| brinzhang_ | mriedem: I am sorry later see, yeah, I missed the log under the below section, got it, thanks. | 00:00 |
|---|---|---|
| *** mdbooth has quit IRC | 00:01 | |
| *** mdbooth has joined #openstack-nova | 00:01 | |
| *** mdbooth_ has quit IRC | 00:02 | |
| brinzhang_ | efried: Hi, are you around? | 00:03 |
| *** betherly has joined #openstack-nova | 00:04 | |
| *** betherly has quit IRC | 00:09 | |
| *** threestrands has joined #openstack-nova | 00:10 | |
| *** artom has joined #openstack-nova | 00:25 | |
| *** tetsuro has joined #openstack-nova | 00:25 | |
| *** betherly has joined #openstack-nova | 00:25 | |
| *** gyee has quit IRC | 00:28 | |
| *** betherly has quit IRC | 00:30 | |
| openstackgerrit | Merged openstack/nova master: Privsep the ebtables modification code. https://review.opendev.org/635435 | 00:38 |
| *** betherly has joined #openstack-nova | 00:46 | |
| *** betherly has quit IRC | 00:51 | |
| *** zhouyao has joined #openstack-nova | 00:51 | |
| zhouyao | https://review.opendev.org/#/c/621646/ Any one can review this? | 00:52 |
| *** tetsuro has quit IRC | 00:56 | |
| *** ricolin_ has joined #openstack-nova | 00:58 | |
| *** bhagyashris has joined #openstack-nova | 01:00 | |
| *** brinzhang has joined #openstack-nova | 01:00 | |
| *** brinzhang_ has quit IRC | 01:02 | |
| *** panda has quit IRC | 01:06 | |
| *** panda has joined #openstack-nova | 01:08 | |
| *** igordc has quit IRC | 01:15 | |
| *** betherly has joined #openstack-nova | 01:18 | |
| *** bbowen has joined #openstack-nova | 01:18 | |
| *** betherly has quit IRC | 01:22 | |
| *** mdbooth_ has joined #openstack-nova | 01:25 | |
| *** mdbooth has quit IRC | 01:28 | |
| *** mdbooth_ has quit IRC | 01:30 | |
| *** tetsuro has joined #openstack-nova | 01:31 | |
| *** zhouyao has quit IRC | 01:36 | |
| *** zhouyao has joined #openstack-nova | 01:43 | |
| *** brinzhang_ has joined #openstack-nova | 01:50 | |
| *** brinzhang_ has quit IRC | 01:51 | |
| *** brinzhang_ has joined #openstack-nova | 01:52 | |
| *** brinzhang_ has quit IRC | 01:52 | |
| *** brinzhang_ has joined #openstack-nova | 01:53 | |
| *** brinzhang has quit IRC | 01:53 | |
| *** brinzhang_ has quit IRC | 01:54 | |
| *** brinzhang_ has joined #openstack-nova | 01:54 | |
| *** brinzhang_ has quit IRC | 01:56 | |
| *** brinzhang_ has joined #openstack-nova | 01:56 | |
| *** brinzhang has joined #openstack-nova | 01:57 | |
| *** brinzhang_ has quit IRC | 02:01 | |
| *** yikun has joined #openstack-nova | 02:04 | |
| *** yaawang has joined #openstack-nova | 02:05 | |
| gmann | efried: i thought it was merged. This is followup nit comments fix of approved spec - https://review.opendev.org/#/c/669196/1 | 02:07 |
| gmann | efried: do i need to follow the exception process for this or i can just restore it ? and and ping melwitt johnthetubaguy to review. | 02:08 |
| openstackgerrit | Merged openstack/nova master: Add functional regression test for bug 1837955 https://review.opendev.org/673357 | 02:12 |
| openstack | bug 1837955 in OpenStack Compute (nova) "MaxRetriesExceeded sometime fails with messaging exception" [Medium,In progress] https://launchpad.net/bugs/1837955 - Assigned to Erik Olof Gunnar Andersson (eandersson) | 02:12 |
| alex_xu | efried: this is an update of existing standrad cpu resource tracking https://review.opendev.org/#/c/666032/ | 02:18 |
| openstackgerrit | Merged openstack/nova master: Cleanup when hitting MaxRetriesExceeded from no host_available https://review.opendev.org/672855 | 02:19 |
| *** sridharg has joined #openstack-nova | 02:23 | |
| *** zhouyao has quit IRC | 02:28 | |
| *** sridharg has quit IRC | 02:50 | |
| *** hongbin has joined #openstack-nova | 02:53 | |
| *** epoojad1 has quit IRC | 02:54 | |
| *** tetsuro has quit IRC | 02:55 | |
| *** brinzhang_ has joined #openstack-nova | 03:01 | |
| *** brinzhang has quit IRC | 03:04 | |
| openstackgerrit | Brin Zhang proposed openstack/nova-specs master: Add flavor group https://review.opendev.org/663563 | 03:25 |
| *** mdbooth has joined #openstack-nova | 03:27 | |
| *** threestrands has quit IRC | 03:30 | |
| *** threestrands has joined #openstack-nova | 03:30 | |
| *** threestrands has quit IRC | 03:31 | |
| *** hongbin has quit IRC | 03:34 | |
| *** psachin has joined #openstack-nova | 03:36 | |
| *** brinzhang_ has quit IRC | 03:43 | |
| *** jaosorior has quit IRC | 04:05 | |
| *** jmlowe has joined #openstack-nova | 04:14 | |
| *** jmlowe has quit IRC | 04:19 | |
| *** udesale has joined #openstack-nova | 04:23 | |
| *** jmlowe has joined #openstack-nova | 04:24 | |
| *** jmlowe has quit IRC | 04:30 | |
| *** gary_perkins has quit IRC | 04:32 | |
| *** gary_perkins has joined #openstack-nova | 04:33 | |
| *** whoami-rajat has joined #openstack-nova | 04:40 | |
| *** ratailor has joined #openstack-nova | 04:47 | |
| *** belmoreira has joined #openstack-nova | 04:54 | |
| *** jaosorior has joined #openstack-nova | 05:04 | |
| *** Luzi has joined #openstack-nova | 05:05 | |
| *** ricolin__ has joined #openstack-nova | 05:21 | |
| openstackgerrit | ya.wang proposed openstack/os-traits master: Add new traits to flag live migration features https://review.opendev.org/673402 | 05:23 |
| *** ricolin_ has quit IRC | 05:24 | |
| *** d34dh0r53 has quit IRC | 05:25 | |
| *** huaqiang has joined #openstack-nova | 05:25 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Multiple API cleanup changes https://review.opendev.org/666889 | 05:38 |
| openstackgerrit | ya.wang proposed openstack/os-traits master: Add new traits to flag live migration features https://review.opendev.org/673402 | 05:48 |
| *** brinzhang has joined #openstack-nova | 05:58 | |
| openstackgerrit | Sundar Nadathur proposed openstack/nova master: ksa auth conf and client for cyborg access https://review.opendev.org/631242 | 05:59 |
| openstackgerrit | Sundar Nadathur proposed openstack/nova master: WIP: Add Cyborg device profile groups to request spec. https://review.opendev.org/631243 | 05:59 |
| openstackgerrit | Sundar Nadathur proposed openstack/nova master: WIP: Create and bind Cyborg ARQs. https://review.opendev.org/631244 | 05:59 |
| openstackgerrit | Sundar Nadathur proposed openstack/nova master: WIP: Get resolved Cyborg ARQs and add PCI BDFs to VM's domain XML. https://review.opendev.org/631245 | 05:59 |
| *** epoojad2 has joined #openstack-nova | 06:06 | |
| *** belmoreira has quit IRC | 06:13 | |
| *** brinzhang_ has joined #openstack-nova | 06:16 | |
| *** d34dh0r53 has joined #openstack-nova | 06:17 | |
| *** brinzhang has quit IRC | 06:19 | |
| openstackgerrit | ZhengMa proposed openstack/nova master: Check container_format in nova, simply abort instance creation directly from image, if container_format of the image is 'compressed'. https://review.opendev.org/673407 | 06:24 |
| openstackgerrit | ZhengMa proposed openstack/nova master: Check container_format in nova before instance creation simply abort instance creation directly from image, if container_format of the image is 'compressed'. https://review.opendev.org/673407 | 06:28 |
| *** maciejjozefczyk has joined #openstack-nova | 06:30 | |
| *** dpawlik has joined #openstack-nova | 06:33 | |
| *** dpawlik has quit IRC | 06:38 | |
| *** dpawlik has joined #openstack-nova | 06:43 | |
| *** slaweq has joined #openstack-nova | 06:53 | |
| *** brinzhang has joined #openstack-nova | 06:54 | |
| *** rpittau|afk is now known as rpittau | 06:58 | |
| *** jaosorior has quit IRC | 07:01 | |
| *** rcernin has quit IRC | 07:02 | |
| *** belmoreira has joined #openstack-nova | 07:08 | |
| *** aojea has joined #openstack-nova | 07:10 | |
| *** tesseract has joined #openstack-nova | 07:11 | |
| *** tssurya has joined #openstack-nova | 07:15 | |
| *** xek has joined #openstack-nova | 07:20 | |
| openstackgerrit | Brin Zhang proposed openstack/nova master: Add delete_on_termination to volume-attach API https://review.opendev.org/673133 | 07:31 |
| *** ricolin__ is now known as ricolin | 07:44 | |
| *** brault has joined #openstack-nova | 07:47 | |
| *** ivve has joined #openstack-nova | 07:49 | |
| openstackgerrit | Yongli He proposed openstack/nova master: Add server sub-resource topology API https://review.opendev.org/621476 | 07:49 |
| *** jangutter has joined #openstack-nova | 07:55 | |
| *** mkrai has joined #openstack-nova | 07:57 | |
| *** ralonsoh has joined #openstack-nova | 07:58 | |
| *** priteau has joined #openstack-nova | 08:08 | |
| *** cdent has joined #openstack-nova | 08:09 | |
| kashyap | aspiers: Morning. On your pastebin, it certainly looks better. | 08:15 |
| kashyap | aspiers: On the absence of 'shim' binaries, yeah, I'm a bit surprised | 08:15 |
| kashyap | aspiers: But you confirm that in `grub` (and `dmesg` - assuming you haven't ran over the ring buffer limit) shows Secure Boot is enabled, that's good enough :-) | 08:16 |
| *** lpetrut has joined #openstack-nova | 08:18 | |
