| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-03-26 | |||
| 10:14:05 | lpetrut | kashyap: checking | |
| 10:18:05 | lpetrut | looks like nova-compute does not start on the Hyper-V node, it's an unrelated issue but thanks for checking. we'll look into it | |
| 10:18:43 | lpetrut | also, I took a look at the patch and it was only changing libvirt driver code, which won't be used on Windows | |
| 10:21:51 | kashyap | Exactly :-) | |
| 10:21:55 | kashyap | Thanks for looking, lpetrut | |
| 10:23:16 | kashyap | stephenfin: When you get another minute, mind taking a look: https://kashyapc.fedorapeople.org/libvirt-cpu-model-extra-flags-a23085f58bd22d27.yaml.txt | |
| 10:23:23 | kashyap | (Before I upload the change.) | |
| 10:23:54 | stephenfin | kashyap: Looks correct, aye | |
| 10:24:25 | kashyap | stephenfin: Even the indentation of two 'dashed' bullet points under item 3.? | |
| 10:24:42 | stephenfin | Yup, that's what I'd expect to see | |
| 10:24:52 | kashyap | Most excellent; thank you! | |
| 10:26:46 | openstackgerrit | Kashyap Chamarthy proposed openstack/nova master: libvirt: Allow to specify granular CPU feature flags https://review.openstack.org/534384 | |
| 10:40:07 | openstackgerrit | Kashyap Chamarthy proposed openstack/nova master: libvirt: Allow to specify granular CPU feature flags https://review.openstack.org/534384 | |
| 10:58:49 | openstackgerrit | sahid proposed openstack/nova-specs master: update: isolate guests emulthreads on CONF.cpu_shared_set https://review.openstack.org/511188 | |
| 10:59:38 | sahid | stephenfin, cfriesen I updated the emultor threads spec, if you have a moment to take a look ^ | |
| 11:00:03 | sahid | much simple | |
| 11:00:26 | alex_xu_ | sean-k-mooney: efried, found one use-case for the granular reqs in images https://review.openstack.org/#/c/554305/2/specs/rocky/approved/glance-image-traits.rst@86 | |
| 11:49:54 | sean-k-mooney | alex_xu_: im sure there are several | |
| 11:51:37 | sean-k-mooney | alex_xu_: im not sure that is a usecase at least not in the normal sense | |
| 11:52:17 | sean-k-mooney | alex_xu_: we had dicussed haveing specific metadata keys for the FPGA usecase not useing the generic resouces and traits workflow | |
| 12:01:17 | gibi | efried: I left some clarification about the numbered request group usage in https://review.openstack.org/#/c/502306/20/specs/rocky/approved/bandwidth-resource-provider.rst@490 | |
| 12:02:23 | gibi | efried: I want to make sure we are on the same page before I clarify the situation in the spec directly | |
| 12:11:30 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: Add --enable and --disable options to nova-manage update_cell https://review.openstack.org/555416 | |
| 12:12:20 | openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: remove not necessary short cut https://review.openstack.org/553122 | |
| 12:12:21 | openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: Fix allocation_candidates not to ignore shared RPs https://review.openstack.org/533396 | |
| 12:12:21 | openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: [WIP] Consider nested RPs in get_all_with_shared https://review.openstack.org/556450 | |
| 12:31:17 | efried | alex_xu_, gibi: looking... | |
| 12:31:29 | efried | ...and good UGT morning! | |
| 12:32:00 | gibi | efried: good morning to you too | |
| 12:34:25 | efried | alex_xu_: You around? I need help understanding the use case. | |
| 12:34:37 | efried | edleafe: Is the sched meeting in 25 or 1:25? | |
| 12:34:53 | cdent | efried: latter | |
| 12:35:14 | cdent | unless my calendar is borked | |
| 12:35:22 | efried | damn, I thought I had my reminders set up to take SFDST into account. | |
| 12:35:49 | cdent | yeah, latter | |
| 12:35:50 | efried | ...last two weeks were | |
| 12:35:52 | efried | weird | |
| 12:36:22 | cdent | yeah, really bad planning by the illumanit, rotschilds and other members of world government | |
| 12:36:49 | efried | oh, I think it's because I've got it set to GMT - because my calendar doesn't do UTC - and GMT switches two weeks after US. So it looked like it was working... but only for a couple of weeks. | |
| 12:36:51 | efried | eff | |
| 12:37:07 | cdent | use iceland as your timezone | |
| 12:37:29 | cdent | I _think_ that may work | |
| 12:38:11 | bauzas | efried: scheduler meeting is planned in 1h 20 mins-ish | |
| 12:38:43 | bauzas | because CET and GMT moved to DST yesterday, not UTC | |
| 12:39:07 | bauzas | and yeah, Iceland TZ == UTC | |
| 12:39:13 | efried | thanks y'all. cdent - Rekjavik worked -- nice. | |
| 12:39:16 | bauzas | because they don't do dayshift | |
| 12:39:20 | efried | Awesome. | |
| 12:39:29 | bauzas | efried: using gcal ? | |
| 12:39:34 | efried | Notes :( | |
| 12:39:48 | openstackgerrit | Merged openstack/osc-placement master: Migrate legacy-osc-placement-dsvm-functional job in-tree https://review.openstack.org/547812 | |
| 12:39:49 | openstackgerrit | Merged openstack/osc-placement master: Add osc-placement-dsvm-functional-py3 job https://review.openstack.org/547815 | |
