_ has joined #openstack-nova | 16:55 | |
| *** haomaiw__ has joined #openstack-nova | 17:02 | |
| johnthetubaguy | mriedem: sdague: how are we looking for a python-novaclient release for the manilla folks, looking like tomorrow or today? | 17:03 |
| *** haomaiwa_ has quit IRC | 17:03 | |
| mriedem | johnthetubaguy: not necessary | 17:04 |
| mriedem | now that we relaxed that json schema validation via dims' change | 17:04 |
| mriedem | https://review.openstack.org/#/c/219696/ | 17:04 |
| dims | yay | 17:04 |
| dims | tdurakov: left a note in https://review.openstack.org/#/c/220207/ | 17:05 |
| johnthetubaguy | mriedem: dims: perfect | 17:06 |
| dims | mriedem: https://bugs.launchpad.net/nova/+bug/1491951 tsk...tsk... | 17:06 |
| openstack | Launchpad bug 1491951 in OpenStack Compute (nova) "missing logging variable in nova network manager deallocate_fixed_ip" [Undecided,New] | 17:06 |
| mriedem | dims: yeah, fixing | 17:06 |
| *** belmoreira has quit IRC | 17:07 | |
| mriedem | i like how the guy from oracle running with patches to stable/juno is asking about how to fix his stuff | 17:07 |
| mriedem | in the ML | 17:07 |
| *** afazekas__ has quit IRC | 17:08 | |
| dims | mriedem: haha | 17:08 |
| dims | mriedem: i had dinner with jogo yday reminiscing about gate failures etc | 17:09 |
| sdague | johnthetubaguy: so, we should still do the python-novaclient release to fix users as soon as the patches merge | 17:10 |
| mriedem | dims: like https://www.youtube.com/watch?v=CZ_3G4xqSDQ ? | 17:10 |
| johnthetubaguy | sdague: yeah, for the /versions stuff right? | 17:10 |
| sdague | johnthetubaguy: yep | 17:10 |
| dims | mriedem: we had a mariachi in the background :) | 17:11 |
| johnthetubaguy | sdague: I wonder if I can proxy that permission to the release team somehow... | 17:11 |
| sdague | johnthetubaguy: I don't know, dhellmann or ttx would | 17:12 |
| *** afazekas__ has joined #openstack-nova | 17:12 | |
| johnthetubaguy | sdague: its just those two patches right? | 17:12 |
| mriedem | johnthetubaguy: sdague: we can do our own release if needed... | 17:12 |
| mriedem | that's how it worked in the long long ago | 17:12 |
| johnthetubaguy | mriedem: I think the permissions have been revoked | 17:13 |
| mriedem | so i guess those 3 will just never be allowed to take vacation or sleep at the same time | 17:13 |
| johnthetubaguy | right | 17:13 |
| mriedem | johnthetubaguy: https://launchpad.net/~nova-drivers | 17:14 |
| sdague | ok, I need to tweak the functional tests as well, we shifted a few things in the devstack service catalog and didn't account for novaclient | 17:14 |
| mriedem | i assume we can still push our own releases | 17:14 |
| *** achanda has joined #openstack-nova | 17:14 | |
| johnthetubaguy | mriedem: I think they changed the permissions behind that, but I could be wrong | 17:15 |
| johnthetubaguy | oh wait, I see what you mean now | 17:15 |
| mriedem | https://review.openstack.org/#/c/220111/ needs to be updated anyway | 17:15 |
| mriedem | i've already -1ed this https://review.openstack.org/#/c/219737/ | 17:16 |
| mtreinish | mriedem: https://review.openstack.org/#/admin/groups/967,members | 17:16 |
| mtreinish | johnthetubaguy: ^^^ | 17:16 |
| mtreinish | that's the group which has permissions to push a novaclient release | 17:16 |
| mriedem | gah | 17:16 |
| *** mpaolino has joined #openstack-nova | 17:16 | |
| mtreinish | which includes sdague | 17:16 |
| *** markvoelker_ has quit IRC | 17:16 | |
| *** rdopiera has quit IRC | 17:16 | |
| *** nic has quit IRC | 17:16 | |
| mriedem | https://review.openstack.org/#/admin/groups/11,members | 17:17 |
| sdague | oh, right, I just don't know how :) | 17:17 |
| mriedem | sdague: https://wiki.openstack.org/wiki/Nova/Client_Release_Process | 17:17 |
| johnthetubaguy | ah, cool | 17:17 |
| dims | mriedem: i see "pushSignedTag = group library-release" in project-config/gerrit/acls/openstack/python-novaclient.config - so need the release team i believe | 17:17 |
| mtreinish | mriedem: that's a bit dated, I think there are common scripts for it now | 17:17 |
| mriedem | mtreinish: last i tried those on novaclient they didn't work | 17:18 |
| mtreinish | although just pushing a signed tag will still work, it just doesn't do the fanciness | 17:18 |
| johnthetubaguy | I send a note in #openstack-relmgr-office saying I am happy with the release, so that should be logged, etc | 17:18 |
| mriedem | because the series wasn't setup the way it wanted it | 17:18 |
| johnthetubaguy | just to delegate my vote to you both | 17:18 |
| mtreinish | mriedem, sdague: just get dims to do it | 17:18 |
| jroll | can local_gb be a float or just an int? | 17:18 |
| *** mpaolino has quit IRC | 17:19 | |
| mriedem | mtreinish: we will | 17:19 |
| mriedem | unless dimsbot is out having drinks with jogo | 17:19 |
| *** erhudy has quit IRC | 17:19 | |
| dims | mtreinish: ack. just need a review in openstack/releases | 17:19 |
| openstackgerrit | Sean Dague proposed openstack/python-novaclient: Don't assume oscomputeversions is correctly deployed https://review.openstack.org/220111 | 17:19 |
| openstackgerrit | Sean Dague proposed openstack/python-novaclient: workaround for RAX repose configuration https://review.openstack.org/220192 | 17:19 |
| openstackgerrit | Sean Dague proposed openstack/python-novaclient: fix novaclient functional tests for new devstack config https://review.openstack.org/220249 | 17:19 |
| *** mpaolino has joined #openstack-nova | 17:19 | |
| sdague | mriedem: there is a test fix in there | 17:19 |
| dims | :) | 17:19 |
| jroll | HostState isn't objectified so I suppose it can | 17:19 |
| sdague | jroll: I'm pretty sure it's an int | 17:19 |
| johnthetubaguy | jroll: yeah, its an int: https://github.com/openstack/nova/blob/master/nova/db/sqlalchemy/models.py#L128 | 17:20 |
| mriedem | sdague: doesn't the functional test fix need to be in the bottom change? | 17:20 |
| jroll | johnthetubaguy: oh duh, db. thanks | 17:20 |
| mriedem | oh it is | 17:20 |
| mriedem | nvm sdague | 17:20 |
| sdague | mriedem: yeh, I was about to say, I was pretty sure I had ordered those right :) | 17:20 |
| mriedem | the way the bot lists them always confuses me | 17:20 |
| *** achanda has quit IRC | 17:21 | |
| sdague | yeh, the bot seems to just be in rando order | 17:21 |
| sdague | I've never actually figured out how it orders the world | 17:21 |
| *** jordantardif has joined #openstack-nova | 17:21 | |
| *** afazekas__ has quit IRC | 17:21 | |
| *** achanda has joined #openstack-nova | 17:22 | |
| mriedem | sdague: questoin in https://review.openstack.org/#/c/220249/ | 17:23 |
| sdague | mriedem: so, it's not even that | 17:24 |
| sdague | it's actually that we changed the service catalog creation in devstack | 17:24 |
| sdague | mriedem: https://review.openstack.org/#/c/219347/2/lib/nova,cm | 17:25 |
| *** danpb has quit IRC | 17:26 | |
| mriedem | ok, get_or_create_endpoint "computev21" \ was removed | 17:26 |
| *** zul has quit IRC | 17:26 | |
| *** zul has joined #openstack-nova | 17:27 | |
| mriedem | +W | 17:27 |
| *** zul has quit IRC | 17:27 | |
| *** zul has joined #openstack-nova | 17:28 | |
| *** penick has joined #openstack-nova | 17:32 | |
| *** annegentle has quit IRC | 17:33 | |
| dansmith | the bot's change is 32 in the gate | 17:33 |
| dansmith | and 4h old | 17:33 |
| *** SourabhP has quit IRC | 17:34 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: nova-net: fix missing log variable in deallocate_fixed_ip https://review.openstack.org/220253 | 17:34 |
| dansmith | mriedem: we could make that a named parameter without breaking string freeze because it's not translated, right? | 17:36 |
| mriedem | dansmith: the log message doesn't change | 17:36 |
| mriedem | but also yes not string freeze | 17:36 |
| mriedem | since it's debug | 17:36 |
| dansmith | mriedem: right, I'm saying it would be more righter to change the %s to %(address)s, right? | 17:37 |
| mriedem | does it mater? | 17:37 |
| mriedem | *matter | 17:37 |
| mriedem | btw, nice usage of 'more righter' | 17:37 |
| *** alexschm has quit IRC | 17:37 | |
| mriedem | that should maybe arguably be INFO level | 17:38 |
| mriedem | since it seems to be an exceptional case | 17:38 |
| mriedem | like, i wanted to see how many times that is logged and a large ops job fails, but i can't | 17:38 |
| dansmith | ah | 17:39 |
| dansmith | yeah | 17:39 |
| dansmith | but we can't make it info at this point right? | 17:39 |
| mriedem | probably not | 17:39 |
| mriedem | *probably* | 17:39 |
| dansmith | on the previous thing, | 17:40 |
| dansmith | I don't know that it matters, but I was thinking someone would come along and -1 my +2 for "this should be a named format" thing | 17:40 |
| mriedem | why would someone do that? | 17:41 |
| mriedem | seems extremely nit picky | 17:41 |
| *** annegentle has joined #openstack-nova | 17:42 | |
| mriedem | when we have log messages with %s all over the place | 17:42 |
