| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-07-05 | |||
| 11:44:01 | sean-k-mooney | stephenfin: should i be useing the change id? | |
| 11:45:32 | sean-k-mooney | stephenfin: the yes i should never mind... | |
| 11:47:52 | mdbooth | sfinucan: I had another look at that errors_out_migration patch. I started by writing a better commit message, and as often happens writing docs leads to better understanding :) I've made a couple of changes to affected functions, but hopefully they're now better described and justified. | |
| 11:48:50 | mdbooth | I was really tempted, btw, to change things around in _finish_resize, because it's weird that we do networky things before sending the start notification. | |
| 11:49:25 | mdbooth | I resisted the urge, though, which has resulted in a slightly untidy return value from that function. | |
| 11:51:53 | efried | jaypipes Why do you prefer set(1, 2, 3) over the {1, 2, 3} notation? | |
| 11:54:00 | stephenfin | mdbooth: Sounds good. I'll check it out after lunch :) | |
| 11:54:32 | stephenfin | Also, I'm using stephenfin instead of sfinucan now. You should have got an away message, but it seems some clients don't show them :( | |
| 11:54:46 | stephenfin | sean-k-mooney: So you're good? | |
| 11:54:58 | sean-k-mooney | stephenfin: i think so one question though | |
| 11:55:06 | stephenfin | sean-k-mooney: it's 'git review -d 466132' just in case | |
| 11:55:08 | stephenfin | hit me | |
| 11:55:24 | sean-k-mooney | stephenfin: when submiting if i use -f will it restore the original topic? | |
| 11:55:57 | sean-k-mooney | stephenfin: it checked it out as review/sean_mooney/kolla-build-external-dockerfiles but normally the branch would be just kolla-build-external-dockerfiles | |
| 11:56:25 | stephenfin | sean-k-mooney: It does some odd stuff. If any of the patches have 'blueprint' mentioned in them, the topic will be set to 'bp/{title}'. Ditto for bug ('bug/{bugID}') | |
| 11:57:02 | stephenfin | sean-k-mooney: In general though, if your branch is called 'review/{username}/{topic}', only the '{topic}' piece is kept (i.e. nothing changes) | |
| 11:57:19 | sean-k-mooney | hum ok well i can change it in the gerrit ui if id does the wronge thing might as well just see what it does. | |
| 11:57:38 | stephenfin | efried: I would guess it's because the set notation is not that common | |
| 11:58:15 | sean-k-mooney | stephenfin: i usually checkout everything manually but i think it about time i start using git review for that too | |
| 11:58:27 | efried | stephenfin I ask because I prefer the latter - and in fact my IDE complains if you use set() in certain circumstances where {...} will do. | |
| 11:58:27 | stephenfin | I would say it's very rare in the nova code base, while the 'set(1, 2, 3)' notation is very common | |
| 11:58:56 | stephenfin | sean-k-mooney: I've used nothing but git-review for the past two years and it's rarely, if ever, let me down | |
| 11:58:57 | efried | stephenfin If it's a "nova codebase consistency" thing, okay. | |
| 11:59:23 | sean-k-mooney | stephenfin: yep git review -f work as i wanted thanks :) | |
| 11:59:25 | efried | sean-k-mooney stephenfin +1 git-review | |
| 11:59:44 | stephenfin | sean-k-mooney: two tips for you too. #1 - if you want to download a single patch with dependencies, use '-x' instead of '-d' | |
| 12:00:16 | stephenfin | (very helpful when someone has uploaded a new revision of a patch in the middle of a series, are you want to rebase the successors) | |
| 12:00:54 | stephenfin | and #2 - use '{change},{revision}' to download a previous revision | |
| 12:01:06 | stephenfin | e.g. 'git review -d 466132,1' | |
| 12:03:23 | openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient master: [ci] Use pseudo-random names for new resources https://review.openstack.org/479777 | |
