Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-01-03
14:45:36 ildikov mriedem: yeah, I saw your ping this morning
14:46:41 ildikov mriedem: I pointed Steve to the patch yesterday, I didn't know you're planning to work on it
14:48:01 mriedem it still needs some additional test coverage, and needs to actually pass first
14:48:53 ildikov we can chat about it tomorrow on the meeting
14:49:05 ildikov I will try to catch Steve later today
14:50:25 openstackgerrit Stephen Finucane proposed openstack/nova master: Modify PciDevice.uuid generation code https://review.openstack.org/530487
14:50:26 openstackgerrit Stephen Finucane proposed openstack/nova master: Add an online migration for PciDevice.uuid https://review.openstack.org/530905
15:03:59 stephenfin jaypipes: Think this is something you could tackle before the end of the cycle? https://review.openstack.org/#/q/status:open+project:openstack/nova+topic:bp/websocket-proxy-to-host-security
15:04:39 jaypipes stephenfin: definitely not my area of expertise, but I will review just because you asked nicely.
15:04:48 stephenfin sdague had volunteered to take the blow at the PTG, but he's not about much these days. It's been about since...Kilo and would be good to finally close out
15:04:58 stephenfin jaypipes: Excellent. Appreciate it :)
15:06:44 openstackgerrit Matthew Booth proposed openstack/nova master: Add test for assignment of uuid to a deleted BDM https://review.openstack.org/530774
15:21:13 mriedem stephenfin: on that series, there is this devstack change https://review.openstack.org/#/c/396186/
15:21:23 mriedem stephenfin: is tempest already running a test to exercise any of this code?
15:21:39 mriedem i thought at one point danpb had a devstack-gate or project-config change to one of the jobs to enable this
15:22:02 mriedem this https://review.openstack.org/#/c/396289/
15:22:12 stephenfin If you're running under libvirt then some of it is exercised
15:22:53 stephenfin You're referring to this job, I imagine? https://review.openstack.org/#/c/396289/
15:23:15 mriedem i see melwitt moved that to a zuulv3 job definition within nova https://review.openstack.org/#/c/513160/4
15:23:33 mriedem however, the nova-next job is still in the openstack-zuul-jobs repo
15:24:09 stephenfin The stuff that's tested is the basic negotiation. The VeNCrypt scheme is untested without that job by melwitt
15:24:38 mriedem so, can we put some effort into getting the new stuff tested before landing the series?
15:24:39 stephenfin I had patches up to move that, I thought?
15:24:52 mriedem you had patches up to move like all jobs into the nova repo, or something close,
15:24:59 mriedem and i'd like to avoid doing a giant job migration right now
15:25:21 mriedem because, e.g. pike is blocked on the tempest-full job migration for master
15:25:53 mriedem so maybe we start with moving the nova-next job into the nova repo,
15:26:03 mriedem then tweak it on top with mel's change to flip the new bit in devstack
15:26:15 mriedem the NOVA_CONSOLE_PROXY_COMPUTE_TLS variable
15:26:33 mriedem would have been easier if https://review.openstack.org/#/c/513160/ was on the same topic branch as the rest of the series
15:27:01 stephenfin Hmm, I just moved nova-lvm and nova-wsgi https://review.openstack.org/#/c/514309/ I don't think nova-next was supposed to move...
15:27:15 stephenfin But I need to investigate. Leave me with it
15:27:49 stephenfin Though, even still, the first two patches _can_ go in. They're tested using existing code
15:28:05 stephenfin As seen by the many CI failures I'd to work through, heh
15:28:53 mriedem well, an easier short-term solution is just push a patch on top of https://review.openstack.org/#/c/396186/ which sets the variable to True in devstack
15:30:00 mriedem btw, not sure why we need the nova-wsgi job in master - everything is running wsgi by default since pike
15:30:19 mriedem so that nova-wsgi job is only useful while working up to wsgi support in pike, right?
15:30:38 stephenfin We don't. I kept it in there so we can backport all the way to Pike. It's removed in a follow-up https://review.openstack.org/#/c/527140/
15:30:54 mriedem but pike also runs wsgi by default in devstack
15:31:01 mriedem so it seems we just don't need it at all
15:31:08 stephenfin Ocata?