| dansmith | well, I can never figure out what is the flavor of the month with this log shit | 17:42 |
| mriedem | ha | 17:42 |
| mriedem | dansmith: so, drive everyone crazy and push up a hacking check for no log messages with %s in it | 17:42 |
| mriedem | i could see why named args are useful in *translated* messages | 17:42 |
| mriedem | for context to the translator | 17:43 |
| *** lpetrut has joined #openstack-nova | 17:43 | |
| openstackgerrit | Ben Nemec proposed openstack/nova: Make ironic ephemeral bdm match default value https://review.openstack.org/205271 | 17:43 |
| mriedem | e.g. %s is %s | 17:43 |
| mriedem | sh*t is f*cked | 17:43 |
| *** vishwanathj has quit IRC | 17:43 | |
| dansmith | well, I guess the reason I'm saying this is that I don't like passing the positional args to the log command instead of formatting them right then and there, | 17:43 |
| *** yassine_ has quit IRC | 17:43 | |
| dansmith | because we don't get the immediate "you formatted three things and passed two" help | 17:43 |
| dansmith | but I don't know that using the named params is any better in that regard | 17:44 |
| mriedem | ok, i see now | 17:44 |
| mriedem | well we have that check in the tests now | 17:44 |
| mriedem | for wonky debug log messages that miss this stuff | 17:44 |
| mriedem | which probably means that this code path is not tested | 17:44 |
| mriedem | which wouldn't surprise me | 17:44 |
| dansmith | yeah | 17:45 |
| mriedem | -or- that check only works on missed named parameters | 17:45 |
| mriedem | i'll dig a bit | 17:45 |
| *** mpaolino has quit IRC | 17:45 | |
| dansmith | well, I +2d because you shamed me | 17:45 |
| mriedem | for saying you were nit picking in a roundabout way? | 17:45 |
| mriedem | you have a point, i just wasn't grasping it | 17:46 |
| *** e0ne has joined #openstack-nova | 17:46 | |
| dansmith | for saying I was nit picking when I was just exposing my insecurity about log policy on Tuesday September 3, 2015 | 17:46 |
| mriedem | today is thursday | 17:46 |
| mriedem | c'mon dan | 17:46 |
| mriedem | ffs | 17:46 |
| dansmith | oh shit, it is | 17:46 |
| dansmith | that's much better than tuesday | 17:47 |
| *** tonytan4ever has quit IRC | 17:47 | |
| *** electrocucaracha has quit IRC | 17:47 | |
| *** ijw_ has quit IRC | 17:48 | |
| *** nic has joined #openstack-nova | 17:49 | |
| *** sayali has quit IRC | 17:50 | |
| *** SourabhP has joined #openstack-nova | 17:53 | |
| *** sayali has joined #openstack-nova | 17:53 | |
| *** bradjones has joined #openstack-nova | 17:53 | |
| *** bradjones has quit IRC | 17:53 | |
| *** bradjones has joined #openstack-nova | 17:53 | |
| *** emagana has quit IRC | 17:54 | |
| *** VikasC has quit IRC | 17:55 | |
| *** sachin_ has joined #openstack-nova | 17:56 | |
| *** VikasC has joined #openstack-nova | 17:57 | |
| *** tonytan4ever has joined #openstack-nova | 18:00 | |
| *** lpetrut has quit IRC | 18:06 | |
| *** mhorban has quit IRC | 18:06 | |
| *** sayali has quit IRC | 18:06 | |
| *** mhorban has joined #openstack-nova | 18:10 | |
| rlrossit | dansmith: you available for some questions on your big chain of for InstanceV2? | 18:13 |
| openstackgerrit | Sean Dague proposed openstack/python-novaclient: workaround for RAX repose configuration https://review.openstack.org/220192 | 18:13 |
| sdague | mriedem: there was a pep8 issue in the top patch | 18:13 |
| sdague | that fixes it | 18:13 |
| mriedem | tsk tsk | 18:14 |
| dansmith | rlrossit: yeah | 18:15 |
| *** vilobhmm_11 has joined #openstack-nova | 18:15 | |
| rlrossit | dansmith: for https://review.openstack.org/#/c/217817/3, why are you only updating tests to use objects.Instance, not anything else? | 18:15 |
| rlrossit | or have I not gotten far enough in the chain yet? | 18:15 |
| *** nic has quit IRC | 18:15 | |
| dansmith | rlrossit: because everything else should be using objects.Instance, but those tests are old, from before we did the objects.$OBJECT thing | 18:16 |
| *** Sree has joined #openstack-nova | 18:16 | |
| rlrossit | dansmith: gotcha | 18:19 |
| sdague | ok, so all the client patches are in the can. We've got the device=None patch in the gate. Last up from this morning was allowing spaces in server names, right? | 18:23 |
| sdague | did alex_xu get one up before he went to bed? | 18:23 |
| sdague | also, thinking about that, what's the implication for dns names? | 18:23 |
| dansmith | s/went to bed/ran out of battery/ | 18:23 |
| dansmith | sdague: this is display name | 18:23 |
| dansmith | sdague: unicode is allowed | 18:23 |
| sdague | oh, just display name | 18:23 |
| dansmith | name is calculated anyway | 18:24 |
| *** gyee has joined #openstack-nova | 18:24 | |
| openstackgerrit | sulochan-acharya proposed openstack/nova: xapi: cleanup volume sr on live migration rollback https://review.openstack.org/217105 | 18:28 |
| *** marzif has joined #openstack-nova | 18:28 | |
| *** rushil has quit IRC | 18:29 | |
| *** harlowja has quit IRC | 18:31 | |
| *** annegentle has quit IRC | 18:31 | |
| sdague | so in the schema name is enforced as a hostname | 18:31 |
| dansmith | sdague: name or display_name? | 18:32 |
| sdague | https://github.com/openstack/nova/blob/b3f07e8c1fa45e9ae2835ed028a092436ae2ad10/nova/api/openstack/compute/schemas/servers.py#L24 | 18:32 |
| *** harlowja has joined #openstack-nova | 18:32 | |
| dansmith | is that actually populated from display_name? | 18:32 |
| sdague | there's no display name passed int | 18:32 |
| mriedem | name is display_name in the api | 18:33 |
| mriedem | hostname there is weird | 18:33 |
| mriedem | thats why it's not happy about spaces | 18:33 |
| mriedem | that should just be string | 18:33 |
| dansmith | yeah, I don't think hostname is right at all | 18:33 |
| mriedem | someone was thinking. server.name == hostname | 18:33 |
| mriedem | https://github.com/openstack/nova/blob/master/nova/objects/instance.py#L155 | 18:34 |
| mriedem | instance.display_name is nullable also | 18:34 |
| mriedem | so it seems like this is the same as the volume.device case | 18:34 |
| mriedem | string or null | 18:34 |
| mriedem | or whatever v2 requires | 18:34 |
| dansmith | http://specs.openstack.org/openstack/nova-specs/specs/kilo/implemented/relax-resource-name-restrictions.html | 18:34 |
| sdague | so clearly that didn't get implemented, or something got reverted | 18:35 |
| mriedem | i think it was just regressed here | 18:35 |
| mriedem | i remmeber looking at this spec ffor server group or flavor names at one point | 18:35 |
| dansmith | yeah, this seems like overzealous api schema work | 18:35 |
| mriedem | like https://github.com/openstack/nova/blob/master/nova/compute/flavors.py#L54 | 18:36 |
| dansmith | https://review.openstack.org/#/c/119741/8/nova/api/openstack/compute/schemas/v3/servers.py,cm | 18:36 |
| dansmith | that's the patch to relax the names | 18:36 |
| dansmith | and it's the one that set it to hostname | 18:36 |
| dansmith | what's the definition of hostname/ | 18:36 |
| sdague | hostname is the restricted set | 18:36 |
| mriedem | https://github.com/openstack/nova/blob/master/nova/api/validation/parameter_types.py#L84 | 18:36 |
| *** _diana_ has joined #openstack-nova | 18:36 | |
| mriedem | 'pattern': '^[a-zA-Z0-9-._]*$', | 18:36 |
| sdague | so this is super confusing | 18:37 |
| sdague | the patch for that spec is - https://review.openstack.org/#/c/119741/8/nova/api/openstack/compute/schemas/v3/servers.py,cm | 18:37 |
| dansmith | I think this was an oversight | 18:37 |
| sdague | which increased the restriction | 18:37 |
| sdague | explicitly | 18:37 |
| mriedem | yeah | 18:37 |
| mriedem | i think it was a mistake | 18:37 |
| dansmith | clearly | 18:37 |
| mriedem | i'd say make it the same as flavor name | 18:37 |
| sdague | yeh, ok, changing back to name seems fine | 18:37 |
| mriedem | this https://github.com/openstack/nova/blob/master/nova/compute/flavors.py#L54 | 18:37 |
| mriedem | or yeah i guess just 'name': parameter_types.name but does that allow unicoe? | 18:38 |
| mriedem | *unicode | 18:38 |
| sdague | mriedem: I'm pretty sure | 18:38 |
| dansmith | https://review.openstack.org/#/c/119741/8/nova/api/validation/parameter_types.py,cm | 18:38 |
| sdague | look at how it's compiled above | 18:38 |
| *** marzif has quit IRC | 18:38 | |
| sdague | mriedem: https://github.com/openstack/nova/blob/b3f07e8c1fa45e9ae2835ed028a092436ae2ad10/nova/api/validation/parameter_types.py#L54-L55 | 18:39 |
| mriedem | yeah | 18:39 |
| dansmith | I wonder if sdague is just re-pasting all my links, | 18:39 |
| mriedem | https://github.com/openstack/nova/commit/f40b6a1d5c91628f78fdbc0417a3a394a69ab716 | 18:39 |
| dansmith | or just these | 18:39 |
| *** hparekh has joined #openstack-nova | 18:39 | |
| sdague | heh, sorry | 18:39 |
| mriedem | which is https://review.openstack.org/#/c/119741/8 | 18:39 |
| mriedem | ha | 18:39 |
| mriedem | jinx! | 18:39 |
| mriedem | so it seems we know what needs to be done | 18:40 |
