| *** zz_dimtruck is now known as dimtruck | 00:00 | |
| *** ljxiash has quit IRC | 00:01 | |
| *** qeelee has joined #openstack-nova | 00:01 | |
| *** thorst has quit IRC | 00:02 | |
| *** akshai has joined #openstack-nova | 00:03 | |
| openstackgerrit | Allen Gao proposed openstack/nova: Use IPOpt and PortOpt in vnc options https://review.openstack.org/300716 | 00:06 |
|---|---|---|
| openstackgerrit | Allen Gao proposed openstack/nova: config options: centralize quota options https://review.openstack.org/300940 | 00:07 |
| *** akshai_ has joined #openstack-nova | 00:10 | |
| aNupoisc | diana_clarke: /usr/bin/ | 00:11 |
| *** zenoway has joined #openstack-nova | 00:12 | |
| *** mriedem has left #openstack-nova | 00:12 | |
| aNupoisc | diana_clarke: sys.path is set to /sys/ | 00:13 |
| *** akshai has quit IRC | 00:13 | |
| *** dimtruck is now known as zz_dimtruck | 00:13 | |
| *** pushkaru has quit IRC | 00:16 | |
| *** zenoway has quit IRC | 00:16 | |
| aNupoisc | diana_clarke: hey got it | 00:17 |
| diana_clarke | aNupoisc: nice! where are they? | 00:17 |
| aNupoisc | diana_clarke: its in /usr/local/lib/python-2.7/dist-packages/ | 00:17 |
| aNupoisc | Yeah actually i got that idea after checking PYTHONPATH | 00:18 |
| aNupoisc | diana_clarke: Thanks :) | 00:19 |
| *** mriedem has joined #openstack-nova | 00:19 | |
| diana_clarke | aNupoisc: I've avoided learning the "why" behind that stuff, but I should really read up on it one day. | 00:19 |
| aNupoisc | diana_clarke: hehe.. I am learning. | 00:20 |
| *** thorst has joined #openstack-nova | 00:25 | |
| *** pushkaru has joined #openstack-nova | 00:25 | |
| *** mtanino__ has joined #openstack-nova | 00:27 | |
| *** mtanino has quit IRC | 00:28 | |
| diana_clarke | aNupoisc: me too :) | 00:28 |
| *** nic has quit IRC | 00:31 | |
| *** ctrath has quit IRC | 00:31 | |
| *** pushkaru has quit IRC | 00:32 | |
| *** vtech has quit IRC | 00:33 | |
| *** mtanino__ has quit IRC | 00:33 | |
| mriedem | wow where does this come from? http://docs.openstack.org/developer/nova/runnova/managing.instance.types.html | 00:34 |
| *** hoangcx has joined #openstack-nova | 00:35 | |
| mriedem | i guess that's just super old devref | 00:35 |
| mriedem | weird | 00:35 |
| tbachman | 2012 | 00:36 |
| tbachman | mriedem: unearthed a time capsule there | 00:36 |
| mriedem | docs.openstack.org search did actually | 00:37 |
| *** erlon has quit IRC | 00:37 | |
| mriedem | was looking for docs that reference the default flavors | 00:37 |
| *** ivase has joined #openstack-nova | 00:37 | |
| mriedem | since the docs need to be updated https://bugs.launchpad.net/openstack-manuals/+bug/1567009 | 00:37 |
| openstack | Launchpad bug 1567009 in openstack-manuals " Remove flavor seeding from the base migration" [Undecided,New] | 00:37 |
| *** akshai_ has quit IRC | 00:37 | |
| *** ivase_ has quit IRC | 00:39 | |
| *** achanda has quit IRC | 00:40 | |
| *** vilobhmm111 has quit IRC | 00:41 | |
| *** sc68cal has quit IRC | 00:42 | |
| *** Marga_ has quit IRC | 00:43 | |
| *** thorst has quit IRC | 00:45 | |
| *** thorst has joined #openstack-nova | 00:46 | |
| *** vilobhmm11 has joined #openstack-nova | 00:46 | |
| *** Marga_ has joined #openstack-nova | 00:47 | |
| *** browne has quit IRC | 00:47 | |
| *** ivase_ has joined #openstack-nova | 00:47 | |
| *** zenoway has joined #openstack-nova | 00:48 | |
| *** Jeffrey4l has joined #openstack-nova | 00:48 | |
| *** tbachman has quit IRC | 00:50 | |
| *** tbachman has joined #openstack-nova | 00:50 | |
| *** ivase has quit IRC | 00:50 | |
| *** dave-mcc_ has quit IRC | 00:51 | |
| *** Marga_ has quit IRC | 00:51 | |
| *** artom has quit IRC | 00:51 | |
| openstackgerrit | Matt Riedemann proposed openstack/python-novaclient: Deprecate image list/show/delete/update CLIs/APIs https://review.openstack.org/302486 | 00:52 |
| *** dave-mccowan has joined #openstack-nova | 00:54 | |
| *** zz_dimtruck is now known as dimtruck | 00:54 | |
| *** thorst has quit IRC | 00:54 | |
| *** zenoway has quit IRC | 00:54 | |
| *** harlowja has quit IRC | 01:00 | |
| *** hieulq has joined #openstack-nova | 01:00 | |
| *** dave-mccowan has quit IRC | 01:03 | |
| *** harlowja has joined #openstack-nova | 01:03 | |
| *** zenoway has joined #openstack-nova | 01:03 | |
| *** rmevans has joined #openstack-nova | 01:04 | |
| *** rmevans has quit IRC | 01:05 | |
| *** rmevans has joined #openstack-nova | 01:06 | |
| *** zenoway has quit IRC | 01:08 | |
| *** Marga_ has joined #openstack-nova | 01:10 | |
| *** ljxiash has joined #openstack-nova | 01:10 | |
| *** Marga_ has quit IRC | 01:14 | |
| mriedem | wow, this is a pretty interesting bug https://bugs.launchpad.net/nova/+bug/1567114 | 01:16 |
| openstack | Launchpad bug 1567114 in OpenStack Compute (nova) "nova boot specifying flavor by name fails intermittently past 1000 flavors" [High,Triaged] | 01:16 |
| mriedem | resource filtering in novaclient is uber inefficient | 01:16 |
| *** achanda has joined #openstack-nova | 01:18 | |
| lifeless | alaski: fixtures 2.0.0 released | 01:20 |
| melwitt | mriedem: I've seen similar things before, if you don't send a filter query param then novaclient will get back a varying 1000 item subset of items based on the osapi_max_limit and not find the thing by name | 01:22 |
| mriedem | melwitt: yup | 01:23 |
| mriedem | melwitt: of course he could just workaround it by sending the flavor id in on boot | 01:23 |
| mriedem | but i didn't realize the lookup by name was so inefficient | 01:23 |
| melwitt | mriedem: when I saw it, it happened because I was implementing something internal that did a server get by hostname and I didn't realize nova silently removes "not allowed" search opts if not admin | 01:24 |
| *** rmevans has quit IRC | 01:24 | |
| melwitt | mriedem: so it was doing a get_all and sometimes the desired server was in the set and sometimes not | 01:25 |
| mriedem | lifeless: thanks, i'll request a g-r update for that | 01:25 |
| mriedem | melwitt: yeah that's what's happening to this guy too | 01:25 |
| mriedem | well, w/o the admin issue | 01:25 |
| *** rmevans has joined #openstack-nova | 01:26 | |
| *** qeelee has quit IRC | 01:27 | |
| melwitt | mriedem: yeah, I thought someone did some improvements a long time ago on the novaclient side for common things like "name" to send the name as a query param. I'd have to go check what that was | 01:28 |
| mriedem | melwitt: it's in the findall code, but only if there is a search_opts in kwargs | 01:28 |
| mriedem | which utils.find_resource doesn't send | 01:28 |
| melwitt | ah, okay | 01:28 |
| *** sdake has quit IRC | 01:28 | |
| *** qeelee has joined #openstack-nova | 01:29 | |
| *** EinstCrazy has joined #openstack-nova | 01:29 | |
| *** aysyd has quit IRC | 01:30 | |
| *** sdake has joined #openstack-nova | 01:32 | |
| *** ducttape_ has joined #openstack-nova | 01:34 | |
| mriedem | i think i see the issue | 01:38 |
| mriedem | FlavorManager.list method has to have a search_opts arg | 01:38 |
| mriedem | because findall in the base manager checks the args spec | 01:38 |
| mriedem | and only adds the name to the filter if search_opts is in the list method's list of formal parameters | 01:39 |
| *** aNupoisc has quit IRC | 01:41 | |
| melwitt | mriedem: I'm trying to see in db api, is it even possible to filter on flavor_get_all by name? maybe I'm missing something but I don't see it | 01:42 |
| *** hieulq has quit IRC | 01:42 | |
| *** hieulq has joined #openstack-nova | 01:42 | |
| melwitt | okay, there's flavor_get_by_name for that | 01:42 |
| *** ivase has joined #openstack-nova | 01:43 | |
| mriedem | melwitt: not from the api https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/flavors.py#L78 | 01:43 |
| mriedem | the flavor keys are in there | 01:43 |
| *** baoli has joined #openstack-nova | 01:44 | |
| melwitt | yeah. that's what I was trying to see if it's even possible for novaclient to do better, if nova api gives a way to query a flavor indicating the name | 01:45 |
| *** ivase_ has quit IRC | 01:45 | |
| mriedem | so, then it's a question of whether or not we even care to fix this, | 01:46 |
| mriedem | it's kind of a busted test scenario, creating 1000+ flavors to boot 1000+ servers | 01:46 |
| mriedem | i'm not sure how many deployments are going to have 1000 flavors | 01:46 |
| mriedem | we certainly assumed it wouldn't be many when doing the online data migration for flavors to the API DB | 01:46 |
| *** baoli has quit IRC | 01:47 | |
| melwitt | hah, yeah. 1000 is more than anything I've heard so far by an order of magnitude | 01:47 |
| *** baoli has joined #openstack-nova | 01:48 | |
| mriedem | alright, marked it won't fix | 01:48 |
| *** ducttape_ has quit IRC | 01:48 | |
| melwitt | I mean, it would be cool in the nova api to be able to give name as a search opt probably. but it seems like that's what it would take to fix this. and I wonder how many other novaclient "get by name" things have the same problem for the same reason, for different apis | 01:48 |
| *** lixiaoy1 has joined #openstack-nova | 01:50 | |
| melwitt | mriedem: if it's a test scenario it feels like they could easily use the uuid to do the boots | 01:50 |
| mriedem | melwitt: that's what i told him | 01:51 |
| mriedem | as the workaround | 01:51 |
| melwitt | I hope that's part of the response payload anyway | 01:51 |
| mriedem | melwitt: i was thinking the same thing about the filtering issue in other apis | 01:51 |
| *** qeelee has quit IRC | 01:51 | |
| mriedem | we'd have to add a microversion and everything, it just doesn't seem like a case we need to pursue for flavors | 01:51 |
| *** thorst has joined #openstack-nova | 01:51 | |
| mriedem | when you can just get it by id | 01:51 |
| melwitt | for creating a flavor I mean | 01:52 |
| *** qeelee has joined #openstack-nova | 01:52 | |
| mriedem | melwitt: it is | 01:53 |
| mriedem | https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/views/flavors.py#L34 | 01:53 |
| mriedem | https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/flavor_manage.py#L95 | 01:53 |
| melwitt | oh, id == flavorid. I wasn't sure what that meant when I saw it in the api ref | 01:54 |
| melwitt | was looking around in the code and not found it yet. thanks | 01:54 |
| mriedem | yes, super confusing | 01:55 |
| mriedem | flavor.id, flavor.flavorid | 01:55 |
| melwitt | sigh. yeah | 01:57 |
| *** lixiaoy11 has joined #openstack-nova | 01:57 | |
| *** thorst has quit IRC | 01:59 | |
| *** lixiaoy1 has quit IRC | 01:59 | |
| smcginnis | flava.flav? | 02:01 |
| *** hoangcx has quit IRC | 02:02 | |
| *** apoorvad has quit IRC | 02:02 | |
| *** salv-orlando has joined #openstack-nova | 02:03 | |
| *** hoangcx has joined #openstack-nova | 02:03 | |
| *** baoli has quit IRC | 02:06 | |
| mriedem | smcginnis: yes, nova doesn't have many flavors because the clock necklace budget is tight | 02:07 |
| *** ircuser-1 has quit IRC | 02:08 | |
| *** wuhg has joined #openstack-nova | 02:13 | |
| mriedem | man, triaged 3 bugs to invalid in the last hour | 02:15 |
| mriedem | markus_z: ^ good times | 02:15 |
| mriedem | need to call it a night | 02:15 |
| *** ivase_ has joined #openstack-nova | 02:15 | |
| *** mriedem has quit IRC | 02:16 | |
| *** browne has joined #openstack-nova | 02:16 | |
| *** ivase has quit IRC | 02:18 | |
| *** sdake_ has joined #openstack-nova | 02:20 | |
| *** ducttape_ has joined #openstack-nova | 02:20 | |
| *** sdake has quit IRC | 02:22 | |
| *** gongysh has quit IRC | 02:22 | |
| *** gongysh has joined #openstack-nova | 02:22 | |
| *** salv-orlando has quit IRC | 02:23 | |
| *** salv-orlando has joined #openstack-nova | 02:28 | |
| *** houming has joined #openstack-nova | 02:28 | |
| *** gongysh has quit IRC | 02:32 | |
| *** gongysh has joined #openstack-nova | 02:32 | |
| *** salv-orlando has quit IRC | 02:33 | |
| openstackgerrit | Zhenguo Niu proposed openstack/nova: virt: add method get_serial_console to ironic driver https://review.openstack.org/302499 | 02:34 |
| *** EinstCrazy has quit IRC | 02:35 | |
| *** EinstCrazy has joined #openstack-nova | 02:36 | |
| *** edmondsw has quit IRC | 02:36 | |
| *** _RuiChen has quit IRC | 02:43 | |
| *** lixiaoy11 has quit IRC | 02:43 | |
| *** RuiChen has joined #openstack-nova | 02:44 | |
| gmann | alex_xu: snikitin : regarding https://review.openstack.org/#/c/128940/127 | 02:45 |
| openstackgerrit | Augustina Ragwitz proposed openstack/nova: Result of RST files from wadl2rst https://review.openstack.org/302500 | 02:45 |
| gmann | alex_xu: snikitin : sorry to put on last moment but i have some query on schema validation part. added my comment inline | 02:46 |
| gmann | will be helpful if anything discussion i missed on those and happy to convert my vote :) | 02:46 |
| *** gcb has quit IRC | 02:47 | |
| alex_xu | gmann: I remember I have similar comments before, but i can't remember clear now, i guess snikitin want to provide a better error message at here. | 02:49 |
| gmann | alex_xu: humm, but from json validation also may be we can get error message for each invalid tag | 02:50 |
| *** ivase has joined #openstack-nova | 02:51 | |
| gmann | alex_xu: and as we have schema validation also there i do not think it will come to those hard coded validation if any invalid tag | 02:51 |
| gmann | so that seems like dead code? | 02:51 |
| alex_xu | gmann: ah, I find some history https://review.openstack.org/#/c/128940/32/nova/api/openstack/compute/plugins/v3/instance_tags.py@71 | 02:53 |
| *** ivase_ has quit IRC | 02:54 | |
| alex_xu | gmann: if expose json-schema into the api ref or expose with other way, it is better validate in the json-schema | 02:56 |
| *** thorst has joined #openstack-nova | 02:56 | |
| gmann | alex_xu: I see, do not json validation error through schema also when having error, i mean max length of array? | 02:58 |
| *** mdnadeem has joined #openstack-nova | 02:59 | |
| gmann | I need to check that tough as we have that max items in attach_interfaces schema | 02:59 |
| *** eliqiao has quit IRC | 03:01 | |
| *** gongysh has quit IRC | 03:01 | |
| *** openstackstatus has quit IRC | 03:01 | |
| alex_xu | gmann: emm...sorry, I didn't get your point | 03:01 |
| *** eliqiao has joined #openstack-nova | 03:02 | |
| gmann | alex_xu: sorry :), I mean for better error message, i think snikitin want to show the max limit | 03:02 |
| gmann | and json schema validation should dump schema with max allowed items, but I am checking that if so | 03:03 |
| alex_xu | gmann: ah, i got you now | 03:03 |
| *** thorst has quit IRC | 03:04 | |
| *** sdake_ has quit IRC | 03:06 | |
| *** sdake has joined #openstack-nova | 03:07 | |
| *** gcb has joined #openstack-nova | 03:08 | |
| *** gongysh has joined #openstack-nova | 03:08 | |
| *** achanda has quit IRC | 03:09 | |
| *** sdake has quit IRC | 03:17 | |
| *** sdake has joined #openstack-nova | 03:18 | |
| *** ducttape_ has quit IRC | 03:19 | |
| *** ducttape_ has joined #openstack-nova | 03:20 | |
| *** ivase_ has joined #openstack-nova | 03:20 | |
| *** ivase has quit IRC | 03:23 | |
| *** sdake has quit IRC | 03:28 | |
| *** ivase has joined #openstack-nova | 03:29 | |
| *** gongysh has quit IRC | 03:30 | |
| *** ivase_ has quit IRC | 03:32 | |
| *** gongysh has joined #openstack-nova | 03:32 | |
| *** dimtruck is now known as zz_dimtruck | 03:32 | |
| *** vtech has joined #openstack-nova | 03:33 | |
| *** efried has quit IRC | 03:39 | |
| *** vtech has quit IRC | 03:41 | |
| *** sheel has joined #openstack-nova | 03:41 | |
| *** efried has joined #openstack-nova | 03:41 | |
| *** houming_ has joined #openstack-nova | 03:41 | |
| *** houming has quit IRC | 03:42 | |
| *** houming_ is now known as houming | 03:42 | |
| *** mylu has quit IRC | 03:45 | |
| openstackgerrit | Allen Gao proposed openstack/nova: config options: centralize 'spice' options https://review.openstack.org/264271 | 03:48 |
| *** ivase_ has joined #openstack-nova | 03:49 | |
| *** salv-orlando has joined #openstack-nova | 03:50 | |
| *** vladikr has quit IRC | 03:50 | |
| *** sekrit has quit IRC | 03:51 | |
| *** houming_ has joined #openstack-nova | 03:51 | |
| *** houming has quit IRC | 03:51 | |
| *** houming_ is now known as houming | 03:51 | |
| allen_gao | alex_xu: would you take a minute to look at this when you find time? thank you. https://review.openstack.org/#/c/300940/ | 03:52 |
| *** salv-orlando has quit IRC | 03:52 | |
| *** ivase has quit IRC | 03:53 | |
| *** salv-orlando has joined #openstack-nova | 03:53 | |
| *** qeelee has quit IRC | 03:55 | |
| *** Marga_ has joined #openstack-nova | 03:56 | |
| *** dave-mccowan has joined #openstack-nova | 03:57 | |
| *** Marga_ has quit IRC | 03:57 | |
| *** Marga_ has joined #openstack-nova | 03:57 | |
| *** salv-orlando has quit IRC | 03:58 | |
| *** dave-mcc_ has joined #openstack-nova | 03:58 | |
| *** ducttape_ has quit IRC | 04:00 | |
| *** dave-mccowan has quit IRC | 04:01 | |
| *** Marga_ has quit IRC | 04:01 | |
| *** thorst has joined #openstack-nova | 04:01 | |
| *** mylu has joined #openstack-nova | 04:02 | |
| *** Marga_ has joined #openstack-nova | 04:02 | |
| *** sekrit has joined #openstack-nova | 04:05 | |
| *** thorst has quit IRC | 04:09 | |
| *** Sukhdev has joined #openstack-nova | 04:11 | |
| *** links has joined #openstack-nova | 04:12 | |
| *** rmevans has quit IRC | 04:17 | |
| *** amotoki has joined #openstack-nova | 04:20 | |
| *** achanda has joined #openstack-nova | 04:22 | |
| *** achanda has quit IRC | 04:37 | |
| *** irenab has quit IRC | 04:37 | |
| *** ivase has joined #openstack-nova | 04:42 | |
| *** eliqiao_ has joined #openstack-nova | 04:43 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova: Pass BDM in swap_volume method of LibvirtDriver https://review.openstack.org/257135 | 04:43 |
| *** eliqiao has quit IRC | 04:44 | |
| *** qeelee has joined #openstack-nova | 04:44 | |
| *** mylu has quit IRC | 04:45 | |
| *** ivase_ has quit IRC | 04:45 | |
| *** xyang1 has quit IRC | 04:48 | |
| alex_xu | allen_gao: done | 04:52 |
| *** salv-orlando has joined #openstack-nova | 04:57 | |
| *** armax has quit IRC | 04:59 | |
| *** salv-orlando has quit IRC | 05:01 | |
| *** achanda has joined #openstack-nova | 05:03 | |
| *** thorst has joined #openstack-nova | 05:06 | |
| *** salv-orlando has joined #openstack-nova | 05:11 | |
| *** jaosorior has joined #openstack-nova | 05:13 | |
| *** wanghao has quit IRC | 05:13 | |
| *** wanghao has joined #openstack-nova | 05:13 | |
| *** thorst has quit IRC | 05:13 | |
| *** gongysh has quit IRC | 05:17 | |
| *** gongysh has joined #openstack-nova | 05:19 | |
| *** zer0c00l has quit IRC | 05:20 | |
| yuywz | dims: can you please remove your -workflow for this if got time? https://review.openstack.org/#/c/287026 | 05:22 |
| *** itamarl has joined #openstack-nova | 05:30 | |
| *** esracelik has joined #openstack-nova | 05:33 | |
| *** achanda has quit IRC | 05:33 | |
| *** irenab has joined #openstack-nova | 05:36 | |
| *** hieulq has quit IRC | 05:37 | |
| *** PaulMurray has joined #openstack-nova | 05:38 | |
| *** Daisy has joined #openstack-nova | 05:39 | |
| *** hieulq has joined #openstack-nova | 05:39 | |
| *** Sukhdev has quit IRC | 05:40 | |
| *** PaulMurray has quit IRC | 05:40 | |
| *** rcernin has joined #openstack-nova | 05:45 | |
| *** zhangjn has quit IRC | 05:45 | |
| *** zhangjn has joined #openstack-nova | 05:46 | |
| *** garyk1 has joined #openstack-nova | 05:46 | |
| *** nkrinner has joined #openstack-nova | 05:49 | |
| openstackgerrit | huan proposed openstack/nova: XenAPI: Do not rely on XenAPI to get device names https://review.openstack.org/213112 | 05:49 |
| *** PaulMurray has joined #openstack-nova | 05:50 | |
| *** moshele has joined #openstack-nova | 05:50 | |
| *** jwcroppe has joined #openstack-nova | 05:50 | |
| *** jwcroppe has quit IRC | 05:57 | |
| *** sridharg has joined #openstack-nova | 05:58 | |
| *** achanda has joined #openstack-nova | 05:59 | |
| *** ircuser-1 has joined #openstack-nova | 06:04 | |
| *** lpetrut has joined #openstack-nova | 06:05 | |
| *** ivase_ has joined #openstack-nova | 06:08 | |
| *** ivase has quit IRC | 06:10 | |
| *** thorst has joined #openstack-nova | 06:11 | |
| *** gongysh has quit IRC | 06:13 | |
| *** unicell has quit IRC | 06:13 | |
| *** gongysh has joined #openstack-nova | 06:14 | |
| *** unicell has joined #openstack-nova | 06:15 | |
| *** paul-carlton1 has joined #openstack-nova | 06:18 | |
| *** gongysh has quit IRC | 06:18 | |
| *** thorst has quit IRC | 06:18 | |
| openstackgerrit | Moshe Levi proposed openstack/nova: ironic driver: Adding InfiniBand support https://review.openstack.org/266540 | 06:19 |
| *** achanda has quit IRC | 06:22 | |
| *** scheuran has joined #openstack-nova | 06:22 | |
| openstackgerrit | Allen Gao proposed openstack/nova: config options: centralize 'spice' options https://review.openstack.org/264271 | 06:23 |
| openstackgerrit | Srushti Gadadare proposed openstack/nova: Fix invalid uuid warnings in cell api testcases https://review.openstack.org/300482 | 06:24 |
| *** moshele has quit IRC | 06:24 | |
| *** pcaruana has joined #openstack-nova | 06:26 | |