| 12:03:44 | sean-k-mooney | efried: ya i use it for upload becase "git push gerrit HEAD:refs/change/466132" is annorying but i havent been useing the other features | |
| 12:03:57 | sean-k-mooney | stephenfin: cool that is good to know | |
| 12:15:58 | openstackgerrit | OpenStack Proposal Bot proposed openstack/nova master: Imported Translations from Zanata https://review.openstack.org/477091 | |
| 12:23:53 | openstackgerrit | Vladyslav Drok proposed openstack/nova master: Implement interface attach/detach in ironic virt driver https://review.openstack.org/419975 | |
| 12:26:34 | openstackgerrit | Vladyslav Drok proposed openstack/nova master: Implement interface attach/detach in ironic virt driver https://review.openstack.org/419975 | |
| 12:39:55 | jaypipes | efried: because {1, 2, 3} is too easy to mistake for a dict. | |
| 12:40:23 | efried | jaypipes It is? | |
| 12:40:35 | jaypipes | cdent: I would moo back but suffering from allergies badly today :( | |
| 12:41:00 | cdent | bummer dood | |
| 12:41:04 | jaypipes | efried: perhaps not for you. for others it is :) | |
| 12:41:38 | efried | jaypipes Okay, good to know. | |
| 12:42:50 | alex_xu | jaypipes: cdent two fixes for AllocationCandidates.get_by_filters https://review.openstack.org/480379 and https://review.openstack.org/480130 | |
| 12:43:05 | openstackgerrit | Gábor Antal proposed openstack/nova master: Transform instance.live_migration_force_complete notification https://review.openstack.org/480561 | |
| 12:43:35 | cdent | I've been pretty good at not paying attention for a while | |
| 12:43:59 | alex_xu | cdent: welcome back :) | |
| 12:44:07 | cdent | thanks alex_xu | |
| 12:44:31 | openstackgerrit | Gábor Antal proposed openstack/nova master: Transform instance.live_migration_force_complete notification https://review.openstack.org/480561 | |
| 12:45:06 | alex_xu | nova api meeting is in 15 mins at #openstack-meeting-4 | |
| 12:47:31 | openstackgerrit | Andrey Volkov proposed openstack/nova master: [placement] Add api-ref for allocations https://review.openstack.org/470933 | |
| 12:47:32 | openstackgerrit | Andrey Volkov proposed openstack/nova master: [placement] Add api-ref for RP allocations https://review.openstack.org/450550 | |
| 12:47:32 | openstackgerrit | Andrey Volkov proposed openstack/nova master: [placement] Add api-ref for RP usages https://review.openstack.org/450105 | |
| 12:47:33 | openstackgerrit | Andrey Volkov proposed openstack/nova master: [placement] Add api-ref for usages https://review.openstack.org/480563 | |
| 12:51:22 | openstackgerrit | Andrey Volkov proposed openstack/nova master: PoC: Select PCI devices with distinct tag values https://review.openstack.org/448008 | |
| 13:18:20 | openstackgerrit | Silvan Kaiser proposed openstack/nova master: [WIP] Add truncate for images if prealloc is true https://review.openstack.org/480464 | |
| 13:19:41 | cdent | avolkov: can you give me a quick summary on where things are with placement api-ref? Has there been any progress on publishing for real? | |
| 13:22:11 | openstackgerrit | Rodolfo Alonso Hernandez proposed openstack/nova master: Read Neutron port 'binding_profile' during boot https://review.openstack.org/449257 | |
| 13:23:50 | bauzas | jaypipes: I'd really love to see your insight in https://review.openstack.org/#/c/480379/2 | |
| 13:35:39 | cdent | you've infected me jaypipes, my allergies are spinning up | |
| 13:46:37 | openstack | Launchpad bug 1698010 in OpenStack Compute (nova) "neutron-based instances should not use the nova-network 'dhcp_domain' option" [High,Confirmed] | |
| 13:46:37 | stephenfin | sean-k-mooney: Any ideas on https://bugs.launchpad.net/nova/+bug/1698010 ? | |
| 13:46:41 | stephenfin | ralonsoh too ^ | |
| 13:48:02 | ralonsoh | stephenfin: let me take a look at this one | |