| mriedem | typey typey as mikal would say | 18:40 |
| sdague | ok, I'll do the typey typey | 18:40 |
| sdague | so, do we have an opinion on whether we need a microversion for this? | 18:41 |
| dansmith | are we going to honor the old behavior if we do? | 18:41 |
| mriedem | i'd say no to microversoin | 18:42 |
| mriedem | b/c we said v2.1 v2.1 is compat with v2.0 | 18:42 |
| mriedem | so saying, except for when it's v2.14 | 18:42 |
| mriedem | seems wrong | 18:42 |
| sdague | ok, v2.1 on kilo if deployed will be different | 18:42 |
| *** chhavi has joined #openstack-nova | 18:42 | |
| mriedem | yeah | 18:42 |
| mriedem | but this makes things less restrictive | 18:42 |
| mriedem | so code that works on v2.1 kilo would also work with this | 18:42 |
| dansmith | I give up trying to figure this api shit out | 18:43 |
| mriedem | sdague: plus you could make the same argument for the device=None thing | 18:43 |
| sdague | but not the other way, between 2 clouds. Which, is probably super low risk given v2.1 is off by default | 18:43 |
| sdague | mriedem: yeh, I think I'm fine calling it bug fix | 18:43 |
| mriedem | i'm always fine with calling a thing a bug fix when it's in my interest to do so :) | 18:43 |
| sdague | we screwed up | 18:43 |
| sdague | we're un screwing | 18:43 |
| sdague | probably means we should backport as well | 18:44 |
| mriedem | yes | 18:44 |
| mriedem | i was going to backport dims' fix also | 18:44 |
| *** ctrath has quit IRC | 18:44 | |
| sdague | yeh, I think that's a plan | 18:44 |
| sdague | just out of curiosity, where's the code that sanitizes names for hostnames? | 18:45 |
| *** jordantardif has left #openstack-nova | 18:45 | |
| mriedem | you mean where is https://github.com/openstack/nova/blob/b3f07e8c1fa45e9ae2835ed028a092436ae2ad10/nova/api/validation/parameter_types.py#L84 used ? | 18:45 |
| sdague | no, I mean at some point we take this data and shove it into dnsmasq | 18:46 |
| sdague | clearly it has to be scrubbed before that | 18:46 |
| mriedem | anywhere there is a host attribute in the db | 18:46 |
| mriedem | service.host | 18:46 |
| *** emagana has joined #openstack-nova | 18:46 | |
| mriedem | network.host | 18:46 |
| *** nic has joined #openstack-nova | 18:47 | |
| mriedem | looks like it's also used for migrate/evacuate/aggregates/cells | 18:47 |
| dansmith | um | 18:47 |
| dansmith | service.host and network.host and internal names, nothing to do with instance.display_name | 18:47 |
| mriedem | i konw | 18:48 |
| *** electrocucaracha has joined #openstack-nova | 18:48 | |
| mriedem | i'm not sure why he's asking | 18:48 |
| dansmith | because we set up dnsmasq for the instance with the hostname | 18:48 |
| dansmith | if we're on nova-net | 18:48 |
| sdague | right, that | 18:48 |
| dansmith | sdague: https://github.com/openstack/nova/blob/master/nova/network/manager.py#L942-L956 | 18:48 |
| dansmith | sdague: however, | 18:49 |
| dansmith | sdague: remember that our definition of hostname was not actually DNS-compliant | 18:49 |
| dansmith | because the neutron folks have a stricter and more righterer definition | 18:49 |
| *** ijw_ has joined #openstack-nova | 18:50 | |
| *** ctrath has joined #openstack-nova | 18:50 | |
| *** yamahata has quit IRC | 18:50 | |
| *** tongli has joined #openstack-nova | 18:51 | |
| *** yamahata has joined #openstack-nova | 18:51 | |
| sdague | yeh, so it seems like we probably need a scrubber in here, otherwise we're going to generate gorp in dnsmasq. Unless I missed it, I'm not seeing a filter anywhere. | 18:51 |
| dansmith | well, we didn't before right? | 18:52 |
| sdague | I guess | 18:52 |
| sdague | maybe, that just becomes another bug that can be fixed | 18:52 |
| *** tongli has quit IRC | 18:52 | |
| *** eharney has quit IRC | 18:52 | |
| dansmith | I'm looking | 18:53 |
| *** tri2sing has quit IRC | 18:53 | |
| *** lpetrut has joined #openstack-nova | 18:53 | |
| dansmith | sdague: so what is your proposed filter? | 18:55 |
| *** ijw_ has quit IRC | 18:55 | |
| dansmith | meaning, | 18:55 |
| dansmith | you can either punt if there is a space, | 18:56 |
| dansmith | or try to coerce to _ | 18:56 |
| dansmith | or fail, etc | 18:56 |
| *** armax has quit IRC | 18:57 | |
| *** jwcroppe_ is now known as jwcroppe | 18:58 | |
| *** Sree has quit IRC | 18:58 | |
| *** Sree has joined #openstack-nova | 18:58 | |
| sdague | so, I was thinking coerce anything that's not a valid dns character to _ | 18:59 |
| sdague | which I realize isn't exactly right, but it would be more of a thing | 18:59 |
| dansmith | and then do the dupe check? | 18:59 |
| sdague | maybe | 18:59 |
| sdague | anyway, I think that's a separate bug honestly | 18:59 |
| *** VikasC has quit IRC | 19:00 | |
| dansmith | you know that _ is not actually a valid character for DNS right? :) | 19:00 |
| mriedem | couldn't you just check it against the parameter_types.hostname regex? | 19:00 |
| mriedem | same as in the api? | 19:00 |
| sdague | dansmith: well, - | 19:00 |
| dansmith | mriedem: well, the question is whether you validate and ignore if it fails, or try to fix it | 19:00 |
| sdague | anyway, different bug... let me finish this commit message | 19:00 |
| mriedem | remember caps and periods | 19:00 |
| *** annegentle has joined #openstack-nova | 19:01 | |
| openstackgerrit | Sean Dague proposed openstack/nova: relax restrictions on server name https://review.openstack.org/220279 | 19:01 |
| mriedem | dansmith: i guess we were just ignoring it before? | 19:01 |
| sdague | yeh, we were | 19:01 |
| dansmith | mriedem: I'm guessing we were writing a hosts file where there was a space in it, | 19:01 |
| sdague | yeh, that's my guess | 19:01 |
| dansmith | such that dnsmasq saw "address foo bar.domain.com | 19:01 |
| sdague | because all I see is a .lower() as transform | 19:02 |
| *** tri2sing has joined #openstack-nova | 19:02 | |
| dansmith | sdague: so this is a bug, but not blocking anyone right? | 19:02 |
| sdague | I believe trove rewrote their tests to not do this | 19:03 |
| sdague | so, yes | 19:03 |
| *** Sree has quit IRC | 19:03 | |
| mriedem | sdague: there are no unit test hits for this? | 19:03 |
| sdague | mriedem: it passed existing tests | 19:03 |
| dansmith | in that case, I think we should let johnthetubaguy and maybe jaypipes comment on this | 19:03 |
| dansmith | alaski: any opinions? | 19:03 |
| sdague | yep, I'm fine with that | 19:03 |
| mriedem | sdague: we should probably at least have a unit test that creates a server against the v2.1 api with a space in the name | 19:03 |
| mriedem | since that's the bug | 19:03 |
| dansmith | mriedem: yes | 19:04 |
| sdague | mriedem: yeh, agree, I'll do that tomorrow morning. I'm about done for the day. | 19:04 |
| dansmith | mriedem: you should -2 with "nobody is above the law, sean." in that condescending voice of yours | 19:04 |
| alaski | reading back | 19:04 |
| mriedem | except dredd | 19:05 |
| dansmith | alaski: I can summarize | 19:05 |
| mriedem | b/ | 19:05 |
| mriedem | *b/c, you know, he is the law | 19:05 |
| dansmith | alaski: when we did that blueprint for juno about relaxing names of things to include unicode, etc, | 19:05 |
| mriedem | sdague: if i get some time this afternoon i'll add the test | 19:05 |
| dansmith | alaski: we apparently made instance.display_name more strict | 19:05 |
| dansmith | alaski: strict to a fake DNS regex, which doesn't include spaces | 19:05 |
| dansmith | alaski: that was a regression reported from v2.0 to v2.1 | 19:05 |
| dansmith | alaski: our feeling is that display name should be a name and should be wide open | 19:06 |
| mriedem | danpbot (tm) | 19:06 |
| alaski | yep, totally agree with that | 19:06 |
| dansmith | alaski: and for nova-network, we should probably just sanitize it to something reasonable | 19:06 |
| alaski | seems reasonable to me | 19:06 |
| alaski | it does have an issue of potential collisions, but I think we have that anyways | 19:07 |
| dansmith | alaski: apparently before we were just writing out a broken hosts file with a space in it and nobody noticed or cared | 19:07 |
| dansmith | alaski: we would filter and then run the usual unique check that we do now | 19:07 |
| *** eharney has joined #openstack-nova | 19:07 | |
| alaski | I'm totally on board with all of that | 19:07 |
| dansmith | okay cool | 19:08 |
| dansmith | just wanted an actually-supports-an-api perspective :) | 19:08 |
| alaski | I've always disliked that display name was linked to hostname anyways, but that's a different convo | 19:08 |
| dansmith | yeah, totes | 19:08 |
| *** electrocucaracha has quit IRC | 19:08 | |
| *** _diana_ has left #openstack-nova | 19:08 | |
| alaski | in my experience every just grabs ips from the server response and does everything with that. it's a pain since they want to hold on to that IP forever, but very few seem to care about DNS/hostnames | 19:10 |
| dansmith | I think uuid.domain.com would make more sense | 19:11 |