| *** eliqiao_ is now known as eliqiao | 06:28 | |
| *** gongysh has joined #openstack-nova | 06:30 | |
| *** ekuris has joined #openstack-nova | 06:30 | |
| PaulMurray | pkoniszewski, hi - you online | 06:34 |
| *** p01nt3r75 has quit IRC | 06:35 | |
| *** andrearosa has joined #openstack-nova | 06:36 | |
| andrearosa | o/ | 06:38 |
| *** apoorv has joined #openstack-nova | 06:38 | |
| *** sridharg has quit IRC | 06:38 | |
| *** vtech has joined #openstack-nova | 06:40 | |
| *** houming has quit IRC | 06:42 | |
| *** houming has joined #openstack-nova | 06:44 | |
| *** tesseract has joined #openstack-nova | 06:45 | |
| *** tesseract is now known as Guest12876 | 06:45 | |
| openstackgerrit | Srushti Gadadare proposed openstack/nova: Fix invalid uuid warnings in api testcases https://review.openstack.org/300477 | 06:46 |
| *** sridharg has joined #openstack-nova | 06:47 | |
| *** sridharg has quit IRC | 06:47 | |
| *** sridharg has joined #openstack-nova | 06:50 | |
| *** salv-orlando has quit IRC | 06:52 | |
| *** qeelee has quit IRC | 06:53 | |
| *** houming has quit IRC | 06:56 | |
| *** qeelee has joined #openstack-nova | 06:56 | |
| *** woodster_ has quit IRC | 06:57 | |
| *** EinstCra_ has joined #openstack-nova | 06:59 | |
| *** ekuris has quit IRC | 07:01 | |
| *** qeelee has quit IRC | 07:02 | |
| *** EinstCrazy has quit IRC | 07:02 | |
| *** qeelee has joined #openstack-nova | 07:04 | |
| *** Marga_ has quit IRC | 07:04 | |
| *** vtech has joined #openstack-nova | 07:04 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Imported Translations from Zanata https://review.openstack.org/301478 | 07:05 |
| *** dave-mcc_ has quit IRC | 07:09 | |
| *** atuvenie has joined #openstack-nova | 07:11 | |
| *** qeelee has quit IRC | 07:11 | |
| *** ihrachys has joined #openstack-nova | 07:11 | |
| *** jlanoux has joined #openstack-nova | 07:13 | |
| *** qeelee has joined #openstack-nova | 07:13 | |
| *** jichen has joined #openstack-nova | 07:15 | |
| *** moshele has joined #openstack-nova | 07:15 | |
| *** thorst has joined #openstack-nova | 07:16 | |
| *** ekuris has joined #openstack-nova | 07:17 | |
| openstackgerrit | Cindia-blue proposed openstack/nova: VMware: Enable Fault tolerance on VM https://review.openstack.org/302552 | 07:19 |
| *** qeelee has quit IRC | 07:19 | |
| *** lpetrut has quit IRC | 07:21 | |
| openstackgerrit | Cindia-blue proposed openstack/nova: VMware: Enable Fault tolerance on VM https://review.openstack.org/251807 | 07:22 |
| *** thorst has quit IRC | 07:24 | |
| *** Daisy has quit IRC | 07:24 | |
| *** Daisy has joined #openstack-nova | 07:25 | |
| *** qeelee has joined #openstack-nova | 07:27 | |
| *** matrohon has joined #openstack-nova | 07:29 | |
| *** Daisy has quit IRC | 07:29 | |
| *** daemontool has joined #openstack-nova | 07:32 | |
| *** ekuris has quit IRC | 07:35 | |
| *** mjura has joined #openstack-nova | 07:36 | |
| *** Oku_OS-away is now known as Oku_OS | 07:38 | |
| *** dedery has joined #openstack-nova | 07:38 | |
| *** jichen has quit IRC | 07:43 | |
| *** jamielennox is now known as jamielennox|away | 07:45 | |
| *** Daisy has joined #openstack-nova | 07:45 | |
| *** browne has quit IRC | 07:47 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Merge tag '13.0.0' https://review.openstack.org/302578 | 07:47 |
| *** jaosorior has quit IRC | 07:47 | |
| *** jaosorior has joined #openstack-nova | 07:47 | |
| *** ekuris has joined #openstack-nova | 07:48 | |
| *** zenoway has joined #openstack-nova | 07:50 | |
| *** matrohon_ has joined #openstack-nova | 07:51 | |
| *** matrohon_ has quit IRC | 07:51 | |
| *** jwcroppe has joined #openstack-nova | 07:54 | |
| *** salv-orlando has joined #openstack-nova | 07:54 | |
| *** houming has joined #openstack-nova | 07:56 | |
| *** jwcroppe has quit IRC | 07:58 | |
| *** qeelee has quit IRC | 07:58 | |
| *** qeelee has joined #openstack-nova | 08:02 | |
| *** haukebruno has joined #openstack-nova | 08:03 | |
| *** atuvenie has quit IRC | 08:04 | |
| *** yassine__ has joined #openstack-nova | 08:07 | |
| *** apoorv has quit IRC | 08:08 | |
| *** ivase has joined #openstack-nova | 08:08 | |
| *** qeelee has quit IRC | 08:09 | |
| *** qeelee has joined #openstack-nova | 08:09 | |
| *** salv-orlando has quit IRC | 08:10 | |
| *** ivase_ has quit IRC | 08:11 | |
| *** ljxiash has quit IRC | 08:14 | |
| *** Jeffrey4l has quit IRC | 08:16 | |
| *** ljxiash has joined #openstack-nova | 08:19 | |
| *** thorst has joined #openstack-nova | 08:21 | |
| *** Jeffrey4l has joined #openstack-nova | 08:22 | |
| *** salv-orlando has joined #openstack-nova | 08:26 | |
| *** openstackstatus has joined #openstack-nova | 08:29 | |
| *** ChanServ sets mode: +v openstackstatus | 08:29 | |
| *** thorst has quit IRC | 08:30 | |
| *** artom has joined #openstack-nova | 08:30 | |
| *** jistr has joined #openstack-nova | 08:30 | |
| -openstackstatus- NOTICE: jobs depending on npm are now working again | 08:34 | |
| paul-carlton1 | when live migrating an instance that is deployed in an environment where the compute nodes have shared storage you don't need to specify --block-migrate, right? | 08:36 |
| *** apoorv has joined #openstack-nova | 08:36 | |
| *** p01nt3r75 has joined #openstack-nova | 08:39 | |
| andrearosa | paul-carlton1: correct, if you specify it with a shared storage you get an exception in one fo the recent changes we landed in M it's better to not specify that flag and nova will work it out for you | 08:40 |
| *** daemontool has quit IRC | 08:41 | |
| andrearosa | paul-carlton1: to use that feature you must have all the env updated to Mitaka code | 08:41 |
| *** gszasz has joined #openstack-nova | 08:43 | |
| *** rmart04 has joined #openstack-nova | 08:45 | |
| pkoniszewski | PaulMurray: hi, just got to my desk | 08:46 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova: Remove nova-manage service subcommand https://review.openstack.org/299307 | 08:47 |
| PaulMurray | pkoniszewski, hi - I put a couple of comments on the etherpad - really busy today | 08:47 |
| PaulMurray | pkoniszewski, andrearosa prefers 8 UTC tomorrow, does that work for you? | 08:48 |
| pkoniszewski | PaulMurray: yeah, it does | 08:48 |
| PaulMurray | ok, lets make it that | 08:48 |
| *** ndipanov has joined #openstack-nova | 08:50 | |
| *** cdent has joined #openstack-nova | 08:50 | |
| andrearosa | PaulMurray: pkoniszewski: works for me as well | 08:50 |
| *** e0ne has joined #openstack-nova | 08:51 | |
| *** EinstCra_ is now known as EinstCrazy | 08:51 | |
| *** danpb has joined #openstack-nova | 08:51 | |
| PaulMurray | pkoniszewski, could you create a hangout and send us the link - might be easier that way | 08:51 |
| pkoniszewski | PaulMurray: both ideas of introduction are good I think, so you can choose yours or andrearosa's idea, im ok with both | 08:52 |
| pkoniszewski | PaulMurray: sure | 08:52 |
| *** salv-orl_ has joined #openstack-nova | 08:52 | |
| PaulMurray | pkoniszewski, I'll try to put a couple of slides together for that part, but might not be able to before we talk tomorrow | 08:53 |
| *** salv-orlando has quit IRC | 08:55 | |
| *** ivase has quit IRC | 08:57 | |
| *** ivase has joined #openstack-nova | 08:57 | |
| *** jaosorior has quit IRC | 08:59 | |
| *** jaosorior has joined #openstack-nova | 09:00 | |
| openstackgerrit | John Garbutt proposed openstack/nova: Move config options from nova/network/linux_net.py https://review.openstack.org/270441 | 09:05 |
| *** derekh has joined #openstack-nova | 09:09 | |
| *** vilobhmm11 has quit IRC | 09:09 | |
| openstackgerrit | Timofey Durakov proposed openstack/nova: Cleanup for post/rollback of live-migrations https://review.openstack.org/280653 | 09:11 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: New field for LibvirtLiveMigrateData https://review.openstack.org/286745 | 09:11 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: move disk_info calc to check methods https://review.openstack.org/291160 | 09:11 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: implement async rpc for live-migration https://review.openstack.org/291161 | 09:11 |
| openstackgerrit | Timofey Durakov proposed openstack/nova: live_migration_cleanup method introduced https://review.openstack.org/302611 | 09:11 |
| *** eglynn has joined #openstack-nova | 09:13 | |
| johnthetubaguy | alex_xu gmann: I was just looking at the comments here about validation: https://review.openstack.org/#/c/128940/127/nova/api/openstack/compute/server_tags.py | 09:22 |
| gmann | johnthetubaguy: hi | 09:22 |
| gmann | johnthetubaguy: actually my main concern is not to add validation in python code | 09:23 |
| johnthetubaguy | we had some patches to improve validation error messages recently, can you remember what we did? | 09:23 |
| johnthetubaguy | I think we changed the validators slightly, custom ones, or something like that? | 09:23 |
| gmann | ohh, i cannot remember may be alex_xu ? | 09:24 |
| johnthetubaguy | https://review.openstack.org/#/c/281143 | 09:24 |
| johnthetubaguy | except we reverted it | 09:24 |
| johnthetubaguy | ah, but then we had this | 09:25 |
| johnthetubaguy | https://review.openstack.org/#/c/288260/ | 09:25 |
| gmann | johnthetubaguy: i can see those in i think not reverted | 09:25 |
| johnthetubaguy | gmann: I am thinking about just merging what they have right now, then getting someone to work, as a follow patch, to add those nice validators? | 09:27 |
| *** thorst has joined #openstack-nova | 09:27 | |
| johnthetubaguy | gmann: if these poor folk hadn't been trying to get this merged for so many years, I would probably ask for a different approach, just feels like its time to make progress | 09:28 |
| *** lpetrut has joined #openstack-nova | 09:28 | |
| *** links has quit IRC | 09:31 | |
| *** alex_xu has quit IRC | 09:32 | |
| *** hoangcx has quit IRC | 09:32 | |
| *** Yingxin has quit IRC | 09:32 | |
| *** Yingxin has joined #openstack-nova | 09:33 | |
| *** thorst has quit IRC | 09:34 | |
| *** hoangcx has joined #openstack-nova | 09:35 | |
| *** alex_xu has joined #openstack-nova | 09:36 | |
| Anticimex | deployer here. can i run multiple nova-scheduler active/active for ha? | 09:38 |
| Anticimex | or is it, like cinder-scheduler, necessary to have max 1 running simultaneously, with e.g. corosync/pacemaker etc | 09:38 |
| *** Daisy has quit IRC | 09:40 | |
| *** mvk has joined #openstack-nova | 09:41 | |
| johnthetubaguy | Anticimex: technically yes, but then the scheduler makes bad decisions, more so as the rate of builds increases, in the next release or so we should have that fixed, so you can do active/active | 09:42 |
| gmann | johnthetubaguy: ok, i can see your point. I am ok with that plan. I will add follow up patch and we can discuss there | 09:42 |
| Anticimex | johnthetubaguy: seems similar to cinder-scheduler situation then. it's not helpful that https://wiki.openstack.org/wiki/HAGuideImprovements/TOC says active/active is fine | 09:43 |
| johnthetubaguy | gmann: OK, cool, normally I would totally pile on with a -1, but I don't think we have a good pattern for folks to copy, easily, yet | 09:43 |
| *** mkrcmari__ has quit IRC | 09:43 | |
| gmann | johnthetubaguy: yea that was pending since too long, even i thought of that while -1 :) | 09:44 |
| johnthetubaguy | Anticimex: hmm, not sure I have not see that wiki before, what does the HA guide say? the same thing? | 09:44 |
| *** links has joined #openstack-nova | 09:44 | |
| openstackgerrit | Gleb Stepanov proposed openstack/nova: Check API versions intersects https://review.openstack.org/301278 | 09:44 |
| johnthetubaguy | Anticimex: ah, so the guide doesn't have all the details get I guess, I am looking at this: http://docs.openstack.org/ha-guide/ | 09:45 |
| Anticimex | no the guides are pretty slim on details | 09:46 |
| *** ptm_away has joined #openstack-nova | 09:47 | |
| Anticimex | lol, saying you use keepalived to move ip addresses from host to host in active/active setup. total confusion | 09:47 |
| Anticimex | johnthetubaguy: at what sort of build rates have people reported the bad decisions? | 09:48 |
| johnthetubaguy | Anticimex: well some things do partial active/active like single active writer, multiple reader, like with galera to avoid some issues | 09:48 |
| *** mvk_ has joined #openstack-nova | 09:49 | |
| Anticimex | johnthetubaguy: hm, example of galera multiple active writers issues? | 09:49 |
| *** gongysh has quit IRC | 09:49 | |
| johnthetubaguy | so select for update causes locking issues, I think someone said you avoid that by having a single node take write requests, or something like that | 09:50 |
| johnthetubaguy | although most projects are working to avoid select for update now | 09:50 |
| *** gongysh has joined #openstack-nova | 09:50 | |
| *** PaulMurray has quit IRC | 09:50 | |
| johnthetubaguy | Anticimex: so on the bad scheduler decisions, its all about how much contention your scheduler rules create, if you spread VMs across your fleet, and their is loads of space, and you turn up the randomness on the pick 1 for top N, you might be just fine with lots of builds and two scheduler nodes | 09:51 |
| *** ptm_away has quit IRC | 09:52 | |
| johnthetubaguy | Anticimex: the work we are doing is really to get rid of some nasty races when there is contention over the same resources | 09:52 |
| Anticimex | ok, good to know | 09:52 |
| *** mvk has quit IRC | 09:53 | |
| *** ndipanov has quit IRC | 09:53 | |
| johnthetubaguy | we do retry around the races, but that has is limits as well, obviously | 09:53 |
| johnthetubaguy | the new stuff is going to aquire a claim on the resource during scheduling, in a way that works fine with active/active, basically | 09:54 |
| *** ndipanov has joined #openstack-nova | 09:54 | |
| johnthetubaguy | Anticimex: its worth reaching out on the operators mailing list, and/or talk to the doc team to get some folks to improve the HA guide, seems like it needs a bit more love | 09:55 |
| Anticimex | johnthetubaguy: ack, thanks | 09:55 |
| Anticimex | will discuss with folks on summit too i suppose | 09:55 |
| *** allen_gao has quit IRC | 09:56 | |
| *** achanda has joined #openstack-nova | 09:56 | |
| *** obondarev has quit IRC | 10:00 | |
| *** allen_gao has joined #openstack-nova | 10:01 | |
| *** achanda has quit IRC | 10:02 | |
| *** femnad has joined #openstack-nova | 10:04 | |
| *** ociuhandu has quit IRC | 10:05 | |
| *** obondarev has joined #openstack-nova | 10:06 | |
| * danpb wonders why when he runs the 'openstack' client (or indeed 'nova' client) it tries to open zope.py 240 times | 10:07 | |
| Anticimex | johnthetubaguy: can you clarify this part, "and you turn up the randomness on the pick 1 for top N" ? | 10:09 |
| Anticimex | is that to help the retry? | 10:09 |
| johnthetubaguy | let me go find the setting | 10:09 |
| Anticimex | "try to build on [list of hypervisors]" - retry thing? | 10:09 |
| Anticimex | awesome, thanks | 10:10 |
| *** obondarev has quit IRC | 10:12 | |
| *** tbachman has quit IRC | 10:14 | |
| *** tbachman has joined #openstack-nova | 10:14 | |
| *** cdent has quit IRC | 10:16 | |
| *** hoangcx has quit IRC | 10:16 | |
| *** hoangcx has joined #openstack-nova | 10:16 | |
| *** sdague has joined #openstack-nova | 10:20 | |
| *** liverpooler has joined #openstack-nova | 10:21 | |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Use common find_server from v2.shell https://review.openstack.org/300587 | 10:21 |
| openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Move all extensions from contrib dir https://review.openstack.org/280726 | 10:21 |
| *** daemontool has joined #openstack-nova | 10:22 | |
| danpb | hmm, that's interesting error "Out of range value for column 'memory_mb' at row 1" | 10:22 |
| danpb | oh whoops, /proc/meminfo says i have 8388607 TB of RAM available | 10:23 |
| allen_gao | roll back to negative? | 10:25 |
| openstackgerrit | Jiajun Liu proposed openstack/nova: config options: centralize cinder options https://review.openstack.org/298191 | 10:25 |
| danpb | yeah, that's the cgroup value for unlimited | 10:25 |
| openstackgerrit | Artom Lifshitz proposed openstack/python-novaclient: Virtual device tagging client support https://review.openstack.org/299222 | 10:27 |
| *** kmartin has quit IRC | 10:28 | |
| *** EinstCrazy has quit IRC | 10:30 | |
| *** obondarev has joined #openstack-nova | 10:31 | |
| *** thorst has joined #openstack-nova | 10:32 | |
| openstackgerrit | Srushti Gadadare proposed openstack/nova: Fix invalid uuid warnings in virt testcases https://review.openstack.org/302655 | 10:32 |
| *** gongysh has quit IRC | 10:32 | |
| *** Yingxin_ has joined #openstack-nova | 10:38 | |
| openstackgerrit | Srushti Gadadare proposed openstack/nova: Fix invalid uuid warnings https://review.openstack.org/302658 | 10:38 |
| *** ivase has quit IRC | 10:39 | |
| *** thorst has quit IRC | 10:39 | |
| *** links has quit IRC | 10:39 | |
| *** Yingxin has quit IRC | 10:40 | |
| *** Yingxin_ is now known as Yingxin | 10:40 | |
| *** xwizard has joined #openstack-nova | 10:44 | |
| *** gongysh has joined #openstack-nova | 10:44 | |
| *** Guest12876 is now known as tesseract | 10:47 | |
| *** mvk_ has quit IRC | 10:47 | |
| *** tesseract is now known as Guest74036 | 10:48 | |
| *** BobBall-1WOL is now known as BobBall | 10:49 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/302016 | 10:51 |
| *** andymaier has joined #openstack-nova | 10:53 | |
| *** alex_xu has quit IRC | 10:55 | |
| *** sridharg has quit IRC | 10:55 | |
| *** links has joined #openstack-nova | 10:56 | |
| *** mfedosin has joined #openstack-nova | 10:56 | |
| *** alex_xu has joined #openstack-nova | 10:57 | |
| *** jlanoux has quit IRC | 10:58 | |
| *** achanda has joined #openstack-nova | 10:58 | |
| openstackgerrit | Sean Dague proposed openstack/nova: remove alembic from requirements.txt https://review.openstack.org/302670 | 11:00 |
| *** p01nt3r75 has quit IRC | 11:00 | |
| sdague | johnthetubaguy: I just noticed the last global requirements update is updating alembic, but we don't directly import that any more | 11:01 |
| sdague | so we should probably remove it instead | 11:01 |
| *** gongysh has quit IRC | 11:02 | |
| *** jlanoux has joined #openstack-nova | 11:03 | |
| BobBall | coreywright: Are you currently working on XenAPI SR naming bug 1566622? | 11:03 |
| openstack | bug 1566622 in OpenStack Compute (nova) "live migration fails with xenapi virt driver and SRs with old-style naming convention" [Undecided,Confirmed] https://launchpad.net/bugs/1566622 | 11:03 |
| *** achanda has quit IRC | 11:03 | |
| openstackgerrit | jichenjc proposed openstack/nova: avoid microversion header in functional test https://review.openstack.org/292612 | 11:04 |
| *** jlanoux_ has joined #openstack-nova | 11:05 | |
| danpb | wow, a single 'nova list' call in a devstack environment causes over 14,500 open() calls for python modules, out of which 13,200 are -ENOENT | 11:05 |
| johnthetubaguy | sdague: I was just doing that patch actually | 11:05 |
| sdague | danpb: right, welcome to python :) | 11:06 |
| danpb | it seems the devstack setup makes it rather worse than it needs to be | 11:06 |
| sdague | danpb: in which way? | 11:06 |
| danpb | by adding a large number of entries to python load path | 11:06 |
| danpb | specifically it adds an entry to /usr/lib/python2.7/site-packages/easy-install.pth for each git checkout you have | 11:07 |
| *** jlanoux has quit IRC | 11:07 | |
| danpb | if it instead creating a symlink from /usr/lib/python2.7/site-packages/nova -> /opt/devstack/nova it'd avoid making python try countless paths which don't have the module it needs | 11:08 |
| sdague | that's a function just of how pip works | 11:08 |
| openstackgerrit | John Garbutt proposed openstack/nova: Drop alembic dependency https://review.openstack.org/302674 | 11:08 |
| danpb | can we make pip suck less | 11:08 |
| sdague | I suppose | 11:08 |
| sdague | how much preformance improvement does it gain ? | 11:09 |
| * danpb still hasn't figured out why we try to import zope.py 270 times | 11:09 | |
| danpb | sdague: i've not measured it yet - i'm just trying to understand why a simple 'nova list' takes 3 seconds to run | 11:09 |
| *** sridharg has joined #openstack-nova | 11:10 | |
| *** sridharg has quit IRC | 11:10 | |
| sdague | danpb: I doubt it has anything to do with the open calls | 11:10 |
| sdague | there are a bunch of API back and forths for that | 11:10 |
| *** sridharg has joined #openstack-nova | 11:10 | |
| *** thorst has joined #openstack-nova | 11:11 | |
| sdague | because of issuing tokens, version discovery | 11:11 |
| danpb | if i shut down the cloud entirely and rnu nova list, it still takes 1.3 seconds to tell me the cloud is not up | 11:11 |
| sdague | that's where I think you'll find all the time | 11:11 |
| danpb | so its not all from api requests | 11:11 |
| *** thorst_ has joined #openstack-nova | 11:11 | |
| sdague | the other thing to look at is the import chain | 11:11 |
| sdague | the failed opens are all going to fs cache I think, however there is some fixed cost associated with every import | 11:12 |
