| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2022-06-17 | |||
| 08:47:29 | bauzas | gibi: do you think we could workaround https://bugs.launchpad.net/nova/+bug/1874664 with some flavor extraspec ? | |
| 08:48:08 | bauzas | it's a placement issue | |
| 08:48:56 | bauzas | if you look at the issue, we try to use the same RP https://paste.opendev.org/show/b9A9TdUl1dzTJaNtbrCd/ | |
| 08:49:53 | bauzas | for both allocations | |
| 09:01:20 | gibi | bauzas: I will check in a minute | |
| 09:05:39 | gibi | bauzas: have you checked that multicreate without vgpu can boot two instance on the same host? (I'm wondering if we do that) | |
| 09:08:38 | bauzas | gibi: good question, I dunno | |
| 09:09:15 | bauzas | gibi: so, my fix I created was using this bug number but I'll then create a separate other | |
| 09:20:36 | gibi | bauzas: I see the limitation | |
| 09:20:39 | gibi | bauzas: it is not placement | |
| 09:20:47 | gibi | bauzas: in that test placement returns 2 allocation candidate | |
| 09:20:55 | gibi | one for each pGPU | |
| 09:21:12 | gibi | but nova only able to consume the first allocation candidate per host | |
| 09:21:55 | gibi | bauzas: https://github.com/openstack/nova/blob/93a65f06df67ce39d65827692150c78013c7f6d5/nova/scheduler/manager.py#L382-L388 | |
| 09:22:12 | gibi | nova only tries to first a_c, for both instance | |
| 09:22:23 | gibi | *tries to claim the first | |
| 09:22:38 | gibi | it succeeds for the first instance but obviously fails for the second | |
| 09:22:48 | gibi | the TODO from jaypipes is valid | |
| 09:23:24 | gibi | if you resolve that TODO then the test_multiple_instance_create_filling_up_capacity test case will be able to boot both instances on the same host | |
| 09:24:24 | bauzas | gibi: looking | |
| 09:33:09 | bauzas | gibi: thanks for finding the RCA | |
| 09:33:18 | gibi | I think we can implement that TODO | |
| 09:33:34 | gibi | without adding to much complexity with sorting | |
| 09:33:51 | gibi | it would not be optimal but would be better than what we have today | |
| 09:39:14 | gibi | I'm affraid the gate is broken on master | |
| 09:39:16 | gibi | https://zuul.opendev.org/t/openstack/builds?job_name=tempest-integrated-compute-centos-9-stream&skip=0 | |
| 09:40:18 | gibi | libvirt.libvirtError: internal error: unable to execute QEMU command 'netdev_add': File descriptor named '(null)' has not been found | |
| 09:49:07 | bauzas | gibi: what a Friday | |
| 09:49:23 | bauzas | Centos 9 Stream /o\ | |
| 09:49:27 | gibi | bauzas: reported a bug https://bugs.launchpad.net/nova/+bug/1979047 please triage it and set it critical I will drop a mail to the ML | |
| 09:50:49 | bauzas | gibi: we need to make this job non-voting | |
| 09:51:14 | gibi | are we give up on this job? :) | |
| 09:51:49 | bauzas | gibi: the priority is to unblock the gate | |
| 09:52:12 | bauzas | want me to write it ? | |
| 09:53:48 | gibi | I sent a mail to the ML | |
| 09:54:06 | bauzas | I'll propose the patch at least | |
| 09:54:31 | bauzas | we're on a Friday | |
| 09:55:01 | gibi | bauzas: my philosophy: sure unblock the gate in 2-3 days if nobody found a quick solution. but keep the pain a bit around to motivate people to try and find a solution first :) | |
| 09:55:42 | bauzas | gibi: my philosophy: remove support from something which breaks and you'll suddently find people who care | |
| 09:56:31 | bauzas | either way, patch is being forged, we need two cores for approving it, so there is no direct impact *yet* | |
| 09:56:37 | gibi | OK, lets go with yours this time :) | |
| 09:56:50 | gibi | I will +2 the non-voting patch | |
| 10:03:12 | opendevreview | Sylvain Bauza proposed openstack/nova master: zuul: Temporarly put Centos9 Stream job non-voting https://review.opendev.org/c/openstack/nova/+/846292 | |
| 10:03:30 | bauzas | gibi: feel free to add a reply to your email with this patch number | |
| 10:04:57 | opendevreview | Sylvain Bauza proposed openstack/nova master: zuul: Temporarly put Centos9 Stream job non-voting https://review.opendev.org/c/openstack/nova/+/846292 | |
| 10:07:14 | gibi | bauzas: thanks +2. done the ML update | |
| 10:17:53 | gibi | bauzas: the grenade job on master also look baaaad https://zuul.opendev.org/t/openstack/builds?job_name=nova-grenade-multinode&project=openstack%2Fnova&branch=master&skip=0 | |
| 10:18:31 | gibi | that seem to be a different error | |
| 10:18:32 | gibi | Details: {'type': 'SecurityGroupInUse', 'message': 'Security Group 40f249d7-9c4e-432c-91d7-d68ef7ac05b0 in use.', 'detail': ''} | |
| 10:18:57 | bauzas | yup, saw | |
| 10:19:00 | bauzas | damn | |