| 13:48:18 | mriedem | stephenfin: that is a duplicate, | |
| 13:48:21 | mriedem | just need to find the other bug | |
| 13:48:24 | stephenfin | I can get network info from the instance using 'Instance.get_network_info', but it just returns a list of 'os_vif.objects.vif.VIFBase' subclass instances | |
| 13:48:29 | stephenfin | mriedem: Oh hai | |
| 13:48:38 | mriedem | the dhcp_domain option is read in the metadata api code and that's how it gets into config drive | |
| 13:48:48 | mriedem | as noted by bnemec in the bug report it's showing up via cloud-init | |
| 13:49:08 | stephenfin | Yup :( I was trying to see where I could get the neutron equivalent from, but we don't seem to cache it anywhere | |
| 13:49:27 | stephenfin | and I don't want to call out to neutron for every instance (assuming that is even an option) | |
| 13:50:07 | mriedem | we already call to neutron for dns info when we create or attach to a port | |
| 13:51:05 | stephenfin | Hmm, so is that a viable path to pursue? | |
| 13:51:37 | openstack | Launchpad bug 1702454 in OpenStack Compute (nova) "Transforming the RequestSpec object into legacy dicts doesn't support the requested_destination field" [High,Confirmed] - Assigned to Sylvain Bauza (sylvain-bauza) | |
| 13:51:37 | bauzas | mriedem: permission to apologize, captain. https://bugs.launchpad.net/nova/+bug/1702454 | |
| 13:52:18 | bauzas | mriedem: tl;dr: evacuations providing a target don't verify that target since newton | |
| 13:53:04 | mriedem | i thought the bug said that was only for legacy specs or whatever? | |
| 13:53:27 | mriedem | someone from suse was just fixing bugs in stable related to evacuate with a targeted host, so something must work | |
| 13:54:17 | bauzas | mriedem: see the description | |
| 13:54:39 | bauzas | mriedem: that's because we turn into dicts only for evacuate | |
| 13:54:41 | mriedem | stephenfin: could we just make dhcp_domain default to None? that would change how https://github.com/openstack/nova/blob/0b9bacb/nova/api/metadata/base.py#L546-L549 behaves | |
| 13:55:00 | bauzas | mriedem: and tbh, the bug was reported internally on our Newton branch | |
| 13:55:35 | bauzas | so I doubt it's fixed upstream but just master because of https://review.openstack.org/#/c/469037/ | |
| 13:55:42 | bauzas | that's where tech debt kills us | |
| 13:56:10 | bauzas | and I wouldn't like to backport https://review.openstack.org/#/c/469037/ | |
| 13:56:20 | bauzas | rather fixing the problem with the transformating | |
| 13:56:58 | stephenfin | mriedem: It would change it, but I don't think it would deliver the desired behavior. bnemec did that, and cloudinit added a default 'localdomain' domain instead of what neutron had configured ('nemebean.com') | |
| 13:58:23 | openstackgerrit | Chris Dent proposed openstack/nova master: Optional separate database for placement API https://review.openstack.org/362766 | |
| 14:05:44 | dansmith | mriedem: are we going to go ahead with this? https://review.openstack.org/#/c/468227/2 | |
| 14:07:37 | mriedem | dansmith: related to bauzas' evacuate in newton clusterfunnery above? | |
| 14:07:53 | mriedem | is there a rhosp newton evacuate red alert going off this morning? | |
| 14:08:20 | bauzas | not really | |
| 14:08:24 | bauzas | AFAIK | |
| 14:08:24 | dansmith | mriedem: is it related to that? I don't think so | |
| 14:08:34 | bauzas | just some internal BZ triage I did | |
| 14:08:44 | dansmith | I don't know about bauzas' thing so I assume unrelated | |
| 14:08:52 | mriedem | dansmith: but yes we should get that in - the evacuate stuff with forced target was broken since newton i think | |
| 14:08:54 | bauzas | yup, unrelated | |
| 14:09:01 | dansmith | mriedem: yeah it's definitely broken | |
| 14:09:01 | bauzas | mriedem: not forced | |
| 14:09:07 | dansmith | mriedem: okay then I'll review | |