| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-05-03 | |||
| 12:54:55 | asettle | And I admit, I'm also very confused. | |
| 12:55:44 | efried | It's working its way through requirements now. The bot proposed u-c bump, which merged; and I proposed g-r bump, which merged. Waiting for the bot to propose nova requirements.txt bump, but I think the nova change set will now verify, cause it uses u-c, right? | |
| 12:55:55 | lbragstad | efried mordred https://pypi.python.org/pypi/keystoneauth1/2.20.0 | |
| 12:58:37 | johnthetubaguy | asettle: good question, sounds like we need a better worked example. We have more docs in here: https://docs.openstack.org/developer/nova/filter_scheduler.html | |
| 13:00:29 | efried | mriedem Do you want me to include links to merged and/or proposed change sets in the use-service-catalog-for-endpoints edit? https://review.openstack.org/#/c/461481/ | |
| 13:00:43 | alex_xu | nova api meeting is running at #openstack-meeting-4 | |
| 13:01:26 | mriedem | efried: nope | |
| 13:01:53 | asettle | johnthetubaguy: yeah I was reading that this morning. I couldn't, admittedly, find too many good examples to transfer across. The only thing I could find for AggregateImagePropertiesIsolation is the literal explanation about halfway down the page. | |
| 13:04:44 | mriedem | dansmith: can you hit this simple grenade change for using wsgi on master https://review.openstack.org/#/c/460543/ ? | |
| 13:05:01 | mriedem | i.e. the devstack change to use uwsgi on master depends on that https://review.openstack.org/#/c/457715/ | |
| 13:06:33 | mriedem | EmilienM: we've said in the past that when we make changes to grenade that we should give the deployment teams a heads up, do you think this change warrants that https://review.openstack.org/#/c/460543/ ? | |
| 13:06:46 | mriedem | that's needed for using uwsgi for nova in devstack on master https://review.openstack.org/#/c/457715/ | |
| 13:08:40 | openstackgerrit | Eric Fried proposed openstack/nova master: nova.utils.get_service_url() https://review.openstack.org/458257 | |
| 13:13:31 | bauzas | sched.org, you <3 | |
| 13:24:49 | dansmith | mriedem: yep got it | |
| 13:26:15 | sfinucan | sdague: One more review here for you, if you would be so kind https://review.openstack.org/460097 | |
| 13:37:25 | openstackgerrit | John Garbutt proposed openstack/nova master: Add policy description for fping https://review.openstack.org/461763 | |
| 13:43:12 | dansmith | sdague: so your native call logging hack suggestion... | |
| 13:43:38 | dansmith | sdague: I dunno if you saw, but we're actually missing logs, and the two requests don't line up. we're not just seeing a big delay | |
| 13:43:43 | dansmith | do you still think it's worth trying that? | |
| 13:46:11 | openstackgerrit | melanie witt proposed openstack/nova master: WIP Use six.text_type() when logging Instance object https://review.openstack.org/461912 | |
| 13:53:44 | openstackgerrit | Dan Smith proposed openstack/nova master: WIP: Debug race in delete-while-booting test https://review.openstack.org/461513 | |
| 13:54:36 | sdague | dansmith: sure | |
| 13:54:47 | dansmith | okay | |
| 13:54:50 | sdague | dansmith: you want me to push it out there? | |
| 13:55:10 | sdague | honestly, the only thing it loses us is the search by request-id, which would be nice | |
| 13:55:12 | dansmith | sdague: I'll just slam it into my patch | |
| 13:55:16 | sdague | but not hugely a thing | |
| 13:55:25 | dansmith | I'm going to put something else to log in the delete path | |
| 13:55:32 | sdague | any idea if there was an exception in this path? | |
| 13:55:33 | dansmith | wait, what? | |
| 13:55:59 | sdague | wait what what? | |
| 13:57:12 | dansmith | commenting the use_journal thing means we don't log request_id? | |
