| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-11-26 | |||
| 15:11:01 | sean-k-mooney | the cheks api is not stable technically in gerrit yet as far as i know | |
| 15:22:21 | stephenfin | it's a plugin so I doubt it ever will be | |
| 15:29:09 | tacco | sean-k-mooney: i foundout something new about my more than 64CPUs issue | |
| 15:31:06 | tacco | i realized with ubuntu i did not had any issues. All 248CPUs up and running with a ubuntu cloud image. strange that fedora and debian buster had the same limitations in here. in debian it 100%is CUSTOM_NR_CPUS_DEFAULT kernel-parameter, could be overrided by grub with nr_cpus=248 but this won't work even if it should work. thats verry strange et all. :D | |
| 15:33:16 | sean-k-mooney | the virt kernels proably have it set smaller or somthign as an optimaisation | |
| 15:36:56 | tacco | yes could be, try to verrify what happens if i build the kernel by myself without that limitations. *crossingFingers* | |
| 15:52:22 | stephenfin | elod: Thank you :) | |
| 15:53:48 | elod | stephenfin: np :) | |
| 15:54:39 | stephenfin | sean-k-mooney: is this a race? If this call fails, we don't roll back the requests on the pools that passed https://github.com/openstack/nova/blob/master/nova/pci/stats.py#L419-L421 | |
| 15:54:41 | elod | I didn't see your recheck, so I've double-stamped it :P | |
| 15:54:43 | stephenfin | a bug | |
| 15:54:50 | stephenfin | elod: all good! | |
| 15:55:10 | stephenfin | I checked. It was that eventlet issue. Nothing to be worried about, thankfully | |
| 15:55:10 | elod | :] | |
| 15:57:38 | gibi | nova meeting starts in 2 minutes | |
| 16:03:40 | sean-k-mooney | gibi: for got to add teh flavor metadata thing to the adgenda but we can cover it at the end | |
| 16:04:19 | sean-k-mooney | stephenfin: we operate on a copy in the filter | |
| 16:04:32 | gibi | sean-k-mooney: sure | |
| 16:04:44 | gibi | sean-k-mooney: I will ping you at the end | |
| 16:04:44 | sean-k-mooney | ill take a look after the nova meeting | |
| 16:09:37 | sean-k-mooney | stephenfin: that is only used in the schduler | |
| 16:10:02 | sean-k-mooney | at least i think that is the case checking | |
| 16:10:29 | sean-k-mooney | stephenfin: https://codesearch.opendev.org/?q=apply_requests&i=nope&files=&excludeFiles=&repos= | |
| 16:10:37 | sean-k-mooney | stephenfin: so its only operating on the copy | |
| 16:10:51 | sean-k-mooney | https://github.com/openstack/nova/blob/master/nova/pci/stats.py#L399 | |
| 16:11:05 | sean-k-mooney | actully not that copy. | |
| 16:12:00 | sean-k-mooney | its called form https://github.com/openstack/nova/blob/5dbfbe692f7de33069746e4dc07bcbf0dd33f656/nova/scheduler/host_manager.py#L279 | |
| 16:14:53 | sean-k-mooney | stephenfin: the ecpetion will propagate all the way up to select destiatnions in the conductor and keep koning | |
| 16:15:14 | stephenfin | okay, so effectively we'll never hit it | |
| 16:15:35 | sean-k-mooney | basically we never save the object to the db in that path | |
| 16:16:28 | sean-k-mooney | for spawn it is caught here https://github.com/openstack/nova/blob/5dbfbe692f7de33069746e4dc07bcbf0dd33f656/nova/conductor/manager.py#L676 | |
| 16:18:10 | sean-k-mooney | look like there is similar logic for unshelve ectra in the other 3 places its called | |
| 16:40:11 | openstackgerrit | Merged openstack/nova stable/train: docs: Rework the PCI passthrough guides https://review.opendev.org/c/openstack/nova/+/764327 | |
| 16:40:26 | openstackgerrit | Merged openstack/nova stable/train: docs: Change order of PCI configuration steps https://review.opendev.org/c/openstack/nova/+/764328 | |
| 16:41:06 | tacco | yes could be, try to verrify what happens if i build the kernel by myself without that limitations. *crossingFingers* | |
| 17:03:31 | openstackgerrit | Lee Yarwood proposed openstack/nova master: nova-live-migration: Disable *all* virt services during negative tests https://review.opendev.org/c/openstack/nova/+/762623 | |
| 18:13:04 | openstackgerrit | Stephen Finucane proposed openstack/nova master: pci: Add a '_filter_pools' helper function https://review.opendev.org/c/openstack/nova/+/764375 | |
