Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-04-24
19:26:57 mriedem smcginnis: the service type and endpoint name are 'placement' right?
19:27:09 coreycb cdent, yep
19:27:54 mriedem cdent: unless the ubuntu packages are going to change to take the port number out of the apache config it seems we should get the install guide updated to create the endpoint with the port number
19:28:05 cdent mriedem: yes
19:28:13 coreycb mriedem, cdent: yes agreed
19:28:23 cdent that's pretty much where asettle and I landed when talking about it last week
19:28:28 mriedem if canonical changed the apache config in a new package release, that would be confusing (at least for ocata)
19:28:49 mriedem you could probably mirate the config somehow at install/update time for the new package, but...
19:28:52 mriedem *migrate
19:28:59 smcginnis Well, if our long term goal is to get rid of ports, shouldn't we make sure new services aren't getting set up that way by default?
19:29:34 mriedem smcginnis: sure, but then canonical has to change the config in their package and provide a migration i'd think
19:29:42 mriedem which is probably a PITA after ocata GA'ed
19:29:57 coreycb mriedem, smcginnis: yeah we'd definitely want to do that on a release boundary if we did
19:32:23 smcginnis mriedem: OK, instance still failing because it's filtering out all my one hosts. I should be able to figure it out from here though. Placement seems to be sorted.
19:32:40 smcginnis mriedem, cdent, coreycb: Thanks for the help.
19:32:45 coreycb smcginnis, maybe cells v2 is needed next
19:33:08 smcginnis coreycb: Any docs.o.o instruction issues with that? I followed the docs on that one too.
19:33:40 coreycb smcginnis, not that i know of
19:34:10 smcginnis coreycb: OK, I'll let you know if I find otherwise.
19:34:13 mriedem yes
19:34:19 coreycb smcginnis, ok thanks
19:34:19 mriedem it's in the compute install guide
19:34:23 mriedem you have to discover the host
19:34:32 mriedem smcginnis: https://docs.openstack.org/ocata/install-guide-ubuntu/nova-compute-install.html
19:34:37 smcginnis mriedem: Yep, did that.
19:34:44 mriedem coreycb: if you can comment in that bug i can move forward with updating the install guide
19:34:48 smcginnis mriedem: Unless that needs to be done again since placement was hosed?
19:34:51 mriedem basically, we create the placement endpoint in keystone with ports for ocata
19:35:01 mriedem and we could reconsider package changes for the apache config in pike
19:35:23 mriedem smcginnis: is the nova-compute log showing that it's creating the resource provider in placement now?
19:36:03 smcginnis mriedem: Created resource provider record via placement API for resource provider with UUID
19:36:11 mriedem ok nice
19:36:34 mriedem smcginnis: are you using a flavor that's too big or something?
19:36:37 mriedem for your compute node?
19:36:53 coreycb mriedem, done
19:36:59 mriedem coreycb: thanks
19:37:04 smcginnis mriedem: Nope, tiniest one. Enabled debug now, so I can see what's going on with the filtering.
19:38:08 smcginnis mriedem: Well, whatever it was, restarting to enable debug seems to have fixed the issue. Got an ACTIVE running instance.
19:38:28 smcginnis mriedem: I'll go back in my own corner now. Thanks!
19:38:51 mriedem hmm
19:39:21 mriedem maybe needed to wait for the inventory to get updated in the compute periodic interval
19:39:36 mriedem or something was stashed in the scheduler worker
19:39:41 mriedem and restarting it flushed that out so we checked again
19:39:50 smcginnis Yeah, I don't think I really changed anything, so maybe just needed to wait or restart.
19:45:54 sean-k-mooney mriedem: hi, im just heading home but i know you had some question regarding the intel cis. ill try and catch up with you later in the week and answer what i can.
19:46:10 mriedem sean-k-mooney: ok
20:00:24 openstackgerrit Eric Fried proposed openstack/nova master: WIP/PoC: nova.utils.get_service_url(group) https://review.openstack.org/458257
20:00:33 openstackgerrit Matt Riedemann proposed openstack/nova master: Deprecate Multinic, floatingip action and os-virtual-interface API https://review.openstack.org/457181
20:02:42 mriedem claudiub: are you around/
20:02:43 mriedem ?
