| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-12-21 | |||
| 10:52:54 | openstackgerrit | Stephen Finucane proposed openstack/nova stable/queens: Fix destination_type attribute in the bdm_v2 documentation https://review.openstack.org/626875 | |
| 10:53:05 | openstackgerrit | Stephen Finucane proposed openstack/nova stable/pike: Fix destination_type attribute in the bdm_v2 documentation https://review.openstack.org/626876 | |
| 11:52:07 | openstackgerrit | Merged openstack/nova master: Add irrelevant-files for grenade-py3 jobs https://review.openstack.org/624997 | |
| 11:52:13 | openstackgerrit | Merged openstack/nova master: nit: Add space to feature support docs https://review.openstack.org/626523 | |
| 11:52:25 | openstackgerrit | Merged openstack/nova master: Fix destination_type attribute in the bdm_v2 documentation https://review.openstack.org/625020 | |
| 12:03:39 | ShilpaSD | stephenfin: stack is failing with error '[ERROR] /home/<NAME>/devstack/lib/placement:179 placement-api did not start', any solution? | |
| 12:33:54 | mdbooth | bauzas: Hey, I just spotted something weird about weighers (in Queens), and I'm wondering if it's known and/or deliberate: minval and maxval for a weigher seem to be global for all requests and hosts weighed by that weigher | |
| 12:35:32 | mdbooth | bauzas: If I haven't misread that, it would mean, for eg, that if any host ever suffered a long succession of build failures, then the BuildFailureWeigher of all subsequent scheduling requests would be normalized against that very negative value. | |
| 12:36:43 | mdbooth | So if 1 host suffers 100 build failures one time, then forever more a single build failure is normalized to 0.99 rather than 0 | |
| 12:37:34 | mdbooth | That seems weird. Not sure what the side-effects would be. | |
| 12:37:51 | mdbooth | Also: it would be really nice if weight values were logged :/ | |
| 13:07:13 | mdbooth | I'm convinced there's something weird going on with our request-id generation | |
| 13:07:20 | mdbooth | I think we're re-using them | |
| 13:48:19 | openstackgerrit | Corey Bryant proposed openstack/nova master: Ensure rbd auth fallback uses matching credentials https://review.openstack.org/626897 | |
| 13:51:16 | stephenfin | ShilpaSD: Are you restacking on a machine where you previously had stacked? | |
| 13:51:38 | ShilpaSD | stephenfin: yes | |
| 13:52:57 | stephenfin | ShilpaSD: Remove anything related to OpenStack in /etc/apache2/sites-(available|enabled)/ and try again | |
| 13:53:21 | stephenfin | ShilpaSD: They changed how placement is deployed so that's all you really need to remove, but I think I saw issues with other services | |
| 13:53:47 | stephenfin | ShilpaSD: If that doesn't work, the easiest thing to do is reinstall the OS. DevStack doesn't properly clean up after itself so upgrades aren't really a thing | |
| 13:54:07 | stephenfin | which it is, if you're restacking with newer code | |
| 13:54:49 | ShilpaSD | stephenfin: ok, the last solution i started, once this done, will check cleanup of /etc/apache2/sites and get back to you | |
| 13:55:06 | ShilpaSD | stephenfin: thanks for solution | |
| 14:28:18 | ShilpaSD | stephenfin: great, stack is successfully created | |
| 14:28:24 | ShilpaSD | thank you | |
| 15:21:19 | openstackgerrit | Stephen Finucane proposed openstack/nova master: doc: Add solution to live migration ssh issues https://review.openstack.org/626931 | |
| 15:21:19 | openstackgerrit | Stephen Finucane proposed openstack/nova master: doc: Switch header styles in support doc https://review.openstack.org/626930 | |
| 15:21:20 | openstackgerrit | Stephen Finucane proposed openstack/nova master: conf: Remove 'disable_libvirt_livesnapshot' option https://review.openstack.org/626932 | |
| 15:54:39 | mriedem | melwitt: huzzah https://review.openstack.org/#/c/509042/ | |
| 16:00:15 | mriedem | comments on the counting quotas from placement spec - mostly just looking to flesh out some details, and optimize the data migration | |
| 16:13:58 | stephenfin | Whoops | |
| 16:21:16 | coreycb | mriedem: does this look ok to you? really small fix. https://review.openstack.org/#/c/626897 | |
| 16:32:24 | mriedem | coreycb: comments inline | |
| 16:34:07 | mriedem | looks like my 'remove all this shit in pike' didn't happen | |
| 16:34:38 | mriedem | heh or queens | |
| 16:49:04 | finucannot | cdent isn't around but I've now answered ShilpaD's "why did placement-api not start when stacking question" for another person. That's three (including me). DevStack should probably just delete that old Apache file | |
| 16:49:08 | finucannot | maybe after Xmas | |
| 17:04:09 | coreycb | mriedem: thanks. i responded but wasn't fully understanding the one comment. | |
| 17:04:18 | coreycb | mriedem: haha yeah | |
| 17:14:32 | mriedem | coreycb: ok i'll check after lunch | |
| 17:14:52 | coreycb | thanks mriedem_afk | |
| 17:16:34 | openstackgerrit | Lee Yarwood proposed openstack/nova master: DNM WIP Workaround known eventlet monkey_patch issue https://review.openstack.org/626952 | |
| 17:59:21 | openstackgerrit | Jack Ding proposed openstack/nova master: Preserve UEFI NVRAM variable store https://review.openstack.org/621646 | |
| 19:03:28 | openstackgerrit | Merged openstack/nova master: Fix best_match() deprecation warning https://review.openstack.org/611204 | |
| 19:09:11 | openstackgerrit | Sylvain Bauza proposed openstack/nova-specs master: Proposes NUMA topology with RPs https://review.openstack.org/552924 | |