| 18:13:04 | openstackgerrit | Stephen Finucane proposed openstack/nova master: pci: Improve docs for '_filter_*' functions https://review.opendev.org/c/openstack/nova/+/764376 | |
| 18:13:05 | openstackgerrit | Stephen Finucane proposed openstack/nova master: pci: Add logging for filtering https://review.opendev.org/c/openstack/nova/+/764377 | |
| 18:13:18 | stephenfin | sean-k-mooney: Cast an eye over that an tell me if you think it's reasonable | |
| 18:13:39 | stephenfin | gibi: You might have thoughts on it | |
| 18:13:42 | stephenfin | ...too | |
| 18:17:11 | stephenfin | lyarwood: If you're around later/tomorrow, could you look at https://review.opendev.org/c/openstack/nova/+/751367/2 again? I don't think I need to anything more on it since elod's comment was just that, rather than a request for change | |
| 18:18:50 | sean-k-mooney | stephenfin: related bug i can proably live with but sure ill take a look | |
| 18:20:58 | sean-k-mooney | stephenfin: ok so you are not printing all the pools that are filtered out | |
| 18:21:14 | sean-k-mooney | i think that is a resonable approch | |
| 18:21:30 | sean-k-mooney | ill review it properly in a bit but skimming it looks fine | |
| 18:42:12 | sean-k-mooney | stephenfin: done. let me know what you think of the generator idea but over all i think it looks fine | |
| #openstack-nova - 2020-11-27 | |||
| 04:35:58 | gmann | lyarwood: seems like tempest test cleanup change did not fix the nova-ceph-multistore job - https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_548/763487/3/check/nova-ceph-multistore/5480be7/testr_results.html | |
| 04:53:21 | gmann | lyarwood: got the issue, its class level vs test level cleanup. fixing in https://review.opendev.org/c/openstack/tempest/+/764407 | |
| 04:55:30 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: DNM: testing fix for nova-ceph-multistore https://review.opendev.org/c/openstack/nova/+/764408 | |
| 04:55:51 | gmann | testing here ^^ | |
| 07:06:23 | gmann | lyarwood: its late here, going to bed. I pinged other tempest core to merger the tempest change once it pass the job in nova testing patch | |
| 07:56:49 | openstackgerrit | Wenping Song proposed openstack/nova master: Nova supports password encrypted VNC https://review.opendev.org/c/openstack/nova/+/622336 | |
| 08:09:10 | bauzas | good morning Nova | |
| 08:36:32 | openstackgerrit | Wenping Song proposed openstack/nova master: Cyborg shelve/unshelve support https://review.opendev.org/c/openstack/nova/+/729563 | |
| 08:55:28 | lyarwood | gmann: ack thanks, I'm not sure if that's going to work tbh, it looks like my patch was working the DELETEs were just racing | |
| 08:59:00 | lyarwood | gmann: on my bad, the DELETE I was looking at was for the volume attachment | |
| 09:10:33 | openstackgerrit | Wenping Song proposed openstack/nova master: Cyborg shelve/unshelve support https://review.opendev.org/c/openstack/nova/+/729563 | |
| 09:12:14 | lyarwood | gmann: okay that's also failing still | |
| 09:13:02 | openstackgerrit | Lee Yarwood proposed openstack/nova master: zuul: Skip test_attach_scsi_disk_with_config_drive in nova-ceph-multistore https://review.opendev.org/c/openstack/nova/+/764424 | |
| 09:13:35 | lyarwood | gibi / stephenfin / bauzas ; ^ lets skip this test in the ceph job for now until we get is passing again | |
| 09:20:55 | lyarwood | that regex was written prior to coffe on a Friday morning after a long week and several night shifts with my child so lets wait to see if it skips the test before +W'ing ;) | |
| 09:23:54 | bauzas | lyarwood: I don't see a problem with the regex itself, and the test name is long so I'm pretty sure we don't remove another test too | |
| 09:24:39 | bauzas | https://github.com/openstack/tempest/search?q=test_attach_scsi_disk_with_config_drive proves me I was right | |
| 09:25:31 | lyarwood | cool thansk | |
| 09:25:33 | lyarwood | thanks* | |
| 09:34:02 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/victoria: zuul: Introduce nova-evacuate https://review.opendev.org/c/openstack/nova/+/764430 | |
