Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-02-09
21:47:04 dansmith which I had never thought of
21:47:10 dansmith but mriedem actually _is_ Eric Foreman
21:47:49 melwitt hahaha
21:47:50 mriedem my hair gets that big sometimes
21:59:18 mriedem god it's been so long, i don't remember how to tell mox to do multiple returns
21:59:26 openstackgerrit Dan Smith proposed openstack/nova master: Remove main_db flavor compatibility https://review.openstack.org/431722
21:59:27 openstackgerrit Dan Smith proposed openstack/nova master: Remove flavor_* functions from db/api https://review.openstack.org/431752
21:59:27 mriedem maybe i need to rewrite this turd of a test in mock first...
21:59:34 dansmith have I mentioned how much I hate flavors?
22:03:53 melwitt has anyone else noticed gorpy printings of the "Skipping periodic task _discover_hosts_in_cells ..." when running unit tests? last time this happened it was because of a KeyError in an exception somewhere, I think
22:04:01 mriedem i've noticed it
22:04:54 dansmith would that be in a unit test that starts n-sch? I haven't seen it
22:06:56 melwitt I dunno. I miss that dumb config option that treated the key errors as exceptions
22:07:31 melwitt I tried to do some hacky thing with monkey patching to get the same thing awhile back but it was crap
22:08:14 dansmith well, the discover_hosts_in_cells periodic is in n-sch and disabled by default,
22:08:20 dansmith so I think that message just means we're running through there
22:09:11 cfriesen dansmith: you and jaypipes
22:09:19 cfriesen both hating flavors
22:10:15 dansmith cfriesen: I hate our flavor implementation, I think he hates the concept
22:16:29 dansmith melwitt: check this out: https://github.com/openstack/nova/blob/master/nova/cells/state.py#L301-L301
22:16:39 dansmith melwitt: pretty sure that breaks cellsv1 in newton once you're migrated
22:17:11 melwitt ouch
22:18:46 openstackgerrit Dan Smith proposed openstack/nova master: Remove flavor_* functions from db/api https://review.openstack.org/431752
22:18:47 openstackgerrit Dan Smith proposed openstack/nova master: Remove main_db flavor compatibility https://review.openstack.org/431722
22:21:05 openstackgerrit Ed Leafe proposed openstack/nova master: placement: adds REST API for nested providers https://review.openstack.org/384807
22:21:06 openstackgerrit Ed Leafe proposed openstack/nova master: placement: allow filter providers in tree https://review.openstack.org/377215
22:21:07 openstackgerrit Ed Leafe proposed openstack/nova master: pci: add uuid field to PciDevice object https://review.openstack.org/414768
22:21:07 openstackgerrit Ed Leafe proposed openstack/nova master: placement: integrate ProviderTree to report client https://review.openstack.org/415921
22:21:08 openstackgerrit Ed Leafe proposed openstack/nova master: placement: SRIOV PF devices as child providers https://review.openstack.org/414769
22:21:09 openstackgerrit Ed Leafe proposed openstack/nova master: placement: adds ProviderTree for nested resources https://review.openstack.org/415920
22:21:09 openstackgerrit Ed Leafe proposed openstack/nova master: placement: update client to set parent provider https://review.openstack.org/385693
22:21:10 openstackgerrit Ed Leafe proposed openstack/nova master: placement: add nested resource providers https://review.openstack.org/377138
22:21:25 edleafe cdent: jaypipes: ^^ rebased
22:22:34 jaypipes edleafe: ty sir
22:29:27 mriedem what's the thing we use when we want to check the type of an arg passed to a mock, but don't have the specific thing and don't want to use mock.ANY?
22:29:51 mriedem test.MatchType
22:29:51 mriedem ah
22:35:55 mordred dansmith, jaypipes: what I hate about flavors are the magical attributes that get attached to metadata of images that should be on flavors
22:36:06 mordred and the fact that I can't define a flavor as a user
22:36:36 mordred and the fact that because of htat I'm stuck with cloud providers deciding that CLEARLY I dont need any local disk because cloud
22:57:57 mriedem what i hate is all of the libvirt unit tests that rely on fake xml dumps
22:58:12 mriedem because the spaces in the xml comparisons ALWAYS takes me an hour to get working in my test
23:13:59 openstackgerrit Dan Smith proposed openstack/nova master: Remove flavor_* functions from db/api https://review.openstack.org/431752
23:34:46 openstackgerrit Chris Friesen proposed openstack/nova-specs master: add spec for instance flavor API https://review.openstack.org/431775
23:35:41 openstackgerrit ChangBo Guo(gcb) proposed openstack/nova master: Use min parameter to restrict live-migration config options https://review.openstack.org/323313
23:39:08 mriedem ugh
23:39:34 mriedem it's great when you try to convert a monster test from mox to mock just to find out that the original test was broken to begin with
23:42:10 cfriesen and then you realize you wrote the original test
23:42:19 mriedem nope
23:43:29 mriedem here she comes
23:43:38 openstackgerrit Matt Riedemann proposed openstack/nova master: libvirt: fix and break up _test_attach_detach_interface https://review.openstack.org/431778
#openstack-nova - 2017-02-10
00:18:04 openstackgerrit Ed Leafe proposed openstack/nova master: placement: RT now adds proper Ironic inventory https://review.openstack.org/404472
00:27:58 openstackgerrit Takashi NATSUME proposed openstack/nova master: Enable cold migration with target host(1/2) https://review.openstack.org/408955
