Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-03-03
14:44:58 openstackgerrit Balazs Gibizer proposed openstack/nova master: Remove **kwargs passing in payload __init__ https://review.openstack.org/418489
14:48:14 mdbooth Next question: is it possible to restart a devstack after the host goes down?
14:48:35 mdbooth i.e. simply restart everything running in screen without re-initialising everything
14:50:13 migi figleaf: re that funct test, with help by lyarwood it's now clear there is issue with older sqlite version, updating to newer one fixes issue, so probably it's broken at oslo.db level
14:50:37 figleaf migi: thanks for digging into that
14:53:13 lyarwood migi / figleaf ; yeah oslo_db just passes back an empty list because of this IndexError with the older sqlite version we have in RHEL - https://github.com/openstack/oslo.db/blob/c61d2d47bd39c31f1cd06faf497ffa85ae66cdd4/oslo_db/sqlalchemy/exc_filters.py#L185
14:54:35 figleaf lyarwood: got it. That jibes with what was causing the test to fail.
14:57:56 openstackgerrit Balazs Gibizer proposed openstack/nova master: Improve assertJsonEqual error reporting https://review.openstack.org/428199
14:59:17 openstackgerrit Matt Riedemann proposed openstack/nova-specs master: Libvirt driver emulator threads placement policy https://review.openstack.org/427066
15:18:26 jeblair mriedem: /win 11
15:18:34 jeblair gah
15:18:46 jeblair mriedem: no we're not running zuulv3 yet. :)
15:18:58 mriedem i saw, osic is down
15:19:32 jeblair mriedem: the gist of those emails is that we have made enough progress to run a test instance, zuulv3 is coming but is a little ways off yet. and we'll make much more noise when it gets closer.
15:20:32 mriedem oh so osic changed a flavor?
15:20:37 mriedem sdague: ^?
15:20:48 sdague mriedem: they changed all their flavors
15:21:05 mriedem and nodepool was hardcoded to use the flavor by id or something?
15:21:09 sdague nope
15:21:13 sdague but it caches flavorid
15:21:19 mriedem oh
15:21:47 sdague so it searches for flavors to lookup, but it caches the flavorid with no expiration
15:22:04 openstackgerrit Artom Lifshitz proposed openstack/nova master: Plumbing for tagged nic attachment https://review.openstack.org/415622
15:22:04 openstackgerrit Artom Lifshitz proposed openstack/nova master: Delete nic metadata when detaching interface https://review.openstack.org/435387
15:22:06 openstackgerrit Artom Lifshitz proposed openstack/nova master: Libvirt support for tagged volume attachment https://review.openstack.org/418100
15:22:06 openstackgerrit Artom Lifshitz proposed openstack/nova master: Delete disk metadata when detaching volume https://review.openstack.org/435082
15:22:09 openstackgerrit Artom Lifshitz proposed openstack/nova master: Plumbing for tagged volume attachment https://review.openstack.org/391941
15:22:09 openstackgerrit Artom Lifshitz proposed openstack/nova master: API support for tagged device attachment https://review.openstack.org/376021
15:22:11 openstackgerrit Artom Lifshitz proposed openstack/nova master: Libvirt support for tagged nic attachment https://review.openstack.org/415520
15:22:28 sdague given that until osic this was working against public for pay clouds, it had never experienced "we deleted the flavor"
15:22:41 sdague because if you do that to customers... it all 'splodes
15:22:56 mriedem cdent: figleaf: are you able to answer my question here? https://review.openstack.org/#/c/440656/5
15:24:05 cdent mriedem: short answer: no
15:24:23 mriedem lyarwood: see the issue in https://review.openstack.org/#/c/242602/ ?
15:24:54 cdent mriedem: but guessing I think that the length check was duplicated from somewhere else and this is just removing that redundancy while adding another confirmation of contents
15:24:57 figleaf mriedem: It's just checking that it's there. Why do you think it's random?
15:25:14 mriedem figleaf: because it could be anything right?
15:25:22 mriedem i'm wondering about the pattern going forward
15:25:33 mriedem if not a length check, should all new namespaces be checked that way in the tests?
15:25:37 mriedem or only the most recently added one?
15:25:53 openstackgerrit Michael Henkel proposed openstack/nova master: Implements: blueprint opencontrail-dpdk-support https://review.openstack.org/441183
15:25:54 figleaf mriedem: wait - are you asking about the STORAGE_DISK_SSD?