| 13:57:14 | dansmith | or you mean it doesn't do the rich field thing? | |
| 13:57:35 | openstackgerrit | Dan Smith proposed openstack/nova master: WIP: Debug race in delete-while-booting test https://review.openstack.org/461513 | |
| 13:58:00 | openstackgerrit | melanie witt proposed openstack/nova master: WIP Use six.text_type() when logging Instance object https://review.openstack.org/461912 | |
| 13:59:20 | sdague | dansmith: it doesn't do the rich field thing | |
| 13:59:35 | dansmith | okay gotcha | |
| 13:59:52 | sdague | so with that on, the log path is through journal.send, which gives us all the extra metadata | |
| 14:00:00 | sdague | with that off, it's just going to spew to stdout | |
| 14:00:12 | dansmith | cool, just wanted to make sure it wasn't going to change more stuff | |
| 14:00:25 | sdague | systemd will send it all to the journal but without any of the fun bits | |
| 14:00:47 | sdague | dansmith: https://review.openstack.org/462163 | |
| 14:01:05 | sdague | I also tweaked the debug line, because processid is going to be in the systemd logs anyway | |
| 14:01:30 | mriedem | melwitt: heh, display_name='tempest-Ãñstáñcé-2027474072' is better than UnicodeDecodeError at least | |
| 14:01:36 | dansmith | sdague: oh i just commented it out of my patch | |
| 14:02:03 | sdague | mriedem: mriedem how did we get to that? | |
| 14:02:16 | mriedem | sdague: https://review.openstack.org/#/c/461912/ | |
| 14:02:40 | sdague | dansmith: sure, honestly, for right now I think it's fine to turn it off if we have concerns, so I'll put the general patch out there as well | |
| 14:02:40 | melwitt | mriedem: true. it's correct when I download the .txt.gz file though and open it with the less command. not sure what's up with that considering it let me copy-paste the name into the review comment from the log file I had downloaded | |
| 14:03:44 | sdague | melwitt: that really feels like something that should be happening at the logging levels below | |
| 14:04:03 | mriedem | sdague: you mean in the oslo.log handlers? | |
| 14:04:07 | jaypipes | sdague, dansmith, mriedem, bauzas: https://hangouts.google.com/call/b6rgkm2kzbh57pdgv6vey2kgxau | |
| 14:04:13 | sdague | mriedem: yeh | |
| 14:04:15 | melwitt | sdague: maybe, if we wanna convert everything to unicode before logging. which we probably do I guess | |
| 14:04:17 | mriedem | sdague: i think melwitt is just trying to figure out how to fix it | |
| 14:04:25 | sdague | mriedem: sure | |
| 14:05:29 | sdague | mriedem: you going to jump on hangout about onboarding room? | |
| 14:05:45 | mriedem | yes | |
| 14:05:57 | openstackgerrit | Gábor Antal proposed openstack/nova master: Add json style checking for sample notifications https://review.openstack.org/443677 | |
| 14:06:45 | dansmith | sdague: okay | |
| 14:06:57 | openstackgerrit | John Garbutt proposed openstack/nova master: Add policy description for instance actions https://review.openstack.org/461757 | |
| 14:09:12 | dansmith | jaypipes: I came to mock, staying to mock | |
| 14:09:55 | abhishek | melwitt: Hi Melanie. I have added the changes which you had suggested. Please review https://review.openstack.org/#/c/460305/ | |
| 14:10:03 | openstackgerrit | John Garbutt proposed openstack/nova master: Add policy description for instance actions https://review.openstack.org/461757 | |
| 14:10:35 | jaypipes | dansmith: :) | |
| 14:10:35 | melwitt | abhishek: ack, will take a look | |
| 14:10:55 | abhishek | melwitt: thanks | |
| 14:10:56 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Use plain routes list for os-attach-interfaces endpoint instead of stevedore https://review.openstack.org/462166 | |
| 14:19:25 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Use plain routes list for os-server-groups endpoint instead of stevedore https://review.openstack.org/462169 | |