| 19:49:44 | mriedem | coreycb: replies inline | |
| 19:49:52 | mriedem | i think i understand the bug now but it could have been better explained | |
| 20:11:40 | lyarwood | mriedem: https://review.openstack.org/#/c/618478/ - Would you mind taking a look at this if you have any bandwidth before calling it for the year? | |
| 20:12:33 | mriedem | ... | |
| 20:12:35 | lyarwood | FWIW I'm going to be around next week and am planning to work on this (if required) and my stable backlog. | |
| 20:12:58 | lyarwood | mriedem: tbh even a NACK at this point would be useful | |
| 20:13:21 | mriedem | i can't really nack it as i don't have better alternatives | |
| 20:13:44 | mriedem | kinda would like dan to look at that approach as well, but he's out until january | |
| 20:14:00 | lyarwood | kk, this can wait until the new year then | |
| 20:14:27 | mriedem | ? | |
| 20:14:27 | mriedem | is there a long-term solution in mind for that/ | |
| 20:15:09 | lyarwood | mdbooth has a suggestion in https://review.openstack.org/#/c/622300/ that I also might work on over the break as I'm out of PTO for the year | |
| 20:15:35 | mriedem | i would feel better about putting in a workaround if we had longer term plans for a solution | |
| 20:16:09 | mriedem | and the root of the issue is the root disk is on shared ceph but the instance files aren't? | |
| 20:16:11 | mriedem | or vice-versa? | |
| 20:16:34 | mriedem | why oh why can't people just use ceph everywhere if they are going to use it at all | |
| 20:16:36 | lyarwood | that's correct, the disks are in rbd and shared, the instance directory is not. | |
| 20:16:54 | lyarwood | well operators don't have an option here, that's our fault | |
| 20:17:04 | mriedem | :/ | |
| 20:17:14 | lyarwood | we leave the console.log, kernel, ramdisks etc in the instance dir | |
| 20:17:22 | mriedem | oh by our you mean nova | |
| 20:17:26 | mriedem | i thought you meant tripleo :) | |
| 20:17:29 | lyarwood | ha | |
| 20:17:38 | lyarwood | it's at fault for alot of things but not this | |
| 20:17:56 | mriedem | but the instance dir could be on shared storage itself | |
| 20:18:01 | mriedem | ceph, nfs, etc | |
| 20:18:09 | lyarwood | correct and nothing checks that at the moment | |
| 20:18:10 | mriedem | mfing gluster | |
| 20:19:33 | mriedem | then why is this solution specific to rbd? just because of the Rbd image backend returning True from "is_shared_block_storage"? | |
| 20:20:05 | mriedem | iow, check_instance_shared_storage_local returns true if the image backend assumes that because disks are on shared storage, everything is on shared storage | |
| 20:20:11 | mriedem | and the only one that does that is the rbd imagebackend | |
| 20:20:34 | mriedem | is that a correct summary? | |
| 20:21:06 | lyarwood | yes correct | |
| 20:21:30 | mriedem | ok, i'll look it over again this afternoon once i wrap up what i'm doing | |
| 20:22:04 | lyarwood | ack thanks, either way as I said I'll start looking at the longer term solution over the break | |
| 20:56:56 | openstackgerrit | Corey Bryant proposed openstack/nova master: Ensure rbd auth fallback uses matching credentials https://review.openstack.org/626897 | |
| 20:58:24 | coreycb | mriedem: thanks again for the input. i think the latest version is on the right track. | |
| 21:03:36 | mriedem | artom: what's going on with the numa live migration blueprint? has anything been done code-wise since the spec was re-approved? i know it's the holidays and all that, but just wondering if that is going to take off in january or... | |
| 21:18:58 | openstackgerrit | Merged openstack/nova-specs master: Change openstack-dev to openstack-discuss https://review.openstack.org/621846 | |
| 21:55:41 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Ensure rbd auth fallback uses matching credentials https://review.openstack.org/626897 | |
| 21:58:27 | mriedem | coreycb: done | |
| 21:59:15 | coreycb | mriedem: awesome thanks a lot | |
| 22:04:02 | coreycb | mriedem: and thanks for updating the things that I said I'd do but didn't do. :) | |
| 22:05:35 | mriedem | my gift to you | |
| 22:08:56 | openstackgerrit | Corey Bryant proposed openstack/nova stable/rocky: Ensure rbd auth fallback uses matching credentials https://review.openstack.org/627009 | |
| 22:09:10 | openstackgerrit | Corey Bryant proposed openstack/nova stable/queens: Ensure rbd auth fallback uses matching credentials https://review.openstack.org/627010 | |
| 22:09:28 | openstackgerrit | Corey Bryant proposed openstack/nova stable/pike: Ensure rbd auth fallback uses matching credentials https://review.openstack.org/627011 | |
| 22:09:42 | openstackgerrit | Corey Bryant proposed openstack/nova stable/ocata: Ensure rbd auth fallback uses matching credentials https://review.openstack.org/627012 | |
| 22:09:56 | openstackgerrit | Matt Riedemann proposed openstack/nova master: libvirt: generalize rbd volume fallback removal statement https://review.openstack.org/627013 | |
| 22:10:10 | mriedem | coreycb: geez you must not expect any other cores to have a problem with that patch | |
| 22:10:14 | mriedem | jumping the gun a bit | |
| 22:10:35 | coreycb | mriedem: oh sorry i can cancel those, sorry | |
| 22:10:35 | mriedem | can you -W those changes while the master branch patch isn't merged yet? | |
| 22:10:41 | mriedem | no need, ust WIP them | |
| 22:11:02 | coreycb | mriedem: need 2 +2 for nova i assume | |
| 22:11:07 | mriedem | heh yes | |