| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2019-07-29 | |||
| 20:09:58 | efried | now we just need someone to send mriedem's test | |
| 20:28:29 | sean-k-mooney | dumb question but we have a bunch of tempest test that end in TestJSON what makes them different form normal tempest tests. are then calling the api with json payloads or somthing? | |
| 20:43:01 | sean-k-mooney | ... i is confused | |
| 20:43:32 | sean-k-mooney | the tempest-fully-py3 job supports depends-on against neutron patches right | |
| 20:44:07 | sean-k-mooney | because my patch failed with an error message i deleted form the neutron code base so it really looks like it didnt work | |
| 20:45:41 | sean-k-mooney | yat http://logs.openstack.org/32/602432/12/check/tempest-full-py3/39b553b/job-output.txt.gz#_2019-07-23_17_51_24_263888 it looks like it didnt check out my patch | |
| 20:46:28 | sean-k-mooney | yep it ran with master ... http://logs.openstack.org/32/602432/12/check/tempest-full-py3/39b553b/job-output.txt.gz#_2019-07-23_17_54_02_352110 | |
| 21:01:33 | openstackgerrit | sean mooney proposed openstack/nova master: libvirt: delegate ovs plug to os-vif https://review.opendev.org/602432 | |
| 21:16:38 | eandersson | btw is it just RamFilter that is being removed? Or is AggregateRamFilter being removed as well? | |
| 21:16:52 | eandersson | Because I see AggregateRamFilter is still a thing in Stein | |
| 21:17:46 | sean-k-mooney | both | |
| 21:18:13 | sean-k-mooney | and the core and disk filter both aggreagte and non aggrate forms | |
| 21:18:52 | sean-k-mooney | eandersson: all 6 filter have been depreacted since ocata | |
| 21:19:10 | sean-k-mooney | we removed the caching scudler i think in stien which was the last valid usecase for them | |
| 21:19:23 | eandersson | I see | |
| 21:20:19 | sean-k-mooney | we still actully have the filter on master but i dont think they work properly anymore | |
| 21:20:55 | eandersson | I did like how you could orchestrate aggregates | |
| 21:21:25 | sean-k-mooney | you can still do that in a sligtly different way via placmenet | |
| 21:21:47 | sean-k-mooney | but its perfromce before did nto scale well and there were a bunch of edgcases | |
| 21:22:04 | sean-k-mooney | it looks like https://github.com/openstack/nova/commit/78645e61c63bf042453d1f822ae8b3f1ee6a311b missed the aggrate versions | |
| 21:22:18 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: [WIP-until-series-is-ready] Introduce live_migration_claim() https://review.opendev.org/635669 | |
| 21:22:18 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: New objects for NUMA live migration https://review.opendev.org/634827 | |
| 21:22:19 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: LM: add support for augmenting migrate_data with info from claims https://review.opendev.org/634828 | |
| 21:22:20 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: LM: add support for updating NUMA-related XML on the source https://review.opendev.org/635229 | |
| 21:22:20 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: RPC changes to prepare for NUMA live migration https://review.opendev.org/634605 | |
| 21:22:20 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: NUMA live migration support https://review.opendev.org/634606 | |
| 21:22:21 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: Deprecate CONF.workarounds.enable_numa_live_migration https://review.opendev.org/640021 | |
| 21:22:21 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: [WIP] Functional test for NUMA live migration https://review.opendev.org/672595 | |
| 21:23:10 | artom | Probably broke a whole bunch of unit tests with that, will let CI tell me what's up. | |
| 21:25:26 | artom | But the problems that were left at the end of Stein have been fixed | |
| 21:32:46 | sean-k-mooney | the main one was saving the updated xml right | |
| 21:33:17 | sean-k-mooney | when i tested it at the end of stien it more or less worked bar the quirk | |