| bhagyashris | stephenfin: Hi, I just encountered in one issue and I have posted comment on your patch https://review.opendev.org/#/c/671793/5 . | 08:22 |
| bhagyashris | stephenfin: Just to info you I am currently working on the upgrade patch https://review.opendev.org/#/c/672224/1 (Fixing review comments given by you) and also testing those change on top of your changes. Once it’s fixed I push the patche soon. | 08:22 |
| *** purplerbot has quit IRC | 08:22 | |
| *** purplerbot has joined #openstack-nova | 08:22 | |
| *** derekh has joined #openstack-nova | 08:27 | |
| *** ttsiouts has joined #openstack-nova | 08:32 | |
| *** jaosorior has joined #openstack-nova | 08:45 | |
| *** ivve has quit IRC | 08:48 | |
| *** johnthetubaguy has quit IRC | 08:51 | |
| *** johnthetubaguy has joined #openstack-nova | 08:53 | |
| *** belmoreira has quit IRC | 08:54 | |
| *** belmoreira has joined #openstack-nova | 08:55 | |
| *** belmoreira has quit IRC | 08:56 | |
| *** tkajinam has quit IRC | 09:03 | |
| openstackgerrit | Brin Zhang proposed openstack/nova master: Add delete_on_termination to volume-attach API https://review.opendev.org/673133 | 09:05 |
| mdbooth | sean-k-mooney: I haven't looked closely, but unlikely. The downstream bug requires a 24 hour delay, which you're unlikely to find in the gate. | 09:05 |
| *** johnthetubaguy has quit IRC | 09:09 | |
| *** johnthetubaguy has joined #openstack-nova | 09:11 | |
| aspiers | kashyap: yeah, once I figured out that I needed a) a newer kernel to avoid some panics and b) qxl video since virtio doesn't work well with SEV yet, I got the guest fully booted and it works beautifully :) | 09:12 |
| kashyap | aspiers: Sweet. You're probably the Second Man(tm) on this channel to have booted SB VM and came out unscathed :D | 09:13 |
| aspiers | kashyap: I didn't even know it was SB until I saw that on the vnc :) | 09:13 |
| kashyap | Hehe | 09:13 |
| aspiers | was only trying for SEV, but I guess q35 enables both | 09:14 |
| aspiers | ah well, q35 only needed for *truly* secure boot IIRC from your spec | 09:14 |
| aspiers | which with hindsight I realise https://review.opendev.org/#/c/673151/ probably accidentally helped quite a lot | 09:14 |
| aspiers | even though it's now abandoned X-D | 09:14 |
| aspiers | kashyap: BTW the whole SEV series is blocked on review of that patch ... hint hint no pressure ;-) | 09:15 |
| kashyap | aspiers: It's on the top of my pile. Just trying to shuffle things with highest prio before I disappear for PTO on 06-Aug (to 23-Aug) | 09:16 |
| aspiers | thanks :) | 09:16 |
| kashyap | aspiers: (And yes, 'q35' is mandatory for SB) | 09:17 |
| kashyap | Hehe, and first time I came across the 'bamboo' canonical machine type for PPC! | 09:17 |
| aspiers | pretty cool huh | 09:17 |
| aspiers | "g3beige" | 09:17 |
| *** ivve has joined #openstack-nova | 09:17 | |
| kashyap | aspiers: A quick question from line-709 here: https://review.opendev.org/#/c/673151/4/nova/virt/libvirt/host.py | 09:21 |
| kashyap | aspiers: SEV "typically requires" 'q35' or it _definitely_ requires 'q35'? | 09:22 |
| aspiers | that's explained in the spec | 09:22 |
| kashyap | Typically implies, you can do away without 'q35' as well | 09:22 |
| kashyap | Ah, I'll open that. | 09:22 |
| aspiers | q35 is required in order to bind all the virtio devices to the PCIe bridge so that they use virtio 1.0 and not virtio 0.9, since QEMU’s iommu_platform feature is added in virtio 1.0 only. | 09:22 |
| aspiers | The gurus went back and forth on this point, and seemed confused themselves | 09:23 |
| aspiers | so I didn't want to write "definitely" | 09:23 |
| aspiers | adding a hint for other reviewers | 09:23 |
| kashyap | (No need to add 'definitely', but to my mind, removing 'typically' will clarify: "in all cases that matter you want 'q35' with SEV") | 09:24 |
| kashyap | aspiers: Yeah, I see the virtio 1.0 vs. 0.9 point. | 09:25 |
| aspiers | hit reload | 09:25 |
| kashyap | And yes, for SEV you do want 1.0-only. | 09:25 |
| kashyap | Ah, thanks. | 09:25 |
| *** mkrai has quit IRC | 09:33 | |
| openstackgerrit | Lee Yarwood proposed openstack/nova master: WIP compute: Lock while rebooting an instance https://review.opendev.org/673463 | 09:42 |
| *** dpawlik has quit IRC | 09:45 | |
| *** dpawlik has joined #openstack-nova | 09:51 | |
| *** dansmith has quit IRC | 09:52 | |
| *** Nick_A has quit IRC | 09:54 | |
| *** ianw has quit IRC | 09:55 | |
| *** dpawlik has quit IRC | 09:57 | |
| *** dansmith has joined #openstack-nova | 09:58 | |
| *** ianw has joined #openstack-nova | 09:59 | |
| *** betherly has joined #openstack-nova | 10:01 | |
| *** sapd1_x has joined #openstack-nova | 10:02 | |
| *** bhagyashris has quit IRC | 10:06 | |
| *** jackyzhu has joined #openstack-nova | 10:07 | |
| *** jackyzhu has quit IRC | 10:07 | |
| *** boxiang_ has joined #openstack-nova | 10:08 | |
| sean-k-mooney | mdbooth: ok we have a persitent proablem that happens ocationally in the gate where we get a disk not found error | 10:13 |
| sean-k-mooney | persumably because the root disk is deleted for some reason | 10:13 |
| *** ttsiouts has quit IRC | 10:14 | |
| *** ttsiouts has joined #openstack-nova | 10:15 | |
| *** ociuhandu has joined #openstack-nova | 10:17 | |
| *** ttsiouts has quit IRC | 10:19 | |
| *** tbachman has quit IRC | 10:21 | |
| *** sapd1_x has quit IRC | 10:22 | |
| *** shilpasd has joined #openstack-nova | 10:22 | |
| *** mkrai has joined #openstack-nova | 10:23 | |
| *** aojea has quit IRC | 10:24 | |
| *** ociuhandu_ has joined #openstack-nova | 10:25 | |
| *** ociuhandu has quit IRC | 10:25 | |
| *** epoojad2 has quit IRC | 10:26 | |
| *** avolkov has joined #openstack-nova | 10:26 | |
| kashyap | aspiers: Also, for future reference you might want to keep your guest XML + the full QEMU command-line with SB somewhere. | 10:27 |
| * kashyap bbiab | 10:27 | |
| *** dpawlik has joined #openstack-nova | 10:29 | |
| *** aojea has joined #openstack-nova | 10:29 | |
| *** belmoreira has joined #openstack-nova | 10:33 | |
| *** udesale has quit IRC | 10:38 | |
| aspiers | kashyap: responding | 10:42 |
| aspiers | kashyap: shall I add the SB bp? | 10:42 |
| aspiers | to the commit message | 10:42 |
| *** tbachman has joined #openstack-nova | 10:43 | |
| *** boxiang_ has quit IRC | 10:47 | |
| *** boxiang_ has joined #openstack-nova | 10:47 | |
| openstackgerrit | Adam Spiers proposed openstack/nova-specs master: Fix URL to blueprint for Secure Boot https://review.opendev.org/673477 | 10:48 |
| *** belmoreira has quit IRC | 10:48 | |
| aspiers | I'm gonna assume the answer to that is yes :) | 10:50 |
| openstackgerrit | Adam Spiers proposed openstack/nova master: Track libvirt host/domain capabilities for multiple machine types https://review.opendev.org/673151 | 10:52 |
| *** aojea has quit IRC | 10:52 | |
| openstackgerrit | Huachang Wang proposed openstack/nova master: Add the support of CPU feature 'AVX512-VNNI' https://review.opendev.org/673480 | 10:52 |
| aspiers | kashyap: https://review.opendev.org/#/c/673151/4..5//COMMIT_MSG (view the other files too) | 10:53 |
| openstackgerrit | Lee Yarwood proposed openstack/nova master: WIP compute: Take an instance.uuid lock when rebooting https://review.opendev.org/673463 | 10:54 |
| *** mkrai has quit IRC | 10:54 | |
| *** aojea has joined #openstack-nova | 10:55 | |
| kashyap | aspiers: Hi, you could, as the SB series will most likely be using bits from there anyway. | 10:59 |
| kashyap | (Ah, you did; just saw) | 10:59 |
| *** bhagyashris has joined #openstack-nova | 11:00 | |
| *** belmoreira has joined #openstack-nova | 11:01 | |
| lyarwood | mdbooth: https://review.opendev.org/#/c/673463/ - thoughts on testing this? I'm stuck between func, unit or something else tbh. | 11:01 |
| * lyarwood -> lunch | 11:01 | |
| *** belmoreira has quit IRC | 11:02 | |
| *** ttsiouts has joined #openstack-nova | 11:04 | |
| openstackgerrit | Merged openstack/nova master: doc: fix links for server actions in api guide https://review.opendev.org/673101 | 11:08 |
| openstackgerrit | Merged openstack/nova master: Fix wrong huge pages in doc https://review.opendev.org/673252 | 11:09 |
| *** belmoreira has joined #openstack-nova | 11:09 | |
| sean-k-mooney | i think we made the sufixs case insensitive by the way but i do prefer uppercase so i guess im ok with ^ | 11:13 |
| *** brinzhang_ has quit IRC | 11:15 | |
| *** brinzhang_ has joined #openstack-nova | 11:16 | |
| *** brtknr has quit IRC | 11:19 | |
| *** brault has quit IRC | 11:20 | |
| huaqiang | ls | 11:24 |
| sean-k-mooney | huaqiang: password: :) | 11:26 |
| *** ociuhandu_ has quit IRC | 11:34 | |
| *** belmoreira has quit IRC | 11:35 | |
| *** ociuhandu has joined #openstack-nova | 11:35 | |
| *** belmoreira has joined #openstack-nova | 11:36 | |
| *** dpawlik has quit IRC | 11:52 | |
| openstackgerrit | Brin Zhang proposed openstack/python-novaclient master: Microversion 2.76: Add delete_on_termination to volume-attach API https://review.opendev.org/673485 | 11:55 |
