Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-04-13
18:57:48 mriedem for smaller things
18:57:53 mriedem i don't think they require a spec for everything
18:58:02 mriedem but i'm not totally familiar with their process
18:58:15 mriedem jangutter: https://docs.openstack.org/developer/neutron/policies/blueprints.html
18:59:48 jangutter mriedem: OK. Should I add that task to the spec or the blueprint?
19:00:14 mriedem jangutter: not the nova one no
19:00:35 mriedem well, you could just call out a dependency on needing to get that support into neutron, but it's kind of implied
19:01:16 jangutter OK. It's a bit of a chicken/egg situation. Both sides need the support at roughly the same time.
19:01:31 artom jangutter, also, and I'm only superficially familiar with this sort of thing, but do you need a new vnic *and* vif type?
19:02:26 jangutter I think sean-k-mooney pointed out that we don't, just new portbinding types.
19:03:36 jangutter It would mean the OVS vif type is now one of the most overloaded ones in the universe.
19:06:39 vladikr what are the chances for anyone looking at the spice spec again? https://review.openstack.org/#/c/442040 ?
19:07:21 artom Heh, while we're doing last-minute spec merged ;) mriedem ^^
19:07:26 artom *merges
19:07:56 artom To be completely honest, I feel like we should have bugged alex_xu for the API part earlier
19:08:25 vladikr the spec was there for a long while - definitely not a last minute spec
19:09:33 mriedem spice-native?
19:09:36 mriedem does anyone use spice?
19:09:49 mriedem spice-uncut
19:09:55 mriedem spice-freerange
19:10:00 mriedem spice-organic
19:10:22 artom The whole of interstellar travel runs on spice;)
19:10:33 mriedem yes i love dune references
19:10:35 artom So clearly we need faster proxies
19:11:56 vladikr mriedem, yes, well, more like 'will', especially with the introduction of the vgpus
19:12:24 vladikr mriedem, it's not just about a faster proxy , but for the ability to connect native clients
19:13:10 mriedem introduction of vgpus? into nova? or in general?
19:13:42 artom mriedem, I think the hope is that eventually we send the vGPU output to the client through SPICE
19:13:50 vladikr mriedem, well, in general and in nova in future :)
19:13:52 artom Ie, full workstation virt
19:16:08 vladikr speaking about vgpu: https://review.openstack.org/#/c/450122/
19:21:29 openstackgerrit Jan Gutter proposed openstack/nova-specs master: Spec to enable Netronome SmartNIC support https://review.openstack.org/456386
19:29:34 sean-k-mooney jangutter: actully the ovs-vif type is really just a special case of the generic tap vif type but ya it has a pretty broad meaning
19:31:54 jangutter In trying to amend the spec, I seem to have plonked in a unicode character somewhere.
19:33:37 clarkb mriedem: is the livemigration stuff happy now with my fixes in?
19:33:48 mriedem clarkb: yeah, i didn't realize that fix merged this morning
19:34:02 clarkb cool just double checking I don't have more to look at now that I am back to a computer
19:36:30 openstackgerrit Jan Gutter proposed openstack/nova-specs master: Spec to enable Netronome SmartNIC support https://review.openstack.org/456386
19:50:39 sean-k-mooney did gerrit just go down
19:50:50 clarkb yes
19:50:53 sean-k-mooney its back
19:50:55 clarkb the notice is on its way out
19:51:03 clarkb but too many irc channels to do it all at once
19:51:58 sean-k-mooney clarkb: is the maitenace still ongoing or was that restart the only one to expect
19:52:42 clarkb hopefully the only one
19:53:39 sean-k-mooney cool ill continue reviewing then
20:00:13 jangutter I'm trying to find out why tox is somehow getting stuck on a unicode character in the spec. It's not happening if I run it locally :-(
20:01:05 sean-k-mooney jangutter: which spec? this one https://review.openstack.org/#/c/456386/5/specs/pike/approved/netronome-smartnic-enablement.rst
20:01:27 sean-k-mooney i can test it locally if you like and see if i get the same error
20:01:31 clarkb jangutter: its likely a locale issue. I think you get C.utf-8 in the jenkins runs
20:01:38 clarkb but maybe not
20:03:58 jangutter Yeah, if I look at the unicode character it's complaining about it's an "en dash" "\u2013".