| sdague | and the way os-client-config and keystoneauth1 pull in things, might have massively ballooned the dependenices | 11:12 |
| sdague | the ML conversation on rootwrap slowness from a couple of years ago has a bunch of analysis on basics of python load time | 11:13 |
| *** mingyu has joined #openstack-nova | 11:14 | |
| *** cdent has joined #openstack-nova | 11:14 | |
| danpb | yeah, it seems only 0.25 seconds are actually spent in the system calls, though 80% of is due to the open+stat calls to find modules | 11:15 |
| *** thorst has quit IRC | 11:15 | |
| danpb | seems like this could be a bit of rathole to investigate | 11:15 |
| *** mvk_ has joined #openstack-nova | 11:16 | |
| sdague | yes, it probably will be | 11:17 |
| sdague | however, I do agree that we should have some better story here other than things are just slow | 11:17 |
| sdague | osc has the same issues | 11:17 |
| danpb | yeah, and given the number of cli calls devstack makes, it has a non-negligible contribution to overall time taken to run devstack | 11:18 |
| sdague | yes | 11:18 |
| *** rtheis has joined #openstack-nova | 11:18 | |
| sdague | I timed out osc overhead in devstack a while ago about this | 11:19 |
| sdague | let me go dig that up | 11:19 |
| kashyap | "osc" == OpenStack Client I suppose. (Interesting discussion.) | 11:20 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova: Improve nova.rpc conf options documentation https://review.openstack.org/302678 | 11:21 |
| sdague | http://lists.openstack.org/pipermail/openstack-dev/2015-April/061347.html | 11:22 |
| sdague | it was about 3.5 minutes total osc time a year ago | 11:22 |
| sdague | my guess is ~ 50% of that is python module load time | 11:22 |
| sdague | as most of osc calls are keystone only, so no complex round trips | 11:23 |
| * danpb wonders if you could do something crazy where by you run osc-daemon where it just opens a unix socket to accept commands, then have a osc-client which writes the command argv to the osc daemon socket for execution | 11:24 | |
| danpb | ... IOW you would only have a 1 time hit for importing modules at the start of devstack | 11:25 |
| cdent | client as a service | 11:25 |
| openstackgerrit | Artom Lifshitz proposed openstack/python-novaclient: Virtual device tagging client support https://review.openstack.org/299222 | 11:26 |
| *** mingyu has quit IRC | 11:26 | |
| danpb | cdent: well the ideal would have been for devstack to be written in python instead of shell :-) | 11:27 |
| cdent | :) | 11:27 |
| cdent | I actually think running a little oscd is a reasonable idea if the lost time is that significant | 11:28 |
| cdent | however, it seems like that masks a real issue for real users | 11:28 |
| danpb | ...if they're foolish enough to use shell for non-trivial scripts | 11:29 |
| *** mingyu has joined #openstack-nova | 11:30 | |
| openstackgerrit | Balazs Gibizer proposed openstack/nova-specs: Versioned notification transformation https://review.openstack.org/286675 | 11:30 |
| cdent | I dunno: it's right there in the name OpenStack Client. I would guess that people are going to assume it's a reasonable choice. | 11:31 |
| openstackgerrit | Sergey Nikitin proposed openstack/nova: Moved tags filtering tests to TestInstanceTagsFiltering test case https://review.openstack.org/258515 | 11:32 |
| *** sfinucan has joined #openstack-nova | 11:33 | |
| *** mfedosin has quit IRC | 11:34 | |
| *** vtech has quit IRC | 11:34 | |
| snikitin | alex_xu, johnthetubaguy: Guys, thank you very very very much for instance tags! My job is finally done! | 11:34 |
| *** vtech has joined #openstack-nova | 11:35 | |
| sdague | danpb: a daemon mode has been talked about for osc... | 11:35 |
| sdague | it just seems to be punting on the problem of importing too many things | 11:35 |
| snikitin | johnthetubaguy: John, could you please reapprove this tags patch? it was rebased https://review.openstack.org/#/c/258515/ | 11:36 |
| snikitin | alex_xu: or Alex. could you please reapprove this tags patch? | 11:36 |
| *** alex_xu has quit IRC | 11:41 | |
| snikitin | johnthetubaguy: ty! | 11:42 |
| johnthetubaguy | snikitin: no | 11:42 |
| johnthetubaguy | oops | 11:42 |
| johnthetubaguy | np | 11:42 |
| snikitin | :) | 11:42 |
| coreywright | BobBall: yes, i'm currently working on bug 1566622 (ie live-migration broken with old sr names) | 11:43 |
| openstack | bug 1566622 in OpenStack Compute (nova) "live migration fails with xenapi virt driver and SRs with old-style naming convention" [Undecided,Confirmed] https://launchpad.net/bugs/1566622 | 11:43 |
| *** alex_xu has joined #openstack-nova | 11:43 | |
| openstackgerrit | jichenjc proposed openstack/nova: Add functional test for v2.7 https://review.openstack.org/292881 | 11:45 |
| coreywright | BobBall: but i'll take all the help i can get from a xenapi (both nova virt driver and dom0-side interface) expert :) | 11:47 |
| coreywright | BobBall: i'm not clever enough about xenapi to think of anything better (yet) than adding both new *and* old style sr names to the migration object | 11:49 |
| *** ChanServ changes topic to "This channel is for Nova development. For support of Nova deployments, please use #openstack. Please see: https://wiki.openstack.org/wiki/Nova/Mitaka_Release_Schedule" | 11:49 | |
| *** Marga_ has joined #openstack-nova | 11:49 | |
| coreywright | BobBall: (whether in sr_uuid_map or a shadow of it: sr_uuid_map_old) | 11:49 |
| johnthetubaguy | coreywright: I think you just need the new name, for the destination side of the map | 11:51 |
| johnthetubaguy | coreywright: as the volume will always be a new add on the destination | 11:51 |
| *** Oku_OS is now known as Oku_OS-away | 11:52 | |
| johnthetubaguy | coreywright: see migrate_send in here: http://docs.vmd.citrix.com/XenServer/6.5.0/1.0/en_gb/api/?c=VM basically it just needs to know which destination SR each source VDI is going to | 11:53 |
| coreywright | johnthetubaguy: correct, the destination might only need the new name, but i need both the old and new names on the source side and i'm not sure given the new name i can generate the old name (given how the new name is essentially a hash of the iscsi info by way of uuid5) | 11:53 |
| *** p01nt3r75 has joined #openstack-nova | 11:53 | |
| johnthetubaguy | coreywright: ah, gotcha, that seems reasonable then, the tricky part is how to do that without breaking rolling upgrades | 11:54 |
| coreywright | correct | 11:54 |
| johnthetubaguy | I think that means adding a new object field for the new names? | 11:54 |
| BobBall | coreywright: See my comment; I think that you can't do that - the destination must have the same UUID as the source | 11:54 |
| *** vladikr has joined #openstack-nova | 11:55 | |
| johnthetubaguy | BobBall: oh, because of the special iSCSI match code that knows it doesn't need to copy the disk? | 11:55 |
| *** cdent has quit IRC | 11:55 | |
| coreywright | BobBall: yeah, i saw your comment, and i'm not trying to discount it, but i'm only smart enough about xenapi to tackle one problem at a time ;) | 11:55 |
| BobBall | coreywright: I think the only option is to get new nova to use the old name when creating the SR in the pre live migrate step, either with a config setting (to use the old style names) or by passing information through from the source nova | 11:55 |
| BobBall | coreywright: My belief is that when you fix that first problem you'll discover that it tries to copy the disk and might cause data corruption - so we need a different solution so XAPI won't try and copy it | 11:56 |
| BobBall | Yes johnthetubaguy. | 11:56 |
| johnthetubaguy | BobBall: yeah, I forgot about the matching uuid hack, +1 we should instead make sure the SR add respects the "matching" uuid scheme | 11:57 |
| BobBall | We will not attempt to mirror the disk to the 'new' SR IFF the SR uuid and VDI uuid on source and destination match | 11:57 |
| coreywright | BobBall: i was waiting until i got to the data corruption (in testing) to have to address it (ie the simple-stupid approach; it's what us known clever people have to do ;) | 11:57 |
| coreywright | "known clever" => "non-clever" | 11:58 |
| coreywright | (hey, it's phonetically close ;) | 11:58 |
| *** cdent has joined #openstack-nova | 11:59 | |
| *** andymaier has quit IRC | 11:59 | |
| BobBall | So if the source UUID was the old style (because it existed prior to upgrading Nova) and the destination UUID is the new style, then XAPI will not believe they are the same SR and it will force a mirror | 11:59 |
| BobBall | heh :) | 11:59 |
| coreywright | BobBall: where's that code in the xenapi nova virt driver that compares the sr and vdi uuid? | 11:59 |
| BobBall | I can't promise you'll get data corruption... | 11:59 |
| BobBall | coreywright: https://github.com/xapi-project/xen-api/blob/master/ocaml/xapi/xapi_vm_migrate.ml#L481 | 12:00 |
| BobBall | it's in XAPI, not the virt driver | 12:00 |
| coreywright | BobBall: AH! | 12:00 |
| coreywright | that's why i haven't stumbled across it yet while digging into this bug! | 12:00 |
| *** BobBall_ has joined #openstack-nova | 12:01 | |
| coreywright | (i thought i just didn't know enough yet about xenapi nova virt driver and/or stumbled across it) | 12:01 |
| *** BobBall_ is now known as Guest30758 | 12:01 | |
| sdague | johnthetubaguy: https://review.openstack.org/#/c/298796 feels like the kind of options we'd want to deprecate, right? | 12:04 |
| *** aysyd has joined #openstack-nova | 12:04 | |
| *** BobBall has quit IRC | 12:04 | |
| *** Guest30758 has quit IRC | 12:05 | |
| *** BobBall has joined #openstack-nova | 12:06 | |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: Config options: base path configuration: improve help text https://review.openstack.org/299236 | 12:06 |
| openstackgerrit | Markus Zoeller (markus_z) proposed openstack/nova: Config options: centralize base path configuration https://review.openstack.org/259406 | 12:06 |
| *** BobBall is now known as Guest34177 | 12:06 | |
| *** EinstCrazy has joined #openstack-nova | 12:06 | |
| *** lpetrut has quit IRC | 12:07 | |
| *** jhesketh has quit IRC | 12:07 | |
| *** lpetrut has joined #openstack-nova | 12:07 | |
| *** achanda has joined #openstack-nova | 12:07 | |
| *** jhesketh has joined #openstack-nova | 12:07 | |
| *** ekuris has quit IRC | 12:07 | |
| *** eglynn has quit IRC | 12:09 | |
| openstackgerrit | Gleb Stepanov proposed openstack/nova: Check API versions intersects https://review.openstack.org/301278 | 12:10 |
| bauzas | could someone tell me why we're resetting the instance attributes here https://github.com/openstack/nova/blob/master/nova/tests/unit/fake_instance.py#L127 ? | 12:10 |
| bauzas | it prevents the flavor to be persisted if I'm calling .create() after this | 12:10 |
| sdague | Guest34177: https://review.openstack.org/#/c/289431/ can you look into this? | 12:10 |
| *** baoli has joined #openstack-nova | 12:13 | |
| bauzas | hmmmm, I should wait for dansmith | 12:13 |
| bauzas | :) | 12:13 |
| *** baoli_ has joined #openstack-nova | 12:13 | |
| *** achanda has quit IRC | 12:15 | |
| openstackgerrit | Roman Podoliaka proposed openstack/nova: db: retry instance_info_cache_update() on deadlock https://review.openstack.org/302714 | 12:16 |
| markus_z | snikitin: Got a minute? This merged some minutes ago and I get a json decode error: https://review.openstack.org/#/c/268932/32/doc/api_samples/os-server-tags/v2.26/server-tags-put-req.json | 12:17 |
| *** raildo-afk is now known as raildo | 12:18 | |
| *** ducttape_ has joined #openstack-nova | 12:18 | |
| markus_z | snikitin: I have no experience with the API samples (so maybe it's a local issue on my laptop), but shouldn't that at least have '{}'? | 12:18 |
| *** baoli has quit IRC | 12:18 | |
| snikitin | markus_z, we can't decode an empty template? | 12:18 |
| sdague | snikitin: no, it needs to be real json | 12:18 |
| markus_z | snikitin: Running "tox -e docs" gives me this error "No JSON object could be decoded" | 12:19 |
| markus_z | snikitin: But it could be my local python version | 12:19 |
| sdague | snikitin: why would there be an empty template? | 12:19 |
| markus_z | snikitin: I can reproduce it locally with echo '' | python -m json.tool | 12:19 |
| markus_z | But the docs gate check was successful, so I'm a little confused | 12:20 |
| sdague | markus_z: we should just delete the empty files | 12:21 |
| sdague | I'm not really sure why they were in that patch | 12:21 |
| snikitin | sdague, sdague: let me check | 12:21 |
| markus_z | sdague: ah, OK, I wasn't sure if that was some kind of convention | 12:21 |
| *** ekuris has joined #openstack-nova | 12:22 | |
| *** edmondsw has joined #openstack-nova | 12:23 | |
| openstackgerrit | Allen Gao proposed openstack/nova: config options: centralize 'spice' options https://review.openstack.org/264271 | 12:24 |
| sdague | oh this is because this API is using put in an unusual way | 12:24 |
| *** doug-fish has joined #openstack-nova | 12:26 | |
| *** akshai has joined #openstack-nova | 12:26 | |
| dedery | hi experts. I'm looking into something at nova's compute/api.py. how can i know/control the timeout of a call to compute_rpcapi? | 12:27 |
| *** mingyu has quit IRC | 12:27 | |
| *** apoorv has quit IRC | 12:27 | |
| dedery | e.g. self.compute_rpcapi.detach_volume(...) | 12:27 |
| *** salv-orl_ has quit IRC | 12:28 | |
| *** wuhg has quit IRC | 12:29 | |
| *** cdelatte has quit IRC | 12:30 | |
| *** ducttape_ has quit IRC | 12:32 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: enhance method to return pointer_model from image prop https://review.openstack.org/174854 | 12:33 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: deprecate use_usb_tablet in favor of pointer_model https://review.openstack.org/176242 | 12:33 |
| *** sahid has joined #openstack-nova | 12:33 | |
| snikitin | m | 12:33 |
| *** pradk has joined #openstack-nova | 12:34 | |
| *** ivase has joined #openstack-nova | 12:35 | |
| markus_z | snikitin: Do you want to delete the empty api sample files? Otherwise I can do it. | 12:36 |
| *** jdurgin1 has joined #openstack-nova | 12:36 | |
| snikitin | markus_z, sdague: According to tags guide this request should have no body. Not empty body, NO body. https://specs.openstack.org/openstack/api-wg/guidelines/tags.html#addressing-individual-tags | 12:37 |
| snikitin | markus_z, sdague: I created an empty template to pass name of file with this template here: | 12:37 |
| snikitin | https://github.com/openstack/nova/blob/master/nova/tests/functional/api_sample_tests/test_server_tags.py#L80 | 12:37 |
| sdague | snikitin: right, I see it | 12:37 |
| *** lucasagomes has quit IRC | 12:37 | |
| snikitin | My local "tox -e docs" doesn't works too :( | 12:37 |
| snikitin | I can fix the test | 12:37 |
| sdague | snikitin: actually I have a patch | 12:37 |
| snikitin | and remove thus template | 12:37 |
| *** lucasagomes has joined #openstack-nova | 12:38 | |
| *** Oku_OS-away is now known as Oku_OS | 12:38 | |
| alaski | lifeless: awesome, thanks! | 12:38 |
| markus_z | snikitin: So "no body" != "empty body"? | 12:38 |
| snikitin | markus_z: yes | 12:38 |
| snikitin | sdague: what patch? | 12:39 |
| openstackgerrit | Sean Dague proposed openstack/nova: allow samples testing for PUT to not have a body https://review.openstack.org/302733 | 12:39 |
| sdague | snikitin: that patch | 12:39 |
| snikitin | sdague: good patch! It solves our problem | 12:40 |
| *** salv-orlando has joined #openstack-nova | 12:40 | |
| sdague | snikitin: yeh, it just took me a minute to figure out the root issue you were hitting | 12:40 |
| snikitin | sdague, markus_z: But I don't know why my patches passed the gate tests | 12:41 |
| openstackgerrit | Alex Szarka proposed openstack/nova: Use assertGreater(A, X) instead of assertTrue(A > X) https://review.openstack.org/302734 | 12:41 |
| *** erlon has joined #openstack-nova | 12:42 | |
| markus_z | snikitin: That's still odd, yeah. | 12:42 |
| sdague | snikitin: I don't know why there is a discrepancy in the doc tests | 12:42 |
| *** doug-fish has quit IRC | 12:42 | |
| *** ababich has quit IRC | 12:42 | |
| sdague | however, it's silly for us to have the empty files | 12:42 |
| *** doug-fish has joined #openstack-nova | 12:42 | |
| *** ababich has joined #openstack-nova | 12:43 | |
| *** aysyd has quit IRC | 12:44 | |
| *** rfolco has joined #openstack-nova | 12:45 | |
| *** lucasagomes has quit IRC | 12:45 | |
| *** mingyu has joined #openstack-nova | 12:45 | |
| *** lucasagomes has joined #openstack-nova | 12:46 | |
| *** doug-fish has quit IRC | 12:47 | |
| *** cdelatte has joined #openstack-nova | 12:47 | |
| *** achanda has joined #openstack-nova | 12:47 | |
| *** aysyd has joined #openstack-nova | 12:50 | |
| *** houming has quit IRC | 12:50 | |
| openstackgerrit | jichenjc proposed openstack/nova: prevent non-admin user search soft-delete server https://review.openstack.org/258472 | 12:51 |
| *** burgerk has joined #openstack-nova | 12:52 | |
| *** delattec has joined #openstack-nova | 12:53 | |
| *** doug-fish has joined #openstack-nova | 12:54 | |
| *** cdelatte has quit IRC | 12:56 | |
| *** doug-fis_ has joined #openstack-nova | 12:56 | |
| *** akshai has quit IRC | 12:56 | |
| *** dedery has quit IRC | 12:57 | |
| *** akshai has joined #openstack-nova | 12:57 | |
| *** diogogmt has quit IRC | 12:58 | |
| *** achanda has quit IRC | 12:58 | |
| *** diogogmt has joined #openstack-nova | 12:59 | |
| *** doug-fish has quit IRC | 12:59 | |
| *** rodrigods has quit IRC | 13:01 | |
| openstackgerrit | sahid proposed openstack/nova: libvirt: handle reserved pages size https://review.openstack.org/292500 | 13:01 |
| openstackgerrit | sahid proposed openstack/nova: virt: reserved number of mempages on compute host https://review.openstack.org/292499 | 13:01 |
| *** rodrigods has joined #openstack-nova | 13:01 | |
| *** rlrossit has joined #openstack-nova | 13:02 | |
| openstackgerrit | Matthew Booth proposed openstack/nova: Rename Raw backend to NoBacking https://review.openstack.org/279626 | 13:02 |
| openstackgerrit | Matthew Booth proposed openstack/nova: Fix signature of copy_image https://review.openstack.org/282432 | 13:02 |
| openstackgerrit | Matthew Booth proposed openstack/nova: Remove fake_imagebackend.Raw and cleanup dependent tests https://review.openstack.org/267661 | 13:02 |
| openstackgerrit | Matthew Booth proposed openstack/nova: Rename Image.check_image_exists to Image.exists() https://review.openstack.org/270998 | 13:02 |
| openstackgerrit | Matthew Booth proposed openstack/nova: Only attempt to inject files if the injection disk exists https://review.openstack.org/250872 | 13:03 |
| openstackgerrit | Artom Lifshitz proposed openstack/python-novaclient: Virtual device tagging client support https://review.openstack.org/299222 | 13:03 |
| *** vtech has quit IRC | 13:04 | |
| *** vtech has joined #openstack-nova | 13:05 | |
| *** Guest34177 is now known as BobBall | 13:05 | |
| openstackgerrit | jichenjc proposed openstack/python-novaclient: move migrations from extension https://review.openstack.org/285213 | 13:05 |
| markus_z | sarafraj: Are you around? | 13:06 |
| *** achanda has joined #openstack-nova | 13:08 | |
| *** diogogmt has quit IRC | 13:08 | |
| *** mvk_ has quit IRC | 13:11 | |
| *** mdrabe has joined #openstack-nova | 13:12 | |
| *** diogogmt has joined #openstack-nova | 13:12 | |
| *** achanda has quit IRC | 13:13 | |
| *** mvk has joined #openstack-nova | 13:13 | |
| *** wanghao has quit IRC | 13:14 | |
| *** akshai has quit IRC | 13:14 | |
| *** jschlueter has quit IRC | 13:14 | |
| *** wanghao has joined #openstack-nova | 13:14 | |
| *** whenry has quit IRC | 13:15 | |
| sahid | markus_z: i did not find time yet to address your comment here https://review.openstack.org/#/c/301247/, but i'll try to do it | 13:15 |
| markus_z | sahid: OK, don't hurry. Would be a good addition to make the use case more clear. | 13:16 |
| *** jschlueter has joined #openstack-nova | 13:17 | |
| *** whenry has joined #openstack-nova | 13:17 | |
| *** daemontool has quit IRC | 13:17 | |
| *** ducttape_ has joined #openstack-nova | 13:18 | |
| *** ducttape_ has quit IRC | 13:18 | |
| *** ducttape_ has joined #openstack-nova | 13:18 | |
| sahid | markus_z: yes, after to re-read response i gave to you, i was not sure if it was clear | 13:19 |
| *** ababich has quit IRC | 13:20 | |
| *** ababich has joined #openstack-nova | 13:21 | |
| *** doug-fis_ has quit IRC | 13:23 | |
| *** akshai has joined #openstack-nova | 13:23 | |
| openstackgerrit | Alex Szarka proposed openstack/nova: Use assertGreater(A, X) instead of assertTrue(A > X) https://review.openstack.org/302734 | 13:23 |
| openstackgerrit | Roman Podoliaka proposed openstack/nova: novnc: make naming of SSL related options consistent https://review.openstack.org/222113 | 13:24 |
| *** mjura has quit IRC | 13:24 | |
| *** ociuhandu has joined #openstack-nova | 13:25 | |
| *** eliqiao1 has joined #openstack-nova | 13:25 | |
| *** mjura has joined #openstack-nova | 13:26 | |