| *** mkrai has joined #openstack-nova | 12:07 | |
| *** jaosorior has quit IRC | 12:07 | |
| *** ociuhandu has quit IRC | 12:14 | |
| *** dpawlik has joined #openstack-nova | 12:16 | |
| openstackgerrit | sean mooney proposed openstack/nova master: Deprecate Aggregate[Core|Ram|Disk]Filters https://review.opendev.org/673496 | 12:21 |
| sean-k-mooney | i just posted to the mailing list about ^ http://lists.openstack.org/pipermail/openstack-discuss/2019-July/008118.html | 12:22 |
| *** eharney has joined #openstack-nova | 12:23 | |
| *** psachin has quit IRC | 12:25 | |
| *** maciejjozefczyk has quit IRC | 12:25 | |
| *** maciejjozefczyk has joined #openstack-nova | 12:27 | |
| yaawang | efried: Can you take a look at this patch(https://review.opendev.org/#/c/673402/) if you have any time? | 12:29 |
| yaawang | efried: This is a part of the implementation of spec expose-auto-converge-post-copy, which is in os-traits project, and the implementation/testing in nova are depend on this patch. | 12:29 |
| openstackgerrit | Merged openstack/nova master: Stop sending bad values from libosinfo to libvirt https://review.opendev.org/669190 | 12:31 |
| *** ricolin has quit IRC | 12:35 | |
| *** bhagyashris has quit IRC | 12:42 | |
| *** bbowen has quit IRC | 12:50 | |
| *** ociuhandu has joined #openstack-nova | 12:50 | |
| *** belmoreira has quit IRC | 12:51 | |
| *** jaosorior has joined #openstack-nova | 12:54 | |
| *** mkrai has quit IRC | 12:55 | |
| *** mkrai_ has joined #openstack-nova | 12:55 | |
| *** belmoreira has joined #openstack-nova | 12:55 | |
| mdbooth | lyarwood: Testing races is really hard. I've written comprehensive race tests before, but the problem is that they're generally hard to understand so I just get moaned at. | 12:59 |
| mdbooth | lyarwood: Fudge it and don't test, or write a test that looks like it's useful, but is simple enough that it isn't. | 13:00 |
| *** mkrai_ has quit IRC | 13:07 | |
| *** boxiang_ has quit IRC | 13:10 | |
| *** boxiang_ has joined #openstack-nova | 13:10 | |
| * cdent hugs mdbooth | 13:11 | |
| *** mriedem has joined #openstack-nova | 13:13 | |
| mriedem | eandersson: were you going to work on backports for these or should i? https://review.opendev.org/#/q/topic:bug/1837955+(status:open+OR+status:merged) | 13:14 |
| *** ttsiouts has quit IRC | 13:14 | |
| *** ttsiouts has joined #openstack-nova | 13:15 | |
| openstackgerrit | Huachang Wang proposed openstack/nova master: doc: correct the information of 'cpu_map' https://review.opendev.org/673272 | 13:18 |
| *** BjoernT has joined #openstack-nova | 13:19 | |
| *** ttsiouts has quit IRC | 13:19 | |
| efried | yaawang: +2, but we need a non-Intel to +A. cdent is looking. | 13:21 |
| efried | brinzhang: I'm here now | 13:22 |
| efried | gmann: Sorry, my search picked up some things it shouldn't have accidentally. Restored. | 13:22 |
| efried | alex_xu: Yup, sorry about that. | 13:24 |
| *** mkrai_ has joined #openstack-nova | 13:24 | |
| artom | efried, btw, is force-abandoning specs past freeze a new thing? IIRC it would be a procedural -2 in the past? | 13:28 |
| efried | artom: There's not a formal/written/official process afaik. | 13:28 |
| artom | I don't really care one way or another, but was surprised by the emails I got last night | 13:28 |
| efried | I decided to abandon because then the author can restore rather than me having to take action. | 13:28 |
| artom | efried, as, as you were ;) | 13:29 |
| artom | *ack | 13:29 |
| efried | Just trying to keep the repo clean-ish. | 13:29 |
| *** bbowen has joined #openstack-nova | 13:30 | |
| *** belmoreira has quit IRC | 13:31 | |
| *** belmoreira has joined #openstack-nova | 13:31 | |
| *** BjoernT_ has joined #openstack-nova | 13:34 | |
| *** BjoernT has quit IRC | 13:35 | |
| *** ratailor has quit IRC | 13:36 | |
| openstackgerrit | Merged openstack/os-traits master: Add new traits to flag live migration features https://review.opendev.org/673402 | 13:37 |
| mriedem | melwitt: how is https://review.opendev.org/#/c/582407/ different from https://review.opendev.org/#/c/562041/ ? | 13:38 |
| mdbooth | lyarwood: Commented on https://review.opendev.org/#/c/673463/2 | 13:40 |
| openstackgerrit | Eric Fried proposed openstack/os-traits master: Followup: spec ref for auto-converge/post-copy https://review.opendev.org/673523 | 13:41 |
| lyarwood | mdbooth: thanks | 13:44 |
| *** ttsiouts has joined #openstack-nova | 13:45 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Add functional recreate test for bug 1764556 https://review.opendev.org/562041 | 13:45 |
| openstack | bug 1764556 in OpenStack Compute (nova) ""nova list" fails with exception.ServiceNotFound if service is deleted and has no UUID" [Medium,In progress] https://launchpad.net/bugs/1764556 - Assigned to melanie witt (melwitt) | 13:45 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Add functional regression test for bug 1778305 https://review.opendev.org/582407 | 13:45 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Don't generate service UUID for deleted services https://review.opendev.org/582408 | 13:45 |
| openstack | bug 1778305 in OpenStack Compute (nova) "Nova may erronously look up service version of a deleted service, when hostname have been reused" [Medium,In progress] https://launchpad.net/bugs/1778305 - Assigned to melanie witt (melwitt) | 13:45 |
| *** mkrai_ has quit IRC | 13:51 | |
| *** yikun has quit IRC | 13:52 | |
| *** mlavalle has joined #openstack-nova | 13:53 | |
| *** factor has joined #openstack-nova | 13:54 | |
| yaawang | efried: cdent Thanks, I saw your patch :) | 13:54 |
| *** Luzi has quit IRC | 13:55 | |
| *** mkrai_ has joined #openstack-nova | 13:56 | |
| yaawang | efried: cdent Will you release a new versoin of os-traits? | 13:56 |
| efried | yaawang: I already updated the hash for the proposed 0.16.0 https://review.opendev.org/673294 | 13:57 |
| efried | yaawang: feel free to +1 that patch. Once cdent re-acks it, the release team should push it by the end of the week. | 13:57 |
| efried | ...which he has now done | 13:58 |
| yaawang | Great, I'll push the nova code after os-traits released, so it'll easy to pass the test :) | 14:00 |
| openstackgerrit | Merged openstack/os-traits master: Followup: spec ref for auto-converge/post-copy https://review.opendev.org/673523 | 14:01 |
| *** rpittau is now known as rpittau|afk | 14:02 | |
| efried | yaawang: You'll need to wait until upper-constraints is bumped. A bot will propose a patch to the openstack/requirements project after the release patch merges. | 14:04 |
| efried | yaawang: Then in your patch, you can update the requirements.txt and lower-constraints.txt to 0.16.0 | 14:05 |
| *** tbachman has quit IRC | 14:07 | |
| *** tbachman has joined #openstack-nova | 14:09 | |
| yaawang | Get you point :) | 14:09 |
| *** brtknr has joined #openstack-nova | 14:10 | |
| *** BjoernT_ has quit IRC | 14:11 | |
| *** BjoernT has joined #openstack-nova | 14:11 | |
| *** BjoernT_ has joined #openstack-nova | 14:15 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova stable/stein: Add functional regression test for bug 1837955 https://review.opendev.org/673532 | 14:16 |
| openstack | bug 1837955 in OpenStack Compute (nova) stein "MaxRetriesExceeded sometime fails with messaging exception" [Medium,Confirmed] https://launchpad.net/bugs/1837955 | 14:16 |
| openstackgerrit | Matt Riedemann proposed openstack/nova stable/stein: Cleanup when hitting MaxRetriesExceeded from no host_available https://review.opendev.org/673533 | 14:16 |
| *** BjoernT has quit IRC | 14:16 | |
| *** mlavalle has quit IRC | 14:17 | |
| *** mlavalle has joined #openstack-nova | 14:19 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova stable/rocky: Add functional regression test for bug 1837955 https://review.opendev.org/673536 | 14:25 |
| openstack | bug 1837955 in OpenStack Compute (nova) stein "MaxRetriesExceeded sometime fails with messaging exception" [Medium,In progress] https://launchpad.net/bugs/1837955 - Assigned to Matt Riedemann (mriedem) | 14:25 |
| sean-k-mooney | mriedem: thanks for reviewing the deprecateion patch. i want to get stephenfin to take a look as i know there is a way i can pull out the repeated content and link it in once. ill look at the other comments you left and adress them after i grab lunch | 14:26 |
| mriedem | eandersson: i've started backports. just fyi the backport of your fix gets tricky in rocky b/c the _cleanup_when_reschedule_fails method doesn't exist in rocky | 14:26 |
| mriedem | sean-k-mooney: you inline it with a doc reference i think, not sure that's really worth it though | 14:27 |
| *** BjoernT_ has quit IRC | 14:27 | |
| sean-k-mooney | ok well i just wanted to get a basic version up to link in the mailing list post | 14:27 |
| sean-k-mooney | im happy to simplfy | 14:27 |
| sean-k-mooney | i more or less based it off your deprecation patch for the non aggreate versions but i grant you its proably too verbose as it is | 14:28 |