| dansmith | displayname.domain.com is a nice ideal, but not in practice I think | 19:11 |
| *** hparekh has quit IRC | 19:13 | |
| alaski | yep | 19:13 |
| *** SourabhP has quit IRC | 19:16 | |
| *** SourabhP has joined #openstack-nova | 19:16 | |
| *** _diana_ has joined #openstack-nova | 19:17 | |
| *** achanda has quit IRC | 19:18 | |
| *** alex_klimov has joined #openstack-nova | 19:23 | |
| *** alexpilotti has quit IRC | 19:26 | |
| *** lpetrut has quit IRC | 19:27 | |
| *** Marga_ has joined #openstack-nova | 19:29 | |
| *** chhavi has quit IRC | 19:29 | |
| *** armax has joined #openstack-nova | 19:29 | |
| *** lpetrut has joined #openstack-nova | 19:30 | |
| *** changbl has quit IRC | 19:30 | |
| melwitt | dansmith alaski: one of the first things I did working on nova internally was untie display_name from hostname. we do either a random hostname generation based on ip or lookup a static name by ip. and people do use the hostnames, we also have a custom hostname option in some clusters where the user can pass in a hostname | 19:31 |
| *** VW_ has quit IRC | 19:32 | |
| *** VW_ has joined #openstack-nova | 19:32 | |
| melwitt | dansmith alaski: in all cases, display_name is just a friendly nickname that appears alongside the hostname in 'nova list' (we also return hostname as a server detail for 'nova list' 'nova show') so you can see for example "dev" and "znc" are a couple I have | 19:32 |
| *** thumpba has joined #openstack-nova | 19:32 | |
| *** ivasev has quit IRC | 19:33 | |
| dansmith | melwitt: right on | 19:33 |
| *** matrohon has joined #openstack-nova | 19:33 | |
| alaski | melwitt: nice. that's a much better way to handle it | 19:34 |
| *** ijw_ has joined #openstack-nova | 19:34 | |
| *** lpetrut has quit IRC | 19:36 | |
| *** VW_ has quit IRC | 19:37 | |
| melwitt | (correction, no random hostname generation but use a template and generate a hostname based on the ip to ensure it's unique) | 19:37 |
| *** VW_ has joined #openstack-nova | 19:39 | |
| *** electrocucaracha has joined #openstack-nova | 19:41 | |
| mriedem | dansmith: weird, i added a test for this logging thing to hit that block, intentionally left the fix out and ran the test, it blows up, but doesn't actually log the failure in the console output | 19:50 |
| mriedem | but my mock assert on the fip.disassociate right after the log statement fails - because it blows up on the bad log.debug call | 19:50 |
| mriedem | seems the log handling stuff in our test fixtures has maybe regressed | 19:50 |
| sdague | so, maybe that's the thing we should do for real in M, split up name and displayname | 19:51 |
| *** tri2sing has quit IRC | 19:53 | |
| *** pixelbeat has quit IRC | 19:54 | |
| *** eharney has quit IRC | 19:55 | |
| *** electrocucaracha has quit IRC | 19:58 | |
| *** eharney has joined #openstack-nova | 20:00 | |
| sdague | mriedem: so for some reason tests didn't run on patch #2 in the series | 20:00 |
| *** emagana has quit IRC | 20:00 | |
| sdague | just did a recheck | 20:00 |
| sdague | with any luck in about 2 hours all that should merge | 20:00 |
| *** Marga_ has quit IRC | 20:01 | |
| melwitt | sdague: I didn't know this until I checked around yesterday and the security groups create optional description issue got addressed by updating the API docs to make it required | 20:02 |
| *** derekh has joined #openstack-nova | 20:02 | |
| *** chutwig is now known as erhud1 | 20:04 | |
| *** erhud1 is now known as erhudy1 | 20:04 | |
| *** emagana has joined #openstack-nova | 20:06 | |
| *** dave-mccowan has quit IRC | 20:06 | |
| *** husanu has quit IRC | 20:07 | |
| *** husanu has joined #openstack-nova | 20:09 | |
| mriedem | sdague: would it be valid to throw this assertion http://git.openstack.org/cgit/openstack/nova/tree/nova/tests/unit/test_fixtures.py#n110 into the test_debug_logging test case below it? | 20:11 |
| *** e0ne has quit IRC | 20:12 | |
| *** baoli_ has quit IRC | 20:13 | |
| mriedem | methinks oslo.log might be regressed | 20:13 |
| mriedem | dims: ^ | 20:13 |
| *** achanda has joined #openstack-nova | 20:18 | |
| dims | mriedem: this one? https://github.com/openstack/oslo.i18n/commit/8d49c279cf28c49666bc7efb0f5ce55eae552841 | 20:19 |
| mriedem | maybe? | 20:20 |
| mriedem | this is a debug log message | 20:20 |
| *** derekh has quit IRC | 20:20 | |
| mriedem | i need to try and see if it fails with a log.info message | 20:20 |
| *** lpetrut has joined #openstack-nova | 20:21 | |
| mriedem | dims: so it can't be that since that's only for _() objects | 20:22 |
| mriedem | and this isn't using a message object | 20:22 |
| mriedem | could be a regression in the upstream fixtures library for all i konw | 20:23 |
| *** achanda has quit IRC | 20:24 | |
| *** e0ne has joined #openstack-nova | 20:25 | |
| mriedem | think i found something | 20:25 |
| *** ZZelle_ has joined #openstack-nova | 20:26 | |
| *** e0ne has quit IRC | 20:26 | |
| *** Marga_ has joined #openstack-nova | 20:27 | |
| *** eharney has quit IRC | 20:27 | |
| *** IanGovett has quit IRC | 20:29 | |
| *** achanda has joined #openstack-nova | 20:29 | |
| *** dave-mccowan has joined #openstack-nova | 20:32 | |
| *** thangp has quit IRC | 20:33 | |
| *** vladikr has quit IRC | 20:36 | |
| openstackgerrit | Chuck Carmack proposed openstack/nova: Make the server group invalid format message more verbose https://review.openstack.org/218468 | 20:36 |
| *** achanda has quit IRC | 20:38 | |
| *** lpetrut has quit IRC | 20:38 | |
| mikal | Morning | 20:38 |
| mriedem | lifeless: given https://github.com/testing-cabal/fixtures/commit/354acf568aa86bb7d43a01c23d73c750f601b335 - isn't this wrong? https://github.com/openstack/nova/blob/master/nova/tests/fixtures.py#L105 | 20:41 |
| *** eharney has joined #openstack-nova | 20:42 | |
| mriedem | mikal: do you know anything about libvirt? :) | 20:42 |
| mriedem | the usual 4 people i'd ask aren't on | 20:42 |
| dansmith | rlrossit: on this: https://review.openstack.org/#/c/217816/ | 20:43 |
| dansmith | rlrossit: are you keeping the -1 on there for the comment thing or is there something else? | 20:44 |
| rlrossit | dansmith: it was for the comment and very slightly for the >= vs. > thing | 20:44 |
| dansmith | rlrossit: did I explain away the >= thing? | 20:45 |
| rlrossit | yep | 20:45 |
| dansmith | I'm trying to figure out how your suggested comment makes more sense | 20:46 |
| dansmith | they seem equivalent to me, except that yours is a little more duplicative of the actual code | 20:46 |
| rlrossit | yeah idk | 20:46 |
| rlrossit | I initially read it and saw we weren't *always* setting it | 20:47 |
| *** achanda has joined #openstack-nova | 20:47 | |
| rlrossit | kinda got confused but the code is pretty straight-forward after I looked at it for a while | 20:47 |
| rlrossit | (I also had to figure out what registration_hook() did in the first place too) | 20:47 |
| dansmith | I was trying to avoid doing this, but see if this is worthy of your +1: | 20:49 |
| openstackgerrit | Dan Smith proposed openstack/nova: Pull out the common bits of InstanceList into _BaseInstanceList https://review.openstack.org/219337 | 20:49 |
| openstackgerrit | Dan Smith proposed openstack/nova: Update objects test infrastructure for multiple versions https://review.openstack.org/217818 | 20:49 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add Instance and InstanceList v2.0 objects https://review.openstack.org/217819 | 20:49 |
| openstackgerrit | Dan Smith proposed openstack/nova: Fix an issue with NovaObjectRegistry hook https://review.openstack.org/217816 | 20:49 |
| openstackgerrit | Dan Smith proposed openstack/nova: Make the conductor fixture use version manifests https://review.openstack.org/217851 | 20:49 |
| openstackgerrit | Dan Smith proposed openstack/nova: Refactor Instance tests to use objects.Instance https://review.openstack.org/217817 | 20:49 |
| openstackgerrit | Dan Smith proposed openstack/nova: Pull out the common bits of Instance into _BaseInstance https://review.openstack.org/217815 | 20:49 |
| openstackgerrit | Dan Smith proposed openstack/nova: Teach conductor to do manifest-based object_class_action() things https://review.openstack.org/219432 | 20:49 |
| rlrossit | holy moly | 20:49 |
| rlrossit | dansmith: that comment explains things much better | 20:51 |
| dansmith | okay | 20:51 |
| mikal | mriedem: I can provide generic confused answers | 20:53 |
| mikal | mriedem: what's up? | 20:53 |
| mriedem | mikal: just in time | 20:53 |
| mriedem | https://bugs.launchpad.net/nova/+bug/1492026 | 20:53 |
| openstack | Launchpad bug 1492026 in OpenStack Compute (nova) "domain.attachDeviceFlags randomly doesn't do anything in libvirt" [Medium,Confirmed] | 20:53 |
| mriedem | for me to write it down | 20:53 |
| mriedem | i'm assuming this is just a crazy bug in libvirt | 20:54 |
| mikal | Ahhh, well I am not a libvirt internals guy, but let me look | 20:54 |