| *** xyang1 has joined #openstack-nova | 13:26 | |
| *** mriedem has joined #openstack-nova | 13:26 | |
| openstackgerrit | Roman Podoliaka proposed openstack/nova: libvirt: delete the last file link in _supports_direct_io() https://review.openstack.org/224764 | 13:26 |
| *** daemontool has joined #openstack-nova | 13:28 | |
| *** ctrath has joined #openstack-nova | 13:28 | |
| BobBall | sdague: Commented | 13:29 |
| BobBall | johnthetubaguy: Could you check my thinking on https://review.openstack.org/#/c/289431/9 | 13:29 |
| johnthetubaguy | BobBall: hmm, I think we did this wrong you know, really we should call the plugins to see what version they support, so they can start responding yes to 1.x and 2.x for a little while | 13:31 |
| openstackgerrit | Chris Dent proposed openstack/nova: WIP Initial support for both microversion headers https://review.openstack.org/300077 | 13:31 |
| openstackgerrit | Chris Dent proposed openstack/nova: Initial use of microversion_parse https://review.openstack.org/300076 | 13:31 |
| openstackgerrit | Sean Dague proposed openstack/nova: allow samples testing for PUT to not have a body https://review.openstack.org/302733 | 13:31 |
| *** ducttape_ has quit IRC | 13:31 | |
| *** cdent has quit IRC | 13:31 | |
| johnthetubaguy | BobBall: thats kinda what we do with RPC | 13:32 |
| *** eliqiao1 has quit IRC | 13:32 | |
| sdague | johnthetubaguy / alaski I missed a return statement so the test failed, it's fixed now | 13:32 |
| rlrossit | mriedem: In the DB API, why does a flavor get by id raise a NotFound exception if it isn't found by the id (https://github.com/openstack/nova/blob/master/nova/db/sqlalchemy/api.py#L5185-L5194), but something like instance actions (https://github.com/openstack/nova/blob/master/nova/db/sqlalchemy/api.py#L6160-L6172) doesn't raise the same exception? | 13:32 |
| rlrossit | that's ^ causing the cells job to still blow up on a negative test for ID get because _from_db_object is getting None instead of an action dict | 13:33 |
| alaski | sdague: got it | 13:33 |
| alex_xu | snikitin: congrats and thanks for your patient :) | 13:34 |
| mriedem | rlrossit: not sure, i guess it depends on how it's used | 13:34 |
| snikitin | alex_xu: :) | 13:34 |
| mriedem | rlrossit: as in we might expect there to not be instance actions for a given request id | 13:34 |
| rlrossit | mriedem: yeah it looks like bw usage also doesn't have the check in the db api, so they have to check the result when the call the db http://git.openstack.org/cgit/openstack/nova/tree/nova/objects/bandwidth_usage.py#n60 | 13:34 |
| mriedem | rlrossit: could also maybe have something to do with bw usage and instance actions not having foreign keys back to the instances table and they don't get soft deleted when an instance is deleted | 13:35 |
| mriedem | rlrossit: thats | 13:35 |
| mriedem | that's an outside guess | 13:36 |
| mriedem | but those are the two tables that have that weirdness in the db schema | 13:36 |
| rlrossit | ok so this just happens to be the 2 outliers that I found? | 13:36 |
| rlrossit | https://github.com/openstack/tempest/blob/master/tempest/api/compute/servers/test_instance_actions_negative.py#L48 well it needs to be raising a 404 somewhere... but hopefully that's higher up | 13:37 |
| mriedem | rlrossit: probably in the rest api | 13:38 |
| rlrossit | but.. but.. why can't we just raise the not found in the db like everyone else :| | 13:38 |
| mriedem | shrug | 13:39 |
| *** artom has quit IRC | 13:39 | |
| rlrossit | I'll check the get in the real rpc api to see how they handle it, and I'll do it the same as that does | 13:39 |
| * rlrossit craves consistency | 13:39 | |
| *** ducttape_ has joined #openstack-nova | 13:40 | |
| rlrossit | https://github.com/openstack/nova/blob/master/nova/objects/instance_action.py#L68-L73 yup that's how the object handles it too | 13:41 |
| BobBall | johnthetubaguy: Sure; that'd work too | 13:41 |
| BobBall | johnthetubaguy: I don't mind which way TBH | 13:42 |
| *** sc68cal has joined #openstack-nova | 13:42 | |
| johnthetubaguy | BobBall: I think that way would stop the need for the version bump to be impacting, and we can add that feature in a way that is not impact, if we wait a release to bump the major version? | 13:43 |
| johnthetubaguy | BobBall: I might be over thinking this though, that feels very complicated | 13:43 |
| *** eliqiao1 has joined #openstack-nova | 13:44 | |
| *** irenab has quit IRC | 13:44 | |
| *** mgoddard_ has joined #openstack-nova | 13:44 | |
| mriedem | moshele: i added https://blueprints.launchpad.net/nova/+spec/ironic-infiniband-support to the nova meeting agenda | 13:45 |
| BobBall | johnthetubaguy: I did propose that back in patch set 4, but I don't mind which way it goes | 13:45 |
| moshele | mriedem: thanks | 13:45 |
| johnthetubaguy | BobBall: I missed that, sorry, I need to find some time to read though that all properly | 13:46 |
| BobBall | johnthetubaguy: It's OK - Stephen uploaded a few new patches and the comment got lost :) | 13:46 |
| *** links has quit IRC | 13:46 | |
| sarafraj | Good morning | 13:47 |
| *** garyk1 has quit IRC | 13:47 | |
| *** mgoddard has quit IRC | 13:48 | |
| sarafraj | markus_z: Hi | 13:48 |
| *** esracelik has quit IRC | 13:48 | |
| *** Oku_OS is now known as Oku_OS-away | 13:48 | |
| markus_z | sarafraj: hey, just wanted to double-check if you have any questions of issues with the bug stuff? | 13:49 |
| sarafraj | markus_z: I see 3 new bugs, for the search query you gave me | 13:50 |
| sarafraj | markus_z: Can you walk me through process with one of these | 13:50 |
| markus_z | sarafraj: yep, let me have a quick look which fits best | 13:51 |
| *** sridharg has quit IRC | 13:52 | |
| mriedem | danpb: ndipanov: easy bug fix for a regression https://review.openstack.org/#/c/295395/ | 13:52 |
| *** mrkz has joined #openstack-nova | 13:53 | |
| markus_z | sarafraj: Maye we use https://bugs.launchpad.net/nova/+bug/1567434 | 13:53 |
| openstack | Launchpad bug 1567434 in OpenStack Compute (nova) "new hypervisor should apper with available resources set to 0" [Undecided,New] | 13:53 |
| ndipanov | mriedem, looking | 13:54 |
| sarafraj | markus_z: yes | 13:54 |
| openstackgerrit | Gleb Stepanov proposed openstack/nova: Check API versions intersects https://review.openstack.org/301278 | 13:55 |
| mriedem | nova meeting in 5 minutes | 13:55 |
| *** ametts has joined #openstack-nova | 13:56 | |
| ndipanov | mriedem, just enough time to look at https://review.openstack.org/301859 | 13:56 |
| markus_z | sarafraj: I'd take a look where the line of code is used in Nova | 13:56 |
| ndipanov | it fixes the feature we added to sriov that was borked | 13:56 |
| ndipanov | so a candidate for backport to stable mitaka I think | 13:56 |
| mriedem | ndipanov: you mean the feature that just landed last week? i saw the 2 bugs that came up from that. | 13:57 |
| mriedem | those landed in newton, not mitaka | 13:57 |
| *** busterswt has joined #openstack-nova | 13:57 | |
| mriedem | would have been nice to test that code before we approved it | 13:57 |
| ndipanov | mriedem, no that feature is fine | 13:58 |
| ndipanov | that is just neutron integration | 13:58 |
| mriedem | which bp? | 13:58 |
| ndipanov | mriedem, sec | 13:58 |
| openstackgerrit | sahid proposed openstack/nova: libvirt: release serial console ports when before to destroy guests https://review.openstack.org/301247 | 13:58 |
| openstackgerrit | sahid proposed openstack/nova: serial: replace log warning to debug when port already used https://review.openstack.org/302777 | 13:58 |
| openstackgerrit | sahid proposed openstack/nova: serial: make acquire port to avoid port conflicts https://review.openstack.org/302778 | 13:58 |
| openstackgerrit | Ryan Rossiter proposed openstack/nova: Remove DictCompat from InstanceAction https://review.openstack.org/301905 | 13:58 |
| *** paul-carlton1 has quit IRC | 13:59 | |
| *** cdent has joined #openstack-nova | 13:59 | |
| ndipanov | mriedem, https://review.openstack.org/#/c/212472/ | 13:59 |
| ndipanov | stupid lp can't find blueprints once they are completed | 14:00 |
| mriedem | https://blueprints.launchpad.net/nova/+spec/sriov-physical-function-passthrough | 14:00 |
| ndipanov | yes | 14:00 |
| mriedem | adrian_otto akuriata alevine alexpilotti aloga andreykurilin anteaya artom auggy | 14:00 |
| mriedem | bauzas belliott belmoreira bobball cburgess claudiub danpb dguitarbite _diana_Â | 14:00 |
| mriedem | diana_clarke dims duncant edleafe efried flip214 funzo garyk gcb gjayavelu | 14:00 |
| mriedem | irina_pov jaypipes jcookekhugen jgrimm jichen jlvillal jroll kashyap klindgren | 14:00 |
| *** Oku_OS-away is now known as Oku_OS | 14:00 | |
| mriedem | krtaylor lbeliveau lxsli macsz markus_z mdorman med_ mikal mjturek mnestratov | 14:01 |
| mriedem | moshele mrda nagyz ndipanov neiljerram nic Nisha PaulMurray raildo rgeragnov | 14:01 |
| mriedem | sc68cal scottda sdague sileht sorrison swamireddy thomasem thorst tjones tonyb | 14:01 |
| mriedem | tpatil tpatzig xyang rdopiera sarafraj woodster | 14:01 |
| mriedem | nova meeting | 14:01 |
| *** yamahata has joined #openstack-nova | 14:01 | |
| markus_z | sarafraj: basically you can add the tag "ironic" as it seems to happen there. | 14:01 |
| markus_z | sarafraj: You can try to reproduce the issue, if information is missing for that, you can contact the reporter | 14:02 |
| sarafraj | markus_z: So the process is to dig a little into code and try to find problematic section and then assign a tag? | 14:03 |
| *** mgoddard has joined #openstack-nova | 14:04 | |
| *** mgoddard_ has quit IRC | 14:04 | |
| markus_z | sarafraj: yep, that's the first step. And ensure that the report has enough information to reproduce the issue. Otherwise it gets hard to solve it. | 14:04 |
| sarafraj | markus_z: Ok got it. thx | 14:05 |
| markus_z | snikitin: Isn't that done with your change from earlier? https://bugs.launchpad.net/nova/+bug/1567369 | 14:06 |
| openstack | Launchpad bug 1567369 in OpenStack Compute (nova) " Added server tags support in nova-api" [Undecided,New] | 14:06 |
| *** artom has joined #openstack-nova | 14:07 | |
| *** sdake has joined #openstack-nova | 14:07 | |
| *** qeelee has quit IRC | 14:07 | |
| *** thomasem_ is now known as thomasem | 14:08 | |
| snikitin | markus_z: What should I do with this bug? | 14:08 |
| *** tinbar has joined #openstack-nova | 14:09 | |
| markus_z | snikitin: Double-check if it has enough information for the manuals team and then reroute to them. | 14:09 |
| *** diogogmt has quit IRC | 14:09 | |
| *** tbachman has quit IRC | 14:10 | |
| *** knikolla has joined #openstack-nova | 14:10 | |
| snikitin | markus_z: Yes, this patch contains enough information | 14:11 |
| *** andrearosa has quit IRC | 14:12 | |
| *** sdake has quit IRC | 14:12 | |
| *** nkrinner has quit IRC | 14:12 | |
| cdent | dansmith: say a nova object (ResourcePool) has a list field (aggregates) which is an AggregatesList, what is the canonical way of adding or removing an aggregate associated with the reosurce pool: calling add/remove_aggregate methods and doing the update internally, or adjusting the value of the field and calling update? | 14:12 |
| *** rcrit has joined #openstack-nova | 14:14 | |
| *** pwnall has joined #openstack-nova | 14:15 | |
| sarafraj | markus_z: BTW I am in UTC-6 timezone. From Austin, TX | 14:16 |
| *** mylu has joined #openstack-nova | 14:17 | |
| *** tinbar has quit IRC | 14:17 | |
| *** ChrisBenson has joined #openstack-nova | 14:17 | |
| *** gagehugo has joined #openstack-nova | 14:18 | |
| *** thorst_ has quit IRC | 14:21 | |
| *** sdake has joined #openstack-nova | 14:21 | |
| rcrit | is there a place one can set injected_files for all instances? E.g. I want to send these files to any instance that gets created? | 14:21 |
| danpb | no,that'd be a feature request | 14:22 |
| * rcrit waves | 14:22 | |
| danpb | the injected files get comes straight down from the API inputs iirc | 14:22 |
| *** thorst_ has joined #openstack-nova | 14:22 | |
| rcrit | ok, just wanted to double-check my reading of the code | 14:23 |
| *** baoli_ has quit IRC | 14:26 | |
| *** thorst__ has joined #openstack-nova | 14:26 | |
| *** Prophidys has quit IRC | 14:26 | |
| *** thorst_ has quit IRC | 14:27 | |
| *** baoli has joined #openstack-nova | 14:30 | |
| *** zz_dimtruck is now known as dimtruck | 14:30 | |
| *** matrohon has quit IRC | 14:31 | |
| *** matrohon has joined #openstack-nova | 14:31 | |
| *** tinbar has joined #openstack-nova | 14:32 | |
| *** eglynn has joined #openstack-nova | 14:34 | |
| *** eharney has joined #openstack-nova | 14:34 | |
| *** tbachman has joined #openstack-nova | 14:34 | |
| openstackgerrit | Sujitha proposed openstack/nova: Config options: Centralize compute options https://review.openstack.org/295938 | 14:35 |
| *** thorst__ has quit IRC | 14:37 | |
| *** mylu has quit IRC | 14:37 | |
| openstackgerrit | Pushkar Umaranikar proposed openstack/nova: Config options: Centralize netconf options https://review.openstack.org/294312 | 14:39 |
| *** edtubill has joined #openstack-nova | 14:39 | |
| *** tbachman_ has joined #openstack-nova | 14:39 | |
| *** tbachman has quit IRC | 14:40 | |
| *** tbachman_ is now known as tbachman | 14:40 | |
| *** thorst_ has joined #openstack-nova | 14:41 | |
| *** mylu has joined #openstack-nova | 14:42 | |
| dims | mriedem : https://review.openstack.org/#/c/298978/ still good to release? | 14:45 |
| dims | (newton: release python-novaclient 3.4.0) | 14:45 |
| mriedem | dims: hell yes it is | 14:46 |
| dansmith | cdent: I think resourcepool should have a ListOfObjectsField, which is just a list, and then you add aggregates to it and call save() on resourcepool | 14:46 |
| markus_z | sarafraj: OK, cool, I will be at the Summit in Austin, we can talk there if you like. | 14:47 |
| cdent | dansmith: thanks, it alread has that ListOfObjectsField, it was just the question on how to manipulate it. | 14:47 |
| dims | mriedem : thanks for your patience :) | 14:47 |
| *** ametts has quit IRC | 14:48 | |
| mriedem | dims: no problem | 14:48 |
| *** ametts has joined #openstack-nova | 14:49 | |
| *** tbachman has quit IRC | 14:49 | |
| sarafraj | markus_z: Sounds good. Looking forward to meet y'all | 14:49 |
| *** kfarr has joined #openstack-nova | 14:49 | |
| openstackgerrit | Zhenguo Niu proposed openstack/nova: [WIP]Add shellinabox console to API https://review.openstack.org/302821 | 14:50 |
| dansmith | cdent: ListOfObjectsField is different than an Object field with an AllocationList | 14:50 |
| dansmith | cdent: but, if what you say, then just add and .save() on the base object yeah | 14:50 |
| dansmith | an actual list object should never have a .save() | 14:51 |
| *** tbachman has joined #openstack-nova | 14:51 | |
| *** salv-orl_ has joined #openstack-nova | 14:52 | |
| *** paul-carlton1 has joined #openstack-nova | 14:52 | |
| *** salv-orl_ has quit IRC | 14:53 | |
| cdent | dansmith: hmmm. maybe you can do some light review on https://review.openstack.org/#/c/284963/ if you get a chance, I feel like I'm getting inconsistent feedback on the right way to do things. At the moment aggregates is an objectField with an AggregateList but it sounds like you're saying that's no good. | 14:54 |
| mriedem | ndipanov: so what happened with that sriov feature that landed in mitaka? was it not tested until the neutron stuff was landed last week to complete the integration? | 14:54 |
| *** salv-orlando has quit IRC | 14:54 | |
| *** thorst__ has joined #openstack-nova | 14:54 | |
| dansmith | cdent: where are the comments suggesting that over the alternative? | 14:55 |
| ndipanov | mriedem, I guess not too deeply | 14:55 |
| dansmith | both will work of course, but I would go the route of not using the list object | 14:55 |
| cdent | dansmith: not all feedback comes on reviews | 14:55 |
| cdent | and in some cases, null feedback is feedback | 14:55 |
| ndipanov | mriedem, I tested it when I was working on the neutron integration | 14:55 |
| ndipanov | but only so as far to boot VMs - I did not look into resource tracking too deeply | 14:56 |
| ndipanov | last week I did a more thorough test and saw that things were not working as expected | 14:56 |
| *** ametts has quit IRC | 14:57 | |
| openstackgerrit | Sylvain Bauza proposed openstack/nova: Add a RequestSpec generation migration script https://review.openstack.org/302828 | 14:57 |
| bauzas | alaski: dansmith: ^ | 14:57 |
| *** cristicalin has joined #openstack-nova | 14:58 | |
| *** thorst_ has quit IRC | 14:58 | |
| bauzas | this is an online data migration script for ensuring all instances having a RequestSpec object | 14:58 |
| *** ametts has joined #openstack-nova | 14:58 | |
| bauzas | kinda cellsv2/scheduler related | 14:58 |
| *** cristicalin has quit IRC | 14:58 | |
| *** cristicalin has joined #openstack-nova | 14:59 | |
| alaski | bauzas: cool, will look in a bit | 14:59 |
| *** ducttape_ has quit IRC | 14:59 | |
| *** mgoddard_ has joined #openstack-nova | 15:00 | |
| *** woodster_ has joined #openstack-nova | 15:00 | |
| *** thorst__ is now known as thorst________ | 15:00 | |
| openstackgerrit | Matt Riedemann proposed openstack/python-novaclient: Deprecate image list/show/delete/update CLIs/APIs https://review.openstack.org/302486 | 15:00 |
| markus_z | dansmith: I don't get your comment here https://review.openstack.org/#/c/302341/1/nova/conf/upgrade_levels.py | 15:00 |
| markus_z | We change the default value from None to "auto". Why doesn't this need a hint in a reno? | 15:01 |
| dansmith | cdent: review feedback not in the reviews surely doesn't help get towards a consistent goal, as you can imagine | 15:01 |
| dansmith | cdent: anyway, I commented in there on this topic, hope that helps | 15:01 |
| cdent | I know right! | 15:01 |
| cdent | thanks dansmith | 15:01 |
| *** markstur_ has joined #openstack-nova | 15:01 | |
| dansmith | markus_z: no, it totes does | 15:01 |
| markus_z | dansmith: that explains everything | 15:01 |
| dansmith | markus_z: I was referring to the '' behavior thing | 15:02 |
| dansmith | markus_z: I'll worry about a reno when I get to the fancy place of having the effing tests working :D | 15:02 |
| markus_z | dansmith: wait, you're giving me right, I wasn't expecting that.. | 15:02 |
| alaski | bauzas: haven't review thoroughly, but why not use the req_spec table for the marker? | 15:02 |
| dansmith | markus_z: giving you right? | 15:02 |
| mriedem | saying he's right | 15:03 |
| dansmith | markus_z: I'm saying you're right, yes | 15:03 |
| openstackgerrit | Lee Yarwood proposed openstack/nova: utils: Move is_volume_backed_instance into compute_utils https://review.openstack.org/281444 | 15:03 |
| openstackgerrit | Lee Yarwood proposed openstack/nova: WIP - libvirt: Implement stable device rescues https://review.openstack.org/270289 | 15:03 |
| openstackgerrit | Lee Yarwood proposed openstack/nova: compute: Allow instances with bdms to be rescued by capable virt drivers https://review.openstack.org/270288 | 15:03 |
| openstackgerrit | Lee Yarwood proposed openstack/nova: libvirt: Add boot ordering to individual disks https://review.openstack.org/270286 | 15:03 |
| openstackgerrit | Lee Yarwood proposed openstack/nova: virt: Provide the bdm of an instance during a rescue https://review.openstack.org/270287 | 15:03 |
| openstackgerrit | Lee Yarwood proposed openstack/nova: image_meta: Add hw_rescue_device and hw_rescue_bus https://review.openstack.org/270285 | 15:03 |
| dansmith | mriedem: I thought maybe he meant "giving me $something" where $something might be "shit" :) | 15:03 |
| markus_z | whops, german-english again, sorry | 15:03 |
| mriedem | du hast Recht! | 15:03 |
| bauzas | alaski: we could indeed | 15:03 |
| *** mgoddard has quit IRC | 15:03 | |
| markus_z | mriedem: Yeah! :) | 15:03 |
| dansmith | heh | 15:03 |
| mriedem | thanks google | 15:03 |
| mriedem | the ! was mine, not googles | 15:03 |
| *** jwcroppe has joined #openstack-nova | 15:04 | |
| markus_z | Finally we switch to a proper language here ;) | 15:04 |
| markus_z | The "!" is implied after each german sentence. | 15:04 |
| mriedem | exactly | 15:04 |
| dansmith | brennen die Tests | 15:04 |
| cdent | that's very useful information in those coments, dansmith. thank you | 15:04 |
| markus_z | dansmith: :D | 15:04 |
| cdent | It doesn't take much to make me not unhappy. | 15:04 |
| bauzas | üdo, üdo, das telefon rüft an | 15:04 |
| markus_z | google translate is not always right. We practice that in Austin. | 15:04 |
| dansmith | lol | 15:05 |
| kfarr | Is there a time period where nova reviewers will be ramping up on reviewing specs for Newton? If people are reviewing specs now, I'd appreciate eyes on https://review.openstack.org/#/c/247561/5 | 15:05 |
| * bauzas ladies and gentlemen, this was the only sentence I'm able to remember | 15:05 | |
| alaski | bauzas: also, check out https://review.openstack.org/#/c/301916/1/nova/cmd/manage.py@1381, there can be an issue with markers and unique constraints | 15:06 |
