| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-03-22 | |||
| 13:59:20 | openstackgerrit | Lee Yarwood proposed openstack/nova master: virt: Provide block_device_info during rescue https://review.opendev.org/700811 | |
| 13:59:21 | openstackgerrit | Lee Yarwood proposed openstack/nova master: api: Introduce microverion 2.83 allowing boot from volume rescue https://review.opendev.org/701430 | |
| 13:59:21 | openstackgerrit | Lee Yarwood proposed openstack/nova master: compute: Report COMPUTE_RESCUE_BFV and check during rescue https://review.opendev.org/701429 | |
| 13:59:22 | openstackgerrit | Lee Yarwood proposed openstack/nova master: libvirt: Support boot from volume stable device instance rescue https://review.opendev.org/701431 | |
| 13:59:22 | openstackgerrit | Lee Yarwood proposed openstack/nova master: compute: Extract _get_bdm_image_metadata into nova.utils https://review.opendev.org/705212 | |
| 13:59:23 | openstackgerrit | Lee Yarwood proposed openstack/nova master: DNM - Test stable device rescue tests with BFV instances https://review.opendev.org/710050 | |
| 14:59:35 | openstackgerrit | Marcin Juszkiewicz proposed openstack/nova master: libvirt: always undefine NVRAM https://review.opendev.org/714311 | |
| 15:22:33 | donnyd | I know there is a facility in nova compute to balance where workloads are placed (affinity), but I don't really know how to use it outside of server groups. Is there a way to set the default to anti-affinity so the workload is balanced across all the hosts I have | |
| 15:26:56 | priteau | donnyd: The Nova scheduler default behaviour is to spread instances across your available hypervisors. See https://docs.openstack.org/nova/latest/admin/configuration/schedulers.html#id23 | |
| 15:27:22 | donnyd | interesting - because mine does not do that | |
| 15:27:54 | donnyd | it will schedule to a host that already has the glance image cached first - then spread the workload | |
| 15:29:30 | donnyd | I brought this up a while back and we tinkered with it a bit - but for the Zuul workload (and really pointed at this specific workload) it seems to use affinity instead of anti-affinity by default | |
| 15:29:58 | donnyd | the really interesting part is most of the other workloads don't do this ... only the CI ones | |
| 15:30:44 | donnyd | https://usercontent.irccloud-cdn.com/file/03J5oioU/image.png | |
| 16:27:08 | donnyd | even more interesting factoid - restarting nova-compute on all the nodes with no instances seems to have cured this behavior | |
| 18:56:15 | hrw | I feel that tomorrow I will have another day of fighting with nova ;D | |
| #openstack-nova - 2020-03-23 | |||
| 02:15:48 | openstackgerrit | mashuting proposed openstack/nova master: Revert "Deprecate the vmwareapi driver" https://review.opendev.org/713401 | |
| 06:30:41 | openstackgerrit | Luyao Zhong proposed openstack/nova master: support live migration with vpmems https://review.opendev.org/687856 | |
| 08:09:31 | openstackgerrit | Kevin Zhao proposed openstack/nova master: Add default cpu model for aarch64 https://review.opendev.org/709494 | |
| 08:37:02 | openstackgerrit | Kevin Zhao proposed openstack/nova master: Add default cpu model for aarch64 https://review.opendev.org/709494 | |
| 08:39:45 | lyarwood | https://review.opendev.org/#/c/714057/ - gibi / stephenfin / any cores, could someone +W this again now it's passing, I'm going to spend the day looking at gate failures btw | |
| 08:40:03 | gibi | lyarwood: ack, will look | |
| 08:40:50 | gibi | done | |
| 08:41:12 | gibi | thanks for looking into the gate issues | |