| mriedem | virDomainAttachDeviceFlags is called but doesn't log or do anything, doesn't fail, doesn't attach, doesn't show up as a call in the libvirt debug logs for the domain | 20:54 |
| mriedem | nothing | 20:54 |
| mikal | For reference, tonyb has done some libvirt internals stuff | 20:54 |
| mriedem | tonyb: ^ | 20:55 |
| mriedem | otherwise i'll be bugging danpb in the morning | 20:55 |
| mikal | Looking at this bug | 20:55 |
| mikal | I read what you wrote as saying nova never calls attach but does generate the xml | 20:55 |
| mikal | Or am I misunderstanding you? | 20:56 |
| mriedem | nova calls domain.attachDeviceFlags | 20:56 |
| mikal | Ok, and libvirt misses it somehow? | 20:56 |
| mriedem | i see that here: http://logs.openstack.org/55/218355/6/check/gate-tempest-dsvm-full/e0da291/logs/screen-n-cpu.txt.gz#_2015-09-02_17_42_10_274 | 20:56 |
| mriedem | yeah | 20:56 |
| mriedem | it's like it goes into a blackhole | 20:56 |
| mikal | Yeah, we need a libvirt dev I would think | 20:56 |
| mikal | tonyb might have some time today now that some of the security stuff is less urgent | 20:57 |
| mikal | I'll ask him | 20:57 |
| mikal | Noting that he's still trying to fix juno too | 20:57 |
| mikal | Or his "tangled web of onions" as he has started calling it | 20:57 |
| mriedem | why is he working on juno btw? | 20:57 |
| mikal | That is a long complicated story | 20:57 |
| mriedem | i have 33 minutes | 20:57 |
| mikal | I think it started with a security backport and broken gating | 20:57 |
| mikal | Its ended up with bloody mindedness and trying to be a good bloke | 20:57 |
| mriedem | oh, the on_completion callback for exe? | 20:58 |
| mikal | I'm not sure, I've been bravely running away | 20:58 |
| flwang1 | mriedem: hi, a quick question, is there any way to detach a volume which server has been deleted? | 20:58 |
| mriedem | flwang1: you shouldn't be able to delete a server instance that has attached volumes w/o nova first detaching those volumes | 20:58 |
| flwang1 | mriedem: i know and I see. but now we have a customer report a bug that he can't detach his volume and after checked, i saw the server has gone | 20:59 |
| *** electrocucaracha has joined #openstack-nova | 20:59 | |
| mikal | mriedem: can you spend some of your 33 minutes looking at https://review.openstack.org/#/c/219980/ please? | 20:59 |
| mikal | mriedem: I've been a good boy | 20:59 |
| *** vilobhmm_11 has quit IRC | 21:00 | |
| *** vilobhmm_11 has joined #openstack-nova | 21:00 | |
| *** rlrossit has left #openstack-nova | 21:00 | |
| anish | flwang1: maybe this is what you want ? https://mcwhirter.com.au/craige/blog/2014/Deleting_Root_Volumes_Attached_to_Non-Existent_Instances/ | 21:00 |
| flwang1 | anish: i would like to give you a big hug | 21:01 |
| *** tjones has joined #openstack-nova | 21:01 | |
| anish | flwang1: INTERNET HUG ! | 21:01 |
| flwang1 | yep, send from New Zealand to wherever you are :) | 21:02 |
| openstackgerrit | Sylvain Bauza proposed openstack/nova-specs: Proposes the RequestSpec object for Mitaka https://review.openstack.org/220312 | 21:02 |
| anish | mikal: "pending_soft_became_hard" quite the name there | 21:03 |
| mikal | anish: thanks! | 21:03 |
| mikal | anish: I was going for descriptive and fits on a line | 21:03 |
| *** jwcroppe_ has joined #openstack-nova | 21:03 | |
| doffm | anish: flwang1: Sounds like a horribe work around. | 21:03 |
| doffm | There seems to be a related bug: | 21:04 |
| doffm | flwang1: https://bugs.launchpad.net/nova/+bug/1335889 | 21:04 |
| openstack | Launchpad bug 1335889 in OpenStack Compute (nova) "Race condition in quickly attaching / deleting volumes" [High,Confirmed] | 21:04 |
| *** dramakri has joined #openstack-nova | 21:04 | |
| *** thorst has quit IRC | 21:04 | |
| *** aysyd has quit IRC | 21:04 | |
| anish | doffm: yes, but I'm not aware of the proper way to fix it, since it shouldn't have happened in the first place | 21:05 |
| anish | more a case of i-ve-accidentally-done-this-on-devstack-too-many-times | 21:05 |
| dims | mriedem: am trying to recreate your scenario without fixtures... seems to be doing the right thing - http://paste.openstack.org/show/444708/ | 21:05 |
| *** dustins has quit IRC | 21:05 | |
| mriedem | anish: doffm: flwang1: i'm pretty sure there is already a patch for that bug | 21:06 |
| *** jwcroppe has quit IRC | 21:06 | |
| *** dramakri has left #openstack-nova | 21:06 | |
| flwang1 | anish: seems the case in that blog is not really same like my case | 21:06 |
| anish | :/ | 21:07 |
| * mikal wanders off to wake his kids before school | 21:07 | |
| doffm | flwang1: mriedem: There is. Cinder patch for forcing a delete of the volume. Now abandoned. | 21:08 |
| *** tdurakov_ has joined #openstack-nova | 21:08 | |
| *** raildo is now known as raildo-afk | 21:08 | |
| flwang1 | doffm: force delete? | 21:09 |
| flwang1 | i think i need a force detach | 21:09 |
| mriedem | there is a separate force delete thing | 21:09 |
| mriedem | for nova-manage | 21:09 |
| doffm | flwang1: Ahem... pardon me. Force detach. | 21:09 |
| flwang1 | IIRC, cinder support that, right? | 21:09 |
| mriedem | it basically deletes the bdm associated with the instance in the nova db | 21:09 |
| mriedem | yes, it's a force detach | 21:10 |
| *** greghaynes has quit IRC | 21:10 | |
| flwang1 | mriedem: i guess it's not in icehouse :( | 21:10 |
| mriedem | it's not in anything yet | 21:10 |
| flwang1 | mriedem: :-/ | 21:10 |
| mriedem | scottda: what happened with the force-detach nova-manage thing? | 21:10 |
| mriedem | did that merge? | 21:10 |
| mriedem | flwang1: you need to tell your customers to move up to non-EOL branches | 21:11 |
| *** Piet has quit IRC | 21:11 | |
| scottda | mriedem: now, it didn't ... | 21:11 |
| flwang1 | mriedem: we're in upgrading | 21:11 |
| mriedem | scottda: i can't seem to find the patch | 21:11 |
| *** tdurakov_ has quit IRC | 21:11 | |
| scottda | andrearosa: came up with a new patch , and I'm actually testing it ATM | 21:11 |
| mriedem | link? | 21:11 |
| scottda | https://review.openstack.org/#/c/184537/9 | 21:12 |
| scottda | But that's not the latest... | 21:12 |
| scottda | he hasn't pushed it up because he didn't have time to test it. | 21:12 |
| *** VW_ has quit IRC | 21:12 | |
| *** tdurakov__ has joined #openstack-nova | 21:13 | |
| *** VW_ has joined #openstack-nova | 21:13 | |
| scottda | hmmm..doesn't seem to work | 21:14 |
| mriedem | oh right, that one, i was looking for 'force' in the title | 21:14 |
| scottda | Yeah, it seems kinda stalled. | 21:15 |
| flwang1 | mriedem: in my case, the volume is not root volume | 21:15 |
| flwang1 | so it's allowed to delete the instance firstly | 21:15 |
| *** tdurakov__ has quit IRC | 21:15 | |
| flwang1 | and the problem is the attach relationship is not cleared as expected | 21:16 |
| mriedem | flwang1: have you recreate this on master? | 21:16 |
| mriedem | *recreated | 21:16 |
| *** rlrossit has joined #openstack-nova | 21:17 | |
| *** VW_ has quit IRC | 21:18 | |
| flwang1 | no, i can't even recreate it on the same env | 21:19 |
| bauzas | melwitt: have you already looked at what we discussed on Tempest/devstack for the cells job ? | 21:19 |
| flwang1 | mriedem: i saw this bp https://blueprints.launchpad.net/nova/+spec/add-force-detach-to-nova | 21:19 |
| bauzas | melwitt: or do you want me to see if I can help with that ? | 21:19 |
| mriedem | flwang1: yeah that's for https://review.openstack.org/#/c/184537/ | 21:20 |
| flwang1 | and john mentioned there should/will be a work around | 21:20 |
| flwang1 | mriedem: ah, got it | 21:20 |
| bauzas | alaski: melwitt: btw. FYI https://review.openstack.org/#/c/218473/ | 21:20 |
| melwitt | bauzas: I haven't done anything with it yet. please feel free :) I have only so far reviewed the shelve one https://review.openstack.org/#/c/220158/ | 21:20 |
| bauzas | alaski: -W'd because I wonder if we should recall the instance creation | 21:21 |
| *** jwcroppe_ has quit IRC | 21:21 | |
| bauzas | melwitt: okay, I think it will need a lot of changes, so I'll probably provide the first one and we'll be able to incrementally provide some | 21:21 |
| melwitt | bauzas: cool, will take a look at your patch | 21:22 |
| alaski | bauzas: I don't think we do, but I need to trace through that | 21:22 |
| *** jwcroppe has joined #openstack-nova | 21:22 | |
| bauzas | alaski: yeah I did trace that | 21:22 |
| melwitt | bauzas: sounds good on the cells job thing | 21:22 |
| bauzas | alaski: but that's still unclear | 21:22 |
| alaski | bauzas: I suspect we want to remove the create() that's there, not add another | 21:22 |
| flwang1 | mriedem: so seems there is no good way to fix it except hacking database? | 21:23 |
| bauzas | alaski: okay, I'll look at that too | 21:23 |
| mriedem | flwang1: well i think you're looking for https://review.openstack.org/#/c/184537/ | 21:24 |
| alaski | bauzas: I mean in a different patch. just based on comstuds comment and wondering when we would actually hit that and want it | 21:24 |