00:29:04 openstackgerrit Takashi NATSUME proposed openstack/nova master: Enable cold migration with target host(2/2) https://review.openstack.org/408964
00:29:38 openstackgerrit Takashi NATSUME proposed openstack/nova master: Add functional tests for cold migration to same host https://review.openstack.org/414926
00:29:44 openstackgerrit Takashi NATSUME proposed openstack/nova master: api-ref: Add parameters in cold migrate action https://review.openstack.org/410042
00:33:46 openstackgerrit Matt Riedemann proposed openstack/nova master: libvirt: wait for interface detach from the guest https://review.openstack.org/349014
00:33:46 openstackgerrit Matt Riedemann proposed openstack/nova master: libvirt: fix and break up _test_attach_detach_interface https://review.openstack.org/431778
00:41:48 openstackgerrit Takashi NATSUME proposed openstack/nova-specs master: Enable cold migration with target host https://review.openstack.org/426636
01:00:31 openstackgerrit yuhui proposed openstack/nova master: Correct some spelling errors https://review.openstack.org/430646
01:02:55 openstackgerrit melanie witt proposed openstack/nova master: Count resources to check quota for cells https://review.openstack.org/416521
01:02:55 openstackgerrit melanie witt proposed openstack/nova master: Make Quotas object favor the API database https://review.openstack.org/410945
01:02:55 openstackgerrit melanie witt proposed openstack/nova master: Add online migration to move quotas to API database https://review.openstack.org/410946
01:30:06 openstackgerrit Li Jingjing proposed openstack/nova master: Fix typo in nova/network/neutronv2/api.py https://review.openstack.org/431796
02:40:49 alex_xu mriedem: our team are working on enable CAT(CPU L3 cache allocation tech) on openstack, currently the kernel support is ready, the team are working the libvirt interface. is there any chance we can submit the spec or raise some discussion in Nova even the libvirt interface not ready? at least some discussion, what kind of usage-model from nova point of view.
02:41:32 mriedem alex_xu: are you going to be at the PTG?
02:41:59 alex_xu mriedem: yea, I will go to the PTG
02:42:26 mriedem that's probably a decent place to discuss it, but i'd recommend starting a thread in the mailing list, or maybe a draft spec since i've never heard of that before
02:43:25 alex_xu mriedem: yeah, got it
02:44:00 mriedem i'm assuming this would be some kind of image metadata thing + minimum libvirt version check in the driver?
02:44:44 mriedem although, if it's not in libvirt we can't even check a minimum version
02:44:52 alex_xu mriedem: yes, but I think the hard thing is how much detail expose to the user, I guess we may not hope we let the user choice 20MB or 30MB l3 cache they need
02:45:13 mriedem that seems like flavor extra specs
02:45:18 mriedem like how qos is handled
02:45:24 alex_xu maybe user just need to specify gold or liver
02:45:32 alex_xu yea
02:45:52 alex_xu s/liver/sliver/
02:47:00 mriedem there is a joke in there somewhere about a liver-flavored flavor
02:48:11 alex_xu I guess we didn't want to add more thing to the extra spec, thinking of whether we can match the propose into resource provider
02:57:12 zhangqiankun I want to use guacamole connection openstack create vm for vnc,what can i do?
03:05:35 mriedem alex_xu: liver-flavored means something that tastes like liver
03:05:46 mriedem alex_xu: so a liver-flavored flavor, is a joke :)
03:07:03 mriedem alex_xu: i think with traits we'd expose the libvirt version, so a request spec could require a certain minimum version of the hypervisor for this capability, or the libvirt driver could just say it has this capability and we decouple it from the version in the request spec
03:07:27 mriedem but i would think you would still need something in the flavor extra specs to tell what size of cache is being requested
03:09:40 alex_xu mriedem: maybe we can create ResourceClass called l3_cache, then track the usage of cache in the scheduler?
03:10:09 alex_xu at least we need figure which host can have enough cache, or need some ways to balance the cache
03:11:20 alex_xu emm...whatever there should be somewhere talk about the size
03:11:29 alex_xu or gold/sliver stuff
03:12:34 alex_xu half years ago, we have initial thought using extra specs, and think about shared/isolate cache, and maybe need to bind with NUMA node. maybe that is over complex
03:12:48 alex_xu try to rethink of all of those, try to find something more easy
03:13:56 alex_xu in China, there are a lot of food made by liver...I guess we may have different understanding of the taste of liver...
03:18:00 alex_xu I remember people won't expose libvirt version into the traits, but the libvirt version will answer the question of capabilitis API
04:53:11 openstackgerrit OpenStack Proposal Bot proposed openstack/python-novaclient master: Updated from global requirements https://review.openstack.org/431906
05:03:07 gmann alex_xu: johnthetubaguy sfinucan it would not raise exception if more than 1 port right ?- https://review.openstack.org/#/c/424931/4/nova/network/neutronv2/api.py@1676
05:03:24 gmann but yea it will return diff port as mentioned by sfinucan
05:03:51 alex_xu gmann: johnthetubaguy prefer to not change anything
05:04:39 alex_xu actually I both to ok for change or not, that is rare case I think
05:04:42 gmann alex_xu: yea that we should keep same and may be previous logic in PS2 is good ?
05:05:17 alex_xu gmann: PS2 is same logic with current one I think?
05:05:38 alex_xu both of them stop the loop when a port found?

Earlier   Later