15:31:14 mriedem we don't support wsgi mode in ocata
15:31:19 stephenfin Oh, well then
15:31:30 mriedem let me make sure i'm not blowing smoke first
15:32:18 mriedem https://github.com/openstack-dev/devstack/blob/stable/pike/lib/nova#L79
15:33:21 stephenfin It's in Ocata too though https://github.com/openstack-dev/devstack/blob/stable/ocata/lib/nova#L69
15:33:34 stephenfin and set to False
15:33:42 mriedem i think that's for grenade
15:33:57 mriedem well,
15:34:00 mriedem https://github.com/openstack-dev/devstack/commit/d5537c1dc835413f1911ab797e3007d85322eace
15:34:11 mriedem that was added when dims added experimental support for running nova-api under wsgi long ago,
15:34:23 mriedem but we don't support it, and with cellsv2 stuff in ocata, the api literally doesn't work in that omde
15:34:24 mriedem *mode
15:34:27 openstackgerrit Chen Hanxiao proposed openstack/nova master: manager: more detailed info of unsupported compute driver https://review.openstack.org/530701
15:34:29 mriedem which tripleo found out the hard way
15:34:43 mriedem https://github.com/openstack/nova/blob/stable/ocata/nova/wsgi/nova-api.py#L15
15:34:52 mriedem https://github.com/openstack/nova/commit/e846c32ce3c3aee2cd83dec7561dc14cb3f0ada8
15:36:51 mriedem stephenfin: commented on https://review.openstack.org/#/c/514309/
15:36:56 mriedem i think you can just squash those two patches
15:38:42 openstackgerrit Jay Pipes proposed openstack/nova-specs master: Support aggregate affinity filters/weighers https://review.openstack.org/529135
15:39:15 mriedem here is that devstack patch to run with the tls stuff https://review.openstack.org/530136
15:40:00 mriedem err this one i mean https://review.openstack.org/530950
15:46:37 hrw morning
15:47:22 mriedem seems we shouldn't need to trace on an expired token... http://logs.openstack.org/60/513160/4/check/tempest-dsvm-neutron-nova-next-full/6458405/logs/screen-n-novnc.txt.gz?level=TRACE#_Dec_08_23_14_32_787200
15:48:11 hrw can someone tell me does https://bugs.linaro.org/attachment.cgi?id=905 have sense? On AArch64 cpu_mode = host-model results in "libvirtError: unsupported configuration: CPU mode 'host-model' for aarch64 kvm domain on aarch64 host is not supported by hypervisor" and instance is not started.
15:48:43 hrw if I set cpu_mode = 'host-passthrough' in [libvirt] section of nova.conf then it works.
15:49:24 hrw but I would prefer to have it as default rather than work around it in every deploy method
15:50:29 stephenfin mriedem: All the above sounds good. I'll squash those patches and and drop that log.exception to...log.debug, I guess
15:50:55 mriedem stephenfin: i don't think that logging is coming from nova
15:50:59 mriedem i think it's coming from websockify
15:52:29 stephenfin I think this is it https://github.com/openstack/nova/blob/master/nova/console/websocketproxy.py#L156-L157
15:52:42 mriedem right that's raising the exception
15:52:49 stephenfin Ohh, gotcha
15:52:49 mriedem and this is catching it https://github.com/novnc/websockify/blob/v0.8.0/websockify/websocket.py#L938
15:53:29 stephenfin Guess we could could configure
15:53:41 stephenfin *websocket proxy's logger in DevStack
15:53:51 stephenfin *websockify's logger
15:54:03 ericyoung with queens-3 coming up... can someone review https://review.openstack.org/#/c/407440/
15:59:52 openstackgerrit Stephen Finucane proposed openstack/nova master: zuul: Move legacy jobs to project https://review.openstack.org/514309
16:07:34 openstackgerrit Jay Pipes proposed openstack/nova-specs master: Support aggregate affinity filters/weighers https://review.openstack.org/529135
16:10:15 openstackgerrit Claudiu Belu proposed openstack/nova master: hyper-v: Deprecates support for Windows / Hyper-V Server 2012 https://review.openstack.org/530953
16:19:15 melwitt mriedem: hey, thanks for all the reviews on consoles
16:21:23 mriedem stephenfin: couple questions inline https://review.openstack.org/#/c/514309/
16:26:29 stephenfin mriedem: All done
16:26:48 stephenfin tl;dr: It's a copy-paste patch. The real work happens once it's in-tree
16:26:56 stephenfin mordred won't know what hit him
16:28:01 mriedem ok just waiting for the new job to hit the experimental queue
16:33:46 mordred stephenfin: uhoh
16:36:45 stephenfin mordred: Just more work for you that you won't have time for. Nothing to see here :)
16:37:13 mordred stephenfin: \o/
16:37:17 mordred stephenfin: my favorite!!!
16:37:25 dims hey mriedem, did you see the OpenStack++ related papers? (ops / edge computing) they have details on how they extended nova for better live migration scenarios
16:37:44 dims mriedem : [8][9] at https://etherpad.openstack.org/p/2017_edge_computing_working_sessions_existing_gaps
16:38:08 dims LOL stephenfin. right on the money :)
16:38:39 melwitt mriedem: do you know if we have a cells meeting today?
16:38:57 openstackgerrit Stephen Finucane proposed openstack/os-vif master: zuul: Enable functional tests in gate https://review.openstack.org/530961
16:39:09 stephenfin ralonsoh_: ^
16:39:15 mriedem dims: nope haven't seen it
16:39:22 mriedem heads down until feature freeze
16:39:26 mriedem melwitt: i assume so

Earlier   Later