| sean-k-mooney | anyway ill be back in 30 mins or so | 14:28 |
| *** mkrai_ has quit IRC | 14:30 | |
| *** dpawlik has quit IRC | 14:34 | |
| *** BjoernT_ has joined #openstack-nova | 14:34 | |
| *** factor has quit IRC | 14:39 | |
| *** factor has joined #openstack-nova | 14:39 | |
| *** tbachman has quit IRC | 14:39 | |
| *** icarusfactor has joined #openstack-nova | 14:43 | |
| *** belmoreira has quit IRC | 14:45 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova stable/rocky: Reproduce bug #1819460 in functional test https://review.opendev.org/673546 | 14:45 |
| openstack | bug 1819460 in OpenStack Compute (nova) rocky "instance stuck in BUILD state due to unhandled exceptions in conductor" [Undecided,New] https://launchpad.net/bugs/1819460 | 14:45 |
| *** factor has quit IRC | 14:46 | |
| *** Nick_A has joined #openstack-nova | 14:48 | |
| *** mlavalle has quit IRC | 14:50 | |
| *** icarusfactor has quit IRC | 14:50 | |
| *** belmoreira has joined #openstack-nova | 14:50 | |
| *** icarusfactor has joined #openstack-nova | 14:50 | |
| *** belmoreira has quit IRC | 14:51 | |
| *** mlavalle has joined #openstack-nova | 14:51 | |
| *** dklyle has quit IRC | 14:52 | |
| *** dklyle has joined #openstack-nova | 14:53 | |
| *** mkrai_ has joined #openstack-nova | 14:53 | |
| *** takamatsu has joined #openstack-nova | 14:54 | |
| *** icarusfactor has quit IRC | 14:55 | |
| *** factor has joined #openstack-nova | 14:56 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova stable/rocky: Handle placement error during re-schedule https://review.opendev.org/673550 | 14:57 |
| *** belmoreira has joined #openstack-nova | 14:59 | |
| melwitt | mriedem: they are similar but one condition causes ServiceTooOld to be raised and the other causes ServiceNotFound to be raised. have to look some more to see exactly what condition results in the different behaviors | 15:03 |
| openstackgerrit | Matt Riedemann proposed openstack/nova stable/rocky: Cleanup when hitting MaxRetriesExceeded from no host_available https://review.opendev.org/673553 | 15:05 |
| mriedem | melwitt: was it the service version? | 15:05 |
| mriedem | that was about the only thing that stood out | 15:05 |
| mriedem | servicetooold would be the version related thing | 15:05 |
| mriedem | eandersson: ^ there is the rocky backport | 15:06 |
| melwitt | mriedem: yeah, when a failure occurs at this code https://github.com/openstack/nova/blob/16.1.4/nova/objects/service.py#L362 | 15:08 |
| melwitt | and it looks like the other bug, the ServiceNotFound is raised from the db.service_update call, after the service version check passed | 15:09 |
| *** tbachman has joined #openstack-nova | 15:10 | |
| mriedem | yeah ok | 15:10 |
| mriedem | i didn't make the too old <> not found connection | 15:11 |
| *** belmoreira has quit IRC | 15:12 | |
| *** hamzy has quit IRC | 15:12 | |
| *** ricolin has joined #openstack-nova | 15:14 | |
| mriedem | bauzas: https://review.opendev.org/#/c/562041/ needed to be rebased and remove the set_nodes/restore_nodes stuff that's been removed on master, otherwise it's the same - can i get a fast re-approve there? | 15:16 |
| mriedem | the rest of the series above it is approved | 15:16 |
| *** belmoreira has joined #openstack-nova | 15:27 | |
| *** mriedem has quit IRC | 15:33 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova stable/rocky: Cleanup when hitting MaxRetriesExceeded from no host_available https://review.opendev.org/673553 | 15:35 |
| *** mriedem has joined #openstack-nova | 15:35 | |
| *** mkrai_ has quit IRC | 15:39 | |
| bauzas | mriedem: sure | 15:39 |
| *** whoami-rajat has quit IRC | 15:39 | |
| bauzas | done | 15:39 |
| *** belmoreira has quit IRC | 15:42 | |
| *** ttsiouts has quit IRC | 15:42 | |
| *** ttsiouts has joined #openstack-nova | 15:42 | |
| *** tssurya has quit IRC | 15:46 | |
| *** ttsiouts has quit IRC | 15:47 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova stable/queens: Add functional regression test for bug 1837955 https://review.opendev.org/673567 | 15:48 |
| openstack | bug 1837955 in OpenStack Compute (nova) stein "MaxRetriesExceeded sometime fails with messaging exception" [Medium,In progress] https://launchpad.net/bugs/1837955 - Assigned to Matt Riedemann (mriedem) | 15:48 |
| *** aojea has quit IRC | 15:50 | |
| *** gyee has joined #openstack-nova | 15:52 | |
| *** aojea has joined #openstack-nova | 15:53 | |
| *** shilpasd has quit IRC | 15:56 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova master: WIP: Add a placement audit command https://review.opendev.org/670112 | 16:03 |
| bauzas | mriedem: quick shot on a total rework of the nova-manage command | 16:05 |
| bauzas | mriedem: I'm rushing to write the func tests before I enter the PTO season | 16:05 |
| *** aojea has quit IRC | 16:07 | |
| *** whoami-rajat has joined #openstack-nova | 16:07 | |
| *** panda has quit IRC | 16:09 | |
| *** lpetrut has quit IRC | 16:11 | |
| *** panda has joined #openstack-nova | 16:13 | |
| *** Sundar has joined #openstack-nova | 16:17 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova stable/rocky: Cleanup when hitting MaxRetriesExceeded from no host_available https://review.opendev.org/673553 | 16:17 |
| *** yonglihe has quit IRC | 16:22 | |
| *** arne_wiebalck has quit IRC | 16:22 | |
| *** yonglihe has joined #openstack-nova | 16:26 | |
| *** arne_wiebalck has joined #openstack-nova | 16:28 | |
| *** rajinir has quit IRC | 16:28 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova stable/queens: Cleanup when hitting MaxRetriesExceeded from no host_available https://review.opendev.org/673576 | 16:29 |
| *** rajinir has joined #openstack-nova | 16:30 | |
| *** factor has quit IRC | 16:31 | |
| *** factor has joined #openstack-nova | 16:31 | |
| sean-k-mooney | efried: by the way i would be totally fine with making the "cpu priortisiation" use case a docs only feature. e.g. in the install guide or admin docs have a topic doc that explains how to use the provider.yaml to add the custom trati and host to configure the cpu_shared_set, cpu_dedicated_set to map to the high and low priority cores defined in your speed select policy | 16:31 |
| sean-k-mooney | efried: rather then being totally downstream only | 16:32 |
| *** trident has quit IRC | 16:32 | |
| *** igordc has joined #openstack-nova | 16:33 | |
| sean-k-mooney | with the cpu resocuse tracking in placmenet spec and provider.yaml everything is there to allow you to staticly configure this without addtional code chagnes | 16:33 |
| sean-k-mooney | we likely will end up looking to support speed select and cache via that mechcanium with train eventuraly for OSP so if we collaberated on upstream does it would both sever as a concreate example of how to use the provider.yaml and de duplicate the work | 16:35 |
| *** rajinir has quit IRC | 16:36 | |
| *** trident has joined #openstack-nova | 16:37 | |
| *** rajinir has joined #openstack-nova | 16:38 | |
| stephenfin | dansmith: I need to determine if the NUMATopology o.vo I'm looking at was created on a pre-Train compute, because some fields I'm adding won't be present. Can I check the version of that object easily? | 16:39 |
| artom | b'reference = NUMAMigrateData(cell_pins={2=set([6,7]),3=set([8,9])},cpu_pins={0=set([0,1]),1=set([2,3])},emulator_pins=set([4,5]),sched_priority=8,sched_vcpus=set([6,7]))' | 16:39 |
| artom | b'actual = NUMAMigrateData(cell_pins={2=set([6,7]),3=set([8,9])},cpu_pins={0=set([0,1]),1=set([2,3])},emulator_pins=set([4,5]),sched_priority=8,sched_vcpus=set([6,7]))' | 16:39 |
| artom | Umm | 16:39 |
| sean-k-mooney | artom: you are forgeting about the reset changed thing | 16:40 |
| sean-k-mooney | when your are comparing two ovo you should convert them to primatives and compare those | 16:40 |
| artom | Maybe. But I still have the right to be damn angry about it. | 16:41 |
| sean-k-mooney | othersize even if they have the same data they will be considered different of one has a files change state different then the other object | 16:41 |
| sean-k-mooney | artom: you have hit this in the past no? | 16:41 |
| artom | But... thanks sean-k-mooney :) | 16:41 |
| sean-k-mooney | we migth have a unitlity funciton for this too | 16:41 |
| artom | sean-k-mooney, don't think so, the past was list ordering and stuch | 16:41 |
| artom | *such | 16:41 |
| sean-k-mooney | ah ok | 16:42 |
| sean-k-mooney | i have hit that in the past so it burn into my memory | 16:42 |
| artom | My memory is apparently resistant to burning :( | 16:42 |
| stephenfin | dansmith: In case it matters, the object is saved in serialized form in the 'ComputeNode.numa_topology' (StringField) field and I'm having to deserialize it using this https://github.com/openstack/nova/blob/master/nova/objects/numa.py#L243-L250 | 16:42 |
| sean-k-mooney | stephenfin: you can check it in two ways one the objec thousl have a version embeded or two you can just check if the field is present | 16:43 |
