Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2019-01-09
15:50:06 openstackgerrit Yikun Jiang proposed openstack/nova master: Per aggregate scheduling weight https://review.openstack.org/628163
16:24:09 melwitt jaypipes: hey, I dunno if you saw my reply on the counting quotas from placement spec https://review.openstack.org/#/c/509042/ but I think we can't consider removing user-specific quotas until we migrate to keystone limits + oslo.limit
16:25:21 jaypipes melwitt: ack, and I wasn't suggesting removing them yet, just that we should plan to just use keystone and the hierarchical projects stuff in future.
16:28:18 melwitt jaypipes: ah, ok. yeah, that is the plan when we move to keystone limits + oslo.limits user-specific quotas will be dropped as part of that. described in johnthetubaguy's spec if you're interested https://review.openstack.org/#/c/602201/
16:31:45 cfriesen so there's a review up for the os-traits portion of the emulated TPM stuff. Am I correct that we'd need a new version of os-traits to pick that up in nova?
16:32:00 jaypipes melwitt: ++
16:32:12 jaypipes cfriesen: yes.
16:32:23 jaypipes cfriesen: and an updated requirements.txt in nova.
16:33:26 cfriesen right. so what's the process for triggering the new package version? is that something we need to request formally somewhere?
16:34:54 melwitt cfriesen: you can request a release of the package via the openstack/releases repo, like this https://review.openstack.org/#/c/612843/
16:35:03 gibi melwitt: hi! kashyap has a bp https://blueprints.launchpad.net/nova/+spec/support-qemu-native-tls-for-live-migration that hasn't been approved but it seems it was discussed before. Do you remember if anything blocks that bp? The implementation seems to be ready
16:37:16 melwitt gibi: yeah, I think everyone is OK with the feature but the discussion was around whether it needed a spec or not. IIRC dansmith suggested the docs and references related to the feature be proposed along with the implementation, which I thought was a good idea,
16:37:58 gibi melwitt: kashyap prepares a followup patch top of the implementation that adds documentation how to set up the compute hosts for tls
16:38:02 melwitt kashyap said he's done that and the bp is on the agenda for the nova meeting tomorrow, where we can get additional ACKs for approving it as specless
16:38:28 dansmith where is the docs patch? it doesn't seem to be in the same series
16:38:37 gibi melwitt: OK, ACKing on the nova meeting works for me
16:38:56 gibi dansmith: patch is not up yet, a perliminary doc is here https://kashyapc.fedorapeople.org/Native-TLS/Setup-for-NBD-and-migration-streams-over-TLS.rst.txt
16:39:22 kashyap` gibi: Actually it is 95% complete doc :-). Converting to rST as we speak...
16:39:32 dansmith gibi: erm, okay, not sure why we wouldn't just have that in a follow-on patch so we can see what that's going to look like :)
16:39:34 dansmith kashyap`: ^
16:39:51 kashyap` dansmith: Hiya, yeah. Follow-up patch indeed
16:39:59 kashyap` Not going to smush it into the first patch
16:40:16 dansmith kashyap: can you have that in gerrit today so we can see it rendered and stuff ahead of the meeting tomorrow?
16:40:20 kashyap` dansmith: Actually gibi said it as much -- a follow-on -- on the review.
16:40:21 gibi dansmith: sure. I just didn't thought that having the doc patch up is a prerequisit to approve the bp
16:40:24 kashyap` Yeap
16:40:35 dansmith gibi: that was the original answer, IIRC :)
16:40:44 dansmith note that tomorrow is the early meeting so I won't be around anyway
16:41:18 kashyap` dansmith: If it soothes your nerves, FWIW, Martin Schuppert from TripleO independently tested this patch and showed me evidence
16:41:22 kashyap` (Which matches mine.)
16:41:55 dansmith kashyap`: nope, all I want to see is docs :)
16:42:02 gibi dansmith: :)
16:42:46 kashyap` dansmith: Sure. It's already there :-) Getting it polished a bit
16:43:21 melwitt ah right, tomorrow's meeting is at 6am for us so I won't be around yet either
16:43:28 kashyap` The high-level thing is: (a) Installer needs to setup all the TLS env; (b) Only _then_ this feature makes sense.
16:44:02 dansmith melwitt: yeah, so let's just see the docs rendered in a patch, and then we can sign off on the blueprint, independent of the meeting
16:44:24 melwitt dansmith: sure, sounds like a plan
16:45:26 melwitt gibi ^
16:46:02 melwitt thanks for helping to get this sorted out
16:46:13 gibi melwitt: works for me
16:46:24 gibi melwitt: I've reviewed the implementation too and looks good to me
16:46:49 melwitt great, thanks for that
16:49:55 melwitt gibi: a note on the bandwidth aware scheduling hangout meeting, I think mriedem had said he'd rather not have the meeting on the 14th, his first day back, if that is possible
16:50:16 gibi melwitt: sure, I can move that to Tuesday if that is better
16:50:39 gibi melwitt: what do you think, 17:00 UTC would work for the team?