| 08:43:05 | bauzas | good morning | |
| 08:43:23 | bauzas | today's a bit hectic, I need to help my children more than latest week | |
| 08:44:28 | brinzhang_ | good morning | |
| 08:44:36 | brinzhang_ | bauzas: Good luck and pay attention to protection. | |
| 08:51:12 | brinzhang_ | gibi: if I remove the NOTE from the nova-runways's bp's link, does need to add history log? | |
| 08:56:55 | gibi | brinzhang: nope | |
| 08:58:02 | brinzhang | gibi: yeah. will remove, but nova-runways doesn't seem to work well enough. | |
| 08:58:17 | gibi | brinzhang: you are right. | |
| 09:11:57 | aarents | morning, lyarwood: I'm considering implementation of finish_unshelve as you suggested in https://review.opendev.org/#/c/696084 I have a question see my last post when you will have time for that | |
| 09:25:00 | brinzhang | gibi, lyarwood: https://review.opendev.org/#/c/682772/ this patch addressed all your comments | |
| 09:57:18 | openstackgerrit | Balazs Gibizer proposed openstack/nova stable/pike: Avoid circular reference during serialization https://review.opendev.org/714148 | |
| 09:58:30 | nightmare_unreal | hello, how can one verify cell-id passed in nova-manage CLI ? | |
| 09:58:59 | nightmare_unreal | basically how to validate passed cellID/UUID | |
| 10:07:44 | nightmare_unreal | found it | |
| 10:15:21 | openstackgerrit | Merged openstack/nova master: nova-live-migration: Only stop n-cpu and q-agt during evacuation testing https://review.opendev.org/714057 | |
| 10:22:02 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/train: nova-live-migration: Ensure subnode is fenced during evacuation testing https://review.opendev.org/713961 | |
| 10:23:00 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/stein: nova-live-migration: Ensure subnode is fenced during evacuation testing https://review.opendev.org/713962 | |
| 10:23:02 | lyarwood | aarents: ack sorry will try to look this afternoon | |
| 10:23:08 | lyarwood | brinzhang_: ack will also try to look this afternoon | |
| 10:23:44 | brinzhang_ | lyarwood: yeah, thanks | |
| 10:43:23 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: Add MigrationTypeField https://review.opendev.org/706013 | |
| 10:46:20 | kashyap | lyarwood: Just looking at the Q35 upstream job - https://review.opendev.org/#/c/708701/ | |
| 10:46:36 | kashyap | While I look at the logs; what was the issue you're alluding to there, that worked for you locally? | |
| 10:47:30 | bauzas | gibi: stephenfin: efried_gone told he was +2 on https://review.opendev.org/#/c/589085/7 if I was writing a specific email | |
| 10:47:37 | bauzas | are you okay with this ? | |
| 10:47:55 | bauzas | I haven't written it until I was providing https://review.opendev.org/#/c/712118/2 | |
| 10:47:58 | lyarwood | kashyap: yeah on F30, I can reproduce the failure on Ubuntu | |
| 10:48:06 | lyarwood | kashyap: I haven't had time to circle back to it yet | |
| 10:48:25 | lyarwood | kashyap: but it looks like the config drive is missing it's LABEL causing the test to fail | |
| 10:48:25 | stephenfin | bauzas: sure | |
| 10:49:15 | kashyap | I see; I'll see if I can poke more to see why Zuul complains | |
| 10:51:20 | lyarwood | kashyap: tbh I think it might be a bug with the version of q35 Ubuntu is using | |
| 10:51:39 | lyarwood | kashyap: well QEMU but you get the point | |
| 10:51:44 | kashyap | Hmm, didn't think of it; you mean version of QEMU? | |
| 10:51:49 | kashyap | Bingo | |
| 10:51:50 | lyarwood | yeah | |
| 10:52:13 | kashyap | lyarwood: BTW, be careful with the term "circle back"; Urban Dictionary has a harsh definition of it. Not gonna link it :D | |