| markus_z | bauzas: This is only useful in extremely specific situations, but I don't speak French, so I don't complain here. | 15:06 |
| * markus_z has to go for today, see you | 15:06 | |
| *** 7JTAALB9C has joined #openstack-nova | 15:06 | |
| *** mjura has quit IRC | 15:06 | |
| bauzas | alaski: well, I was thinking of using a separate project | 15:07 |
| *** ababich has quit IRC | 15:07 | |
| bauzas | alaski: but lemme think of using request_spec table instead | 15:07 |
| *** j_king_ is now known as j_king | 15:07 | |
| *** ababich has joined #openstack-nova | 15:08 | |
| *** thorst________ is now known as thorst_ | 15:08 | |
| alaski | okay. but even with a separate project there's still a constraint on instance_uuid | 15:08 |
| mriedem | kfarr: http://lists.openstack.org/pipermail/openstack-dev/2016-March/090370.html | 15:08 |
| bauzas | okay, I should just reconsider using instance_mapping for this | 15:08 |
| bauzas | alaski: what I only need is a place to store an uuid | 15:09 |
| kfarr | thanks mriedem! | 15:09 |
| *** ababich has quit IRC | 15:09 | |
| *** diogogmt has joined #openstack-nova | 15:09 | |
| *** ababich has joined #openstack-nova | 15:09 | |
| bauzas | argh, UC on instance_uuid for req_spec | 15:09 |
| alaski | bauzas: yeah. I think it makes sense to use the table that's being migrated to whenever possible | 15:09 |
| *** jwcroppe has quit IRC | 15:10 | |
| alaski | the UC isn't too hard to work around, it just needs to be taken into account | 15:10 |
| bauzas | double-argh, we store the uuid correctly in the table (string(36) | 15:10 |
| *** Guest74036 has quit IRC | 15:10 | |
| bauzas | so I couldn't prefix that | 15:10 |
| alaski | I replaced '-' with ' ' | 15:10 |
| alaski | easy enough to change back | 15:10 |
| *** tbachman has quit IRC | 15:11 | |
| bauzas | alaski: well, I need to consider the problem | 15:13 |
| bauzas | using uuids having ':' could work, but that sounds very hacky | 15:14 |
| bauzas | and errorprone if we begin to support that as an uuid | 15:14 |
| ildikov | mriedem: dansmith: hi | 15:15 |
| ildikov | mriedem: dansmith: I plan to upload a new version for the multiattach spec for which it would be nice to agree on the service version check approach if that's still what we want to follow | 15:16 |
| *** tbachman has joined #openstack-nova | 15:17 | |
| *** anush_ has joined #openstack-nova | 15:19 | |
| *** baoli has quit IRC | 15:19 | |
| mriedem | ildikov: that is mostly out of my memory | 15:19 |
| *** burt has joined #openstack-nova | 15:20 | |
| mriedem | ildikov: did jgriffith and hemna come to a conclusion on which approach to take for detach? | 15:20 |
| ildikov | mriedem: I have to admit I don't have too many memories either, so some refreshing would be nice | 15:20 |
| ildikov | mriedem: they continued the discussion on the cinder channel and I think they analysed more the differences and I think it turned out that the proposals are not really different | 15:21 |
| mriedem | which one is a little bit country and which one is a little bit rock & roll? | 15:22 |
| ildikov | mriedem: but I did not read the whole discussion carefully | 15:22 |
| mriedem | https://www.youtube.com/watch?v=_AfXznngjGw | 15:22 |
| * mriedem imagines that's what the cinder team does after resolving issues | 15:23 | |
| *** jwcroppe has joined #openstack-nova | 15:23 | |
| ildikov | mriedem: they talked about the db entries and what is available where and then it turned out they have misunderstanding because the new Gerrit UI messed up visualisation :) | 15:23 |
| smcginnis | lol | 15:23 |
| mriedem | oh sure, it's all gerrit's fault | 15:23 |
| * smcginnis thinks with a wig, mriedem could pull off a mean Donny | 15:23 | |
| *** achanda has joined #openstack-nova | 15:24 | |
| dansmith | hah | 15:24 |
| *** eliqiao1 has quit IRC | 15:24 | |
| dansmith | I would pay money to see that | 15:24 |
| ildikov | mriedem: lol, I'm too visual for that video :D | 15:24 |
| mriedem | smcginnis: i haven't had a haircut in months so it's getting pretty feathery | 15:24 |
| ildikov | mriedem: well, I have to admit I don't like the new UI either, so that excuse works for me :) | 15:25 |
| *** alex_xu has quit IRC | 15:25 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/302016 | 15:26 |
| openstackgerrit | Dan Smith proposed openstack/nova: Default compute upgrade_level to auto https://review.openstack.org/302341 | 15:26 |
| openstackgerrit | Dan Smith proposed openstack/nova: Add AllServicesCurrent fixture https://review.openstack.org/302869 | 15:26 |
| *** yamahata has quit IRC | 15:26 | |
| *** kmartin has joined #openstack-nova | 15:27 | |
| jgriffith | smcginnis: I was thinking Marie... but ok :) | 15:27 |
| dansmith | mriedem: sc68cal owes me something for this can of worms that is changing that tiny default | 15:27 |
| smcginnis | jgriffith: Didn't want to go there. :) | 15:27 |
| mriedem | dansmith: how about a can of worms? | 15:27 |
| jgriffith | ha! | 15:27 |
| dansmith | no, have that already | 15:28 |
| dansmith | want something better | 15:28 |
| mriedem | a hook? | 15:28 |
| sc68cal | dansmith: a sushi platter? | 15:28 |
| dansmith | and not one of those mythical beers that nobody ever collects on | 15:28 |
| dansmith | sc68cal: too close to a can of worms | 15:28 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-novaclient: Updated from global requirements https://review.openstack.org/302884 | 15:28 |
| mriedem | ildikov: so we probably have to find the review comments that talked about the service version checking, it had to have been something that was related to passing things from the api to the compute | 15:28 |
| *** Swami has joined #openstack-nova | 15:28 | |
| mriedem | because we had to make sure the computes were all up to date | 15:28 |
| *** alex_xu has joined #openstack-nova | 15:29 | |
| mriedem | was it something like the attachment id? i want to say it was related to the issue with the BDMs | 15:29 |
| ildikov | mriedem: we discussed it mostly here | 15:29 |
| mriedem | which we put a fix in for that dansmith worked on | 15:29 |
| *** cristicalin has quit IRC | 15:29 | |
| ildikov | I think it was more about how to ensure that we only allow multiattach if the service version matches the minimum required | 15:30 |
| mriedem | did you ever update a spec for that while we were talking about it? | 15:30 |
| *** armax has joined #openstack-nova | 15:30 | |
| *** jwcroppe has quit IRC | 15:31 | |
| sdague | anyone else want to kick this very easy change in - https://review.openstack.org/#/c/302670/ ? | 15:31 |
| mriedem | ildikov: thought it was related to https://review.openstack.org/#/c/267169/ | 15:31 |
| dansmith | sdague: yes, yes I do | 15:31 |
| ildikov | mriedem: not really as we had the cache issue with that and I don't think we ever finalised how it should look like | 15:32 |
| dansmith | sdague: mtreinish hit that multinode devstack fix for me yesterday evening by the way | 15:32 |
| dansmith | sdague: so we should be good there | 15:32 |
| sdague | dansmith: yeh, I saw this morning. Good stuff | 15:32 |
| dansmith | sdague: also, this morning I woke up, and with my first breath, I could tell the world was a better place | 15:32 |
| dansmith | sdague: because we're no longer seeding flavors in our original migration | 15:33 |
| sdague | heh | 15:33 |
| ildikov | mriedem: partially but we also had to ensure we pass attachment_id at detach time | 15:33 |
| sdague | dansmith: idk, maybe we should bring back that hard coded struct from change #1 | 15:33 |
| ildikov | mriedem: and that was added in Mitaka, so we would've been in trouble with older computes IIRC | 15:33 |
| dansmith | sdague: I'd prefer that over the migration approach :D | 15:33 |
| *** links has joined #openstack-nova | 15:36 | |
| *** sc68cal has quit IRC | 15:36 | |
| ildikov | mriedem: ah and we also did not want inconsistent API behaviour in the sense of one compute supports multiattach and another one in the system does not | 15:36 |
| mriedem | ildikov: well, that is related to the api capabilities stuff that will be discussed at the summit | 15:36 |
| *** EinstCrazy has quit IRC | 15:36 | |
| ildikov | mriedem: so basically I think the aim was/is to have a fully upgraded environment in order to use multiattach to avoid inconsistencies and issues | 15:36 |
| mriedem | this was the change to use that new bdm objects method to avoid races https://review.openstack.org/#/c/267464/3/nova/compute/manager.py | 15:37 |
| *** mtanino has joined #openstack-nova | 15:37 | |
| *** achanda has quit IRC | 15:37 | |
| mriedem | but we never set a service version, i guess we'd just use whatever was the latest compute service version from mitaka | 15:37 |
| mriedem | when checking that in the newton api | 15:37 |
| ildikov | mriedem: ah, ok, I missed the API capabilities thing, but will follow up from now on | 15:38 |
| *** nelsnelson has joined #openstack-nova | 15:38 | |
| ildikov | yeah, I remember the BDM one with the new methods | 15:39 |
| *** moshele has quit IRC | 15:39 | |
| mriedem | dansmith: do you remember at all what we were talking about wrt multiattach and blocking from the api on the compute version? | 15:39 |
| *** mdrabe has quit IRC | 15:39 | |
| dansmith | I remember very little, but I thought if we deferred to newton then we didn't have to do that | 15:39 |
| dansmith | oh, no | 15:39 |
| *** nelsnels_ has quit IRC | 15:39 | |
| ildikov | I know we had issues with the version is cached and is not updated properly | 15:39 |
| *** ababich has quit IRC | 15:40 | |
| *** ababich1 has joined #openstack-nova | 15:40 | |
| dansmith | yeah, I dunno, that requires real though to remember the details and I'm in the middle of something else | 15:40 |
| dansmith | was it captured in a spec? | 15:40 |
| ildikov | dansmith: hmm, if we end up with not doing it that would be perfectly fine with me :) | 15:40 |
| mriedem | there is a ML thread from january | 15:41 |
| ildikov | dansmith: we never really agreed about that fully so it wasn't | 15:41 |
| mriedem | i'm hoping the details are in there | 15:41 |
| mriedem | http://lists.openstack.org/pipermail/openstack-dev/2016-January/084031.html | 15:41 |
| mriedem | "We talked about adding a check in the API layer to see if there are any liberty computes running and if not, then we could allow the 2nd attach on a multi-attach volume, but that is racy and not fail-safe, which could mean a user could get a multi-attach request to pass in one case but fail in another (it's a problem related to how the service versions are cached in-memory per API worker and isn't something that can be resolved be | 15:42 |
| ildikov | mriedem: cool, thanks, I totally forgot about the mail thread | 15:42 |
| * mriedem thanks my past self | 15:43 | |
| mriedem | i think it was directly related to the potential for races without https://review.openstack.org/#/c/267169/ | 15:43 |
| ildikov | mriedem: just wanted to say the same :) | 15:43 |
| mriedem | so we needed https://review.openstack.org/#/c/267464/3/nova/compute/manager.py in the mitaka computes | 15:43 |
| *** tinbar has quit IRC | 15:44 | |
| *** afranc has joined #openstack-nova | 15:44 | |
| ildikov | mriedem: we were also discussing capabilities and how to communicate that the feature is now available | 15:47 |
| *** jaosorior has quit IRC | 15:47 | |
| *** scheuran has quit IRC | 15:47 | |
| ildikov | the other part is the race and the need of the attachment_id in detach | 15:47 |
| *** jaosorior has joined #openstack-nova | 15:47 | |
| *** ababich1 has quit IRC | 15:48 | |
| *** links has quit IRC | 15:48 | |
| *** ababich has joined #openstack-nova | 15:48 | |
| *** baoli has joined #openstack-nova | 15:49 | |
| *** matrohon has quit IRC | 15:49 | |
| *** ababich has quit IRC | 15:50 | |
| mriedem | passing cinder the attachment_id during detach landed in mitaka https://review.openstack.org/#/c/193134/ | 15:50 |
| *** ababich has joined #openstack-nova | 15:50 | |
| mriedem | the ML thread said that https://review.openstack.org/#/c/256273/ was also needed for detach in mitaka, but that was reverted | 15:51 |
| mriedem | b/c older cinder couldn't handle it | 15:51 |
| *** sfinucan has quit IRC | 15:52 | |
| mriedem | and the libvirt driver didn't get the multiattach changes in mitaka either, so in newton, the api can't do multiattach to any mitaka computes | 15:53 |
| *** browne has joined #openstack-nova | 15:53 | |
| *** pcaruana has quit IRC | 15:53 | |
| *** derekh has quit IRC | 15:53 | |
| *** mdrabe has joined #openstack-nova | 15:53 | |
| mriedem | well, the api can attach a volume to an instance on a mitaka compute, | 15:53 |
| mriedem | but it can't attach an already attached volume to a 2nd instance on a mitaka compute | 15:54 |
| mriedem | maybe that's where we got to with the service version checking | 15:54 |
| *** 7JTAALB9C has quit IRC | 15:54 | |
| openstackgerrit | Waldemar Znoinski proposed openstack/nova: virt/hardware: Fix 'isolate' case on non-SMT hosts https://review.openstack.org/285321 | 15:55 |
| mriedem | ildikov: the capabilities stuff, like not being able to do multiattach on xen, is hopefully something that gets sorted out during the api design summit session on api discoverability/capability which i think alaski and sdague are running | 15:55 |
| mriedem | although we said in the mailing list that today we'd just fail the multiattach request with NoValidHost | 15:55 |
| mriedem | otherwise it seems we'd need to add a capabilities filter to the scheduler for this | 15:56 |
| *** ducttape_ has joined #openstack-nova | 15:56 | |
| openstackgerrit | Waldemar Znoinski proposed openstack/nova: virt/hardware: Add diagnostic logs for scheduling https://review.openstack.org/285322 | 15:56 |
| mriedem | which is what i think cinder has for multiattach, a capabilities filter in the scheduler b/c not all volume drivers support multiattach | 15:57 |
| ildikov | yeah, it's correct that attach couldn't make it even into the computes in Mitaka | 15:57 |
| ildikov | so if we don't want inconsistent behaviour we will need the service version or smth like | 15:58 |
| mriedem | and i seem to remember dansmith yielding a bit on the api caching/racing issue with the service version checking during the mitaka midcycle, but i could have misinterpreted that | 15:59 |
| ildikov | I assume we can add a capabilities filter in an iteration as well, right? | 15:59 |
| *** apoorvad has joined #openstack-nova | 16:00 | |
| *** ccarmack has joined #openstack-nova | 16:00 | |
| ildikov | I'm not sure whether it was on the midcycle or before, but I remember as well he was | 16:00 |
| *** ametts has quit IRC | 16:01 | |
| *** mtanino has quit IRC | 16:02 | |
| *** armax has quit IRC | 16:02 | |
| *** armax has joined #openstack-nova | 16:02 | |
| mriedem | the more i think about the nova capabilities filter in the scheduler, that doesn't make sense in this case, | 16:02 |
| mriedem | when attaching a volume to an instance, we aren't using the scheduler, we're just casting to the compute that the instance is on | 16:02 |
| *** mtanino has joined #openstack-nova | 16:03 | |
| ildikov | boot from volume? | 16:03 |
| mriedem | boot from volume with a multiattach volume would be an issue for scheduling, but we said we aren't supporting BFV with multiattach right? | 16:03 |
| ildikov | live migration? | 16:03 |
| ildikov | we are not supporting booting from the actual multiattach volume | 16:03 |
| ildikov | but we're supporting booting with that volume | 16:03 |
| *** dave-mccowan has joined #openstack-nova | 16:03 | |
| *** angdraug has joined #openstack-nova | 16:04 | |
| openstackgerrit | Dan Smith proposed openstack/nova: Default compute upgrade_level to auto https://review.openstack.org/302341 | 16:05 |
| dansmith | bauzas: ^ | 16:05 |
| mriedem | you'd need a scheduler capabilities filter in that case unless we accept that you could just fail with a NoValidHost | 16:05 |
| ildikov | this how I meant to add that filter later and for now just fail | 16:05 |
| ildikov | if that's acceptable | 16:06 |
| *** gszasz has quit IRC | 16:06 | |
| ildikov | or just simply fail | 16:06 |
| *** FL1SK has quit IRC | 16:06 | |
| mriedem | i honestly don't remember what we talked about for the BFV case, didn't you have a change up that amended the spec for that? | 16:06 |
| *** itamarl has quit IRC | 16:06 | |
| *** mgoddard_ has quit IRC | 16:06 | |
| ildikov | I amended | 16:06 |
| mriedem | i don't want to even think about live migration with multiattach volumes... | 16:07 |
| ildikov | when the source and target is volume and the bootindex or how it is called is 0 we fail | 16:07 |
| *** vtech has quit IRC | 16:07 | |
| ildikov | live migration is an issue because of the detach misery isn't it? | 16:07 |
| *** yamahata has joined #openstack-nova | 16:07 | |
| *** akshai has quit IRC | 16:08 | |
| *** vtech has joined #openstack-nova | 16:08 | |
| *** mgoddard has joined #openstack-nova | 16:08 | |
| mriedem | can you remind me why we even want volume multiattach in the first place again? what's the use case/benefit? | 16:08 |
| *** e0ne has quit IRC | 16:08 | |
| ildikov | but well, I think the amended spec says we don't support it with multiattach volume | 16:08 |
| ildikov | use it as a hot standby for instance with clustered FS | 16:09 |
| *** e0ne has joined #openstack-nova | 16:09 | |
| *** syjulian has joined #openstack-nova | 16:09 | |
| *** zenoway has quit IRC | 16:09 | |
| dansmith | sdague: is there an award for "added the most test fixtures this cycle" ? | 16:09 |
| *** zenoway has joined #openstack-nova | 16:10 | |
| *** vtech has quit IRC | 16:10 | |
| ildikov | mriedem: the spec: https://specs.openstack.org/openstack/nova-specs/specs/mitaka/approved/multi-attach-volume.html | 16:11 |
| *** zenoway has quit IRC | 16:11 | |
| *** zenoway has joined #openstack-nova | 16:11 | |
| sarafraj | markus_z: Is this a valid bug https://bugs.launchpad.net/nova/+bug/1567232? | 16:13 |
| openstack | Launchpad bug 1567232 in OpenStack Compute (nova) "migrate vm, do not check neutron-openvswitch-agent status" [Undecided,New] | 16:13 |
| *** ekuris has quit IRC | 16:13 | |
| *** ababich has quit IRC | 16:13 | |
| *** unicell has quit IRC | 16:13 | |
| *** ababich has joined #openstack-nova | 16:14 | |
| *** unicell has joined #openstack-nova | 16:15 | |
| mriedem | so the BFV limitations are in https://specs.openstack.org/openstack/nova-specs/specs/mitaka/approved/multi-attach-volume.html#rest-api-impact | 16:15 |
| *** zenoway has quit IRC | 16:15 | |
| mriedem | we don't really say anything about landing on a compute that doesn't support multiattach, and even if we did, if it's the first attachment for the volume and it's libvirt, it would be fine, | 16:16 |
| femnad | sarafraj: he left a while ago | 16:16 |
| *** zenoway has joined #openstack-nova | 16:16 | |
| femnad | sarafraj: but that seems to be a valid bug if you ask me | 16:16 |
| mriedem | the problem is BFV with a multiattach volume that is already attached (i think) and you land on another host (e.g. xen) that doesn't support it | 16:16 |
| femnad | sarafraj: you need multiple hosts to test it though | 16:16 |
| ildikov | mriedem: at a certain point we said we will discuss this further so I didn't amend the spec more | 16:17 |
| *** ametts has joined #openstack-nova | 16:17 | |
| *** PaulMurray has joined #openstack-nova | 16:17 | |
| mriedem | so what happens if i boot from volume with a multiattach volume to an instance on a kvm compute node, and then later try to boot from that same volume to another host that's on a xen compute? | 16:18 |
| ildikov | mriedem: I can reach out to the hypervisor teams and check how it works without libvirt if it does at all | 16:18 |
| sarafraj | femnad: Thanks. I am trying to find what release or branch this bug was file against. | 16:18 |
| *** mylu has quit IRC | 16:19 | |
| femnad | sarafraj: yeap, that seems missing what is otherwise a great report | 16:19 |
| mriedem | i guess we said, "Only the libvirt driver is modified to support multi-attach, for all other virt drivers this capability is disabled, the information is stored among the virt driver capabilities dict in the base ComputeDriver." | 16:19 |
| danpb | mriedem: iirc, if one host has got it in mutli-attach mode, cinder shoudl reject request from any other host to use it in exclusive mode | 16:19 |
| ildikov | mriedem: if not being able to boot the instance without the volume then today the whole boot process fails | 16:19 |
| *** jlanoux_ has quit IRC | 16:19 | |
| danpb | IOW, other virt drivers should not need todo anything special - it should be handled by cinder | 16:19 |
| mriedem | danpb: ok, i wasn't sure if the compute manager needed to check a virt driver capabliities when doing attach_volume with a multiattach volume and a driver that didn't support multiattach | 16:20 |
| *** mylu has joined #openstack-nova | 16:20 | |
| *** zenoway has quit IRC | 16:20 | |
| mriedem | so if vol A is attached to instance B on kvm host X, and we try to attach vol A to instance C on xen host Y, cinder should puke during the os-initialize_connection or os-detach call... | 16:21 |
