Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-02-13
17:56:57 BobBall It's therefore possible that this is just the "start" of random breakages until we have upgraded to Xenial
17:57:07 mriedem ok, it's -WIP for now
17:59:34 BobBall http://dd6b71949550285df7dc-dda4e480e005aaa13ec303551d2d8155.r49.cf1.rackcdn.com/01/432001/8/check/dsvm-tempest-neutron-network/7e28db3/logs/localrc.txt.gz contains FORCE=yes (although the comment is ages old because the original CI was using saucy!)
18:00:10 BobBall so I think it probably won't break the CI... but definitely something that increases the urgency of upgrading to Xenial.
18:03:49 sdague mriedem: should probably remove f23 at the same time
18:03:57 sdague that's no longer distro supported
18:08:10 sdague mriedem: I tweaked it
18:09:53 mriedem sdague: ok, sort of unrelated, along with adding zesty, but i can mention that in the commit message
18:10:29 sdague yeh, it's just mostly an adjustment to the current cycle distros
18:12:33 mriedem melwitt: dansmith: jaypipes: how much are we going to try and get this into rc2? or can it be backported? https://review.openstack.org/#/q/topic:fix-bfv-boot-resources
18:12:58 jaypipes mriedem: happy to review those, sure.
18:14:10 dansmith mriedem: they wouldn't be reasonable for a backport, IMHO, because the scheduler will send instances to computes that wouldn't take them because their claims code isn't updated
18:14:25 dansmith mriedem: haven't looked to see if they're accounting for newton computes even
18:14:32 dansmith seems like a large and risky thing for an rc2 to me
18:15:26 openstackgerrit John Garbutt proposed openstack/nova-specs master: Add policy-project-checks spec https://review.openstack.org/433037
18:15:36 openstackgerrit John Garbutt proposed openstack/nova-specs master: WIP: spec about having observer role by default https://review.openstack.org/427872
18:16:32 mriedem these patches aren't accounting for newton computes
18:16:46 johnthetubaguy sdague: I split up the policy spec into more manageable chunks
18:17:16 johnthetubaguy sdague: if you get chance pre PTG for a quick check, that would be awesome
18:18:23 openstackgerrit Stephen Finucane proposed openstack/nova master: hardware: Flatten functions https://review.openstack.org/367470
18:18:24 openstackgerrit Stephen Finucane proposed openstack/nova master: Rename '_numa_get_constraints_XXX' functions https://review.openstack.org/385072
18:18:24 openstackgerrit Stephen Finucane proposed openstack/nova master: Standardize '_get_XXX_constraints' functions https://review.openstack.org/385071
18:18:25 openstackgerrit Stephen Finucane proposed openstack/nova master: De-duplicate _numa_get_flavor_XXX_map_list https://review.openstack.org/385074
18:18:38 macsz gcb: are you working on python3 support and removing mox?
18:19:12 mriedem johnthetubaguy: i see this backport popped up for ocata https://review.openstack.org/#/c/432954/ is that something we should get into rc2?
18:19:19 mriedem it adds a config option for the placement endpoint
18:19:31 openstackgerrit Stephen Finucane proposed openstack/nova master: Fix suggested database migration command https://review.openstack.org/420060
18:20:13 johnthetubaguy mriedem: I am tempted to add it, although it is a bit feature-like, it does mean users can use their service catalog in their normal way for placement
18:21:19 mriedem johnthetubaguy: checking config options, i guess you can do this same thing more or less for connecting to glance/cinder/neutron/ironic too
18:22:02 johnthetubaguy mriedem: the other ones have URL over-rides that do a similar (but worse) thing
18:22:39 mriedem yeah
18:23:27 johnthetubaguy mriedem: its a bit like this I guess: https://github.com/openstack/nova/blob/8de0c648c218dd748c2076551d1d9819dbeb3bd6/nova/conf/cinder.py#L25
18:23:52 mriedem yeah
18:24:15 mriedem it will be fun when someone wants to not call the endpoint "placement" in their service catalog
18:25:02 mriedem oh also, this missed nova-status
18:25:18 mriedem https://github.com/openstack/nova/blob/master/nova/cmd/status.py#L182
18:25:25 johnthetubaguy mriedem: there is some follow up work from that service token stuff, where we should fix up what we do with all that
18:25:46 johnthetubaguy mriedem: dang, that bit got missed
18:25:50 mriedem johnthetubaguy: as far as pike bp?
18:26:41 johnthetubaguy mriedem: I haven't created a BP for that, probably worth doing (tells raj_singh so there are two folks trying to remember that ^)
18:26:45 mriedem johnthetubaguy: https://blueprints.launchpad.net/nova/+spec/use-service-tokens-pike
18:26:48 mriedem it's already done
18:26:55 johnthetubaguy its more follow on from that I guess
18:27:18 johnthetubaguy like glance doesn't use the service catalog at all right now
18:27:24 mriedem yeah i know
18:27:33 mriedem the patch to use ksa was approved but tests were failing
18:27:41 mriedem b/c of the global variables being mocked
18:27:58 mriedem anyway, the pike bp is approved
18:28:22 mriedem no unit test for this? https://review.openstack.org/#/c/426163/
18:28:23 mriedem gah
18:28:43 raj_singh mriedem: I think we fixed the tests for that, let me take a look.