| 14:21:22 | mriedem | dansmith: https://etherpad.openstack.org/p/BOS-forum-nova-project-onboarding | |
| 14:22:56 | peter-hamilton | hi all, the bulk of the cert validation support in cursive is done: https://review.openstack.org/#/c/357202/ | |
| 14:23:23 | peter-hamilton | i'd like to get additional nova folks eyes on it before we go further (thanks again for the review jaypipes) | |
| 14:29:42 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: use plain routes list for os-virtual-interfaces https://review.openstack.org/462173 | |
| 14:34:05 | openstackgerrit | John Garbutt proposed openstack/nova master: Add policy description for os-quota-classes https://review.openstack.org/461996 | |
| 14:39:50 | efried | johnthetubaguy jaypipes mordred mriedem edmondsw FYI, the necessary ksa and requirements pieces for use-service-catalog-for-endpoints are in place, so the nova change https://review.openstack.org/#/c/458257/ is passing Jenkins and (I believe) ready for a solid review hammering. | |
| 14:39:59 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Use plain routes list for os-consoles and os-console-auth-tokens endpoint instead of stevedore https://review.openstack.org/462175 | |
| 14:40:27 | jaypipes | efried: will try to review this afternoon. | |
| 14:40:34 | efried | jaypipes Thanks! | |
| 14:40:40 | edmondsw | efried roger that | |
| 14:43:54 | openstackgerrit | John Garbutt proposed openstack/nova master: Add description to policies in extended_az and extend_ser_attrs https://review.openstack.org/461600 | |
| 14:46:48 | edmondsw | mriedem edleafe is there spec for the capabilities work that was discussed at the PTG? | |
| 14:46:53 | edmondsw | All I'm finding is https://blueprints.launchpad.net/nova/+spec/expose-host-capabilities which is languishing and doesn't talk about what was discussed at PTG | |
| 14:47:15 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Use plain routes list for block_device_mapping instead of stevedore https://review.openstack.org/462180 | |
| 14:48:43 | jaypipes | edmondsw: are you referring to the API capabilities or are you referring to the resource provider traits stuff? | |
| 14:48:56 | edmondsw | jaypipes API | |
| 14:49:26 | edleafe | edmondsw: The api-wg spec for capabilities was abandoned: https://review.openstack.org/#/c/386555/ | |
| 14:50:47 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Use plain routes list for user_data instead of stevedore https://review.openstack.org/462183 | |
| 14:51:01 | edmondsw | edleafe the abandon comment talks about "first fixing the immediate problems | |
| 14:51:06 | edmondsw | where is that happening? | |
| 14:51:40 | openstackgerrit | Gábor Antal proposed openstack/nova master: Transform aggregate.add_host notification https://review.openstack.org/396210 | |
| 14:52:19 | edleafe | edmondsw: as far as I know no one has picked that up | |
| 14:52:41 | edmondsw | k | |
| 14:53:01 | edleafe | edmondsw: so if you've got some spare cycles... :) | |
| 14:53:11 | edmondsw | edleafe wish I did... | |
| 14:53:24 | edleafe | heh, welcome to the club! | |
| 14:56:43 | melwitt | sdague: when I was debugging locally, I found this is where it blows up in the log formatting, it's trying to convert the message to unicode, but with the Instance object I found it only works if the encoding 'utf-8' is specified to six.text_type(), and that requires the instance arg be a string, not an Instance object type | |
| 14:56:58 | melwitt | https://github.com/openstack/oslo.log/blob/3ee14a1/oslo_log/formatters.py#L83-L93 | |
| 14:58:14 | mriedem | sdague: jaypipes: dansmith: bauzas: i assume none of us have a touchscreen laptop to project if we don't have a flip chart or whiteboard right? | |
| 14:58:34 | mriedem | that would be easier to see than a flip chart if we're expecting 40 people | |