| stephenfin | sean-k-mooney: Is it embedded? I'm looking at the 'VersionedObject' base class and I can't see anything obvious | 16:45 |
| stephenfin | there's VERSION but that's a class variable, not an instance variable | 16:45 |
| sean-k-mooney | its in the primitave dict | 16:45 |
| efried | sean-k-mooney: what you said | 16:45 |
| sean-k-mooney | https://github.com/openstack/oslo.versionedobjects/blob/master/oslo_versionedobjects/base.py#L408 | 16:45 |
| stephenfin | That's only present in the serialized form though, right? | 16:46 |
| *** jmlowe has joined #openstack-nova | 16:46 | |
| sean-k-mooney | stephenfin: that is what we retire form the db no? | 16:46 |
| *** ociuhandu has quit IRC | 16:46 | |
| stephenfin | Which I guess I can use since I'm deserializing things as noted above | 16:47 |
| *** ociuhandu has joined #openstack-nova | 16:47 | |
| *** yonglihe_ has joined #openstack-nova | 16:47 | |
| stephenfin | It is, but it feels a bit wrong to be modifying that 'obj_from_db_obj' method to do this stuff | 16:47 |
| *** yonglihe has quit IRC | 16:47 | |
| *** yonglihe_ is now known as yonglihe | 16:47 | |
| stephenfin | I'd much rather have done it where I'm actually using it, but I guess beggars can't be choosers, heh | 16:47 |
| sean-k-mooney | what are you trying to do | 16:48 |
| stephenfin | sec | 16:48 |
| stephenfin | sean-k-mooney: I need to work around this https://review.opendev.org/#/c/671800/8/nova/virt/hardware.py@1912 | 16:49 |
| stephenfin | back in 2 mins | 16:49 |
| sean-k-mooney | we defien a cpu_usage property on teh ovo that know how to retrun teh correct value | 16:50 |
| sean-k-mooney | for the old object cpu_usage would be in the object field for the new value it would be the sum of shared_cpu_usage+dedicated_cpu_usage | 16:52 |
| *** mkrai_ has joined #openstack-nova | 16:52 | |
| stephenfin | Yup, correct. I don't think we actually need to worry about supporting stuff reading old NUMATopology objects though | 16:52 |
| sean-k-mooney | a property can just use "in" to determin if the field is set | 16:52 |
| stephenfin | I was trying to figure out what actually sets NUMAToplogy.cpu_usage | 16:53 |
| stephenfin | I figured it would be driver, but it seems it's not | 16:53 |
| sean-k-mooney | its set by the resouce tracker | 16:53 |
| stephenfin | Instead, the driver (libvirt) returns the object with those fields set to 0 | 16:53 |
| stephenfin | Yeah | 16:53 |
| stephenfin | which seems a little dumb, tbh, when the compute node could be doing this for us | 16:53 |
| sean-k-mooney | its update by a perodic task | 16:53 |
| sean-k-mooney | well the resouce tracker runs on the compute node | 16:54 |
| stephenfin | Sorry, the driver | 16:54 |
| sean-k-mooney | and this code is ment to be virt driver independent | 16:54 |
| sean-k-mooney | its update by https://github.com/openstack/nova/blob/master/nova/compute/resource_tracker.py#L676 | 16:55 |
| stephenfin | I mean, why can't we the driver populate this field when it creates it https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L6614 | 16:55 |
| stephenfin | Yeah, via a call to '_update_available_resource' | 16:55 |
| stephenfin | which calls '_update_usage_from_instances' (plural), which calls '_update_usage_from_instance' (singular) multiple times for each instance | 16:56 |
| sean-k-mooney | well we start by calling https://github.com/openstack/nova/blob/master/nova/compute/resource_tracker.py#L695 | 16:56 |
| sean-k-mooney | resources = self.driver.get_available_resource(nodename) | 16:56 |
| sean-k-mooney | which gets the resouce form the driver | 16:57 |
| sean-k-mooney | _update_usage_from_instances is to take care of cleaning up after/during migrations | 16:57 |
| sean-k-mooney | at least i thinkt that is what that does | 16:58 |
| stephenfin | self.driver.get_available_resource returns the NUMATopology object with the usage fields set to 0 | 16:58 |
| stephenfin | and the '_update_usage_from_instance' is what actually sets those | 16:59 |
| stephenfin | which isn't intuitive at all | 16:59 |
| stephenfin | At least, to me | 16:59 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Remove Request Spec Migration upgrade status check https://review.opendev.org/673580 | 17:01 |
| *** derekh has quit IRC | 17:01 | |
| sean-k-mooney | we proably could optimise it so the first call calulated everythign but i think this is called in a perodic task | 17:02 |
| sean-k-mooney | and im not sure if just the inner fucntion is called _update_available_resource | 17:02 |
| sean-k-mooney | or if update_available_resource is called | 17:02 |
| *** BjoernT_ has quit IRC | 17:03 | |
| *** markvoelker has quit IRC | 17:03 | |
| openstackgerrit | Merged openstack/nova-specs master: Fix URL to blueprint for Secure Boot https://review.opendev.org/673477 | 17:04 |
| *** mkrai_ has quit IRC | 17:05 | |
| *** cdent has quit IRC | 17:05 | |
| *** ociuhandu has quit IRC | 17:07 | |
| sean-k-mooney | efried: by the way "what you said" were you agreeing with adding a topic doc for speed select using provider.yaml ectra | 17:08 |
| sean-k-mooney | or was that in relation to something else | 17:08 |
| efried | sean-k-mooney: I was basically +1ing the concept that the way we're looking at mixed shared+dedicated and providers.yaml is, at least from Intel's perspective, all about enabling those use cases in a way that keeps nova from having to know too much. | 17:09 |
| efried | sean-k-mooney: I would be fully supportive of the idea of having an in-tree doc that described the CPU priority setup as an *example* of how you could piece together providers.yaml etc. | 17:10 |
| efried | But I wouldn't think we should have it as a suggested, supported, etc. thing. | 17:10 |
| sean-k-mooney | ah right, well i could keep my pm happy by wrighign some doc to show how to support those usecase with trian | 17:10 |
| *** ricolin has quit IRC | 17:11 | |
| sean-k-mooney | do you know if there has been any progress on making RMD listent to libvirt events and read the flaovr info | 17:12 |
| sean-k-mooney | we could also enable cache and memory bandwith in a similar way if RMD has been updateded as we spoke about at the ptg | 17:13 |
| efried | sean-k-mooney: They're working on it, but I don't think there's anything to show yet. | 17:14 |
| efried | FYI I think they're planning to hack the domain xml file.... | 17:14 |
| sean-k-mooney | so it will never be supported got it | 17:15 |
| sean-k-mooney | why? | 17:15 |
| sean-k-mooney | the instance and flavor uuid are in the xml and they can do the enforcement without needing to modify the xml | 17:16 |
| efried | sean-k-mooney: I thought they needed to update the xml to assign cacheways to the VM? | 17:16 |
| sean-k-mooney | no they can do it out of band via the resouce control file system | 17:17 |
| efried | okay, neat. | 17:17 |
| sean-k-mooney | they basically have two options they can tell the kernel to apploy the policy to a set of host cores. e.g. the ones the vm is pinned too or they can use the process id | 17:18 |
| sean-k-mooney | at least for cache | 17:18 |
| *** jmlowe has quit IRC | 17:18 | |
| *** ralonsoh has quit IRC | 17:23 | |
| stephenfin | sean-k-mooney: Should we prevent one thread sibling being used for VCPUs and another for PCPUs? | 17:28 |
| stephenfin | I think yes, but I'd like a sanity check :) | 17:29 |
| stephenfin | s/and another/and the other/ | 17:29 |
| stephenfin | Or others, for something like POWER | 17:29 |
| sean-k-mooney | we could | 17:29 |
| sean-k-mooney | we dont have too | 17:29 |
| sean-k-mooney | dedicated cpus does not imply isolate | 17:30 |
| stephenfin | Nope, but the isolate policy does | 17:30 |
| sean-k-mooney | we could issue a warning | 17:30 |
| sean-k-mooney | not any more | 17:30 |
| sean-k-mooney | isolate means find a host with SMT disabled | 17:30 |
| stephenfin | True | 17:30 |
| sean-k-mooney | so it wont be an issue | 17:30 |
| sean-k-mooney | we shoudl also document how to add/remove the hyperthreind trait using provider.yaml | 17:31 |
| stephenfin | True. I guess I need to decide what I'm doing with '(un)pin_cpus_with_siblings' on that front | 17:31 |
| stephenfin | Yeah, I haven't actually given that part of the spec much thought yet | 17:31 |
| sean-k-mooney | actully we cant do that with provider.ymal | 17:32 |
| *** brault has joined #openstack-nova | 17:32 | |
| *** brault has quit IRC | 17:32 | |
| sean-k-mooney | because its a standard trait | 17:32 |
| efried | not because it's a standard trait | 17:32 |
| efried | because we're not supporting trait removal in this pass | 17:32 |
| efried | removing a standard trait is totally something you will eventually be able to do with providers.yaml. | 17:33 |
| stephenfin | I was going to automatically report it via the libvirt driver | 17:33 |
| stephenfin | Can we not do that? | 17:33 |