15:25:58 mriedem yeah
15:26:03 cdent mriedem: that test isn't namespace
15:26:15 figleaf Ah - I was reading your comment a few lines up
15:26:15 cdent it's just asserting it's been defined
15:26:17 lyarwood mriedem: yup, I'll sort it out shortly
15:26:22 mriedem yeah sorry symbol name
15:26:30 figleaf mriedem: FWIW, I'm not a fan of tests like that
15:26:44 figleaf but it seems that Jay is
15:26:44 mriedem i mean i get that it's totally duplicating a constant
15:27:07 mriedem i guess i was trying to see the value
15:27:12 mriedem anyway, need to move rooms
15:32:29 openstackgerrit Artom Lifshitz proposed openstack/nova master: Plumbing for tagged nic attachment https://review.openstack.org/415622
15:32:29 openstackgerrit Artom Lifshitz proposed openstack/nova master: Delete nic metadata when detaching interface https://review.openstack.org/435387
15:32:30 openstackgerrit Artom Lifshitz proposed openstack/nova master: Libvirt support for tagged volume attachment https://review.openstack.org/418100
15:32:31 openstackgerrit Artom Lifshitz proposed openstack/nova master: Delete disk metadata when detaching volume https://review.openstack.org/435082
15:32:33 openstackgerrit Artom Lifshitz proposed openstack/nova master: Plumbing for tagged volume attachment https://review.openstack.org/391941
15:32:33 openstackgerrit Artom Lifshitz proposed openstack/nova master: API support for tagged device attachment https://review.openstack.org/376021
15:32:35 openstackgerrit Artom Lifshitz proposed openstack/nova master: Libvirt support for tagged nic attachment https://review.openstack.org/415520
15:53:16 mriedem cdent: figleaf: i'm sure there has been an api wg discussion before about specific error codes in the api, do you happen to remember if anything was ever proposed
15:53:18 mriedem like a spec
15:53:25 mriedem talking about it in -infra
15:53:40 cdent mriedem: yeah, one sec
15:55:43 leakypipes superdan: won't be until this afternoon that I can get to splitting that Ironic inventory patch. I'd still like to do it, though. OK with ya?
15:57:12 leakypipes superdan: also, wallaby-bait: https://review.openstack.org/#/c/439187/
15:58:36 superdan leakypipes: yar
15:58:47 superdan leakypipes: I could even cater that activity, depending
16:02:02 mriedem leakypipes: https://review.openstack.org/#/c/440656/5/os_traits/tests/test_os_traits.py@35
16:02:10 leakypipes superdan: :)
16:03:29 leakypipes mriedem: answered.
16:03:34 mriedem leakypipes: so remember the regex you have to parse in the scheduler report client to tell if you've hit an inventory in use error?
16:03:41 mriedem which will break if translated or the error msg changes?
16:04:10 mriedem turns out there is a spec for returning a body with error details, which would include a code you can key off of http://specs.openstack.org/openstack/api-wg/guidelines/errors.html
16:04:21 mriedem i wonder if we should start toying with the idea of using that in placement
16:04:51 mriedem while it's small, since trying to retrofit that into nova at this point would be tough
16:04:51 leakypipes mriedem: yes, I'm aware of that spec and encouraged someone to work on that in Pike. can't remember who though...
16:05:04 leakypipes mriedem: agreed.
16:05:23 leakypipes mriedem: I shall put out a call to arms on the ML.
16:05:36 leakypipes mriedem: just as soon as I finish this damn placement PTG wrapup email
16:05:50 mriedem leakypipes: ok meh all around on tha traits test thing :) +2
16:06:01 leakypipes mriedem: yeah, I know :) it's meh
16:06:13 mriedem bauwser: want to get https://review.openstack.org/#/c/440656/ and then i can request a release for os-traits?
16:06:19 leakypipes mriedem: not much to test in "hey, here's some constants" :)
16:06:33 mriedem leakypipes: yeah i knwo
16:06:35 mriedem *know
16:09:17 openstackgerrit Alex Szarka proposed openstack/nova master: Reduce code complexity in vmops._spawn method https://review.openstack.org/438502
16:15:59 openstackgerrit Lee Yarwood proposed openstack/nova master: libvirt: Reconnect volumes and encryptors during a hard reboot https://review.openstack.org/400384
16:16:41 openstackgerrit Lee Yarwood proposed openstack/nova master: libvirt: Check if domain is persistent before detaching volume https://review.openstack.org/441204
16:18:33 openstackgerrit Arnaud Morin proposed openstack/nova master: Add warning on setting secure_proxy_ssl_header https://review.openstack.org/423375
16:18:36 AJaeger could another core +2A the nova translation import, please? The change is fine with the deletions... https://review.openstack.org/#/c/439925/
16:21:31 superdan mriedem: merged
16:22:23 openstackgerrit Maciej Szankin proposed openstack/nova master: conf: fix netconf, my_ip and host are unclear https://review.openstack.org/441210
16:32:36 mriedem lyarwood: have you ever seen anything like this during a swap volume? this is mitaka http://paste.openstack.org/show/601357/
16:32:44 mriedem with https://review.openstack.org/#/c/433793/
16:33:06 mriedem i don't have the libvirt/qemu instance logs so kind of blind
16:35:09 lyarwood looking
16:36:02 lyarwood mriedem: nope, is this from a downstream customer report?
16:36:57 mriedem yeah
16:37:15 mriedem found https://bugzilla.redhat.com/show_bug.cgi?id=1002813 but that's pretty old

Earlier   Later