| bauzas | alaski: of course the above change removes the exception from the parent cell | 21:24 |
| mriedem | flwang1: or whatever that's going to end up being | 21:24 |
| bauzas | alaski: but I really need to understand why it's raising | 21:24 |
| bauzas | alaski: yeah agreed, but I still wonder why comstud provided that | 21:25 |
| *** tri2sing has joined #openstack-nova | 21:25 | |
| mriedem | mikal: https://review.openstack.org/#/c/219859/ - i've seen your name in some of that code | 21:25 |
| alaski | bauzas: probably being cautious | 21:25 |
| mriedem | mikal: would help with debugging lxc failures when running devstack + tempest + lxc | 21:25 |
| bauzas | alaski: well, I need to see if that commit was before the 5 times DB check | 21:25 |
| *** eharney has quit IRC | 21:26 | |
| bauzas | alaski: because if by 5 times we can't save the instance, there is no reason to try again a 6th time IMHO | 21:26 |
| flwang1 | mriedem: ok, thank you very much | 21:26 |
| mriedem | ccarmack: why did you abandon this? https://review.openstack.org/#/c/219778/ | 21:28 |
| lifeless | mriedem: its not wrong, it is stale | 21:28 |
| mriedem | ccarmack: for https://review.openstack.org/#/c/220158/ ? | 21:28 |
| lifeless | mriedem: if its a resource that *could* leak, then its missing the leak protection you'd get if you used _setUp | 21:28 |
| ccarmack | mriedem: for some reason, vim kept putting a linefeed at the end of file, so I started over | 21:28 |
| mriedem | lifeless: yeah so it's supported, but we should be using _setUp | 21:28 |
| mriedem | oh vim | 21:29 |
| ccarmack | mriedem then I used sed to remove the 3 shelve lines... | 21:29 |
| *** jwcroppe_ has joined #openstack-nova | 21:29 | |
| mriedem | mikal: i didn't completely understand your init reboot patch, put some comments inline | 21:29 |
| *** penick has quit IRC | 21:29 | |
| mriedem | mikal: but it's also been a long day and me brain no work very good | 21:30 |
| *** jwcroppe has quit IRC | 21:31 | |
| ccarmack | mriedem: if there's anything else I can do for the cells tempest changes, let me know | 21:31 |
| mriedem | ccarmack: there are probably others with the same pattern | 21:31 |
| mriedem | see if there are tests we're blacklisting which already have feature toggles in tempest | 21:32 |
| *** penick has joined #openstack-nova | 21:32 | |
| mriedem | i think there was a resize one | 21:32 |
| mriedem | alaski: cells supports resize thoug hright? | 21:32 |
| mriedem | just couldn't pass the test? | 21:32 |
| *** jcoufal has quit IRC | 21:32 | |
| mriedem | ccarmack: aggregates might be a better one to start on | 21:32 |
| ccarmack | mriedem: was there a list of features supported by cells? | 21:33 |
| alaski | mriedem: yes it supports resize | 21:33 |
| mriedem | ccarmack: i don't think so | 21:33 |
| mriedem | just tribal knowledge of what it doesn't support | 21:33 |
| alaski | mriedem: I may be misremembering but resize might have failed due to using security groups or something else in it | 21:34 |
| mriedem | oh, yeah, maybe | 21:34 |
| mriedem | heh, compute-feature-enabled.security_groups = False | 21:34 |
| melwitt | alaski mriedem: I think resize will work as long as the flavor to size to is available in the cell db. if it's not, it will fail | 21:36 |
| alaski | yeah, that too | 21:36 |
| melwitt | alaski mriedem: in tempest it fails because devstack creates a new flavor through the API "m1.nano" and then uses it for the resize tests IIRC, so it won't work | 21:37 |
| alaski | right, so I was remembering incorrectly | 21:38 |
| mriedem | we cant get that flavor into the cells db? | 21:38 |
| melwitt | (because nothing propagates flavor from API to cell) | 21:38 |
| mriedem | hrm | 21:38 |
| alaski | we looked at doing that work and it was kind of decided we wouldn't touch it now | 21:38 |
| *** annegentle has quit IRC | 21:38 | |
| mriedem | we should be able to just tell tempest which flavors to use | 21:38 |
| mriedem | there are 2 of them | 21:38 |
| mriedem | mtreinish: ^ right> | 21:38 |
| melwitt | I think only instances are synced API <--> cell. everything else not | 21:38 |
| *** VW_ has joined #openstack-nova | 21:39 | |
| mtreinish | mriedem: yeah tempest is given 2 flavors | 21:39 |
| *** burgerk has quit IRC | 21:39 | |
| mtreinish | I think you can get away with 1, but I'd have to double check | 21:39 |
| mriedem | mtreinish: but tmepest doesn't like create it's own and use it does it? for resize tests? | 21:39 |
| mtreinish | mriedem: no it resizes between the 2 provided | 21:39 |
| mriedem | ok, | 21:39 |
| *** marzif has joined #openstack-nova | 21:39 | |
| mtreinish | mriedem: the flavors tests create a flavor, but they're not used for anything | 21:39 |
| mriedem | so melwitt - seems like we should just not do that thing in devstack if running cells | 21:39 |
| mtreinish | except to verify that the flavors calls worked | 21:40 |
| mriedem | flavor_ref=1 | 21:40 |
| mriedem | flavor_ref_alt=2 | 21:40 |
| mriedem | yar | 21:40 |
| mriedem | melwitt: alaski: seems we could just have a bug that says that resize test fails with cells b/c devstack assumes a wrong thing | 21:40 |
| mtreinish | mriedem: well I think devstack uses 42 and 84 by default | 21:41 |
| mriedem | my point is, if cells, don't auto-create flavors and use those in tempest.conf | 21:41 |
| mtreinish | someone's a fan of the hitchhikers guide | 21:41 |
| mriedem | use whatever nova flavor-list provides | 21:41 |
| alaski | mriedem: that's good with me | 21:41 |
| alaski | I think devstack creates flavors properly for cells though | 21:42 |
| alaski | at least initially | 21:42 |
| melwitt | or have devstack create the flavor in the cell db too if it's done via nova-manage. I'd have to look at it again | 21:42 |
| alaski | if something is getting created later we missed it | 21:42 |
| melwitt | I don't remember how it's doing it | 21:42 |
| mriedem | right here https://github.com/openstack-dev/devstack/blob/master/lib/tempest#L202 | 21:43 |
| melwitt | oh, API then, not nova-manage | 21:43 |
| alaski | ahh yeah, that kills the cells | 21:43 |
| *** jwcroppe has joined #openstack-nova | 21:43 | |
| mriedem | heh | 21:43 |
| mriedem | can we start referring to cells as 'the cells'? | 21:43 |
| rlrossit | dansmith: I thought I was a genius on https://review.openstack.org/#/c/217818/5 but I just confused myself | 21:43 |
| melwitt | mriedem: +1 | 21:43 |
| ccarmack | melwitt, alaski, mriedem: should the API be fixed to pass the flavor to cells? (I'm sure I'm missing something) | 21:44 |
| dansmith | rlrossit: no comments there, did you mean to comment? | 21:44 |
| mriedem | ccarmack: i think we're talking about just a devstack change | 21:45 |
| mriedem | but i have to run | 21:45 |
| rlrossit | dansmith: no, because I found a super awesome way to improve it, but I tried it and it didn't work | 21:45 |
| mriedem | ccarmack: melwitt: alaski: sounds like we could file a bug for this thought | 21:45 |
| mriedem | *though | 21:45 |
| mtreinish | alaski: so it needs another if is_service_enabled n-cells in that block | 21:45 |
| mtreinish | ? | 21:45 |
| alaski | ccarmack: it's non trivial. The plumbing is fairly easy but we never took the time to find a good solution to what to do if a cell is down during that | 21:45 |
| *** jwcroppe_ has quit IRC | 21:45 | |
| dansmith | rlrossit: ....okay, and you're just telling me? :) | 21:45 |
| rlrossit | dansmith: I was thinking instead of doing '%s%s' % (obj_name, version) you could just do the name of the class (because most of the time they will be the same as obj_name) | 21:46 |
| rlrossit | but my normal methods of getting the class name return not what I want | 21:46 |
| ccarmack | alaski: ok, in the meantime I can look for item to unblacklist (aggregates?) | 21:46 |
| ccarmack | aggregates are like cells of cells? | 21:46 |
| rlrossit | dansmith: my apprentice python wizardry is pathetic | 21:47 |
| alaski | mtreinish: it has to be done outside of the api. the original flavors get in via nova-manage at https://github.com/openstack-dev/devstack/blob/master/lib/nova#L675 | 21:47 |
| dansmith | rlrossit: well, that's not really what we want though I think | 21:47 |
| alaski | mtreinish: so we would need a way to load them in Nova that's not a hack | 21:47 |
| dansmith | rlrossit: because I could create a subclass of Instance that stores things in a different way, and I'd want to still be held to the same versioning hash | 21:47 |
| dansmith | rlrossit: to make sure I didn't change it | 21:47 |
| dansmith | rlrossit: but foo.__class__.__name__ should get you what you want | 21:48 |
| *** jwcroppe_ has joined #openstack-nova | 21:48 | |
| alaski | ccarmack: aggregates don't work yet afaik. there's been some work to get them working but I don't think it all merged | 21:48 |
| rlrossit | dansmith: but in the end it would have to be registered as a different class wouldn't it? So it would need its own fingerprint? | 21:48 |
| mtreinish | alaski: ah ok | 21:48 |