| sean-k-mooney | efried: the virt diver does not own the hypertreading trait | 17:33 |
| sean-k-mooney | so its not allowed to report it | 17:33 |
| sean-k-mooney | per the cpu resouce trackign spec | 17:33 |
| stephenfin | sean-k-mooney: Can we make it own it? | 17:33 |
| efried | who owns it? | 17:33 |
| sean-k-mooney | the operator | 17:33 |
| stephenfin | I forget the context on this, tbh | 17:33 |
| efried | okay, if the operator owns it, they can use providers.yaml to have it added to the compute. | 17:33 |
| efried | or they can just ... add it. | 17:33 |
| efried | standard/custom isn't the dividing line for what you can do out of band | 17:34 |
| sean-k-mooney | it has its own section https://specs.openstack.org/openstack/nova-specs/specs/train/approved/cpu-resources.html#add-hw-cpu-hyperthreading-trait | 17:34 |
| *** tesseract has quit IRC | 17:34 | |
| efried | If we haven't declared this trait to be in the set of ones compute "owns" then the operator can do whatever they want with it. | 17:35 |
| sean-k-mooney | yes | 17:35 |
| stephenfin | efried: We can still report it though, right? | 17:35 |
| stephenfin | We just don't override it | 17:35 |
| stephenfin | Not sure that makes sense but I don't really know how virt-reported traits work yet | 17:35 |
| sean-k-mooney | stephenfin: since we dont have a way to remove direver reported tratis we should not unless we have a config option to disable it | 17:35 |
| sean-k-mooney | stephenfin: the reason that this exits is two fold. one peole wanted to be able to leave hypertreaind enabled on the host for the shared cpus and not list the hyper tread in the dedicated set | 17:37 |
| sean-k-mooney | the other reason is with numa peopl wanted to have teh flexablity to decler the trait diffently per socket/numa node | 17:38 |
| sean-k-mooney | it was motivated by small edge deployment of 1-5 nodes | 17:39 |
| sean-k-mooney | where its not resonable to dedicate entire host to one confiuration or the other | 17:39 |
| sean-k-mooney | since we can nolnger do this per instance | 17:39 |
| sean-k-mooney | via cpu tread polices | 17:39 |
| stephenfin | How about we report the trait but only if all hyperthreads are enabled/disabled | 17:39 |
| stephenfin | i.e. there's either no cpu set configuration | 17:39 |
| stephenfin | or there is, and hyper threads aren't listed in either option | 17:40 |
| stephenfin | or the host doesn't have hyperthreads | 17:40 |
| sean-k-mooney | all hypertreads? e.g. you whitelist both thread siblings | 17:40 |
| sean-k-mooney | i think that overly complex to debug | 17:40 |
| stephenfin | Hmm, that's true | 17:41 |
| sean-k-mooney | i think a config to enable auto reporting or just making it work with provider.yaml is the simplest | 17:41 |
| stephenfin | Yeah, config option wfm | 17:41 |
| stephenfin | *disable* auto-reporting though | 17:42 |
| sean-k-mooney | sure | 17:42 |
| stephenfin | I'd rather cater to the 90% than the 10% | 17:42 |
| sean-k-mooney | but that should be in the driver section | 17:42 |
| stephenfin | Yup | 17:42 |
| sean-k-mooney | and ya im fine with disableing it if you know what your doing and want to specfy something custum | 17:43 |
| sean-k-mooney | stephenfin: you realise however that if its auto enabeld that isolated vms will not be moveable if you have hypertreading on everything | 17:43 |
| sean-k-mooney | or bootable | 17:43 |
| sean-k-mooney | i guess the same is true of require | 17:44 |
| stephenfin | Yup | 17:44 |
| stephenfin | Just go toggle the config option and you're sorted | 17:44 |
| efried | just note that if virt reports it, you can't disable it. | 17:44 |
| stephenfin | efried: wdym? | 17:44 |
| sean-k-mooney | efried: you can if there is a config option in the libvirt section | 17:44 |
| efried | I mean, you could, like with a conf option or something, but that would be a new and different way of doing things. | 17:44 |
| efried | ^ | 17:44 |
| sean-k-mooney | right but nothing reports this trait currently | 17:45 |
| efried | Sorry, let me backpedal a bit | 17:45 |
| sean-k-mooney | we standardised it for this spec and have not implemented using it yet | 17:45 |
| efried | I think it would be fine if, within the actual libvirt driver (in update_provider_tree), there was a CONF-based conditional which, if true (and other conditions are met), would report the hyperthreading trait on the compute node RP. | 17:46 |
| efried | But it should not be one of the automatic traits controlled by the RT outside of the virt driver. | 17:46 |
| sean-k-mooney | right | 17:47 |
| sean-k-mooney | i agree with that | 17:47 |
| efried | clear as mud stephenfin? | 17:47 |
| stephenfin | What's the difference between virt-driver reported traits and resource tracker-provided traits? | 17:48 |
| sean-k-mooney | stephenfin: he is saying don do it here https://github.com/openstack/nova/blob/master/nova/compute/resource_tracker.py#L1001-L1003 | 17:48 |
| sean-k-mooney | which but do it in the driver update provider tree | 17:49 |
| efried | exactly | 17:49 |
| efried | it sucks how messy this has gotten | 17:50 |
| stephenfin | Oh, so do it here instead? https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L6797-L6805 | 17:50 |
| stephenfin | Got it, if so | 17:50 |
| efried | Yes, more or less. | 17:50 |
| sean-k-mooney | yes | 17:50 |
| sean-k-mooney | like im doing here https://review.opendev.org/#/c/666914/5/nova/virt/libvirt/driver.py | 17:50 |
| *** francoisp has quit IRC | 17:50 | |
| sean-k-mooney | for my storage bus traits | 17:50 |
| efried | Not sure I would do it from within _get_cpu_traits, but that's the right neighborhood. | 17:51 |
| sean-k-mooney | i would not | 17:51 |
| sean-k-mooney | add a seperate function | 17:51 |
| efried | I guess it's a CPU trait... | 17:51 |
| efried | anyway, we're on the same page. | 17:51 |
| sean-k-mooney | im adding 3 more traits functions to the libvirt driver | 17:52 |
| stephenfin | Yeah, but that general area | 17:52 |
| sean-k-mooney | traits.update(self._get_storage_bus_traits()) | 17:52 |
| stephenfin | Cool, I'll do just that | 17:52 |
| sean-k-mooney | and just calling them in sequen to build up the dict of tratis | 17:52 |
| sean-k-mooney | it makes the code simpler | 17:53 |
| efried | Doesn't bother me to expand trait logic in a given upt. What's bugging me is the growing number of different sources-of-truth for traits, and trying to keep track of their precedence. | 17:53 |
| efried | virt driver, RT, direct placement, and soon providers.yaml (which will funnel through the RT) | 17:54 |
| sean-k-mooney | stephenfin: the follow up patch is a better example https://review.opendev.org/#/c/666915/6/nova/virt/libvirt/driver.py@6820 | 17:54 |
| sean-k-mooney | efried: does direct placmeent actully work? | 17:54 |
| sean-k-mooney | or do we overwrite them still | 17:54 |
| efried | we're not supposed to | 17:55 |
| efried | unless you muck with ones the compute supposedly "owns". | 17:55 |
| efried | so like, you can't switch a CPU capability trait on or off | 17:55 |
| sean-k-mooney | that is why we pass the bool for if the trait should be present in the dict of tratis right | 17:55 |
| efried | or like MULTI_ATTACH | 17:55 |
| efried | partly, yeah - the bool is so we can get both sides of the "I own it" equation. Otherwise we would only be able to switch on ones that were off. | 17:56 |
| sean-k-mooney | efried: we could remove one of the soucrces of truth if we wanted too | 17:56 |
| stephenfin | sean-k-mooney++ efried++ | 17:56 |
| sean-k-mooney | e.g. we could move the traits that are set in the RT into the driver | 17:56 |
| stephenfin | and with that, I'm off home o/ | 17:56 |
| efried | o/ stephenfin | 17:56 |
| sean-k-mooney | by impleemnting the common code in a function that all driver call | 17:56 |
| sean-k-mooney | on the other hand i dont realy want to change that right now | 17:57 |
| *** Sundar has quit IRC | 17:59 | |
| sean-k-mooney | efried: by the way are you implementing provider.yaml or is that being done by someone else | 17:59 |
| openstackgerrit | Andreas Jaeger proposed openstack/os-resource-classes master: Update api-ref location https://review.opendev.org/673604 | 17:59 |
| sean-k-mooney | it looks like there are two implementations https://review.opendev.org/#/q/topic:bp/provider-config-file+(status:open+OR+status:merged) | 18:00 |
| efried | sean-k-mooney: dustinc is doing it. | 18:00 |
| efried | ...based on some stuff I did last year, yeah. | 18:00 |
| sean-k-mooney | ah ok | 18:01 |
| dustinc | sean-k-mooney: yeah I am working on it as we chat | 18:01 |
| dustinc | the newer one | 18:01 |
| sean-k-mooney | https://review.opendev.org/#/c/673341/1 | 18:01 |
| sean-k-mooney | is the other one still relevent | 18:01 |
| sean-k-mooney | or can it be abandoned | 18:01 |
| sean-k-mooney | https://review.opendev.org/#/c/622622/1 | 18:01 |
| dustinc | defer to efried, but I imagine it can be abandoned | 18:02 |