20:02:53 claudiub o/
20:03:02 mriedem claudiub: can you get https://review.openstack.org/#/c/458843/ so we can unblock stable/ocata changes?
20:03:24 claudiub ack
20:03:27 mriedem thanks
20:07:20 claudiub mriedem: can you also push this in? :D https://review.openstack.org/#/c/445540/
20:07:35 mriedem i'll take a look
20:12:37 mriedem no unit tests huh
20:12:37 mriedem nice
20:12:41 mriedem real nice
20:13:28 mriedem claudiub: approved. let me know if you need me to do a patch release for stable/ocata os-vif
20:13:56 claudiub that would actually help. :)
20:14:13 mriedem ok i'll submit one
20:15:02 claudiub anyways, just an fyi, there is a silent hyper-v + ovs ci running on neutron. currently collecting history, and checking for stability. in time, it will replace the current neutron hyper-v ci.
20:15:10 mriedem ok
20:15:22 mriedem btw, i'm not sure about this https://review.openstack.org/#/c/431683/3/requirements.txt
20:17:34 mriedem looks like we missed ovo 1.17.0 in os-vif before the final ocata release
20:17:55 mriedem if we release that in stable/ocata we need to bump the minor version, which could conflict with what's on master
20:18:14 claudiub yeah we did. the last requirements updates before that in os-vif was in november
20:18:39 claudiub and we need that pbr update too
20:18:49 mriedem we could separate out the pbr update
20:18:54 claudiub in ocata, all projects are <2.0.0
20:19:16 mriedem we haven't done an os-vif release in pike yet, and the last release for os-vif in ocata was 1.4.0
20:19:29 mriedem we could bring the ovo change in and release it in stable/ocata as 1.5.0,
20:19:31 mriedem but that's kind of odd
20:19:51 claudiub true, but it works, imo.
20:19:52 mriedem unless we actually need it on stable/ocata, which we don't, i think we should just ignore that one and do pbr separately
20:20:45 claudiub hm, i was going to say that upper constraints are being installed anyways, but apparently in ocata they aren't for os-vif
20:21:13 claudiub that is actually going to be a problem in the future
20:21:36 claudiub as the requirements will eventually start to conflict at some point
20:21:39 mriedem hmm yeah
20:21:53 claudiub it isn't in master either
20:22:09 mriedem jaypipes: is there any reason not to be using upper-constraints in os-vif tox.ini?
20:22:26 jaypipes mriedem: no, just an oversight likely.
20:26:49 mriedem claudiub: here is the pbr sync by itself https://review.openstack.org/#/c/459442/
20:30:23 claudiub left comment
20:32:08 mriedem claudiub: replied. that's not what the bot does though.
20:32:41 openstackgerrit OpenStack Proposal Bot proposed openstack/os-vif master: Updated from global requirements https://review.openstack.org/451049
20:33:14 claudiub hm, weird. I've seen the bot do it one time though a long time ago.
21:01:09 mriedem mgagne: i've given you more information than you ever wanted to know https://review.openstack.org/#/c/454322/
21:03:20 mgagne mriedem: thanks, will look right now
21:22:57 mriedem jaypipes: you're doing the get_inventory changes under some random topic branch right?
21:23:07 mriedem jaypipes: looks like you did get_inventory for ironic under this https://review.openstack.org/#/q/topic:bp/custom-resource-classes-pike,n,z
21:23:17 jaypipes mriedem: correct.
21:23:21 mriedem jaypipes: or should we consider that bp closed?
21:23:28 mriedem and the get_inventory for the other drivers is just currency?
21:23:28 jaypipes that bp is closed yet.
21:23:33 mriedem s/yet/yes/?
21:24:07 jaypipes mriedem: correct. I would like to get the get_inventory() methods implemented for the other virt drivers before the nested resource providers is done. will make my life easier.
21:24:14 mriedem ok
21:24:14 jaypipes mriedem: not having to redo a bunch of stuff.

Earlier   Later