| mriedem | is what i'm hearing | 16:21 |
| * mriedem gets lunch quick | 16:22 | |
| danpb | mriedem: that's what I'd expect to happen - of course someone ought to validate that it does :-) | 16:22 |
| *** tbachman_ has joined #openstack-nova | 16:23 | |
| *** anush_ has quit IRC | 16:23 | |
| ildikov | danpb: I can only validate that i would be happy with this scenario :) | 16:23 |
| *** anush_ has joined #openstack-nova | 16:24 | |
| *** tbachman has quit IRC | 16:25 | |
| *** tbachman_ is now known as tbachman | 16:25 | |
| ildikov | hemna: can you confirm the scenario above? ^^ | 16:26 |
| hemna | hey | 16:26 |
| *** ekuris has joined #openstack-nova | 16:26 | |
| *** nic has joined #openstack-nova | 16:26 | |
| hemna | Cinder doesnt know anything about what virt layer is being used anywhere | 16:27 |
| hemna | all it knows is host, instance_uuid and the state of the volume | 16:27 |
| ildikov | hmm then it's still on Nova side | 16:28 |
| *** Bjoern_ has joined #openstack-nova | 16:28 | |
| *** femnad has left #openstack-nova | 16:28 | |
| danpb | hemna: cinder should know about whether the volume is used in shared mode, or exclusive mode though | 16:29 |
| ildikov | hemna: tnx | 16:29 |
| *** Bjoern_ is now known as BjoernT | 16:30 | |
| hemna | cinder doesn't | 16:30 |
| hemna | only the driver really knows currently | 16:30 |
| hemna | and that's specific to the array/backend | 16:30 |
| hemna | ideally, it would be great to have every attach to it's own separate target (non shared) | 16:30 |
| hemna | but some arrays/backends can't do that | 16:31 |
| hemna | jgriffith, seemed to have some magic to make unique /dev/disk/by-path entries show up via udev for shared mode targets | 16:31 |
| *** unicell has quit IRC | 16:31 | |
| hemna | if that's possible, then that really does solve everything. | 16:31 |
| hemna | but it has to work for every transport (iSCSI, FC, etc) | 16:31 |
| danpb | the code seems to show that cinder *is* checking at the api level | 16:32 |
| hemna | it would eliminate the need for a flag for shared mode | 16:32 |
| hemna | danpb, ? | 16:33 |
| danpb | ie in cinder/volume/manager.py the attach_volume method does | 16:33 |
| danpb | if (volume['status'] == 'in-use' and not volume['multiattach'] | 16:33 |
| danpb | and not volume['migration_status']): | 16:33 |
| danpb | raise exception.InvalidVolume( | 16:33 |
| danpb | reason=_("volume is already attached")) | 16:33 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Move config options from nova/api directory https://review.openstack.org/258180 | 16:33 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve the help text for the API options (4) https://review.openstack.org/297237 | 16:33 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve the help text for the API options (2) https://review.openstack.org/260587 | 16:33 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve the help text for the API options (1) https://review.openstack.org/260586 | 16:33 |
| openstackgerrit | Ed Leafe proposed openstack/nova: Improve the help text for the API options (3) https://review.openstack.org/260588 | 16:33 |
| hemna | that multiattach flag is simply a flag put on the volume that allows it to be multiattached | 16:33 |
| hemna | there could be some confusion in terms here. | 16:33 |
| *** moshele has joined #openstack-nova | 16:34 | |
| hemna | volume['multiattach'] is added to the volume at user create time, to allow it to be attached more than once (if the backend supports it) | 16:34 |
| hemna | what I was referring to earlier with 'shared/non shared' mode is | 16:35 |
| *** haukebruno has quit IRC | 16:35 | |
| hemna | for multi-attach, when the driver creates a target for the initiator, is it always the same target for the host regardless of how many times its attached (shared mode), or a brand new target for every attach (non-shared) | 16:35 |
| hemna | this is the trickiest part of it all | 16:36 |
| hemna | so in shared mode | 16:36 |
| *** mylu has quit IRC | 16:36 | |
| *** david-lyle has quit IRC | 16:36 | |
| hemna | the same n-cpu host will always see the same /dev/disk/by-path/<volume> path for every attach on that host. | 16:36 |
| hemna | vs. | 16:36 |
| *** haplo37 has joined #openstack-nova | 16:37 | |
| hemna | a brand new /dev/disk/by-path/<volume> path for each attach. | 16:37 |
| *** browne has quit IRC | 16:37 | |
| hemna | if it's shared nova has to be careful and not call os-brick.disconnect_volume unles it's the very last usage of that /dev/disk/by-path/<volume> | 16:37 |
| hemna | I hope I explained that ok | 16:38 |
| *** Oku_OS is now known as Oku_OS-away | 16:38 | |
| mriedem | ildikov: hemna: i think we probably need a set of scenarios to work everything through when we talk about this stuff, like mixed level nova computes, different virt drivers, shared/non-shared volume drivers, boot from volume, etc, it gets too easy to miss something otherwise | 16:39 |
| hemna | yah | 16:39 |
| hemna | it's very complex | 16:39 |
| *** Oku_OS-away is now known as Oku_OS | 16:39 | |
| mriedem | and with that we can figure out what we think should happen, like if it's a failure or if it's ok, | 16:40 |
| hemna | let me work on the detach diagram | 16:40 |
| mriedem | and then the code has to be written against those | 16:40 |
| hemna | and live migration | 16:40 |
| mriedem | after lunch i can at least write up some of the nova scenarios and then we can fill it in | 16:40 |
| ildikov | mriedem: do we have room for a design session with a whiteboard? | 16:41 |
| ildikov | mriedem: or maybe use the contributors' meetup on Friday? | 16:41 |
| mriedem | ildikov: we have a fishbowl room for the design session, i'm assuming there will be a whiteboard | 16:42 |
| ildikov | not as a starting point but to finalise | 16:42 |
| mriedem | i'm expecting some overlap into the friday meetup | 16:42 |
| mriedem | 40 min in a design session isn't going to be enough time to work this out probably, so i expect more time on friday | 16:42 |
| *** ekuris has quit IRC | 16:42 | |
| *** mylu has joined #openstack-nova | 16:42 | |
| ildikov | ok, I wasn't sure how much time we have for this topic | 16:42 |
| *** e0ne has quit IRC | 16:42 | |
| ildikov | but I agree we will prolly need Friday as well | 16:43 |
| mriedem | this is why i said i think multiattach is going to be the only thing we have time for in the 40 min session | 16:43 |
| *** mingyu has quit IRC | 16:43 | |
| *** hydhu has joined #openstack-nova | 16:43 | |
| mriedem | we won't have time to talk about extended volumes etc | 16:43 |
| *** mingyu has joined #openstack-nova | 16:43 | |
| ildikov | ah, ok I thought we share the session with other topics this is why I asked about Friday | 16:44 |
| *** daemontool has quit IRC | 16:44 | |
| hemna | yah I think we need more time in a room | 16:44 |
| ildikov | yeap | 16:46 |
| ildikov | we can spend some time with collecting scenarios on one of the meetings | 16:47 |
| *** xyang1 has quit IRC | 16:47 | |
| *** mingyu has quit IRC | 16:48 | |
| *** hydhu has quit IRC | 16:48 | |
| *** Oku_OS is now known as Oku_OS-away | 16:48 | |
| *** aNupoisc has joined #openstack-nova | 16:50 | |
| ildikov | I added a scenarios section to the etherpad | 16:52 |
| *** rcernin has quit IRC | 16:52 | |
| openstackgerrit | Roman Podoliaka proposed openstack/nova: wip: remove redundant API call to neutron https://review.openstack.org/302954 | 16:53 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/302016 | 16:54 |
| *** ZZelle_ has joined #openstack-nova | 16:55 | |
| dansmith | what do people think? next april 1st should we make the "OpenStack Proposal Bot" propose some interestinger things? Like marriage, or world peace, etc? | 16:56 |
| dansmith | proposing just requirements seems really boring | 16:56 |
| rpodolyaka | lol | 16:56 |
| *** jwcroppe has joined #openstack-nova | 16:58 | |
| *** PaulMurray has quit IRC | 16:58 | |
| *** angdraug has quit IRC | 16:58 | |
| artom | A proposal bot sounds like something a love-sick telemarketer would build... | 16:58 |
| dansmith | or a 30 year old that still lives in his parents' basement | 16:59 |
| artom | And an awkward silence befell the room. | 16:59 |
| dansmith | heh | 16:59 |
| * dansmith makes a note to use "befell" more often | 17:00 | |
| artom | I'll start trawling your submissions looking for "befell" in the comments | 17:00 |
| *** lpetrut has quit IRC | 17:00 | |
| dansmith | And a -2 befell your silly patch | 17:00 |
| artom | Oh please please please use that somewhere :D | 17:01 |
| *** jistr has quit IRC | 17:01 | |
| *** mylu has quit IRC | 17:02 | |
| *** xyang1 has joined #openstack-nova | 17:02 | |
| dansmith | heh | 17:02 |
| *** jwcroppe has quit IRC | 17:02 | |
| *** mylu has joined #openstack-nova | 17:03 | |
| *** danpb has quit IRC | 17:03 | |
| *** yamahata has quit IRC | 17:06 | |
| *** david-lyle has joined #openstack-nova | 17:06 | |
| *** ametts has quit IRC | 17:07 | |
| *** rmart04 has quit IRC | 17:07 | |
| *** pcaruana has joined #openstack-nova | 17:08 | |
| *** ccarmack has quit IRC | 17:09 | |
| *** BobBall is now known as BobBall_AWOL | 17:10 | |
| *** unicell has joined #openstack-nova | 17:10 | |
| *** aNupoisc has quit IRC | 17:11 | |
| *** david-lyle has quit IRC | 17:11 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/302016 | 17:12 |
| *** salv-orlando has joined #openstack-nova | 17:12 | |
| *** jwcroppe has joined #openstack-nova | 17:13 | |
| *** aNupoisc has joined #openstack-nova | 17:14 | |
| mriedem | wth http://logs.openstack.org/70/302670/1/gate/gate-nova-python27-db/77945a3/console.html#_2016-04-07_16_01_24_219 | 17:15 |
| openstackgerrit | Matt Riedemann proposed openstack/nova: remove alembic from requirements.txt https://review.openstack.org/302670 | 17:15 |
| *** salv-orlando has quit IRC | 17:15 | |
| *** p01nt3r75 has quit IRC | 17:15 | |
| *** baoli has quit IRC | 17:16 | |
| *** baoli has joined #openstack-nova | 17:17 | |
| *** david-lyle has joined #openstack-nova | 17:17 | |
| sneti | Hi johnthetubaguy, can you reapprove this patch. I rebased it. | 17:18 |
| *** Marga_ has quit IRC | 17:19 | |
| *** cdent has quit IRC | 17:20 | |
| *** ametts has joined #openstack-nova | 17:20 | |
| *** browne has joined #openstack-nova | 17:21 | |
| *** Prophidys has joined #openstack-nova | 17:21 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/302016 | 17:22 |
| openstackgerrit | Artom Lifshitz proposed openstack/nova: Fix max concurrent builds's unlimited semaphore https://review.openstack.org/239054 | 17:22 |
| *** ccarmack has joined #openstack-nova | 17:22 | |
| *** baoli has quit IRC | 17:23 | |
| *** ljxiash has quit IRC | 17:23 | |
| *** david-lyle has quit IRC | 17:23 | |
| *** ljxiash has joined #openstack-nova | 17:24 | |
| sdague | um, what just happened to all our unit tests - http://logs.openstack.org/33/302733/2/gate/gate-nova-python27-db/f8f9374/console.html ? | 17:25 |
| sdague | ovo release? | 17:25 |
| *** kfarr has quit IRC | 17:25 | |
| *** slagle has quit IRC | 17:26 | |
| dansmith | uh | 17:26 |
| mriedem | sdague: yeah i saw that on yours | 17:26 |
| mriedem | i rebased it but didn't check dep releases | 17:26 |
| *** Marga_ has joined #openstack-nova | 17:26 | |
| *** mingyu has joined #openstack-nova | 17:27 | |
| dansmith | last ovo release was 3/27, AFAICT | 17:28 |
| dansmith | which contained no code changes | 17:28 |
| *** ljxiash has quit IRC | 17:28 | |
| sdague | TypeError: _doit() takes at least 1 argument (0 given) | 17:29 |
| sdague | that's all over the place | 17:29 |
| dansmith | yeah | 17:29 |
| mriedem | http://logstash.openstack.org/#dashboard/file/logstash.json?query=message%3A%5C%22TypeError%3A%20_doit%5C%5C(%5C%5C)%20takes%20at%20least%201%20argument%20%5C%5C(0%20given%5C%5C)%5C%22%20AND%20project%3A%5C%22openstack%2Fnova%5C%22%20AND%20tags%3A%5C%22console%5C%22&from=7d | 17:29 |
| mriedem | logstash seems to be behind, only showing hits on one change | 17:29 |
| dansmith | there is no "doit" in o.vo | 17:30 |
| diana_clarke | is it related to the fixtures release? | 17:30 |
| dansmith | that's in our fixture | 17:30 |
| alaski | it's fixtures | 17:30 |
| mriedem | fixtures 2.0.0 was released last night | 17:30 |
| diana_clarke | lifeless popped into chat yesterday to warn us about that | 17:31 |
| alaski | I'm working on a fix, but the short story is that it's now really strict about methods being patched with like methods | 17:31 |
| dansmith | ah, so in monkeypatch? | 17:31 |
| alaski | dansmith: yeah | 17:32 |
| dansmith | awesome | 17:32 |
| alaski | if you want to patch a class method, you need to patch with a class method | 17:32 |
| mriedem | https://github.com/openstack/requirements/commit/9e1ee3b3ac20371b8556d10d1ae0bb0ede3c0048 | 17:32 |
| alaski | and on python2 it must be from the same class even | 17:32 |
| alaski | it's a major version bump because it's not compatible | 17:32 |
| alaski | how did it get pulled in? | 17:32 |
| *** markstur_ has quit IRC | 17:32 | |
| mriedem | the bot | 17:32 |
| *** anush_ has quit IRC | 17:32 | |
| mriedem | see ^ | 17:32 |
| mriedem | b/c the bot doesn't run unit tests | 17:32 |
| mriedem | with fixtures | 17:33 |
| mriedem | well, the requirements repo doesn't | 17:33 |
| *** anush_ has joined #openstack-nova | 17:33 | |
| *** rcernin has joined #openstack-nova | 17:33 | |
| *** moshele has quit IRC | 17:33 | |
| *** rlrossit has quit IRC | 17:33 | |
| dansmith | alaski: oh, it's patching the class, but because we want to intercept the instance method | 17:33 |
| alaski | right. it used to be that it just patched it and it always became an instancemethod | 17:34 |
| alaski | now it checks if it should be a staticmethod and creates a patched staticmethod | 17:35 |
| alaski | but a side effect of how it was done is that things have to match really closely now | 17:35 |
| alaski | http://pastebin.com/LDEMth7R fixes the _doit thing | 17:36 |
| alaski | which is ugly and not what we want | 17:36 |
| alaski | it that just gets it done to ~400 failing tests from a few other monkeypatches that need to be addressed | 17:36 |
| alaski | I'm currently fighting py2/3 differences to get a patch up to fixtures | 17:37 |
| dansmith | alaski: beautifully hacky | 17:38 |
| *** aNupoisc has quit IRC | 17:39 | |
| *** ctrath has quit IRC | 17:40 | |
| *** ctrath has joined #openstack-nova | 17:40 | |
| *** PaulMurray has joined #openstack-nova | 17:40 | |
| *** vtech has joined #openstack-nova | 17:41 | |
| alaski | :) | 17:41 |
| *** ihrachys has quit IRC | 17:41 | |
| *** vtech has quit IRC | 17:41 | |
| alaski | I'm learning a lot more than I ever expected about bound vs unbound methods | 17:41 |
| tbachman | lol | 17:42 |
| alaski | and this is one thing that py3 makes nicer | 17:42 |
| *** vtech has joined #openstack-nova | 17:42 | |
| dansmith | s/expected/wanted/ | 17:43 |
| alaski | yeah :) | 17:43 |
| *** mvk has quit IRC | 17:45 | |
| *** rlrossit has joined #openstack-nova | 17:48 | |
| *** anush_ has quit IRC | 17:50 | |
| *** anush_ has joined #openstack-nova | 17:50 | |
| *** pwnall has quit IRC | 17:51 | |
| *** ducttape_ has quit IRC | 17:51 | |
| *** aginwala has joined #openstack-nova | 17:52 | |
| *** rmart04 has joined #openstack-nova | 17:52 | |
| *** e0ne has joined #openstack-nova | 17:53 | |
| *** ducttape_ has joined #openstack-nova | 17:53 | |
| *** david-lyle has joined #openstack-nova | 17:56 | |
| *** david-lyle has quit IRC | 17:56 | |
| *** mdrabe has quit IRC | 17:57 | |
| *** ducttape_ has quit IRC | 17:57 | |
| *** david-lyle has joined #openstack-nova | 17:58 | |
| *** pushkaru has joined #openstack-nova | 17:58 | |
| *** ctrath has quit IRC | 17:58 | |
| *** tbachman has quit IRC | 17:58 | |
| openstackgerrit | Nikola Dipanov proposed openstack/nova: pci: Make sure PF is 'available' when last VF is freed https://review.openstack.org/303012 | 17:59 |
| *** anush_ has quit IRC | 18:02 | |
| *** e0ne has quit IRC | 18:03 | |
| *** ociuhandu has quit IRC | 18:04 | |
| *** haplo37 has quit IRC | 18:08 | |
| *** aginwala has quit IRC | 18:10 | |
| openstackgerrit | Ed Leafe proposed openstack/nova: Remove unused config option 'fake_call' https://review.openstack.org/282419 | 18:11 |
| *** zenoway has joined #openstack-nova | 18:12 | |
| *** sahid has quit IRC | 18:12 | |
| *** mingyu has quit IRC | 18:12 | |
| *** vilobhmm11 has joined #openstack-nova | 18:13 | |
| *** e0ne has joined #openstack-nova | 18:13 | |
| *** vilobhmm111 has joined #openstack-nova | 18:14 | |
| *** mdrabe has joined #openstack-nova | 18:14 | |
| *** e0ne has quit IRC | 18:14 | |
| *** aginwala has joined #openstack-nova | 18:14 | |
| *** aginwala has quit IRC | 18:16 | |
| *** zenoway has quit IRC | 18:16 | |
| *** efoley has joined #openstack-nova | 18:17 | |
| *** vilobhmm11 has quit IRC | 18:18 | |
| *** anush_ has joined #openstack-nova | 18:19 | |
| *** jichen has joined #openstack-nova | 18:19 | |
| *** artom has quit IRC | 18:20 | |
| jichen | oomichi: hi, could you please help to take a look at https://review.openstack.org/#/c/288153/? I added some comments ,thanks | 18:20 |
| *** ociuhandu has joined #openstack-nova | 18:21 | |
| *** crobertsrh has joined #openstack-nova | 18:22 | |
| *** crobertsrh has left #openstack-nova | 18:22 | |
| *** Marga_ has quit IRC | 18:24 | |
| *** madorn has joined #openstack-nova | 18:24 | |
| *** madorn has left #openstack-nova | 18:24 | |
| *** sdake_ has joined #openstack-nova | 18:26 | |
| *** doug-fish has joined #openstack-nova | 18:26 | |
| *** sdake has quit IRC | 18:26 | |
| *** paul-carlton1 has quit IRC | 18:29 | |
| *** zenoway has joined #openstack-nova | 18:30 | |
| *** ssurana has joined #openstack-nova | 18:31 | |
| sdague | is there a fix up for nova for that fixtures change? | 18:31 |
| alaski | sdague: not that I know of, I'm working on the fixtures side right now | 18:31 |
| sdague | alaski: as in a fix to fixtures? | 18:32 |
| dansmith | alaski: your hack seemed okay to unblock us for now, no? | 18:32 |
| alaski | sdague: but the form of the Nova change needs to be like http://pastebin.com/LDEMth7R I think | 18:32 |
| alaski | sdague: yes | 18:32 |
| sdague | do we consider it regression? | 18:32 |
| dansmith | sdague: it's surely a change in behavior | 18:32 |
| *** salv-orlando has joined #openstack-nova | 18:32 | |
| alaski | It's a problem that we bumped major versions in fixtures without testing first | 18:32 |
| alaski | we went 1.4->2.0 | 18:32 |
| dansmith | ah | 18:32 |
| alaski | because there is a backwards incompatible change here | 18:32 |
| dims | alaski : that's the 2nd report i have for fixtures.. | 18:33 |
| dims | lifeless : should we have capped fixtures in g-r/u-c? ^^ | 18:33 |
| alaski | dansmith: and yes, my fix should unblock us. It's just needed in a few more places | 18:33 |
| alaski | I didn't track them all down yet | 18:33 |
| dansmith | alaski: ugh | 18:33 |
| sdague | dims: neutron is definitely blowing up in a similar way | 18:33 |
| dims | sdague : y ihar lined up 2 patches | 18:34 |
| sdague | so I think we might want to move the constraint back to 1.4.0 | 18:34 |
| dims | just u-c sdague? | 18:34 |
| sdague | well, I guess the question is, can the fix worth on both sides of this? | 18:34 |
| sdague | or do we just have to bite the bullet and fix things now | 18:34 |
| dansmith | the fix will make 2.0.1 doable for us I think | 18:35 |
| *** BjoernT has quit IRC | 18:35 | |
| dansmith | it will make our fix nicer | 18:35 |
| dansmith | I would think capping now and working on the nova changes would be better than blocking until we apply his hack everywhere | 18:35 |
| sdague | dims: sound good to you? | 18:36 |
| alaski | +1 | 18:36 |
| dims | sdague : we may have to coordinate with neutron, not sure if doing the cap will break them again | 18:36 |
| sdague | right, that's my concern, if people are actually fixing this already then block breaks new people | 18:36 |
| *** ctrath has joined #openstack-nova | 18:37 | |
| dims | at least neutron, last i checked they were going ahead | 18:37 |
| sdague | ok, then I guess we just go ahead probably | 18:38 |
| sdague | neutron and nova seem to be the only ones affected by it | 18:38 |
| *** salv-orlando has quit IRC | 18:38 | |
| dansmith | I don't see why rolling back will be a problem | 18:38 |
| sdague | dansmith: because if they land a fix that's not compatible with 1.4.0 | 18:39 |
| dims | right | 18:39 |
| *** vilobhmm111 has quit IRC | 18:39 | |
| *** slagle has joined #openstack-nova | 18:39 | |
| dansmith | sdague: yeah, I understand what incompatible means, I'm saying I don't know why alaski's change for example would not be compatible with 1.4 | 18:39 |
| sdague | dansmith: I don't know, that was my question | 18:40 |
| dansmith | it's a valid change regardless of whether fixtures is being overly pedantic | 18:40 |
| sdague | are our changes going to work on 1.4 and 2.0? | 18:40 |
| *** cdent has joined #openstack-nova | 18:40 | |
| mriedem | and will neutron's? | 18:40 |
| mriedem | i have a patch up to make fixtures>=2.0 in g-r | 18:40 |