| 09:34:03 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/victoria: zuul: Merge nova-evacuate into nova-multinode-live-migration https://review.opendev.org/c/openstack/nova/+/764431 | |
| 09:34:04 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/victoria: zuul: Replace nova-live-migration with zuulv3 jobs https://review.opendev.org/c/openstack/nova/+/764432 | |
| 09:34:19 | lyarwood | elod: ^ okay here's my attempt to unfsck the stable/victoria evacuation stuff, one of the squashed changes hasn't merged on master yet but I just wanted to start testing | |
| 09:37:52 | elod | lyarwood: wow, sounds good! thanks for the info (and for working on the issue :)), I'll look at them | |
| 09:38:05 | elod | \o/ | |
| 09:38:48 | lyarwood | elod: np and apologies it took so long | |
| 09:39:02 | Yumeng | hi bauzas gibi and sean-k-mooney: about the cyborg-owned vgpu support spec, I've discussed all your concerns with the team and updated the spec https://review.opendev.org/c/openstack/nova-specs/+/750116/ please see my update at your convenience. | |
| 09:59:25 | lyarwood | stephenfin: random one, any idea why we trigger a full gate run with docs changes? | |
| 10:00:18 | lyarwood | https://review.opendev.org/c/openstack/nova/+/764329 - I guess it isn't full but still that's odd. | |
| 10:02:24 | lyarwood | oh fail | |
| 10:02:30 | lyarwood | you changed nova/conf/pci.py | |
| 10:02:36 | lyarwood | that triggers things, ignore me | |
| 10:07:57 | stephenfin | yeah | |
| 10:08:03 | stephenfin | it's the conf change that triggers things | |
| 10:08:13 | stephenfin | I considered changing the regex to skip changes to that but that's silly | |
| 10:10:00 | lyarwood | yeah I'm not sure it's worth it | |
| 10:14:57 | stephenfin | lyarwood: I've held off on +W of that, as requested. Feel free to stick it on if it works as expected | |
| 10:24:10 | openstackgerrit | Alexandre arents proposed openstack/nova master: libvirt: Abort live-migration job when monitoring fails https://review.opendev.org/c/openstack/nova/+/764435 | |
| 10:24:14 | lyarwood | stephenfin: ack will do thanks | |
| 10:33:29 | stephenfin | holy s***, I always forget just how rough nova.pci is /o\ | |
| 10:33:39 | stephenfin | some serious TLC needed there | |
| 11:35:27 | openstackgerrit | Stephen Finucane proposed openstack/nova master: pci: Improve docs for '_filter_*' functions https://review.opendev.org/c/openstack/nova/+/764376 | |
| 11:35:27 | openstackgerrit | Stephen Finucane proposed openstack/nova master: pci: Add logging for filtering https://review.opendev.org/c/openstack/nova/+/764377 | |
| 11:35:28 | openstackgerrit | Stephen Finucane proposed openstack/nova master: pci: Improve testing of 'nova.pci.request' module https://review.opendev.org/c/openstack/nova/+/764446 | |
| 11:40:02 | sean-k-mooney | Yumeng: thanks im onloy around for a few hours today but ill try to review it before i leave | |
| 11:41:10 | sean-k-mooney | stephenfin: it is rough but it works. we very very rarely have any bugs or issue related to nova.pci we the bugs related to sriov/pci are almost always in the compute manager | |
| 11:41:58 | Yumeng | sean-k-mooney: That's great! Thanks sean. Just take your time. ^^ | |
| 12:08:28 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/victoria: zuul: Replace nova-live-migration with zuulv3 jobs https://review.opendev.org/c/openstack/nova/+/764432 | |
| 12:23:04 | openstackgerrit | Stephen Finucane proposed openstack/nova stable/train: Validate id as integer for os-aggregates https://review.opendev.org/c/openstack/nova/+/764455 | |
| 12:26:47 | openstackgerrit | Stephen Finucane proposed openstack/nova stable/stein: Validate id as integer for os-aggregates https://review.opendev.org/c/openstack/nova/+/764308 | |
| 12:28:06 | stephenfin | hmm, cherry-picking via the web UI is curiously slow now :-\ | |
| 12:28:39 | openstackgerrit | Stephen Finucane proposed openstack/nova stable/rocky: Validate id as integer for os-aggregates https://review.opendev.org/c/openstack/nova/+/764309 | |