| 10:23:12 | sean-k-mooney | i can take a look at it too i guess | |
| 10:23:49 | opendevreview | Merged openstack/placement master: Add WA about resource_providers.can_host removal https://review.opendev.org/c/openstack/placement/+/845730 | |
| 10:24:31 | sean-k-mooney | so i guessl libvirt need to be udated for a new qemu release? | |
| 10:24:40 | sean-k-mooney | to fix interface attach | |
| 10:24:54 | sean-k-mooney | or they broke there api compat and we need to adress that | |
| 10:25:33 | gibi | sean-k-mooney: I added the libvirt and qemu versions to the bugreport I don't know if any of those are fresh | |
| 10:26:27 | tkajinam | gibi, the bug title reminds me this https://bugzilla.redhat.com/show_bug.cgi?id=2092856 | |
| 10:27:11 | gibi | tkajinam: good find. That very much like it | |
| 10:27:17 | tkajinam | we pinned libvirt to workaround this in tripleo | |
| 10:28:14 | sean-k-mooney | ya i was about to ping danpb and ask about that bug | |
| 10:29:23 | gibi | bauzas: I'm going to open a bug on tempest about the security group issue | |
| 10:29:49 | bauzas | gibi: cool | |
| 10:33:51 | gibi | https://bugs.launchpad.net/tempest/+bug/1979052 | |
| 10:35:52 | bauzas | I need to go off | |
| 10:36:01 | bauzas | (fortunately, wasn't doing gym today) | |
| 10:36:14 | bauzas | but I have my kid with me so I need to cook :) | |
| 10:36:41 | gibi | let me know what did you cook when you are back | |
| 10:36:46 | gibi | I need inspiration | |
| 10:45:07 | sean-k-mooney | im considerign firing up my new(ish) BBQ this evening and cookig somethign or at least doing that at the weekend | |
| 10:45:47 | gibi | sounds nice :) | |
| 10:47:43 | sean-k-mooney | since i invested in a "kamado joe big joe 3" i might as well make use of it. i had left over chicken form tha tlast time in the freezer which i had last night. it was pretty tasty | |
| 10:48:55 | sean-k-mooney | i did a prime rib roast on it before and it was amazing, also did really nice potatos corn and fish | |
| 10:50:26 | sean-k-mooney | its been 50:50 for bread sofar. the first loaf was nice the second i didnt have the dome heat soaked enouch so it went way to high then too low so the bread did not cook to the center but im pretty happy all in all | |
| 10:51:19 | gibi | I need ribs :) | |
| 12:35:35 | bauzas | sean-k-mooney: hah, I'm a BBQ lover | |
| 12:36:14 | bauzas | kamados <3 | |
| 12:37:22 | bauzas | I have a Weber Genesis 2 (gas) at home with a plancha, so usually, I cook both vegetables and potatoes while the meat is grilling | |
| 12:37:47 | bauzas | in general, sausages as it's French thing | |
| 12:39:29 | bauzas | https://www.bestfranceforever.com/you-cant-have-a-french-barbecue-without-these-10-ingredients/ | |
| 12:39:39 | bauzas | (sorry for the URL name... :) | |
| 12:47:33 | sean-k-mooney | yep | |
| 12:47:43 | sean-k-mooney | there is a lot of variation | |
| 13:27:38 | opendevreview | Sylvain Bauza proposed openstack/nova master: Add a functest for verifying multiple VGPU allocations https://review.opendev.org/c/openstack/nova/+/845747 | |
| 13:27:39 | opendevreview | Sylvain Bauza proposed openstack/nova master: Support multiple allocations for vGPUs https://review.opendev.org/c/openstack/nova/+/845757 | |
| 13:50:34 | Uggla | bauzas, looking at your BBQ damned it is connected ! | |
| 13:50:58 | bauzas | well, yes and no | |
| 13:51:11 | bauzas | my BBQ *can* be connected | |
| 13:52:04 | bauzas | and indeed I use a grill temperature sensor that's connected for my meat cooking | |
| 13:52:08 | bauzas | but not a weber | |
| 13:52:34 | Uggla | bauzas, meat cooking pro ! | |
| 13:57:48 | bauzas | Uggla: unfortunately, my middle name isn't Rafa :) | |
| 13:58:08 | bauzas | only French people looking at YT can understand :) | |
| 14:01:35 | Uggla | bauzas, I don't have the ref... | |
| 14:01:58 | bauzas | Uggla: https://www.youtube.com/c/LeBarbecueDeRafa | |
| 14:02:55 | sean-k-mooney | i went with a ceramic grill basically because i wanted the charcol taste but didnt want to have to worry about tending a fire or unstable temperatures | |
| 14:03:36 | bauzas | sean-k-mooney: I have charcoil tasting with my Weber flavorizers tho | |
| 14:03:38 | sean-k-mooney | that and it can be used for low an slow smokes or you can use it as a pizza oven | |
| 14:04:13 | bauzas | sean-k-mooney: and eventually after 15 years of grilling, I gave up playing with charcoil | |
| 14:42:32 | opendevreview | Artom Lifshitz proposed openstack/nova stable/victoria: fake: Ensure need_legacy_block_device_info returns False https://review.opendev.org/c/openstack/nova/+/843947 | |