| 21:34:47 | artom | The two problems were: 1. not saving the new instance_numa_topology 2. If the resource audit ran during the LM, it wouldn't pick up the migrating instance, and would remove its resources from usage | |
| 21:35:29 | sean-k-mooney | oh ok and they are fixed in the patches you just pushed | |
| 21:36:18 | sean-k-mooney | im going to redeploy my 2 node ovn setup to something more more useful | |
| 21:36:23 | artom | Yeah. But like I said, I probably broke other things, just by changing some method names and removing that PinMapping object | |
| 21:36:33 | sean-k-mooney | should i pull down your latest patches and give them a try? | |
| 21:36:45 | artom | Not yet :) | |
| 21:36:56 | artom | I'll email the ML when it's ready (should be this week) | |
| 21:37:17 | artom | I know Windriver (and Cisco too, apparently?) would like to test it out | |
| 21:37:19 | sean-k-mooney | ok i might try and test numa + sriov next week | |
| 21:37:51 | sean-k-mooney | i havent powered up my physical server since the sriov stuff merged but it would be nice to test both together | |
| 21:39:20 | sean-k-mooney | ideally if we can get your numa stuff merged soon ish i can then test stephens cpu pinning change too | |
| 21:41:17 | artom | His thing was surprisingly quick | |
| 21:41:25 | artom | Given the upgrade impact, I was expecting it to take longer | |
| 21:41:56 | sean-k-mooney | the PCUs in placement changes | |
| 21:42:13 | artom | Yeah | |
| 21:42:44 | sean-k-mooney | well the fact we dont have numa this cycle signifcatly redues the upgrade impact | |
| 21:42:59 | sean-k-mooney | but i als dont know if the reshap is fully working yet | |
| 21:43:20 | sean-k-mooney | i think it is but there was still one edgecase that stephen was looking into | |
| 21:43:55 | sean-k-mooney | not with the reshap but with the schduler handeling old instance or host topologies | |
| 21:51:12 | mriedem | sean-k-mooney: eandersson: the Aggregate filters are not deprecated | |
| 21:51:35 | mriedem | otherwise the docs would say so https://docs.openstack.org/nova/latest/admin/configuration/schedulers.html#compute-filters | |
| 21:52:15 | sean-k-mooney | mriedem: the aggreate filters dont work anymore because the limmits set in the aggreate metadata are ignored by placmenet | |
| 21:52:37 | sean-k-mooney | so you have to set the limits on all the compute nodes or viat the placement api instead | |
| 21:52:52 | mriedem | i realize, i wrote these docs https://docs.openstack.org/nova/latest/admin/configuration/schedulers.html#allocation-ratios | |
| 21:52:57 | eandersson | They kinda work | |
| 21:53:09 | eandersson | I mean they work for us because computes are misconfigured | |
| 21:53:45 | mriedem | i don't know why we didn't deprecate the aggregate filters when the others were deprecated, probably different timelines | |
| 21:54:46 | sean-k-mooney | when we found that bug and melwitt sent http://lists.openstack.org/pipermail/openstack-dev/2018-January/126283.html i thought that was use signeling the depreacation | |
| 21:54:58 | mriedem | sean-k-mooney: tempest TestJson is a remnant from when we had xml api testing | |
| 21:55:02 | sean-k-mooney | we also have that gian warning https://docs.openstack.org/nova/latest/admin/configuration/schedulers.html#scheduling-considerations | |
| 21:55:10 | sean-k-mooney | mriedem: ah ok | |
| 21:55:29 | mriedem | sean-k-mooney: well i guess no one bothered to follow up and actually deprecate those filters | |
| 21:55:30 | sean-k-mooney | i rembere there was an xml api and i never used it | |
| 21:55:45 | mriedem | i deprecated the core/ram/disk filters since those were clearly no longer useful outside of the caching scheduler | |
| 21:55:53 | sean-k-mooney | mriedem: ok am can we do it in train and kill them in U then? | |
| 21:56:14 | mriedem | i would ask in the ML b/c that is a bunch of context i don't have loaded into my head right now | |