| dustinc | there is more to that one than what I am doing, dunno if it will be relevant for something else | 18:02 |
| sean-k-mooney | dustinc: well it will live in gerrit for ever. | 18:02 |
| *** priteau has quit IRC | 18:02 | |
| efried | I can abandon it. Can always restore later (or more likely, steal stuff from it) | 18:02 |
| efried | yeah | 18:02 |
| sean-k-mooney | ill star https://review.opendev.org/#/c/673341/1 so i know to review that one instead | 18:03 |
| dustinc | thanks! | 18:03 |
| efried | abandoned the other. | 18:03 |
| sean-k-mooney | on of the things i do like about gerrit is 1 the history but more so if you use topic properly its easy to find stuff that is reslated | 18:04 |
| sean-k-mooney | *related | 18:04 |
| *** hamzy has joined #openstack-nova | 18:11 | |
| aspiers | sean-k-mooney: agreed. I was thinking earlier that it should be possible to make a human-readable Change-Id | 18:22 |
| aspiers | then more useful metadata could be written to express logical dependencies between commits | 18:23 |
| *** BjoernT has joined #openstack-nova | 18:24 | |
| *** BjoernT has quit IRC | 18:28 | |
| *** BjoernT has joined #openstack-nova | 18:32 | |
| *** BjoernT has quit IRC | 18:35 | |
| *** BjoernT_ has joined #openstack-nova | 18:35 | |
| *** bbowen has quit IRC | 18:36 | |
| *** BjoernT_ has quit IRC | 18:40 | |
| openstackgerrit | Merged openstack/os-resource-classes master: Update api-ref location https://review.opendev.org/673604 | 18:46 |
| openstackgerrit | melanie witt proposed openstack/nova master: Add functional recreate test for bug 1764556 https://review.opendev.org/562041 | 18:49 |
| openstack | bug 1764556 in OpenStack Compute (nova) ""nova list" fails with exception.ServiceNotFound if service is deleted and has no UUID" [Medium,In progress] https://launchpad.net/bugs/1764556 - Assigned to melanie witt (melwitt) | 18:49 |
| openstackgerrit | melanie witt proposed openstack/nova master: Add functional regression test for bug 1778305 https://review.opendev.org/582407 | 18:49 |
| openstackgerrit | melanie witt proposed openstack/nova master: Don't generate service UUID for deleted services https://review.opendev.org/582408 | 18:49 |
| openstack | bug 1778305 in OpenStack Compute (nova) "Nova may erronously look up service version of a deleted service, when hostname have been reused" [Medium,In progress] https://launchpad.net/bugs/1778305 - Assigned to melanie witt (melwitt) | 18:49 |
| openstackgerrit | Merged openstack/nova master: Run 'tempest-ipv6-only' job in gate https://review.opendev.org/671900 | 18:49 |
| openstackgerrit | Eric Fried proposed openstack/nova master: DNM: Where have all the global_request_id gone? https://review.opendev.org/672986 | 18:51 |
| mriedem | just debugged a unit test for nearly an hour and it was so dumb, my brain must be mush | 19:11 |
| *** eharney has quit IRC | 19:15 | |
| *** mriedem has quit IRC | 19:20 | |
| *** BjoernT has joined #openstack-nova | 19:21 | |
| *** tbachman has quit IRC | 19:22 | |
| *** amodi has quit IRC | 19:25 | |
| *** tbachman has joined #openstack-nova | 19:29 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Add blocker migration for completing services.uuid migration https://review.opendev.org/673620 | 19:32 |
| *** mriedem has joined #openstack-nova | 19:32 | |
| aspiers | kashyap: I just tried the same SEV experiment on Leap 15.0 with the same guest image (SLES15) and it gets stuck at the UEFI boot prompt, any ideas? | 19:40 |
| openstackgerrit | Merged openstack/nova master: Replace "integrated-gate-py3" template with new "integrated-gate-compute" https://review.opendev.org/671551 | 19:42 |
| *** jmlowe has joined #openstack-nova | 19:57 | |
| *** whoami-rajat has quit IRC | 19:57 | |
| eandersson | Is it safe to change overcommit ratios from 16 to 1 on a compute (it was never intended to have 16). | 20:10 |
| eandersson | Will it update the placement inventory db automatically when the service is restarted? | 20:10 |
| eandersson | *updaate | 20:11 |
| *** hamzy has quit IRC | 20:14 | |
| sean-k-mooney | eandersson: if you have overcommited unintentially all that will happen is no vm willl be allowed on that host untill you drop below the new overcommit limit | 20:14 |
| sean-k-mooney | so yes | 20:14 |
| eandersson | Perfect - I assumed as much, but my lab env was acting weird (probably needs to be rebuilt). | 20:14 |
| * eandersson needs a better lab. | 20:15 | |
| *** xek has quit IRC | 20:15 | |
| *** BjoernT has quit IRC | 20:15 | |
| *** BjoernT has joined #openstack-nova | 20:16 | |
| mriedem | sean-k-mooney: remember last week when i was having problems with an ubuntu 18.04 devstack vm, weird libvirt hang issues and such? | 20:18 |
| mriedem | i've got another one today with lots of random uwsgi crashes, | 20:18 |
| sean-k-mooney | yep | 20:18 |
| mriedem | i realized i'm running on py27 | 20:18 |
| mriedem | which we probably don't run on the gate much anywhere anymore | 20:18 |
| mriedem | so that might have something to do with it | 20:18 |
| sean-k-mooney | oh maybe | 20:18 |
| mriedem | uwsgi seg faults | 20:18 |
| sean-k-mooney | we still run in the gate with py27 i think | 20:19 |
| sean-k-mooney | at least i think we have at lest 1 tempest job | 20:19 |
| *** BjoernT_ has joined #openstack-nova | 20:19 | |
| mriedem | yeah we do, tempest-full is py27 | 20:19 |
| sean-k-mooney | i have been using python 3 however for ubuntu 18.04 | 20:20 |
| sean-k-mooney | i actuly had the opisite issue not that long ago trying to run some py3 test on centos 7 | 20:20 |
| *** BjoernT has quit IRC | 20:20 | |
| sean-k-mooney | we might need to cap something in upper constratits | 20:21 |
| sean-k-mooney | we have started to see more and more upstream libs drop py2 support | 20:22 |
| mriedem | heh n-cpu just crashed too | 20:22 |
| mriedem | wtf | 20:22 |
| *** jmlowe has quit IRC | 20:23 | |
| *** BjoernT has joined #openstack-nova | 20:23 | |
| sean-k-mooney | i can restack with py 2.7 and see if it breaks | 20:23 |
| sean-k-mooney | i assume your running master | 20:24 |
| *** BjoernT_ has quit IRC | 20:24 | |
| *** dpawlik has joined #openstack-nova | 20:25 | |
| *** jmlowe has joined #openstack-nova | 20:25 | |
| *** BjoernT_ has joined #openstack-nova | 20:27 | |
| *** BjoernT has quit IRC | 20:29 | |
| sean-k-mooney | my dev vm is restacking now with Python 2.7.15 on kernel 4.18.0-25-generic | 20:34 |
| *** BjoernT_ is now known as BjoernT | 20:34 | |
| sean-k-mooney | ill let you know if it breaks | 20:34 |
| mriedem | hallelujah i finally created a vm | 20:35 |
| mriedem | yes this is master, devstack created today | 20:35 |
| *** dpawlik has quit IRC | 20:35 | |
| *** bbowen has joined #openstack-nova | 20:36 | |
| *** jmlowe has quit IRC | 20:36 | |
| *** kashyap has quit IRC | 20:37 | |
| *** jmlowe has joined #openstack-nova | 20:38 | |
| *** BjoernT_ has joined #openstack-nova | 20:39 | |
| *** BjoernT has quit IRC | 20:40 | |
| *** belmoreira has joined #openstack-nova | 20:44 | |
| aspiers | we're going to need a new version of that T-shirt: "OMG it worked on devstack" | 20:46 |
| *** eharney has joined #openstack-nova | 20:46 | |
| sean-k-mooney | haha it always works with devstack. devstack is life :) | 20:47 |
| sean-k-mooney | just stop running it in production | 20:47 |
| *** hamzy has joined #openstack-nova | 20:47 | |
| *** igordc has quit IRC | 20:50 | |
| sean-k-mooney | well it stacked and i seam to be able to boot vms ok. ill leave my setup on python2 for a week and see if i hit any issues | 20:54 |
| *** jmlowe has quit IRC | 20:56 | |
| sean-k-mooney | mriedem: by the way are you using nested virt. i am but maybe that is related | 21:02 |
| sean-k-mooney | but the first level vm where devstack is running and the second level vm booted by nova are using kvm | 21:03 |
| mriedem | yes nested virt with qemu | 21:03 |
| mriedem | oh wait, | 21:04 |
| mriedem | i assumed qemu, but looks like nova-cpu.conf is using kvm | 21:04 |
| mriedem | that's likely the reason for the weird failures and hangs with libvirtd | 21:04 |
| mriedem | wtf | 21:04 |
| sean-k-mooney | right well if you have nested vert enable i would expect ti to use kvm | 21:04 |
| mriedem | devstack should be detecting and using qemu i think in most cases - devstack-gate (and i think zuulv3 now) force to qemu | 21:04 |
| sean-k-mooney | they hardcode that | 21:05 |
| sean-k-mooney | devstack prefers kvm | 21:05 |
| sean-k-mooney | i think | 21:05 |
| mriedem | http://logs.openstack.org/40/673340/1/check/tempest-full/f05f8ae/controller/logs/etc/nova/nova-cpu_conf.txt.gz | 21:05 |
| mriedem | virt_type = qemu | 21:05 |
| sean-k-mooney | i hardcode virt_type=kvm in my local.conf | 21:05 |
| mriedem | no i'm pretty sure we use qemu exclusively in the gate b/c of weird shit failing randomly with nested virt and kvm | 21:05 |
| sean-k-mooney | yes | 21:06 |
| sean-k-mooney | nested vert is very kenel specfic | 21:06 |
| *** bbowen has quit IRC | 21:06 | |