| rlrossit | (different meaning an addition to the obj_classes() list) | 21:48 |
| dansmith | rlrossit: no, I could have an Instance and and MongoDBInstance, both should have the same hash, right? | 21:49 |
| *** mriedem has quit IRC | 21:49 | |
| *** matrohon has quit IRC | 21:49 | |
| dansmith | rlrossit: specifically, the objname is the symbolic name for a given schema, so we want to check and store that | 21:50 |
| rlrossit | dansmith: yes, but with the way you do the hash stuff right now, it would make an Instance2 and store its hash again right? | 21:50 |
| *** jwcroppe has quit IRC | 21:50 | |
| rlrossit | oh wait no you would overwrite the existing instance in the registry | 21:50 |
| *** otter768 has joined #openstack-nova | 21:50 | |
| rlrossit | right? | 21:50 |
| dansmith | I'm sure the existing test would not do the right thing | 21:50 |
| dansmith | my point is, using the class name as the hash key is not right | 21:51 |
| dansmith | we should probably collect all the things that call themselves foo, and make sure their hashes are all equal to the stored hash | 21:51 |
| dansmith | to support that case | 21:51 |
| dansmith | however, Instance and Instance1 are different schemas | 21:51 |
| rlrossit | but then InstanceV1 and InstanceV2 will be different | 21:51 |
| dansmith | but Instance and MongoInstance are not | 21:51 |
| *** unicell has quit IRC | 21:52 | |
| rlrossit | dansmith: yeah basically my class name proposal will work the same as what you have in the patch right now (I think) and, in my opinion, makes updating hashes a lot easier, because I don't know if Instance is v1 or v2, and likewise I don't know which on Instance1 is in the tree | 21:53 |
| rlrossit | I don't know how to handle your different objects, same hash problem | 21:53 |
| openstackgerrit | Bartosz Górski proposed openstack/python-novaclient: Adds missing internationalization for help message https://review.openstack.org/219768 | 21:54 |
| dansmith | okay, but I specifically don't want to store the hashes by class name, I want them by object name | 21:54 |
| openstackgerrit | dane-fichter proposed openstack/nova: Nova support for image-signing https://review.openstack.org/189843 | 21:54 |
| *** haomaiwa_ has joined #openstack-nova | 21:54 | |
| dansmith | because that's the symbolic name of the schema we're trying to ensure never changes (without a version bump) | 21:54 |
| rlrossit | dansmith: but... class name and object name are (almost) always the same thing aren't they? | 21:54 |
| *** ccarmack has left #openstack-nova | 21:55 | |
| dansmith | rlrossit: they are, but there are no class names on the RPC wire, they're the symbolic name | 21:55 |
| *** haomaiw__ has quit IRC | 21:56 | |
| dansmith | rlrossit: so I could change Service to ServiceFoo and make obj_name() return Service and that would be fine | 21:56 |
| dansmith | nothing would break | 21:56 |
| dansmith | nor would the hash change | 21:56 |
| dansmith | and then I could have a ServiceBar that also implements the symbolic Service schema | 21:56 |
| dansmith | those two Service objects, for a given version, have to behave the same (i.e. have the same schema and thus the same hash) | 21:57 |
| *** annegentle has joined #openstack-nova | 21:57 | |
| dansmith | if we stored class names in the table, | 21:57 |
| rlrossit | dansmith: so, I get that | 21:57 |
| dansmith | and we had two Service implementations, | 21:57 |
| rlrossit | https://review.openstack.org/#/c/217818/5/nova/tests/unit/objects/test_objects.py,cm this does not do that though.... right? | 21:57 |
| dansmith | then people would let one service object become different from the other, but that's not okay | 21:57 |
| rlrossit | ServiceFoo would be Service and ServiceBar would be Service1 in the table | 21:57 |
| dansmith | rlrossit: correct, that addresses the other axis of supporting two major versions of the schema at once | 21:58 |
| dansmith | rlrossit: no | 21:58 |
| dansmith | I see what's confusing you | 21:58 |
| rlrossit | oh wait those implement the same version | 21:58 |
| rlrossit | so Bar would overwrite Foo in the registry | 21:59 |
| dansmith | right | 21:59 |
| dansmith | so right now the registry doesn't have room to support two implementations of the same object at the same version | 21:59 |
| rlrossit | so I agree with you there | 21:59 |
| rlrossit | then I guess I don't get why you disagree my class name thought for things that have the same base, but are implementing different schemas | 22:00 |
| dansmith | I don't understand what you just said.. same base but different schemas? | 22:00 |
| rlrossit | InstanceV1 and InstanceV2... they both are pretending to be Instance in the registry | 22:01 |
| rlrossit | they cause 2 entries in the registry | 22:01 |
| rlrossit | so they need 2 hashes | 22:01 |
| dansmith | okay, but no, they're not the same schem | 22:01 |
| dansmith | because they are different major versions | 22:01 |
| dansmith | there's a difference between Instance V2 which has different fields from Instance V1, | 22:02 |
| dansmith | that's a different axis of difference from SqlInstance and MongoInstance | 22:02 |
| rlrossit | yeah | 22:02 |
| dansmith | the former two are different major versions of the same thing, so the schemas are different | 22:02 |
| dansmith | the latter are two identical schemas with different implementations | 22:02 |
| rlrossit | and my suggestion hits on the table in the case of the former, not the latter | 22:02 |
| *** annegentle has quit IRC | 22:03 | |
| dansmith | rlrossit: but if we do that, then it also affects the latter, right? | 22:03 |
| dansmith | where we're recording the hash of the implementation and not the hash of the schema | 22:04 |
| rlrossit | dansmith: shouldn't | 22:04 |
| rlrossit | if we have 2 entries in the registry, log both hashes, with the names being the name of the implementing class. If we only have one, just use the name of the obj_name | 22:04 |
| dansmith | I understand that the mechanics are easier if you do that | 22:05 |
| dansmith | I think it's mixing concepts and wrong on principle :) | 22:05 |
| rlrossit | dansmith: it's just a different name.... it's not that bad... it's easier to understand than <object><n> naming scheme in my opinion | 22:06 |
| rlrossit | pwwwweeeeeease | 22:07 |
| dansmith | but it's wrong! | 22:07 |
| dansmith | sorry, but I feel extremely strongly about this | 22:07 |
| *** mdrabe has quit IRC | 22:07 | |
| rlrossit | I didn't before, but now I do too! | 22:08 |
| rlrossit | dansmith: let's shelve this until tomorrow, I need to head out | 22:09 |
| rlrossit | I'll think about it some more | 22:09 |
| dansmith | alright | 22:09 |
| *** emagana has quit IRC | 22:10 | |
| tonyb | Phooey I was going to talk to mriedem about the libvirt bug :( | 22:11 |
| *** rlrossit has left #openstack-nova | 22:12 | |
| *** sachin_ has quit IRC | 22:13 | |
| *** penick has quit IRC | 22:14 | |
| *** thorst has joined #openstack-nova | 22:16 | |
| *** sachin_ has joined #openstack-nova | 22:16 | |
| *** Piet has joined #openstack-nova | 22:17 | |
| *** jwcroppe has joined #openstack-nova | 22:20 | |
| *** ctrath has quit IRC | 22:21 | |
| *** jwcroppe_ has quit IRC | 22:22 | |
| *** jwcroppe_ has joined #openstack-nova | 22:24 | |
| *** jwcroppe has quit IRC | 22:26 | |
| *** VW_ has quit IRC | 22:26 | |
| *** jwcroppe has joined #openstack-nova | 22:27 | |
| *** alex_klimov has quit IRC | 22:28 | |
| *** jwcroppe_ has quit IRC | 22:29 | |
| *** penick has joined #openstack-nova | 22:31 | |
| *** greghaynes has joined #openstack-nova | 22:32 | |
| *** Marga_ has quit IRC | 22:35 | |
| *** sachin_ has quit IRC | 22:37 | |
| *** Marga_ has joined #openstack-nova | 22:37 | |
| *** sachin_ has joined #openstack-nova | 22:38 | |
| *** sachin__ has joined #openstack-nova | 22:39 | |
| *** sachin_ has quit IRC | 22:39 | |
| *** thumpba has quit IRC | 22:40 | |
| *** Marga_ has quit IRC | 22:43 | |
| *** Marga_ has joined #openstack-nova | 22:44 | |
| *** penick has quit IRC | 22:45 | |
| *** pixelbeat has joined #openstack-nova | 22:46 | |
| openstackgerrit | Michael Still proposed openstack/nova-specs: Mark liberty specs as implemented https://review.openstack.org/220338 | 22:47 |
| *** penick has joined #openstack-nova | 22:47 | |
| *** tonytan4ever has quit IRC | 22:49 | |
| *** edmondsw has quit IRC | 22:53 | |
| *** mriedem has joined #openstack-nova | 22:58 | |
| *** jecarey has quit IRC | 22:59 | |
| *** jeblair has quit IRC | 22:59 | |
| ZZelle_ | mriedem, hi | 22:59 |
| *** jeblair has joined #openstack-nova | 23:00 | |
| *** jeblair has quit IRC | 23:01 | |
| *** jeblair has joined #openstack-nova | 23:02 | |
| *** gyee has quit IRC | 23:04 | |
| *** zzzeek has quit IRC | 23:10 | |
| *** Marga_ has quit IRC | 23:12 | |
| *** markvoelker has joined #openstack-nova | 23:13 | |
| *** markvoelker has quit IRC | 23:17 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova: nova-net: fix missing log variable in deallocate_fixed_ip https://review.openstack.org/220253 | 23:18 |