| sdague | right | 18:40 |
| alaski | anyone have a test name for the _doit failure? | 18:40 |
| alaski | I can check real quick | 18:41 |
| mriedem | nova.tests.unit.api.openstack.compute.legacy_v2.test_servers.ServersControllerCreateTest.test_create_instance_pass_disabled | 18:41 |
| sdague | dims: their fix just removes the use of fixtures for that thing - https://review.openstack.org/#/c/302997/2 | 18:42 |
| mriedem | sdague: is there a bug up for tracking this yet? | 18:42 |
| sdague | mriedem: not that I know of | 18:42 |
| mriedem | i guess we could pile on https://bugs.launchpad.net/neutron/+bug/1567295 | 18:42 |
| openstack | Launchpad bug 1567295 in neutron "tox -e py27 is failing " [High,In progress] - Assigned to Gary Kotton (garyk) | 18:42 |
| *** akshai has joined #openstack-nova | 18:42 | |
| *** aginwala has joined #openstack-nova | 18:43 | |
| *** ducttape_ has joined #openstack-nova | 18:43 | |
| alaski | okay, my hack works on 1.4 | 18:44 |
| *** angdraug has joined #openstack-nova | 18:44 | |
| dims | sdague : gary seems to have W-1's his patch https://review.openstack.org/#/c/302737/ | 18:44 |
| sdague | dims: I think ihar's is the fix, right? | 18:44 |
| dims | yep | 18:45 |
| sdague | https://review.openstack.org/#/c/302997/ | 18:45 |
| sdague | that should work with a block as well | 18:45 |
| dims | ok | 18:45 |
| dansmith | that's broken | 18:45 |
| sdague | though the time to land that is probably just as long as landing the nova fix, right? | 18:45 |
| dansmith | because it doesn't ever stop the mock | 18:45 |
| sdague | heh, yeh | 18:46 |
| sdague | it will mean a ton of mocks stacked up right? | 18:46 |
| sdague | it might actually work, but for the wrong reason | 18:46 |
| mriedem | one of you guys better -1 that | 18:46 |
| sdague | just did | 18:47 |
| dansmith | I did | 18:47 |
| sdague | ok, where is alaski's fix? | 18:48 |
| *** hemna is now known as hemnafk | 18:48 | |
| alaski | not pushed up yet. I have a partial fix, just an example of how to fix | 18:48 |
| dansmith | sdague: someone has to apply it to all the other places where it's broken | 18:49 |
| *** delattec has quit IRC | 18:49 | |
| dansmith | it's just one part of the fix.. the pattern | 18:49 |
| mriedem | neutron-lbaas is also busted | 18:49 |
| sdague | how many places is it busted in nova? | 18:50 |
| mriedem | shouldn't we cap for now? | 18:50 |
| *** aginwala has quit IRC | 18:50 | |
| alaski | I can do the fix, I just need a few minutes to finish up the fixtures commit message I'm working on to describe everything | 18:50 |
| dansmith | mriedem: I think we should | 18:50 |
| alaski | sdague: not sure, the first place I applied it fixed about 4000 failures leaving ~400. probably a handful of places after that | 18:50 |
| mriedem | from what i see on http://status.openstack.org/elastic-recheck/data/uncategorized.html it's mostly just neutron and nova right now, there are some other projects' py27 jobs failing but not in the same way | 18:50 |
| sdague | mriedem: you want to push that in parallel, and we can then decide what to do with patches that have test results on them | 18:50 |
| mriedem | sdague: sure | 18:51 |
| sdague | mriedem: logstash hints it's just nova and neutron (stadium) | 18:51 |
| *** vtech has quit IRC | 18:51 | |
| mriedem | yup | 18:51 |
| dims | mriedem : so we can just adjust u-c back to the old one? | 18:51 |
| sdague | dims: that's just going to be overriden tomorrow by the proposal though | 18:52 |
| sdague | we probably want to temporarily cap | 18:52 |
| dims | sdague : y, i have been holding back things i know that break like webob | 18:52 |
| *** salv-orlando has joined #openstack-nova | 18:52 | |
| sdague | or temporarily block | 18:52 |
| sdague | fortunatelly the mitaka release is actually out, as this could have fubared that | 18:52 |
| sdague | mitaka is frozen on uc, right? | 18:53 |
| *** pwnall has joined #openstack-nova | 18:53 | |
| dims | sdague : yes, mriedem and i talked that the u-c is a defacto cap | 18:53 |
| dims | mriedem : want to change your review to block 2.0 temporarily? | 18:54 |
| mriedem | yeah https://review.openstack.org/#/c/303034/ | 18:54 |
| *** aginwala has joined #openstack-nova | 18:57 | |
| mriedem | logstash workers spiked pretty hard in the last 24 hours it looks like so logstash isn't showing us the full failure | 18:58 |
| *** PaulMurray has quit IRC | 18:58 | |
| syjulian | hi mriedem, I changed the BP according to your comment. Please let me know what you think https://blueprints.launchpad.net/nova/+spec/nova-create-notification-scheduler-hints | 18:58 |
| *** ctrath has quit IRC | 18:59 | |
| *** e0ne has joined #openstack-nova | 19:00 | |
| *** salv-orlando has quit IRC | 19:01 | |
| openstackgerrit | jichenjc proposed openstack/nova: Remove stub_compute config options for console https://review.openstack.org/234217 | 19:01 |
| openstackgerrit | jichenjc proposed openstack/nova: Use instance object instead of db record https://review.openstack.org/234218 | 19:01 |
| *** pushkaru has quit IRC | 19:03 | |
| *** cdelatte has joined #openstack-nova | 19:03 | |
| *** ccarmack has quit IRC | 19:05 | |
| *** aginwala has quit IRC | 19:05 | |
| *** nikhil has quit IRC | 19:07 | |
| *** rlrossit has quit IRC | 19:10 | |
| alaski | https://github.com/testing-cabal/fixtures/pull/26 btw | 19:10 |
| *** nikhil has joined #openstack-nova | 19:11 | |
| *** mingyu has joined #openstack-nova | 19:12 | |
| *** david-lyle_ has joined #openstack-nova | 19:13 | |
| *** salv-orlando has joined #openstack-nova | 19:14 | |
| *** Marga_ has joined #openstack-nova | 19:14 | |
| *** hdd has joined #openstack-nova | 19:14 | |
| *** vtech has joined #openstack-nova | 19:14 | |
| *** hdd has left #openstack-nova | 19:15 | |
| *** david-lyle has quit IRC | 19:15 | |
| *** mingyu has quit IRC | 19:17 | |
| *** aginwala has joined #openstack-nova | 19:17 | |
| *** Sukhdev has joined #openstack-nova | 19:18 | |
| *** david-lyle_ has quit IRC | 19:19 | |
| *** pushkaru has joined #openstack-nova | 19:21 | |
| *** ccarmack has joined #openstack-nova | 19:23 | |
| *** ihrachys has joined #openstack-nova | 19:24 | |
| *** tbachman has joined #openstack-nova | 19:25 | |
| *** mdrabe has quit IRC | 19:26 | |
| *** mdrabe has joined #openstack-nova | 19:26 | |
| *** jaosorior has quit IRC | 19:27 | |
| *** david-lyle has joined #openstack-nova | 19:29 | |
| *** mdrabe has quit IRC | 19:29 | |
| *** mdrabe has joined #openstack-nova | 19:29 | |
| *** ctrath has joined #openstack-nova | 19:30 | |
| *** vilobhmm11 has joined #openstack-nova | 19:30 | |
| *** gagehugo has quit IRC | 19:31 | |
| *** david-lyle has quit IRC | 19:31 | |
| *** david-lyle has joined #openstack-nova | 19:31 | |
| *** cdelatte has quit IRC | 19:31 | |
| *** Jeffrey4l_ has joined #openstack-nova | 19:32 | |
| *** mvk has joined #openstack-nova | 19:32 | |
| *** Jeffrey4l has quit IRC | 19:34 | |
| lifeless | dims: g-r? no. u-c - u-c is exact fits | 19:34 |
| lifeless | alaski: has to be from the same class? Thats really odd | 19:35 |
| *** rlrossit has joined #openstack-nova | 19:35 | |
| lifeless | alaski: surely an unintended side effect? | 19:35 |
| *** mvk_ has joined #openstack-nova | 19:35 | |
| *** mylu has quit IRC | 19:35 | |
| alaski | lifeless: seems to be only py2 for that one. If it's bound to another class you get a TypeError because self doesn't match the expected class | 19:36 |
| *** david-lyle has quit IRC | 19:37 | |
| alaski | it's explained in http://stackoverflow.com/questions/14298844/method-objects-vs-function-objects-python-class-instances-vs-class pretty well | 19:38 |
| *** mylu has joined #openstack-nova | 19:38 | |
| alaski | in the answer | 19:38 |
| *** mvk has quit IRC | 19:39 | |
| *** baoli has joined #openstack-nova | 19:39 | |
| lifeless | alaski: ack; I need to get daughter off to kindy, then I'll have full attention on this | 19:39 |
| alaski | okay. I should run that change through the full Nova test suite as well | 19:40 |
| dims | thanks lifeless | 19:40 |
| lifeless | alaski: AIUI it constructs the class we're patching onto? we can't do that... | 19:42 |
| lifeless | alaski: looking @ inspect.signature might be useful, or some of the internals of mock | 19:42 |
| alaski | lifeless: okay I'll take a look. I just need to know if it's static or not, which isn't clear from getattr unless you pull if from the class | 19:43 |
| alaski | constructed class | 19:43 |
| *** aginwala has quit IRC | 19:45 | |
| lifeless | alaski: look at im_self to tell the difference | 19:48 |
| mriedem | hmmm, odd, if n-cpu isn't running and you try to list instances, you get a 500 error | 19:49 |
| melwitt | wat | 19:49 |
| mriedem | at least on my neutron devstack | 19:50 |
| dansmith | that' | 19:50 |
| dansmith | is broken | 19:50 |
| melwitt | I guess I haven't tried that in a long time but I'm surprised | 19:50 |
| dansmith | that's definitely something needing fixing if it's real | 19:50 |
| *** efoley has quit IRC | 19:50 | |
| melwitt | yeah | 19:50 |
| dansmith | both perf wise and availability wise | 19:51 |
| sdague | mriedem: you got more details on that? | 19:51 |
| mriedem | checking it out, trying to get logs | 19:51 |
| sdague | because... I don't even understand how that could be a thing | 19:51 |
| dansmith | yeah | 19:51 |
| mriedem | hmm, neutron-server isn't running here either | 19:51 |
| dansmith | performance would be so terrible even when it's working | 19:51 |
| sdague | list instances should be n-api <-> db | 19:51 |
| sdague | oh... if neutron-server is down, I can believe it | 19:52 |
| *** diogogmt has quit IRC | 19:52 | |
| dansmith | hopefully we don't ping neutron on a list either | 19:52 |
| dansmith | but maybe indicates more trouble | 19:52 |
| sdague | security groups | 19:52 |
| dansmith | on a list? | 19:52 |
| *** diogogmt has joined #openstack-nova | 19:52 | |
| sdague | networks are in the base list | 19:52 |
| *** rlrossit has quit IRC | 19:53 | |
| dansmith | we get networks from info_cache | 19:53 |
| sdague | I guess security groups aren't | 19:53 |
| sdague | ok, idk | 19:53 |
| melwitt | well, I think most of the apis call compute.api get first before doing anything | 19:53 |
| dansmith | list would be get_all | 19:53 |
| dansmith | but still, that should be fine without neutron or n-cpu | 19:54 |
| mriedem | i have to restack this, it's all fubar | 19:54 |
| melwitt | oh, yeah. | 19:54 |
| *** alex_xu has quit IRC | 19:54 | |
| melwitt | I was thinking of the "do something to an instance" which is different | 19:54 |
| dansmith | even still, it shouldn't hit either of those things until/unless you need to do something involving the compute | 19:54 |
| dansmith | if we are, then we need to fix it | 19:54 |
| melwitt | yeah | 19:55 |
| *** alex_xu has joined #openstack-nova | 19:55 | |
| melwitt | looks like self.compute_api.get_all for list all | 19:56 |
| dansmith | right | 19:57 |
| melwitt | doesn't seem to do anything other than query the db | 19:58 |
| *** anush_ has quit IRC | 19:58 | |
| mriedem | yeah don't waste time on this, i'm restacking to see if i can recreate it | 19:59 |
| oomichi | jichen: are you still here? | 20:00 |
| *** ctrath has quit IRC | 20:00 | |
| *** yamahata has joined #openstack-nova | 20:00 | |
| *** anush_ has joined #openstack-nova | 20:00 | |
| jichen | oomichi:yes, I am here | 20:00 |
| *** ametts has quit IRC | 20:01 | |
| oomichi | jichen: I mean it is nice to store operated keys into a list and return the list to api user when a problem happens in the middle | 20:02 |
| *** tbachman_ has joined #openstack-nova | 20:02 | |
| *** DCWilliams_VA has joined #openstack-nova | 20:02 | |
| *** ccarmack has quit IRC | 20:02 | |
| oomichi | jichen: that is a temporary list and returned to api user as an api response | 20:02 |
| *** tbachman has quit IRC | 20:02 | |
| *** tbachman_ is now known as tbachman | 20:02 | |
| vilobhmm11 | alaski : ping | 20:03 |
| alaski | vilobhmm11: hey | 20:03 |
| oomichi | jichen: when receving the list, api users can know what keys are operated already | 20:03 |
| oomichi | jichen: before a problem happens | 20:03 |
| vilobhmm11 | alaski : need you insights on this topic http://www.gossamer-threads.com/lists/openstack/dev/53889 | 20:03 |
| *** baoli has quit IRC | 20:03 | |
| vilobhmm11 | if you can reply to the ML it would be helpful for the cross project effort | 20:04 |
| *** baoli has joined #openstack-nova | 20:04 | |
| alaski | okay. It may not be today but I'll respond | 20:04 |
| *** tbachman has quit IRC | 20:04 | |
| vilobhmm11 | alaski : cool thanks! no rush | 20:04 |
| oomichi | jichen: that is my thinking on https://review.openstack.org/#/c/288153 | 20:05 |
| vilobhmm11 | alaski : meanwhile do you have some etherpad docs from previous summit where this topic "need for reservation" was discussed in detail | 20:05 |
| alaski | lifeless: I was overcomplicating this because I was pdb debugging and looking at the printout of the methods not the type. the types distinguish between function and instancemethod even when the printout differs on bound or not | 20:05 |
| mriedem | oomichi: i just realized that the os-simple-tenant-usage testing in tempest is not validating the datetime format in the response for started_at/ended_at | 20:05 |
| mriedem | it just checks that it's a string | 20:05 |
| jichen | oomichi: um..we don't raise exception ,just catach and print debug information, at line 158, we still still return the dict ? | 20:05 |
| jichen | oomichi: of this file https://review.openstack.org/#/c/288153/4/nova/api/openstack/compute/quota_sets.py | 20:06 |
| mriedem | but if nova started sending a different datetime format back, we'd be breaking the api w/o knowing it | 20:06 |
| alaski | vilobhmm11: I'm not aware of it being discussed in detail before | 20:06 |
| vilobhmm11 | alaski : okay | 20:06 |
| oomichi | mriedem: let me check tempest code now | 20:06 |
| mriedem | https://github.com/openstack/tempest-lib/blob/master/tempest_lib/api_schema/response/compute/v2_1/tenant_usages.py#L34 | 20:07 |
| *** mkrcmari__ has joined #openstack-nova | 20:07 | |
| oomichi | mriedem: ah, so fast. thanks for pointing it out | 20:07 |
| mriedem | i was just looking at it | 20:07 |
| mriedem | maybe enforcing the datetime format in the response is too strict? i'm not sure. | 20:08 |
| oomichi | mriedem: then, which is nova side patch? | 20:08 |
| mriedem | but it looks like novaclient is expecting a certain format | 20:08 |
| oomichi | mriedem: it is fine to make tempest side validation strict for me to block backwads incompatible changes | 20:09 |
| mriedem | oomichi: the response value for started_at is here https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/simple_tenant_usage.py#L132 | 20:09 |
| mriedem | which is instance.launched_at, which is an oslo.versionedobject DateTimeField | 20:09 |
| mriedem | which gets it's string version here https://github.com/openstack/oslo.versionedobjects/blob/master/oslo_versionedobjects/fields.py#L403 | 20:09 |
| mriedem | and the format for that is here https://github.com/openstack/oslo.versionedobjects/blob/master/oslo_versionedobjects/_utils.py#L21 | 20:10 |
| *** mvk_ has quit IRC | 20:10 | |
| *** aginwala has joined #openstack-nova | 20:11 | |
| alaski | lifeless: except py34 makes this hard. py27: getattr(obj, 'foo') returns a method if obj.foo is not static, py34: it returns a function | 20:12 |
| *** jed56 has quit IRC | 20:13 | |
| *** lpetrut has joined #openstack-nova | 20:13 | |
| *** efoley has joined #openstack-nova | 20:13 | |
| oomichi | mriedem: I see, do we have already a bug report on LP? | 20:13 |
| mriedem | oomichi: no | 20:14 |
| oomichi | mriedem: if not, I am doing that | 20:14 |
| oomichi | mriedem: ok, I wil | 20:14 |
| *** vishwanathj has quit IRC | 20:14 | |
| mriedem | ok | 20:15 |
| oomichi | mriedem: https://bugs.launchpad.net/tempest/+bug/1567640 | 20:16 |
| openstack | Launchpad bug 1567640 in tempest "Need to validate date and time format of a response of tenant_usage API" [Medium,New] | 20:16 |
| *** tbachman has joined #openstack-nova | 20:16 | |
| oomichi | mriedem: maybe gmann or me will do that | 20:17 |
| *** ametts has joined #openstack-nova | 20:17 | |
| mriedem | oomichi: thanks, i added some details | 20:17 |
| *** ayoung has quit IRC | 20:18 | |
| oomichi | mriedem: related to this, how about having these schemas on nova tree instead of tempest tree? we will be able to implement/maintain these schemas | 20:18 |
| mriedem | i kind of like having tempest keeping nova honest | 20:19 |
| mriedem | because if you broke this in nova, you'd have to intentionally get it changed in tempest to work | 20:19 |
| oomichi | mriedem: these schemas are nova specific and this discussion happened 2 years ago when started implementing these schama | 20:19 |
| oomichi | mriedem: yeah, nice point. we can block such kind of change with multiple patches across repos | 20:20 |
| oomichi | actually, I saw this kind of situations before | 20:20 |
| *** DCWilliams_VA has quit IRC | 20:22 | |
| oomichi | jichen: sorry for my late response. yeah, that you pointed is common case in nova tree now | 20:22 |
| *** eharney has quit IRC | 20:22 | |
| *** sdake_ is now known as sdake | 20:23 | |
| *** ljxiash has joined #openstack-nova | 20:24 | |
| mriedem | dansmith: melwitt: sdague: ok i don't know what was going on before, but after restacking and stopping nova-compute, and waiting for it to be shown as down in service-list, i'm still able to list instances | 20:25 |
| mriedem | so, well, that's good :) | 20:25 |
| jichen | oomichi: np, so I think common case already handle the expectation you have? or you want me to remove the 'deleted' one from the list? | 20:26 |
| *** eharney has joined #openstack-nova | 20:26 | |
| *** pwnall has quit IRC | 20:28 | |
| melwitt | mriedem: did you try stopping the neutron-server too | 20:28 |
| *** ljxiash has quit IRC | 20:29 | |
| mriedem | melwitt: i can quick | 20:29 |
| dansmith | mriedem: yeah | 20:29 |
| melwitt | okay. was just thinking about it since you mentioned both were down | 20:29 |
| *** tongli has joined #openstack-nova | 20:29 | |
| *** mrkz has quit IRC | 20:30 | |
| *** aginwala has quit IRC | 20:30 | |
| mriedem | yikes... | 20:31 |
| mriedem | 500 | 20:31 |
| mriedem | wtf | 20:31 |
| *** mrkz has joined #openstack-nova | 20:31 | |
| dansmith | mriedem: can you see what fails from the trace or something? | 20:31 |
| dansmith | 'cause that's terribad | 20:31 |
| dansmith | IMHO | 20:31 |
| mriedem | http://paste.openstack.org/show/493416/ | 20:31 |
| dansmith | I'm kinda shocked no big clouds have complained | 20:31 |
| mriedem | security groups | 20:31 |
| *** tbachman has quit IRC | 20:31 | |
| dansmith | well, | 20:32 |
| dansmith | yes and ports, but maybe that's how they have to find them | 20:32 |
| dansmith | gawd | 20:32 |
| mriedem | yes, it gets the ports for hte instances and the security groups from the ports | 20:32 |
| dansmith | so we ping neutron once for each instance in list? | 20:32 |
| *** tbachman has joined #openstack-nova | 20:32 | |
| dansmith | ah, no | 20:32 |
| dansmith | looks like it's a batch operation at least | 20:32 |
| dansmith | but still.. ughomatic | 20:32 |
| mriedem | yes, it's not great | 20:32 |
| mriedem | i was recently looking at that code | 20:32 |
| mriedem | https://bugs.launchpad.net/nova/+bug/1567655 | 20:34 |
| openstack | Launchpad bug 1567655 in OpenStack Compute (nova) "500 error when trying to list instances and neutron-server is down" [Undecided,New] | 20:34 |
| dansmith | if this is at thing we need we should probably cache this stuff in info_cache | 20:34 |
| dansmith | alaski: you know anything about this? | 20:34 |
| alaski | I don't recall ever seeing that | 20:35 |
| mriedem | melwitt: btw, thanks for reminding me to try that, my brain is fried today :) | 20:35 |
| dansmith | oh | 20:35 |
| dansmith | heh | 20:35 |
| dansmith | no secgroups in cells right? | 20:35 |
| mriedem | right | 20:35 |
| alaski | oh yeah, that's true | 20:35 |
| dansmith | that's probably why big clouds don't hurt on this | 20:35 |
| mriedem | and cern uses nova-net | 20:35 |
| dansmith | and cells | 20:36 |
| *** vtech has quit IRC | 20:36 | |
| dansmith | well, as much as I know I'll regret saying this | 20:36 |
| dansmith | I might be a good person to work on this | 20:36 |
| dansmith | object change and all | 20:37 |
| *** ducttape_ has quit IRC | 20:37 | |
| dansmith | or at least do some work on it | 20:37 |
| melwitt | dansmith: awesome! you can't take it back now | 20:38 |
| dansmith | the hell I can't | 20:38 |
| *** vtech has joined #openstack-nova | 20:38 | |
| mriedem | i think anyone but me right now is good because i'm leaving early | 20:38 |
| dansmith | well, this is going to be big | 20:38 |
| dansmith | I won't even start on it until monday | 20:38 |
| mriedem | storing security groups in info_cache is weird though | 20:38 |
| dansmith | why? | 20:38 |
| melwitt | mriedem: thanks for unearthing that. that's a crazy bug | 20:38 |