20:04:27 jangutter I've been bitten by them before: cutting and pasting from IM apps that believe the world should be a prettier place.
20:05:29 openstackgerrit Matt Riedemann proposed openstack/nova master: Fix 404 for no bdm.device_name when showing volume attachment details https://review.openstack.org/290196
20:05:32 mriedem gmann: jaypipes: oomichi: ^ updated this very old change, i don't think requires a microversion
20:05:35 clarkb oh regardless of locale it seems like the docs build expect ascii for some reason
20:05:40 mriedem jaypipes: could you drop the -2 on that?
20:11:39 jangutter sean-k-mooney: I've updated it to https://review.openstack.org/#/c/456386/7/specs/pike/approved/netronome-smartnic-enablement.rst
20:12:01 jangutter sean-k-mooney: would you mind checking? tox completes fine on my side...
20:12:11 peter-hamilton jaypipes: do you think you'll be able to get to the cert validation spec today?
20:12:33 peter-hamilton jaypipes: don't want to be pest, obv
20:12:38 mriedem i think he's under the weather
20:12:45 peter-hamilton mriedem: ah, ok
20:13:17 peter-hamilton mriedem: do you mind giving it a look over then? he's been a +2 for the past couple versions
20:13:35 mriedem i've got some other things queued up first
20:15:15 peter-hamilton mriedem: gotcha
20:16:04 sean-k-mooney jangutter: sure im just pulling it down now
20:17:44 mriedem me too
20:20:29 sean-k-mooney jangutter: ya i get the same error as the gate
20:21:01 mriedem me too
20:21:03 mriedem after i rebased
20:21:07 mriedem and did a tox -r -e docs
20:22:15 mriedem i'm wondering if it has something to do with this https://github.com/openstack/nova-specs/commit/0a4c1bcaeb600229103cc61d574d57cac320ba84
20:22:19 openstackgerrit Dan Smith proposed openstack/nova master: Fix hypervisors api missing HostMappingNotFound handlers https://review.openstack.org/456227
20:22:25 clarkb oh chardet updated
20:22:38 clarkb and I know openstack-manuals ran into a similar problem
20:22:54 mriedem ah yeah
20:22:56 mriedem 3.0.2
20:23:09 jangutter Are those the dashes we are looking for?
20:23:25 mriedem *2.3.0
20:23:37 sean-k-mooney so we have to cap chardet or is there a way to fix it?
20:24:08 mriedem oh wait no 3.0.2 is new, but u-c is using 2.3.0
20:24:48 mriedem clarkb: i don't see chardet used in this job
20:24:51 mriedem http://logs.openstack.org/86/456386/7/check/gate-nova-specs-docs-ubuntu-xenial/9c8b9dd/tox/venv-3.log
20:25:07 jangutter I've had _fun_ with pbr and tox before.
20:27:02 mriedem someone remind me next time i mark my entire nova folder as read in thunderbird, that i need to not set view > threads > threads with unread
20:27:07 mriedem because i got scared
20:27:32 sdague heh
20:28:35 sean-k-mooney mriedem: clarkb downgradeing chardet to 2.3.0 has no effect
20:29:05 mriedem yeah we have chardet 2.3.0 in u-c already
20:29:08 mriedem that's not new
20:29:52 mriedem reverting https://github.com/openstack/nova-specs/commit/0a4c1bcaeb600229103cc61d574d57cac320ba84 makes no difference either
20:30:14 mriedem nova meeting in 30 minutes
20:31:01 sean-k-mooney mriedem: looking at https://github.com/openstack/nova-specs/blob/master/tox.ini i dont think we are using the upper constraints in nova-specs
20:31:17 mriedem oh that's right
20:31:25 mriedem and i can't remember why
20:31:49 sean-k-mooney well it has no "code" so i kindof makes sense that it would not be needed
20:32:43 melwitt mriedem: for adding "upgrade impact" to the spec template, should I do it only for the queens template or for pike and queens?
20:32:53 mriedem melwitt: quenes
20:32:55 mriedem *queens

Earlier   Later