Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2019-08-15
00:47:25 melwitt ok. sorry for all the questions, this is confusing to me
00:47:45 efried the right solution, long-term, and mordred is working on this, is for openstacksdk to publish fixtures that we can just import and use. Because openstacksdk will always know what it's doing internally and how.
00:48:39 efried This - like the previous iteration of the fixture that mocked Adapter.get_major_api_version - is just a stopgap to let us move forward until that happens.
00:49:20 efried and it's only affecting unit/func tests. The real proof that sdk is working for us is in the fact that all the tempest tests - which all use placement, which uses sdk - pass.
00:49:41 melwitt yeah, I get that part, I just didn't want to fast approve something I had zero understanding of. I'll look at it some more and if I can grok it, I'll send it on
00:50:06 melwitt thanks for explaining all that
00:50:32 efried Thanks for looking.
00:50:32 efried I think alex_xu might be arriving on the scene soon, so if you prefer not to +W it, he might be able to do so while we sleep.
00:52:50 melwitt it might just take me a bit longer to get this
00:53:13 efried It's not the end of the world if it has to wait for morning.
00:53:55 efried It would just be nice to be able to recheck the u-c patch and have it merged by morning so I can move on with the stuff that builds on it.
00:54:48 melwitt yeah, I understand. I'll keep looking at it, just need a bit more time
00:55:00 efried thanks melwitt
01:43:18 openstackgerrit Dongcan Ye proposed openstack/nova master: Libvirt: Rename _guest_add_memory_balloon https://review.opendev.org/676534
01:45:56 openstackgerrit Takashi NATSUME proposed openstack/nova master: Fix 'has_calls' method calls in unit tests https://review.opendev.org/676536
01:53:16 openstackgerrit Mohammed Naser proposed openstack/nova master: cleanup: remove osapi_v21.project_id_regex https://review.opendev.org/676537
01:58:38 melwitt efried: +W (finally)
02:02:21 openstackgerrit Takashi NATSUME proposed openstack/nova master: Fix 'has_calls' method calls in unit tests https://review.opendev.org/676536
02:06:35 yonglihe melwitt: Hi, I hope the comments address your concern, very nice to have you review that, thank. https://review.opendev.org/#/c/627765/
02:14:01 openstackgerrit Mohammed Naser proposed openstack/nova master: config: remove deprecated checksum options https://review.opendev.org/676539
02:29:54 openstackgerrit Mohammed Naser proposed openstack/nova master: Remove nova.image.download.modules extension point https://review.opendev.org/676540
02:32:39 melwitt yonglihe: I'll try to take a look tomorrow, sorry for the delay
03:13:50 openstackgerrit CHENGLEI proposed openstack/nova stable/pike: Insert the description of the change. https://review.opendev.org/676541
05:17:36 openstackgerrit ZhengMa proposed openstack/nova master: Check container_format on instance creation https://review.opendev.org/673407
07:52:34 gibi sean-k-mooney: thanks for the devstack conf with the mixed sriov ovs setup
08:11:40 gibi sean-k-mooney: did you set up connectivity between br-ex and enp1s0f1 on the physical host?
08:24:41 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Pass RequestContext to oslo_policy https://review.opendev.org/674038
08:28:30 openstackgerrit ZhengMa proposed openstack/nova master: Check container_format on instance creation https://review.opendev.org/673407
08:28:35 sean-k-mooney gibi: that config does it automatically
08:28:55 sean-k-mooney it add enp1sof1 to the br-ex
08:29:01 sean-k-mooney *adds
08:30:00 sean-k-mooney i also put the echo to allocate the vf in the local.conf and set the pci whitelist based on the vendor id and procut id of the vf so it automaticl allows the correect vfs
08:31:14 sean-k-mooney when stacking compltes you are basiaclly ready to boot a vm with an sriov port and it will get a dhcp adress for the dhcp agent attached to ovs
08:32:18 sean-k-mooney i say basiacally ready as cirros may not have your nic driver so you might need to upload a new image first
08:45:51 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Add test coverage of existing os-services policies https://review.opendev.org/669181
08:49:28 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Add test coverage of existing os-services policies https://review.opendev.org/669181
08:50:27 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Introduce scope_types in os-services https://review.opendev.org/645427
08:51:06 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Add new default roles in os-services API policies https://review.opendev.org/648480
08:54:37 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Add test coverage of existing admin_actions policies https://review.opendev.org/657698
08:54:55 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Ensure we pass a target in admin actions https://review.opendev.org/663095
09:03:54 gibi sean-k-mooney: thanks
09:14:06 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Add test coverage of existing os-agents policies https://review.opendev.org/669578
09:19:28 gibi gmann: hi! If you have time, could you give your oppinion in https://review.opendev.org/#/c/675371/ ?
09:23:10 openstackgerrit Merged openstack/nova master: Enhance SDK fixture for 0.34.0 https://review.opendev.org/676495
09:39:14 gmann gibi: sure, i will check.