| 21:56:30 | sean-k-mooney | right and as you noted ironic stoped reporting the cpus ram and disk so they activly were broken with ironic as of stien | |
| 21:56:47 | mriedem | noted in my commit message you mean | |
| 21:56:51 | sean-k-mooney | yes | |
| 21:57:16 | mriedem | eandersson: also fyi https://review.opendev.org/#/c/640898/ - an osc command to update allocation ratios on provider aggregates | |
| 21:57:39 | mriedem | it's been hung up in how to actually write the command using osc conventions | |
| 22:01:48 | sean-k-mooney | speaking of osc my patch for adding evacuate to osc https://review.opendev.org/#/c/643578/ | |
| 22:02:21 | sean-k-mooney | i havent looked at it in a while but ill try to get back to it soon ish | |
| 22:02:24 | mriedem | there are unanswered comments in there | |
| 22:02:42 | mriedem | mainly that dansmith wanted to continue calling it evacuate to avoid different names all over the place | |
| 22:02:43 | sean-k-mooney | i think ill mail the list and ask for feedback on the way forward vs nameing | |
| 22:02:51 | sean-k-mooney | ya | |
| 22:03:21 | sean-k-mooney | i am ok to do that i did care more earlier but i would prefer to land a command rather then none | |
| 22:03:59 | sean-k-mooney | if people want too keep using evacuate then i guest that is fine. | |
| 22:04:22 | sean-k-mooney | but ill email the list tomorrow. | |
| 22:04:43 | sean-k-mooney | if i rememebr ill email the list about deprecating those filter too | |
| 22:06:36 | sean-k-mooney | hum we are still sometimes getting "Could not open '/opt/stack/data/nova/instances/69509dd0-a86f-4efe-997c-986a076e4a5d/disk': No such file or directory\n"" issue in gate jobs | |
| 22:20:15 | sean-k-mooney | mdbooth: is http://logs.openstack.org/58/640258/10/check/neutron-tempest-dvr/922e7e9/controller/logs/screen-n-cpu.txt.gz#_Jul_29_21_20_26_895955 the same as the downstream bug https://bugzilla.redhat.com/show_bug.cgi?id=1445391 i noticed that we fail to resize the image way earlier | |
| 22:20:15 | openstack | sean-k-mooney: Error: Error getting bugzilla.redhat.com bug #1445391: NotPermitted | |
| 22:20:16 | sean-k-mooney | http://logs.openstack.org/58/640258/10/check/neutron-tempest-dvr/922e7e9/controller/logs/screen-n-cpu.txt.gz#_Jul_29_21_20_05_626653 which makes me think it might be related. | |
| 22:21:33 | sean-k-mooney | that bug is private but the upstream version is https://bugs.launchpad.net/nova/+bug/1686703 | |
| 22:21:33 | openstack | Launchpad bug 1686703 in OpenStack Compute (nova) "Error in finish_migration results in image deletion on source with no copy" [Undecided,Won't fix] | |
| 22:25:29 | sean-k-mooney | i think that ^ can happen for just normal spawns too rather then requrieing resizes/migrations as it implies | |
| 22:26:32 | sean-k-mooney | at least if the upstream gate failure is the same thing it was not a move operations that triggered it so i think it can happen anytime a resize of the image fails | |
| 22:26:43 | sean-k-mooney | anyway night all o/ | |
| 22:27:24 | efried | Having gone on a spec abandoning rampage, /me calls it quits. o/ | |
| 22:27:40 | sean-k-mooney | ya i saw hehe o/ | |
| #openstack-nova - 2019-07-31 | |||
| 02:00:15 | openstackgerrit | Merged openstack/nova master: Add functional recreate test for bug 1764556 https://review.opendev.org/562041 | |
| 02:00:15 | 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) | |
| 05:36:30 | openstackgerrit | Sundar Nadathur proposed openstack/nova master: ksa auth conf and client for Cyborg access https://review.opendev.org/631242 | |
| 05:36:31 | openstackgerrit | Sundar Nadathur proposed openstack/nova master: WIP: Add Cyborg device profile groups to request spec. https://review.opendev.org/631243 | |
| 05:36:31 | openstackgerrit | Sundar Nadathur proposed openstack/nova master: WIP: Create and bind Cyborg ARQs. https://review.opendev.org/631244 | |
| 05:36:32 | 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 | |