| 10:53:25 | kashyap | I'll check the QEMU version of Ubuntu; sometimes they miss to backport some bug fixes. FWIW, we've pointed them out before | |
| 10:53:38 | kashyap | (I mean, in a different scenario) | |
| 11:08:10 | openstackgerrit | Marcin Juszkiewicz proposed openstack/nova master: libvirt: change UEFI check to handle AArch64 better https://review.opendev.org/714311 | |
| 11:09:21 | hrw | this patch should pass tests (does 'tox -e py38') and better shows issue | |
| 11:10:07 | kashyap | hrw: Hi | |
| 11:10:13 | hrw | hi kashyap | |
| 11:10:18 | hrw | others | |
| 11:10:52 | kashyap | hrw: Good point on the undefine with NVRAM thing | |
| 11:10:55 | hrw | now I have one of those nasty days when I look at nova code... | |
| 11:10:57 | kashyap | Will look post-lunch | |
| 11:11:03 | kashyap | :D | |
| 11:11:31 | hrw | speaking of q35... when it will be finally default? | |
| 11:12:44 | hrw | x86 users will get hit hard with it. pleasure to look would be | |
| 11:13:43 | bauzas | gibi: stephenfin: done http://lists.openstack.org/pipermail/openstack-discuss/2020-March/013507.html | |
| 11:14:15 | gibi | bauzas: thanks | |
| 11:26:01 | hrw | "kernel doesn't support AMD SEV" message is funny | |
| 11:26:18 | hrw | as it gets checked on all archs instead of just x86 | |
| 11:53:04 | hrw | I wonder how much it will be rewritten once it will be supported somewhere else than amd cpus | |
| 11:54:44 | stephenfin | gibi: did you ever do a blog post on the bandwidth-aware scheduling feature? | |
| 11:55:05 | stephenfin | I recall there being one floating around before a past PTG but heck if I can find it now | |
| 11:55:09 | gibi | stephenfin: rubasov did one in the early days. let me find the link | |
| 11:55:38 | gibi | stephenfin: is this the one you remember? https://rubasov.github.io/2018/09/21/openstack-qos-min-bw-demo.html | |
| 11:56:12 | sean-k-mooney | stephenfin: there was two summit presentations too | |
| 11:56:36 | stephenfin | gibi: That's the one, yup. Thanks! | |
| 11:56:51 | sean-k-mooney | one descirbiing the poc and then one describing the final feature | |
| 11:57:22 | gibi | sean-k-mooney: I only remember showing the PoC | |
| 11:58:04 | sean-k-mooney | yes showed the poc in vancovour and the final version in denver i think | |
| 11:58:24 | gibi | https://www.youtube.com/watch?v=ls9RXpG6Is0 | |
| 11:58:48 | gibi | there was some PTG presentation, true. The above is a summit presentation | |
| 11:59:02 | sean-k-mooney | sorry yes that is waht im thinking of | |
| 11:59:23 | gibi | I think we don't have any recording, other than the blog post, about the PTG presentation | |
| 12:01:34 | hrw | https://paste.centos.org/view/abe36617 - does it look ok to send for review? skipping AMD SEV check on !x86-64 | |
| 12:09:47 | hrw | fixed missing import, running tox, will send | |
| 12:11:44 | openstackgerrit | Balazs Gibizer proposed openstack/nova stable/pike: Avoid circular reference during serialization https://review.opendev.org/714148 | |
| 12:12:20 | openstackgerrit | Marcin Juszkiewicz proposed openstack/nova master: libvirt: check for AMD SEV only on x86-64 https://review.opendev.org/714425 | |
| 12:18:43 | kashyap | hrw: Upstream there's a CI job to test it: https://review.opendev.org/#/c/708701/ | |
| 12:19:32 | kashyap | hrw: "it" == q35; and aiming to switch to it in the 'V' release, or early, if possible. | |
| 12:20:51 | hrw | kashyap: cool ;) | |
| 12:21:41 | hrw | kashyap: I start to think how to add aarch64 ci job ;D | |