09:42:24 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Add new default roles and mapping in policy base class https://review.opendev.org/645452
09:42:57 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Add new default roles in os-services API policies https://review.opendev.org/648480
09:54:17 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Add test coverage of existing admin_actions policies https://review.opendev.org/657698
09:54:38 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Ensure we pass a target in admin actions https://review.opendev.org/663095
10:14:45 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Introduce scope_types in Admin Actions https://review.opendev.org/657823
10:19:49 openstackgerrit Guo Jingyu proposed openstack/nova master: define and catch InstanceGuestAgentNotRunning for set-password failure https://review.opendev.org/676667
10:39:11 openstackgerrit Guo Jingyu proposed openstack/nova master: define and catch InstanceAgentNotRunning for set-password failure https://review.opendev.org/676667
10:42:40 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Add policy deprecation fixture and Suppress warnings in tests https://review.opendev.org/676670
10:44:15 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Add new default roles in os-services API policies https://review.opendev.org/648480
10:44:27 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Add new default roles in os-services API policies https://review.opendev.org/648480
10:48:42 openstackgerrit Ghanshyam Mann proposed openstack/nova master: Add new default roles in os-services API policies https://review.opendev.org/648480
10:53:45 openstackgerrit Chris Dent proposed openstack/nova master: Add a "Caveats" section to the eventlet profiling docs https://review.opendev.org/676672
10:56:39 sean-k-mooney what time does matt normally log on? about 1-2 hours form now?
10:57:22 stephenfin sean-k-mooney: We have a conflict with cpu-resources
10:57:23 sean-k-mooney i started early today so it feels weird to have been online for a while and not see the us folks about
10:57:34 sean-k-mooney we do?
10:57:44 sean-k-mooney with what
10:58:17 stephenfin There's a workaround option to disable VCPU -> PCPU flavor translation and we've decided to report both PCPUs and VCPUs for compute nodes without explicit CPU pinset configuration
10:58:33 sean-k-mooney yes
10:58:42 sean-k-mooney i was talking about this yesterday
10:58:47 stephenfin If I reshape, I move every allocation for pinned instances on that host from VCPUs to PCPUs
10:58:59 sean-k-mooney right
10:59:19 sean-k-mooney which is why i raised that workaround option as an issue
10:59:23 stephenfin But if you enable the workaround option (so your pinned instances continue to request VCPUs), more pinned instances will land on that host that use VCPUs instead of PCPUs
10:59:39 sean-k-mooney yes
10:59:48 stephenfin so those two things can't coexist
11:00:05 sean-k-mooney yes you cant reshap until you disabel the workaround
11:00:27 stephenfin yup, and the workaround needs to be set on every single host, and toggled on every host at the same time
11:00:41 sean-k-mooney well currently its only set on the schduler
11:00:44 stephenfin compute nodes and controllers (well, scheduler hosts)
11:00:47 sean-k-mooney but it would have to be if you were to use
11:00:54 sean-k-mooney it to contol the reshape
11:01:09 stephenfin I'm tempted to drop the config option, to be honest
11:01:26 sean-k-mooney well the translation is on by default
11:01:43 stephenfin Yup. I'd rather we didn't support turning it off again
11:01:44 sean-k-mooney so im ok with saying that it should nver be set on a new host
11:02:15 stephenfin You just need to make sure you upgrade sufficient hosts at once or you'll see degraded scheduling performance due to lack of viable hosts
11:02:27 stephenfin *a sufficient number of compute nodes at once
11:02:29 sean-k-mooney e.g. you should only use it if and only if you want to first upgrade to train then when all node are on train enable this feature
11:03:17 stephenfin so you'd rather keep the config option and make it global?
11:03:21 sean-k-mooney we could remove it but it there for host with vms that cant be migrated
11:03:40 stephenfin why not? The hosts just have to be updated
11:03:44 sean-k-mooney it cant really be global unless you mean set in ever host
11:03:58 stephenfin I do. I think that's the only way this could work
11:04:06 sean-k-mooney ya
11:04:17 sean-k-mooney so if we keep it i think it needs to be set on all host
11:04:24 sean-k-mooney and your reshap code needs to check it
11:04:32 stephenfin It would need to toggle (a) flavor translation, (b) whether we report PCPUs (since that's a heuristic we rely on to detect if reshapes are needed) and (c) whether we reshape
11:04:44 stephenfin Is it worth the effort?
11:04:48 sean-k-mooney and we need to document that it must be set too the same value on all hosts
11:04:53 stephenfin Yup
11:05:10 sean-k-mooney can you support inplacem upgrade without live migration without it
11:05:23 stephenfin and turned off again on the controller first
11:06:30 stephenfin otherwise the compute nodes will have reshaped everything but the scheduler will still be making requests for VCPUs
11:06:51 sean-k-mooney the workflow you be. set workatound on all hosts-> upgrade->disbale workaound on all hosts. and yes it would have to toggel all those things

Earlier   Later