18:29:14 johnthetubaguy mriedem: :S yeah, no tests, not sure how I missed that, grrr
18:29:28 mriedem i can add tests and also fix the nova-status thing,
18:29:41 mriedem but if we're going to backport the config option change, we should have the other stuff (unit test and nova-status
18:29:58 mriedem plus i'm also procrastinating on dealing with https://etherpad.openstack.org/p/nova-ptg-pike
18:30:11 johnthetubaguy good to get both fixed for sure
18:30:13 raj_singh mriedem: https://review.openstack.org/#/c/412634/ Can you take your -2 off.
18:30:31 mriedem raj_singh: done
18:30:47 raj_singh thx, you are welcome to replace it with -1 :)
18:30:51 johnthetubaguy the others could be post release, worst case, but thats probably a bad idea
18:31:19 mriedem well if we land the one but nova-status doesn't use it, nova-status could fail since it's pusing the wrong endpoint
18:31:26 mriedem *using
18:31:37 mriedem assuming people are going to use nova-status, but i hope they do
18:32:21 johnthetubaguy yeah, it sucks to have that broken
18:32:29 openstackgerrit Anusha Unnam proposed openstack/nova master: Reno for additional-notification-fields-for-searchlight https://review.openstack.org/428835
18:32:40 mriedem raj_singh: did you see this gem? https://review.openstack.org/#/c/431713/
18:33:01 raj_singh mriedem: you rock!
18:35:42 openstackgerrit Pushkar Umaranikar proposed openstack/nova master: Placement api: set custom json_error_formatter in aggregate and usage https://review.openstack.org/433224
18:39:08 melwitt dansmith, mriedem: what's "newton computes" mean?
18:39:27 dansmith melwitt: compute nodes still running n-cpu from newton
18:39:41 melwitt oh, okay
18:42:08 sterdnotshaken With regard to regions, is it required to have replica's of Cinder, Neutron, Nova, etc within each region? I know that all the separate regions share Horizon and Keystone instances, but I'm just wondering if we can get away with not duplicating some of these services in each region…
18:43:13 openstackgerrit John Garbutt proposed openstack/nova-specs master: Add policy-project-checks spec https://review.openstack.org/433037
18:47:08 oomichi sdague: johnthetubaguy: hi, about api-ref thing: Which do you prefer: "Including object name in parameter" or not (like addFixedIp.networkId vs networkId)
18:47:44 oomichi sdague: johnthetubaguy: I am reviewing https://review.openstack.org/#/c/338606/18/api-ref/source/servers-action-fixed-ip.inc
18:48:23 oomichi addFixedIp.networkId seems easy to understand the structures of requrest/response of APIs
18:49:12 johnthetubaguy oomichi: I thought we only did that if it was unclear which bit of the request/response we were talking about
18:49:47 johnthetubaguy oomichi: so for host.host, we needed it for all of them
18:50:02 oomichi johnthetubaguy: like multiple paramters in a single object?
18:50:17 johnthetubaguy oomichi: like "host" appears twice
18:50:18 oomichi johnthetubaguy: ah, nice sample :)
18:50:36 mriedem we have that all over, like networks.port
18:50:42 mriedem block_device_mapping_v2.boot_index
18:51:03 sterdnotshaken Anyone happen to know the answer to my question regarding regions?
18:51:16 mriedem sterdnotshaken: see the channel topic please,
18:51:29 mriedem that's a question better asked in ask.openstack.org, or #openstack, or in the general mailing list
18:52:21 openstackgerrit John Garbutt proposed openstack/nova-specs master: Add policy-project-checks spec https://review.openstack.org/433037
18:53:51 oomichi mriedem: yeah, that aslo is a nice sample. "create a server" is super complex API, then it is nice to clarify which object the parameters belong to
19:03:20 openstackgerrit John Garbutt proposed openstack/nova-specs master: WIP: spec about having observer role by default https://review.openstack.org/427872
19:04:23 EmilienM mriedem: hey! in https://review.openstack.org/#/c/432954/ - do you mean the patch is currently noop?
19:04:49 mriedem EmilienM: no
19:05:16 openstack Launchpad bug 1664334 in OpenStack Compute (nova) "CONF.placement.os_interface is not used in nova-status upgrade check" [Medium,Confirmed] - Assigned to Matt Riedemann (mriedem)
19:05:16 mriedem EmilienM: https://bugs.launchpad.net/nova/+bug/1664334
19:06:19 EmilienM mriedem: ok, so it's lower priority than nova-compute trying to reach placement, right?
19:07:06 mriedem it's not low priority if you use the nova-status upgrade check command to verify the upgrade and talking to 'public' doesn't work
19:08:58 EmilienM mriedem: I see. Thanks
19:21:30 sdague oomichi: yeh, that was the convention we came up with x.y
19:21:47 sdague we never really explained it a ton, I'm sure the cover docs could be better
19:23:54 sneti mriedem, additional-notification-fields-for-searchlight is reapproved for Pike. can you remove -2 on https://review.openstack.org/#/c/407228/
19:23:55 cfriesen question for resourcetracker types....is it possible for a sched filter to determine whether a host is running with KVM support vs software qemu?

Earlier   Later