16:51:39 melwitt gibi: he didn't mention specific favorable days, but I expect tuesday should be better. and I expect 1700 UTC should work
16:52:01 gibi melwitt: thanks, I will send and update to the ML
16:52:07 melwitt ok, thanks
16:53:32 kashyap` Thanks, folks!
16:54:02 kashyap` I'll get some dinner, and then get the rendering going. I'm writing some Nova-specific context. Because most people would use installer tools to take care of all the CA setup
16:54:37 melwitt ok, thanks kashyap`
17:31:45 openstackgerrit sean mooney proposed openstack/nova master: [WIP] libvirt: auto detach/attach sriov ports on migration https://review.openstack.org/629589
17:41:54 openstackgerrit Stephen Finucane proposed openstack/nova stable/queens: Fail to live migration if instance has a NUMA topology https://review.openstack.org/629597
18:11:47 sean-k-mooney adrianc: by the way this https://review.openstack.org/#/c/609166/4 is not needed an i can abandon it right
18:14:03 adrianc sean-k-mooney: yep, you can abandon
18:14:17 sean-k-mooney cool its replaced by the code in https://review.openstack.org/#/c/619929/7
18:15:29 kashyap Any rST gurus here, what is wrong here?
18:15:58 kashyap Assuming a doc called 'live-migration-usage.rst' exists, if I do: "Refer to the :doc:`live-migration-usage`" from another .rst doc, it should Just Work?
18:16:02 sean-k-mooney kashyap: i just ask stephenfin
18:16:30 sean-k-mooney kashyap: im not sure if you need to use a relivie path
18:16:33 stephenfin kashyap: the path is relative, so if it's in the same directory, yes
18:16:39 stephenfin otherwise get those dots out
18:16:47 adrianc sean-k-mooney: yes, also noticed you added a commit for the direct mode, ill take a look tomorrow
18:16:57 stephenfin adrianc: been reviewing all those. On the last one
18:17:10 stephenfin adrianc: Mostly +2 except for minor changes, mostly nits
18:17:13 stephenfin just fyi :)
18:17:18 kashyap stephenfin: You mean get those colons out, if it's (it is) in the same dir?
18:17:36 sean-k-mooney adrianc: ya its still a work in progress ill need to test it properly tomorrow and then add unit tests and a release note
18:17:43 adrianc stephenfin: :) great!
18:17:47 kashyap - Refer to the :doc:`live-migration-usage`
18:17:47 kashyap Is this the `diff`?
18:17:48 kashyap + Refer to the doc`live-migration-usage`
18:17:48 stephenfin kashyap: heh, no. :doc:`../live-migration-usage` (if not same directory)
18:18:03 kashyap It _is_ in the same dir
18:18:03 stephenfin :doc:`path` is correct
18:18:07 stephenfin then you're good
18:18:49 kashyap secure-live-migration-with-qemu-native-tls.rst:138: (ERROR/3) Unknown interpreted text role "doc".
18:18:49 kashyap $> rst2html secure-live-migration-with-qemu-native-tls.rst foo.html
18:18:49 kashyap stephenfin: Hmm, but look:
18:19:01 kashyap That's the _only_ error between me and dinner.
18:19:11 stephenfin doc8?
18:19:17 stephenfin Ignore it, if so
18:19:35 stephenfin doc8 uses docutils, which doesn't know about doctrees and therefore doesn't have a doc role
18:19:48 stephenfin doctrees are a purely Sphinx thing
18:20:33 kashyap Aah, so it should be fine.
18:20:48 stephenfin yup, all good
18:21:25 kashyap Thanks!
18:43:49 openstackgerrit Kashyap Chamarthy proposed openstack/nova master: docs: Secure live migration with QEMU-native TLS https://review.openstack.org/629627
18:45:48 kashyap melwitt: dansmith ^ First cut
18:45:59 dansmith cool
18:46:21 melwitt kashyap: thanks
18:46:31 dansmith I expect a review from stephenfin on that too
18:46:46 dansmith which will yield a $number_of_lines count of nit changes
18:47:19 dansmith I found an issue in the first line already :P
18:47:22 kashyap Haha, yeah. Happy to get stephenfin's rST eyes
18:47:35 kashyap dansmith: What is it?
18:47:46 dansmith kashyap: you have to wait :)
18:48:00 kashyap Nerves are wracking, dammit. I go better eat
19:23:52 openstackgerrit Inbar Stolberg proposed openstack/nova master: Author: Inbar Stolberg https://review.openstack.org/629646
19:42:54 dansmith kashyap fails the build

Earlier   Later