| sean-k-mooney | is this still on the vexhost cloud | 21:08 |
| sean-k-mooney | they use centos7 host i think and they have been quite stable with nested virt from what i have heard but i know the default ubunut 18.04 kernel has some issues | 21:09 |
| *** BjoernT has joined #openstack-nova | 21:10 | |
| sean-k-mooney | kernel 4.19+ has the nested virt fixes | 21:10 |
| mriedem | yes this is on vexxhost | 21:10 |
| mriedem | 18.04 image | 21:10 |
| mriedem | $ uname -r -v | 21:12 |
| mriedem | 4.15.0-20-generic #21-Ubuntu SMP Tue Apr 24 06:16:15 UTC 2018 | 21:12 |
| mriedem | anyway, i haven't had issues before with these images, but i'm guessing they were properly detecting and configuring nova to use qemu rather than kvm | 21:12 |
| sean-k-mooney | im using 4.18.0-25-generic | 21:12 |
| mriedem | so it might be something in the devstack logic with 18.04 | 21:12 |
| *** BjoernT_ has quit IRC | 21:12 | |
| mriedem | i think the gate jobs are setup to just hard-code virt_type=qemu | 21:12 |
| mriedem | fungi: ^ do you know off hand? | 21:12 |
| mriedem | clarkb probably does | 21:12 |
| sean-k-mooney | mriedem: it does | 21:12 |
| sean-k-mooney | mriedem: https://github.com/openstack/devstack/blob/master/.zuul.yaml#L223 | 21:13 |
| mriedem | ah yup | 21:13 |
| mriedem | just found that too | 21:14 |
| sean-k-mooney | anyway i woudl advise doing an apt-get dist upgrade and grab a new kernel if you want to fix the nested virt issue | 21:14 |
| sean-k-mooney | oh im running the ubuntu Hardware Enablement Stack (HWE) kernel which is why mine is newer | 21:17 |
| *** gyee has quit IRC | 21:23 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: neutron: log something more useful in _get_instance_nw_info https://review.opendev.org/673694 | 21:24 |
| mriedem | it's not a contest sean gawd | 21:24 |
| mriedem | having all sorts of fun, | 21:28 |
| mriedem | like hanging to create a server, | 21:28 |
| mriedem | allocations are removed and task_state is reset to None on error, but there is no traceback | 21:28 |
| mriedem | Jul 30 21:15:11 devstack nova-compute[29645]: DEBUG nova.compute.manager [None req-a4247c77-5e4c-4e48-ba09-00f5d2903f23 admin admin] [instance: 33f64d9c-0a1a-410c-b5ba-e32ab501098d] Starting instance... {{(pid=29645) _do_build_and_run_instance /opt/stack/nova/nova/compute/manager.py:1955}} | 21:28 |
| mriedem | Jul 30 21:16:11 devstack nova-compute[29645]: INFO nova.compute.manager [None req-a4247c77-5e4c-4e48-ba09-00f5d2903f23 admin admin] [instance: 33f64d9c-0a1a-410c-b5ba-e32ab501098d] Successfully reverted task state from None on failure for instance. | 21:28 |
| mriedem | Jul 30 21:16:12 devstack nova-compute[29645]: INFO nova.scheduler.client.report [None req-a4247c77-5e4c-4e48-ba09-00f5d2903f23 admin admin] Deleted allocation for instance 33f64d9c-0a1a-410c-b5ba-e32ab501098d | 21:28 |
| mriedem | so we must be leaking an error somewhere and not logging it | 21:28 |
| mriedem | definitely hitting something like https://review.opendev.org/#/c/666857/ as a result | 21:30 |
| sean-k-mooney | ya maybe | 21:31 |
| *** belmoreira has quit IRC | 21:32 | |
| sean-k-mooney | but in all seriousnes i used ot have nested virt issue before i changed to the hwe kernel so if you keep having issue give it a try and see if it helps | 21:33 |
| sean-k-mooney | i have been hitting some gate isseu where vms just stay in building too | 21:33 |
| sean-k-mooney | so i suspec your are right about leaking another error like that one | 21:34 |
| mriedem | i really haven't had weird issues like this in forever, just started happening but i don't think the image i'm using has changed (my devstack startup script has the image id hard-coded) | 21:34 |
| sean-k-mooney | ya its strange | 21:36 |
| *** BjoernT has quit IRC | 21:36 | |
| mriedem | yeah, | 21:36 |
| mriedem | | d5474e0f-3a32-4036-a89f-8a57dfbd237e | Ubuntu 18.04 LTS (x86_64) [2018-05-09] | active | | 21:36 |
| mriedem | so that tells me it's devstack/nova that changed | 21:36 |
| mriedem | or dependent packages | 21:36 |
| *** belmoreira has joined #openstack-nova | 21:36 | |
| sean-k-mooney | well you said you are runnign python 2 and like sphinx other libs are starting to kill there python2 support | 21:37 |
| sean-k-mooney | so we might have an uncapped dep that is not happy on python 2 | 21:37 |
| mriedem | heh of course when i restart with more debug logging it works | 21:38 |
| mriedem | yeah i updated my local.conf to make sure the next one i use is py3 and uses qemu | 21:38 |
| sean-k-mooney | i need to actully check python3.7 agian | 21:39 |
| sean-k-mooney | i got simmilar issue to use using 3.7 back in january or febuary | 21:39 |
| sean-k-mooney | the nova compute agent would hang when spawning a vm | 21:40 |
| sean-k-mooney | even though the functional tests and unit tests ran fine | 21:40 |
| *** maciejjozefczyk has quit IRC | 21:41 | |
| *** gyee has joined #openstack-nova | 21:42 | |
| sean-k-mooney | anyway sorry i couldnt help narrow it down the next vm i create ill stack before i update the kernel and ill let you know if i run into the same issues | 21:44 |
| sean-k-mooney | i need to set a vm to fix a dpdk issue so ill proably do that later in the week | 21:45 |
| sean-k-mooney | night o/ | 21:45 |
| *** belmoreira has quit IRC | 21:53 | |
| *** slaweq has quit IRC | 21:53 | |
| *** boxiang_ has quit IRC | 21:54 | |
| *** boxiang_ has joined #openstack-nova | 21:54 | |
| efried | Why do we have user/flavors.rst and admin/flavors.rst? Does "user" encompass "admin" in this sense? | 22:01 |
| eandersson | sean-k-mooney, testing the patch out in the lab now, looks good so far | 22:04 |
| eandersson | (for rocky) | 22:04 |
| *** panda has quit IRC | 22:15 | |
| *** panda has joined #openstack-nova | 22:18 | |
| *** ivve has quit IRC | 22:23 | |
| *** eharney has quit IRC | 22:26 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: rt: remove misleading log in _remove_deleted_instances_allocations https://review.opendev.org/673706 | 22:27 |
| *** READ10 has joined #openstack-nova | 22:36 | |
| mriedem | efried: it's from the great docs migration in pike | 22:39 |
| mriedem | there are a few things like that, like with quotas | 22:39 |
| mriedem | i think one of the flavor docs say they should be combined - or maybe i'm thinking of quotas | 22:39 |
| mriedem | they are different audiences, like by default only admins can create flavors - there are other admin-only things in /users but i think over time that probably moves to /admin, not sure | 22:40 |
| mriedem | e.g. https://docs.openstack.org/nova/latest/user/#maintenance | 22:40 |
| efried | yeah, that's what I thought | 22:40 |
| efried | admins create flavors, so why would there be flavors docs in the user section? | 22:41 |
| efried | maybe donnyd has an opinion, and/or wants to be the one to make it so | 22:41 |
| mriedem | users need to know things about flavors like for vcpu/ram/disk, how flavors define swap and ephemeral disk, stuff like that | 22:41 |
| mriedem | getting into extra specs is a bit more of a stretch | 22:41 |
| donnyd | Yea makes no sense for users to learn about flavors other than they need to pick one | 22:42 |
| donnyd | so i agree efried | 22:42 |
| efried | I guess if they need to `show` a flavor and understand what they're looking at in order to pick one, maybe it makes sense for there to be a user doc at that level. Is that what you were talking about mriedem? | 22:42 |
| mriedem | something like that | 22:43 |
| donnyd | They should know what a flavor means in the user guide, like if you pick a flavor that translates to you get x / y / and z in your instance | 22:43 |
| efried | Okay. This went from "cut/paste user/flavors.rst into admin/flavors.rst" to "do something harder" so I'll punt, certainly at 5:45pm | 22:44 |
| mriedem | i think the overview in the user/flavors docs are ok, and the extra specs could maybe move to admin/flavors | 22:44 |
| mriedem | yeah this isn't really something i'm going to put a lot of brain power in right now | 22:45 |
| efried | I'm outta here. See y'all tomorrow. o/ | 22:45 |
| mriedem | users might care about extra specs though for advanced vms, sriov, stuff like that | 22:45 |
| mriedem | so idk | 22:45 |
| mriedem | o/ | 22:45 |
| mriedem | same here | 22:45 |
| *** mriedem has quit IRC | 22:45 | |
| openstackgerrit | Dustin Cowles proposed openstack/nova master: WIP: Provider config file https://review.opendev.org/673341 | 22:52 |
| *** READ10 has quit IRC | 22:56 | |
| *** tkajinam has joined #openstack-nova | 22:57 | |
| *** rcernin has joined #openstack-nova | 23:01 | |
| *** brault has joined #openstack-nova | 23:06 | |
| *** brault has quit IRC | 23:10 | |
| *** slaweq has joined #openstack-nova | 23:11 | |
| *** slaweq has quit IRC | 23:15 | |
| *** avolkov has quit IRC | 23:35 | |
| *** mlavalle has quit IRC | 23:40 | |
| *** betherly has quit IRC | 23:41 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!