| 12:39:50 | bauzas | because TB/Lightning has a UTC TZ | |
| 12:39:58 | bauzas | hah | |
| 12:40:44 | bauzas | FWIW, for the next 6 months, I will only be possible to be in the scheduler meeting for the first 20 mins :( | |
| 12:40:55 | bauzas | efried: cdent: edleafe: ^ | |
| 12:41:08 | efried | ack | |
| 12:41:37 | openstackgerrit | Roman Dobosz proposed openstack/nova master: allow compute nodes to be associated with host agg https://review.openstack.org/526753 | |
| 12:42:35 | openstackgerrit | Yikun Jiang (Kero) proposed openstack/nova master: Add host to API and Conductor https://review.openstack.org/556513 | |
| 12:43:15 | openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: [WIP] shared and nested allocation candidates https://review.openstack.org/556514 | |
| 12:47:08 | cdent | gerrit seems to be dragging a bit? | |
| 12:47:33 | kashyap | It's always dragging a bit | |
| 12:47:52 | kashyap | FFS, releasenotes still fails for reasons I can't see: http://logs.openstack.org/84/534384/14/check/build-openstack-releasenotes/4507437/job-output.txt.gz#_2018-03-26_10_58_14_063277 | |
| 12:48:01 | kashyap | Can anyone tell? (The original change: https://review.openstack.org/#/c/534384/) | |
| 12:48:29 | kashyap | (In unreleased.rst; it seems to be moaning about a "Literal block expected", thus treating a warning as an error.) | |
| 12:49:12 | efried | gibi: Responded. Let me know if you want to talk through the alternatives. | |
| 12:53:14 | kashyap | bauzas: You seem to have added the unreleased.rst file. I don't see anyone using it. Can it be deleted? | |
| 12:53:37 | kashyap | (That is the Nova commit: 4c2dfa5) | |
| 12:53:38 | bauzas | please, no. | |
| 12:53:44 | bauzas | it's used for good reasons | |
| 12:53:56 | bauzas | https://docs.openstack.org/releasenotes/nova/unreleased.html | |
| 12:54:06 | stephenfin | kashyap: The reno extension isn't reporting your issues correctly | |
| 12:54:09 | kashyap | bauzas: Do you spot what is the failure here: http://logs.openstack.org/84/534384/14/check/build-openstack-releasenotes/4507437/job-output.txt.gz#_2018-03-26_10_58_14_063277 | |
| 12:54:13 | stephenfin | *reno sphinx extension | |
| 12:54:26 | kashyap | stephenfin: Hmm, how can I put myself out of my misery, then? | |
| 12:54:33 | bauzas | I guess some relnote had a problem | |
| 12:54:41 | bauzas | hence the sphinx build error | |
| 12:54:47 | bauzas | but the rst page itself is empty | |
| 12:55:04 | bauzas | that's only reno which generates that file using the sphinx ext | |
| 12:55:25 | stephenfin | kashyap: This is what I did for oslo.config. The reno extension works the same way so the fix should be similar https://review.openstack.org/#/c/554632/ | |
| 12:55:53 | gibi | efried: thanks, looking... | |
| 12:56:44 | stephenfin | kashyap: tl;dr: you need to keep an offset counter as you emit rST lines, and it's worth writing what you generate to temporary file so you can view it | |
| 12:57:01 | kashyap | Yeah, was reading this change: https://review.openstack.org/#/c/554632/1/oslo_config/sphinxext.py | |
| 12:57:04 | stephenfin | kashyap: I'm pretty sure there's a better way to do it (maybe dumping the offending lines to the terminal) but I haven't figure it out | |
| 12:57:11 | kashyap | Thanks for the summary. | |
| 12:57:53 | gibi | efried: thanks for the namespacing alternative. Will it mean that Neutron can say in the port that 'resources-neutron-1': {'BW': 10000} ? | |
| 12:58:11 | efried | gibi: Yeah, something like that. | |
| 12:58:25 | kashyap | stephenfin: Right now I don't have time to fix 'reno', I wonder what I can do to figure out what needs to be fixed in this:https://review.openstack.org/#/c/534384/14/releasenotes/notes/libvirt-cpu-model-extra-flags-a23085f58bd22d27.yaml | |
| 12:58:53 | efried | gibi: Since we haven't connected the final dots on granular even for intra-nova cases, we probably still have the flexibility to flip to using resources-nova-N there. | |
| 12:59:15 | edleafe | efried: Looks like you got your answer. I'm just dragging myself to start working at 8am. | |
| 12:59:20 | gibi | efried: so to keep ports separate it should look more like resources-neutron-<port-uuid> | |
| 12:59:23 | efried | or resources-flavor-N / resources-image-N | |
| 12:59:43 | gibi | or resources-port-<port-uuid> | |
| 13:00:01 | stephenfin | kashyap: It's ugly, but I'd strip it down and start adding parts until it stops working :/ | |
| 13:01:00 | kashyap | stephenfin: Hmm, is there a way locally verify that? Because, 'reno lint .` isn't catching what I need | |
| 13:01:34 | kashyap | (Lest I'll be stuck with this yak) | |
| 13:01:42 | stephenfin | 'tox -e releasenotes' should do the trick | |
| 13:01:47 | kashyap | stephenfin: But did I do anything "fancy" there? I'll see what I can "trip down" | |
| 13:01:50 | kashyap | stephenfin: I did run that | |
| 13:01:56 | kashyap | The _only_ thing that gave me was: | |