| *** sachin__ has quit IRC | 23:20 | |
| *** marzif has quit IRC | 23:21 | |
| *** Sukhdev_ has joined #openstack-nova | 23:23 | |
| *** dannywilson has quit IRC | 23:23 | |
| alex_xu | sdague: mriedem we should alow space at begining and end of name, right? as the v2 api support that. | 23:24 |
| *** markvoelker has joined #openstack-nova | 23:25 | |
| mriedem | alex_xu: https://review.openstack.org/#/c/220279/ | 23:26 |
| mriedem | needs a unit test | 23:26 |
| mriedem | if you want to do that | 23:26 |
| mriedem | ZZelle_: if you have a question please post it | 23:26 |
| alex_xu | mriedem: yea, I notice that | 23:26 |
| alex_xu | mriedem: just working on that | 23:26 |
| alex_xu | mriedem: but that still not allow space at beginning and end | 23:26 |
| mriedem | i think sdague is done for the day | 23:26 |
| alex_xu | mriedem: ok, let me update that first, then see people what say | 23:27 |
| ZZelle_ | mriedem, any chance to get https://review.openstack.org/186716 merged in liberty? | 23:27 |
| mriedem | ZZelle_: that doesn't look like a bug fix | 23:27 |
| ZZelle_ | mriedem, yes but it's related to a change release in icehouse glance | 23:28 |
| *** tjones has quit IRC | 23:28 | |
| *** penick has quit IRC | 23:29 | |
| *** annegentle has joined #openstack-nova | 23:29 | |
| tonyb | mriedem: ping? | 23:29 |
| *** nic has quit IRC | 23:30 | |
| mriedem | ZZelle_: i' mnot sure that matters | 23:31 |
| mriedem | we're in feature freeze | 23:31 |
| *** bradjones has quit IRC | 23:31 | |
| mriedem | changing nova to use a feature in glance, regardless of when it was added, does not seem like a bug fix | 23:31 |
| mriedem | especially at this point in the release | 23:31 |
| *** achanda has quit IRC | 23:31 | |
| mriedem | tonyb: word up | 23:32 |
| *** sachin_ has joined #openstack-nova | 23:32 | |
| tonyb | mriedem: I think I may have found your libvirt bug | 23:32 |
| mriedem | omfg | 23:32 |
| *** kse has joined #openstack-nova | 23:32 | |
| *** y_sawai has joined #openstack-nova | 23:32 | |
| tonyb | mriedem: do you knwo anything about the python <-> C API? | 23:33 |
| tonyb | mriedem: http://paste.openstack.org/show/444807/ | 23:34 |
| jroll | mtreinish: heh, look what I found https://bugs.launchpad.net/ironic/+bug/1334033 | 23:35 |
| openstack | Launchpad bug 1334033 in OpenStack Core Infrastructure "Some Tempest tests require booting multiple instances at once, but we only deploy 1 Ironic node" [Medium,Triaged] | 23:35 |
| tonyb | So my hypothesis is that "PyArg_ParseTuple" is sometimes confused by the fact that flags is a kwarg not positional | 23:35 |
| tonyb | mriedem: That *might* explain why the to_xml() function isn't called. | 23:36 |
| * tonyb grabs the CPython code | 23:37 | |
| mriedem | tonyb: well to_xml() is getting called | 23:38 |
| mriedem | conf.to_xml() | 23:38 |
| mriedem | i know that b/c the generated xml is logged in the n-cpu logs | 23:38 |
| tonyb | mriedem: but not in the failing case right? | 23:39 |
| *** oomichi_away is now known as oomichi | 23:39 | |
| mriedem | it is in the failing case | 23:39 |
| tonyb | Oh okay. | 23:39 |
| mriedem | so i think n-cpu is definitely calling domain.attachDeviceFlags | 23:39 |
| mriedem | or whatever it's called | 23:39 |
| tonyb | mriedem: Yeah. | 23:39 |
| mriedem | the problem is between that call, there is no virDomainAttachDevice showing for that guest in libvirtd.logs | 23:39 |
| mriedem | but there is for other domains where it works | 23:40 |
| mriedem | maybe i should check syslog to see if something just completely dies in there | 23:40 |
| mriedem | i didn't check those yet | 23:40 |
| tonyb | mriedem: Right and PyArg_ParseTuple is in the call chain and is an easy target for why we wouldn't actuyally hity libvirt | 23:40 |
| *** electrocucaracha has quit IRC | 23:40 | |
| mriedem | so attachDeviceFlags is supposed to return an int | 23:40 |
| mriedem | i wonder what it's returning | 23:40 |
| *** y_sawai has quit IRC | 23:41 | |
| *** ZZelle_ has quit IRC | 23:42 | |
| mriedem | tonyb: let me check before that guest/domain refactor change to see how flags were passed | 23:42 |
| * tonyb guesses None | 23:42 | |
| mriedem | it should be 0 | 23:42 |
| mriedem | it's really really odd though that this would only ever fail in this one test case that i'm trying to re-enable | 23:43 |
| mriedem | and it fails every time | 23:43 |
| mriedem | there shouldn't really be any difference between the test i'm enabling and normal tests that attach a volume in tempest by the time they get to this code | 23:43 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/nova: Enable options for oslo.reports https://review.openstack.org/216917 | 23:43 |
| tonyb | Are all the failures whet flags!=0 | 23:43 |
| mriedem | the only difference in this test is we don't specify the device to attach to | 23:43 |
| *** dims has quit IRC | 23:44 | |
| mriedem | this was the change https://review.openstack.org/#/c/188763/ | 23:44 |
| *** ijw_ has quit IRC | 23:45 | |
| mriedem | tonyb: hmm, you're right about kwargs, it wasn't passing flags as kwargs before | 23:45 |
| mriedem | if that was the issue it'd be totally bonkers | 23:45 |
| mtreinish | jroll: heh, leave it to adam_g to preempt what we're seeing | 23:45 |
| mriedem | but i can push a change to nova to test that out and make the tempest change depend on it | 23:45 |
| jroll | mtreinish: ikr... also the WONTFIX from tempest is helpful | 23:46 |
| adam_g | wah? | 23:46 |
| jroll | adam_g: https://bugs.launchpad.net/ironic/+bug/1334033 | 23:46 |
| openstack | Launchpad bug 1334033 in OpenStack Core Infrastructure "Some Tempest tests require booting multiple instances at once, but we only deploy 1 Ironic node" [Medium,Triaged] | 23:46 |
| mriedem | sdague: johnthetubaguy: novaclient 2.28 release request https://review.openstack.org/#/c/220349/ | 23:47 |
| mriedem | dhellmann: ^ | 23:47 |
| mtreinish | jroll: heh, but talk about kinda a useless cloud | 23:47 |
| mtreinish | if it only lets you deploy one guest at a time :) | 23:47 |
| jroll | mtreinish: well, a cloud with 8gb of usable RAM in the entire thing will be pretty useless anyway :) | 23:48 |
| jroll | mtreinish: if infra is cool with us using 16-32gb nodes for this job, we're all good | 23:48 |
| *** changbl has joined #openstack-nova | 23:48 | |
| *** shoutm has joined #openstack-nova | 23:50 | |
| mtreinish | jroll: heh, you just need smaller ramdisks :) | 23:50 |
| * tonyb really shoudl have finished the "use non-dirsto libvirt/qemu in devstack" plugin | 23:50 | |
| mtreinish | heh, now that I think about it the flavors we use for the regular dsvm jobs actually have about as much ram as my first computer | 23:51 |
| jroll | mtreinish: let me know when you can fit python into less than 512mb :P | 23:51 |
| jroll | (coreos uses half the ram for the disk, half for usable memory) | 23:51 |
| gmann | mriedem: please look once you get time. VIF backporting to Kilo - https://review.openstack.org/#/c/198944/ | 23:51 |
| jroll | oh, python and an operating system because that's also useful | 23:51 |
| tonyb | mriedem: I'll patch my libvirt locally and see what I see. | 23:52 |
| *** dims has joined #openstack-nova | 23:53 | |
| *** derekh has joined #openstack-nova | 23:53 | |
| mtreinish | jroll: well I guess cirros doesn't count, we boot that with 64mb of ram for the gate testing | 23:54 |
| mtreinish | I think that's as small as we could get it | 23:55 |
| adam_g | jroll, might it be possible to start using multi-node devstack for this? with a second node that is just an empty host running libvrt? | 23:55 |
| jroll | adam_g: that's a reasonable idea | 23:55 |
| mtreinish | and the python install on my system is 449M | 23:55 |
| jroll | and probably fairly straightforward | 23:55 |
| jroll | mtreinish: :D | 23:55 |
| mtreinish | adam_g: yeah we could try that | 23:55 |
| *** tpeoples has quit IRC | 23:56 | |
| adam_g | actually i was palying with that a while back when jogo started workign on the mulit-node devstack for nova | 23:56 |
| *** VW_ has joined #openstack-nova | 23:56 | |
| mtreinish | adam_g: there might be some stability problems | 23:56 |
| mtreinish | its still not quite good enough to be voting | 23:56 |
| adam_g | i had to get creative with GRE tunneling to make it work, which required some MTU configuration in places we were unable to push to | 23:56 |
| jroll | would be good to investigate at least | 23:57 |
| adam_g | https://bugs.launchpad.net/ironic/+bug/1401298 | 23:57 |
| openstack | Launchpad bug 1401298 in Ironic "extra dhcp options not exposed to pxe deploy ramdisk" [Wishlist,In progress] - Assigned to Adam Gandelman (gandelman-a) | 23:57 |
| jroll | heh, so many old bugs | 23:57 |
| * adam_g removes assignee | 23:57 | |
| adam_g | :) | 23:57 |
| mtreinish | slacker | 23:57 |
| *** aix has joined #openstack-nova | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!