| mriedem | given we have....instance.security_groups | 20:38 |
| mriedem | info_cache is the list of vifs right? | 20:39 |
| dansmith | info_cache is a cache of shit that is actually neutron's | 20:39 |
| dansmith | no | 20:39 |
| mriedem | is there secgroup stuff in there? | 20:39 |
| dansmith | info_cache just only has one property | 20:39 |
| dansmith | which is network_info | 20:39 |
| *** aginwala has joined #openstack-nova | 20:40 | |
| mriedem | yabbut https://github.com/openstack/nova/blob/master/nova/network/model.py#L467 | 20:40 |
| dansmith | that's networkinfo | 20:40 |
| dansmith | https://github.com/openstack/nova/blob/master/nova/objects/instance_info_cache.py#L43 | 20:40 |
| dansmith | which is just there | 20:40 |
| *** ametts has quit IRC | 20:41 | |
| *** ericksonsantos has joined #openstack-nova | 20:41 | |
| *** baoli has quit IRC | 20:41 | |
| mriedem | which https://github.com/openstack/nova/blob/master/nova/objects/fields.py#L608 | 20:41 |
| *** cdelatte has joined #openstack-nova | 20:41 | |
| mriedem | and https://github.com/openstack/nova/blob/master/nova/objects/fields.py#L623 | 20:41 |
| dansmith | right | 20:41 |
| mriedem | and there is no security group stuff in there | 20:41 |
| dansmith | I'm not sure what your point is | 20:42 |
| mriedem | not saying you can't add it... | 20:42 |
| oomichi | jichen: can you take a look at https://review.openstack.org/#/c/288153/ ? I wrote a comment | 20:42 |
| *** ducttape_ has joined #openstack-nova | 20:42 | |
| dansmith | I'm not adding secgroups to network_info | 20:42 |
| dansmith | I'm saying a new property on info_cache to store secgroup cache info as well | 20:42 |
| mriedem | oh | 20:43 |
| mriedem | well why didn't you just say that?! :P | 20:43 |
| * dansmith looks for a line to quote | 20:43 | |
| jichen | oomichi: ok, will do , thanks! | 20:43 |
| dansmith | [13:34:36] <dansmith>if this is at thing we need we should probably cache this stuff in info_cache | 20:43 |
| dansmith | there we go :) | 20:43 |
| mriedem | that made 0 sense to me | 20:44 |
| mriedem | obviously | 20:44 |
| mriedem | so it doesn't count | 20:44 |
| mriedem | anyway, | 20:44 |
| mriedem | i'm heading out | 20:44 |
| dansmith | we could also just store it in instance_extra | 20:44 |
| dansmith | and then we can cache the actual instance.security_groups | 20:44 |
| mriedem | yeah......but.... | 20:44 |
| dansmith | either way, we have to have a db migration | 20:44 |
| dansmith | info_cache kinda seems like the better place given the similarity to us caching the network info | 20:44 |
| mriedem | seems like it would make more sense living closer to info_cache that is already for networky things | 20:45 |
| mriedem | yes | 20:45 |
| dansmith | yeah | 20:45 |
| mriedem | instance.extra should just be renamed instance.dump | 20:45 |
| *** baoli has joined #openstack-nova | 20:45 | |
| dansmith | cool, so who is up for doing this work? | 20:45 |
| mriedem | i'm up for you doing it | 20:45 |
| lifeless | alaski: the model is different in 2 vs 3; it may well need a conditional. nearly with you | 20:45 |
| * dansmith looks around for "you" | 20:45 | |
| mriedem | dansmith: actually you're doing inventory stuff | 20:45 |
| mriedem | and need to review cells things | 20:46 |
| mriedem | so it probably shouldn't be you | 20:46 |
| *** baoli_ has joined #openstack-nova | 20:46 | |
| dansmith | \o\ /o/ \o/ | 20:46 |
| mriedem | make rlrossit do it | 20:46 |
| *** mvk_ has joined #openstack-nova | 20:46 | |
| mriedem | let's talk more tomorrow | 20:46 |
| * mriedem is leaving | 20:46 | |
| *** pwnall has joined #openstack-nova | 20:46 | |
| *** mriedem has quit IRC | 20:46 | |
| * dansmith works up an escape plan | 20:46 | |
| *** vishwana_ has joined #openstack-nova | 20:47 | |
| *** Sukhdev has quit IRC | 20:47 | |
| *** burgerk has quit IRC | 20:47 | |
| *** rfolco has quit IRC | 20:47 | |
| *** rwmjones_hols is now known as rwmjones | 20:49 | |
| *** mkrcmari__ has quit IRC | 20:50 | |
| *** baoli has quit IRC | 20:50 | |
| *** mkrcmari__ has joined #openstack-nova | 20:50 | |
| *** baoli_ has quit IRC | 20:50 | |
| *** baoli has joined #openstack-nova | 20:51 | |
| openstackgerrit | Andrew Bogott proposed openstack/nova: Add the libvirt.ephemeral_image_format config option https://review.openstack.org/301861 | 20:51 |
| alaski | lifeless: I worked out something that's not great but works | 20:52 |
| alaski | although it fails on latest python because getargspec is apparently going away | 20:52 |
| *** mvk_ has quit IRC | 20:54 | |
| *** mylu has quit IRC | 20:54 | |
| dims | alaski : inspect.getargspec? | 20:54 |
| alaski | dims: yeah, replaced with getfullargspec | 20:54 |
| *** mylu has joined #openstack-nova | 20:54 | |
| alaski | or inspect.signature | 20:54 |
| *** aginwala has quit IRC | 20:55 | |
| *** pwnall has quit IRC | 20:55 | |
| *** mvk_ has joined #openstack-nova | 20:55 | |
| *** aginwala has joined #openstack-nova | 20:56 | |
| *** david-lyle has joined #openstack-nova | 20:56 | |
| dims | alaski : interesting, we haven't started using getfullargspec in oslo.* | 20:57 |
| alaski | dims: it's in 3.4, but apparently not in the nightly build https://travis-ci.org/testing-cabal/fixtures/builds/121535942 | 20:57 |
| dims | our CI is 3.4 | 20:57 |
| dims | whoa. looks like we may have to rush that :) | 20:57 |
| dims | harlowja : ^^ | 20:58 |
| harlowja | sup | 20:59 |
| *** mkrcmari__ has quit IRC | 20:59 | |
| harlowja | fixtures broken? | 20:59 |
| harlowja | wasn't me, lol | 20:59 |
| dims | harlowja : we use inspect.getargspec a lot, it's going way in 3.5 | 21:00 |
| harlowja | ya, funcsigs to the rescue | 21:01 |
| harlowja | and/or code in https://github.com/openstack/oslo.utils/blob/master/oslo_utils/reflection.py | 21:01 |
| harlowja | via get_signature () there | 21:01 |
| *** pcaruana has quit IRC | 21:01 | |
| *** davideagnello has quit IRC | 21:02 | |
| harlowja | although i thought getargspec was deprecated, not removed | 21:02 |
| *** davideagnello has joined #openstack-nova | 21:02 | |
| *** mingyu has joined #openstack-nova | 21:02 | |
| harlowja | https://docs.python.org/3/library/inspect.html#inspect.getargspec | 21:02 |
| dims | removed in nightlies already. alaski was pointing at https://travis-ci.org/testing-cabal/fixtures/builds/121535942 | 21:02 |
| harlowja | kk, weird | 21:02 |
| harlowja | that seems ummm, broken | 21:03 |
| alaski | that looks to be running 3.6 | 21:03 |
| harlowja | lifeless yt | 21:03 |
| harlowja | or other python contributor (haypo not here) | 21:03 |
| *** lucasagomes has quit IRC | 21:03 | |
| *** mylu has quit IRC | 21:04 | |
| *** e0ne has quit IRC | 21:04 | |
| *** eharney has quit IRC | 21:05 | |
| harlowja | alaski https://hg.python.org/cpython/rev/32c8bdcd66cc | 21:05 |
| harlowja | 'Resurrect inspect.getargspec in 3.6. Backout a565aad5d6e1.' | 21:05 |
| harlowja | dims ^ | 21:05 |
| harlowja | seems like that's being reverted | 21:05 |
| dims | :) | 21:05 |
| harlowja | phew | 21:05 |
| harlowja | when we move to py4 maybe it will be an issue? | 21:06 |
| *** jichen has quit IRC | 21:06 | |
| dims | LOL, you mean python8? (https://mail.python.org/pipermail/python-dev/2016-March/143603.html) | 21:06 |
| *** mingyu has quit IRC | 21:07 | |
| harlowja | victor... | 21:08 |
| harlowja | lol | 21:08 |
| openstackgerrit | Sujitha proposed openstack/nova: Config options: Centralize servicegroup options https://review.openstack.org/303074 | 21:09 |
| harlowja | dims yes though, we should start to look at not using getargspec anymore wherever it is used | 21:10 |
| *** tbachman has quit IRC | 21:10 | |
| dims | +1 | 21:10 |
| harlowja | and use the signature stuff (or other helper routines in oslo.utils) | 21:10 |
| harlowja | thats the story and i'm sticking to it :-P | 21:10 |
| *** pwnall has joined #openstack-nova | 21:11 | |
| *** lucasagomes has joined #openstack-nova | 21:11 | |
| *** raildo is now known as raildo-afk | 21:11 | |
| *** e0ne has joined #openstack-nova | 21:11 | |
| *** thorst_ has quit IRC | 21:12 | |
| *** ndipanov has quit IRC | 21:13 | |
| *** efoley_ has joined #openstack-nova | 21:17 | |
| *** tongli has quit IRC | 21:17 | |
| *** shakamunyi has joined #openstack-nova | 21:17 | |
| *** thorst_ has joined #openstack-nova | 21:18 | |
| *** efoley has quit IRC | 21:18 | |
| *** ivase has quit IRC | 21:18 | |
| *** salv-orl_ has joined #openstack-nova | 21:18 | |
| *** e0ne has quit IRC | 21:18 | |
| *** ivase has joined #openstack-nova | 21:19 | |
| *** mylu has joined #openstack-nova | 21:19 | |
| *** mylu has quit IRC | 21:20 | |
| *** thorst__ has joined #openstack-nova | 21:21 | |
| lifeless | alaski: ok, I'm on deck | 21:22 |
| lifeless | harlowja: hi | 21:22 |
| *** thorst_ has quit IRC | 21:22 | |
| *** salv-orlando has quit IRC | 21:22 | |
| harlowja | lifeless all good, i think i found the issue | 21:22 |
| harlowja | and/or someone else reverted it in upstream python nightly | 21:23 |
| *** ivase has quit IRC | 21:23 | |
| *** mylu has joined #openstack-nova | 21:24 | |
| lifeless | harlowja: what was it? | 21:24 |
| harlowja | https://hg.python.org/cpython/rev/32c8bdcd66cc getargspec going away | 21:24 |
| harlowja | but then it came back | 21:24 |
| *** shakamunyi has quit IRC | 21:24 | |
| *** Sukhdev has joined #openstack-nova | 21:25 | |
| *** thorst__ has quit IRC | 21:25 | |
| *** mylu has quit IRC | 21:29 | |
| alaski | lifeless: I pushed up some additional revs on the patch. The latest is much better, but I'm still finding some test failures in Nova with it. Trying to see which side is doing something wrong | 21:30 |
| *** rtheis has quit IRC | 21:31 | |
| *** ayoung has joined #openstack-nova | 21:32 | |
| lifeless | harlowja: ah, well its replaced by inspect.signature, no ? | 21:32 |
| harlowja | lifeless yes, except for a bunch of people using getargspec i think | 21:32 |
| lifeless | harlowja: sure, but deprecated :) | 21:32 |
| harlowja | ya | 21:32 |
| lifeless | harlowja: anyhow - how did you end up looking at that? | 21:32 |
| harlowja | alaski and dims were doing some stuffs :-P | 21:33 |
| harlowja | and they saw https://travis-ci.org/testing-cabal/fixtures/builds/121535942 | 21:33 |
| harlowja | and getargspec go away in nightly | 21:33 |
| *** ducttape_ has quit IRC | 21:33 | |
| lifeless | oh | 21:34 |
| lifeless | heh | 21:34 |
| lifeless | alaski: ^ so yeah, use inspect.signature (and we can drag in funcsigs to get signature on < 3.3) | 21:34 |
| alaski | okay | 21:35 |
| lifeless | funcsigs is a dep of mock already, so its in u-c etc etc | 21:35 |
| harlowja | oslo.utils.reflection should do that | 21:36 |
| alaski | does funcsigs just need to be in requirements? | 21:36 |
| harlowja | should already be there | 21:36 |
| lifeless | alaski: for fixtures? yeah. funcsigs>1, then in the module do something like | 21:36 |
| harlowja | https://github.com/openstack/oslo.utils/blob/master/oslo_utils/reflection.py#L37 | 21:36 |
| harlowja | oh | 21:36 |
| harlowja | for fixtures, ya, not there, ha | 21:36 |
| lifeless | if not getattr(inspect, 'signature', None): from funcsigs import signature as inspectsignature else: inspectsignature = inspect.signature | 21:37 |
| *** lucasagomes has quit IRC | 21:37 | |
| *** baoli has quit IRC | 21:37 | |
| *** baoli has joined #openstack-nova | 21:38 | |
| *** edtubill has quit IRC | 21:39 | |
| alaski | okay, with two small fixes to Nova the latest patch to fixtures works for me | 21:41 |
| alaski | on py27 at least | 21:41 |
| lifeless | \o/ | 21:41 |
| lifeless | what do those fixes look like ? | 21:41 |
| *** syjulian has quit IRC | 21:43 | |
| *** thorst_ has joined #openstack-nova | 21:44 | |
| alaski | http://paste.openstack.org/show/493423/ | 21:44 |
| alaski | I'm actually surprised that was working before now | 21:45 |
| *** edtubill has joined #openstack-nova | 21:45 | |
| alaski | I mean working now | 21:45 |
| lifeless | alaski: those look odd | 21:45 |
| *** lucasagomes has joined #openstack-nova | 21:45 | |
| lifeless | alaski: I'd expect a @classmethod to have a cls parameter | 21:45 |
| alaski | they did | 21:45 |
| alaski | let me get the full context | 21:45 |
| lifeless | alaski: I mean, its odd that this is a fix :) | 21:45 |
| *** efried has quit IRC | 21:46 | |
| alaski | http://git.openstack.org/cgit/openstack/nova/tree/nova/tests/unit/db/fakes.py#n455 | 21:46 |
| alaski | the issue is that the classmethod are being referenced in the class definition, apparently a no-no in python | 21:46 |
| alaski | I'm really not sure why it's working now | 21:47 |
| *** erlon has quit IRC | 21:47 | |
| alaski | I guess it's not really a class definition, they're just within a test method | 21:47 |
| lifeless | so, I'm smelling something wrong here | 21:48 |
| *** thorst_ has quit IRC | 21:48 | |
| lifeless | alaski: if those were @staticmethod | 21:48 |
| lifeless | alaski: that would make sense to me | 21:48 |
| lifeless | alaski: but a classmethod expects to be called with a class parameter | 21:49 |
| alaski | classmethod is wrong anyways since they're patching staticmethods | 21:49 |
| lifeless | alaski: they're patching functions AFAICT | 21:49 |
| lifeless | flavor.py line 600 | 21:50 |
| *** lucasagomes has quit IRC | 21:50 | |
| alaski | line 305 I think | 21:50 |
| lifeless | alaski: no | 21:51 |
| lifeless | nova.objects.flavor._flavor_get_all_from_db | 21:51 |
| lifeless | fakes.py line 453 | 21:51 |
| alaski | that one works fine | 21:51 |
| lifeless | nova.objects.Flavor._flavor_get_by_name_from_db | 21:51 |
| alaski | it's _get and _get_by_name that that errored | 21:52 |
| *** mylu has joined #openstack-nova | 21:52 | |
| *** Sukhdev has quit IRC | 21:52 | |
| lifeless | oh, I see, capital | 21:53 |
| lifeless | and magic | 21:53 |
| alaski | I tried changing to staticmethod and I get TypeError: 'staticmethod' object is not callable | 21:53 |
| lifeless | ok, so we need to dig a little more | 21:53 |
| lifeless | I need to do a call (urgh sorry) and then I have no more interrupts at all, so 100% on this | 21:54 |
| alaski | no worries, though I may have to head out in a bit. it's about the end of my day | 21:54 |
| *** eliqiao1 has joined #openstack-nova | 21:54 | |
| lifeless | what are you using to reproduce? | 21:54 |
| *** edtubill has quit IRC | 21:55 | |
| alaski | tox -epy27 nova.tests.unit.virt.xenapi.test_xenapi.XenAPIVMTestCase | 21:55 |
| alaski | with latest fixtures patch | 21:56 |
| *** lucasagomes has joined #openstack-nova | 21:56 | |
| openstackgerrit | Sujitha proposed openstack/nova: Config options: Centralize network options https://review.openstack.org/303091 | 21:57 |
| *** ducttape_ has joined #openstack-nova | 22:00 | |
| *** xyang1 has quit IRC | 22:00 | |
| *** ayoung has quit IRC | 22:01 | |
| mikal | . | 22:02 |
| *** ihrachys has quit IRC | 22:05 | |
| *** efoley_ has quit IRC | 22:06 | |
| *** dave-mccowan has quit IRC | 22:06 | |
| *** itlinux has joined #openstack-nova | 22:08 | |
| *** mylu has quit IRC | 22:08 | |
| *** pushkaru has quit IRC | 22:08 | |
| itlinux | hello all, I have not done much with Docker, but I am looking for some info about Docker with Nova, advantages of using it over a default spin up instances | 22:08 |
| itlinux | anyone has some links ? | 22:08 |
| *** mdrabe has quit IRC | 22:09 | |
| *** pwnall has quit IRC | 22:11 | |
| *** mrkz has quit IRC | 22:11 | |
| *** mrkz has joined #openstack-nova | 22:11 | |
| *** ducttape_ has quit IRC | 22:11 | |
| *** penick has joined #openstack-nova | 22:13 | |
| *** NightKhaos has quit IRC | 22:13 | |
| *** busterswt has quit IRC | 22:15 | |
| *** ducttape_ has joined #openstack-nova | 22:16 | |
| *** shakamunyi has joined #openstack-nova | 22:18 | |
| *** mikal has quit IRC | 22:18 | |
| *** mikal has joined #openstack-nova | 22:18 | |
| *** dave-mccowan has joined #openstack-nova | 22:18 | |
| *** knikolla has quit IRC | 22:19 | |
| *** barra204 has joined #openstack-nova | 22:20 | |
| *** sneti_ has joined #openstack-nova | 22:21 | |
| *** Sukhdev_ has joined #openstack-nova | 22:21 | |
| *** NightKhaos has joined #openstack-nova | 22:22 | |
| *** aunnam_ has joined #openstack-nova | 22:22 | |
| *** doug-fish has quit IRC | 22:23 | |
| *** shakamunyi has quit IRC | 22:24 | |
| *** barra204 has quit IRC | 22:25 | |
| *** burt has quit IRC | 22:25 | |
| *** ayoung has joined #openstack-nova | 22:26 | |
| *** Sukhdev_ has quit IRC | 22:26 | |
| *** aginwala has quit IRC | 22:28 | |
| *** aginwala has joined #openstack-nova | 22:28 | |
| *** dimtruck is now known as zz_dimtruck | 22:28 | |
| *** rcernin has quit IRC | 22:29 | |
| *** jwcroppe has quit IRC | 22:30 | |
| *** aginwala_ has joined #openstack-nova | 22:31 | |
| *** ducttape_ has quit IRC | 22:31 | |
| *** lpetrut has quit IRC | 22:32 | |
| *** angdraug has quit IRC | 22:32 | |
| *** aginwala has quit IRC | 22:32 | |
| *** rmart04 has quit IRC | 22:32 | |
| *** sneti_ has quit IRC | 22:35 | |
| *** aunnam_ has quit IRC | 22:35 | |
| *** aunnam_ has joined #openstack-nova | 22:37 | |
| *** yassine__ has quit IRC | 22:37 | |
| *** pumarani- has joined #openstack-nova | 22:42 | |
| *** eliqiao1 has quit IRC | 22:42 | |
| *** aginwala_ has quit IRC | 22:43 | |
| alaski | lifeless: think I got it. When patching a staticmethod with a static/classmethod we don't want to rewrap the fake static/classmethod with staticmethod which is what was happening | 22:46 |
| *** sneti_ has joined #openstack-nova | 22:47 | |
| alaski | setattr(obj, name, staticmethod(func)) works but it was essentially ending up with setattr(obj, name, staticmethod(staticmethod(func)) which fails | 22:47 |
| alaski | I pushed up a fix, and am running on Nova now | 22:48 |
| *** daemontool has joined #openstack-nova | 22:48 | |
| *** baoli has quit IRC | 22:52 | |
| *** mylu has joined #openstack-nova | 22:52 | |
| *** aginwala has joined #openstack-nova | 22:53 | |
| *** sdake has quit IRC | 22:54 | |
| *** p01nt3r75 has joined #openstack-nova | 22:54 | |
| alaski | lifeless: yep, that did it | 22:55 |
| alaski | latest patch works on nova py27 for me | 22:55 |
| *** aginwala has quit IRC | 22:56 | |
| alaski | but I'm way past my end of day so I will get to any comments tomorrow, or feel free to clean/fix anything you see | 22:56 |
| *** zz_dimtruck is now known as dimtruck | 22:56 | |
| *** sdague has quit IRC | 22:57 | |
| *** sdake has joined #openstack-nova | 22:58 | |
| *** mrkz has quit IRC | 22:59 | |
| *** aginwala has joined #openstack-nova | 23:00 | |
| *** akshai has quit IRC | 23:09 | |
| lifeless | alaski: \o/ with no Nova changes? | 23:10 |
| *** sdake has quit IRC | 23:12 | |
| *** aginwala has quit IRC | 23:13 | |
| *** aginwala has joined #openstack-nova | 23:13 | |
| *** tbachman has joined #openstack-nova | 23:14 | |
| *** mriedem has joined #openstack-nova | 23:14 | |
| *** vilobhmm111 has joined #openstack-nova | 23:15 | |
| alaski | lifeless: correct. that ran on Nova master and passed | 23:15 |
| *** vilobhmm11 has quit IRC | 23:17 | |
| *** aginwala has quit IRC | 23:17 | |
| mriedem | we're still planning to cap fixtures<2.0 until something is released right? | 23:19 |
| mriedem | or will fixtures 2.0.1 fix the issue? | 23:19 |
| *** flwang has joined #openstack-nova | 23:20 | |
| *** aginwala has joined #openstack-nova | 23:20 | |
| alaski | a new release with this code will fix Nova at least | 23:21 |
| flwang | mriedem: hi guys, i have a question about nova's metadata service | 23:21 |
| *** aysyd has quit IRC | 23:21 | |
| alaski | but I'm out now until tomorrow | 23:21 |
| flwang | when i run "curl http://169.254.169.254/latest/meta-data/hostname" in the instance, so where does nova get the info? | 23:22 |
| *** aunnam_ has quit IRC | 23:22 | |
| *** sneti_ has quit IRC | 23:22 | |
| *** pumarani- has quit IRC | 23:22 | |
| flwang | mriedem: ^ | 23:22 |
| *** akrish9 has joined #openstack-nova | 23:23 | |
| *** akrish9 has quit IRC | 23:23 | |
| *** akrish9 has joined #openstack-nova | 23:23 | |
| *** salv-orl_ has quit IRC | 23:24 | |
| mriedem | flwang: from https://github.com/openstack/nova/tree/master/nova/api/metadata | 23:24 |
| flwang | mriedem: so for example, if i have a instance named 'a' and after changed the name to 'b', will the metadata i see in the instance be changed as well? | 23:24 |
| *** ayoung has quit IRC | 23:26 | |
| mriedem | flwang: looks like it should https://github.com/openstack/nova/blob/master/nova/api/metadata/base.py#L337 | 23:26 |
| mriedem | inside meta_data.json | 23:26 |
| flwang | just tried with devstack, it didn't change | 23:26 |
| flwang | i asked, because we're seeing some problem caused by this | 23:27 |
| mriedem | i think there is a cache in there | 23:27 |
| *** akrish9 has quit IRC | 23:27 | |
| flwang | a renamed instance is renamed back to its original name magically | 23:27 |
| flwang | mriedem: is there any chance the cloun-init re-read the 'cached' name after some actions, like live migration? | 23:28 |
| mriedem | idk | 23:28 |
| mriedem | there is a cache in the metadata request handler https://github.com/openstack/nova/blob/master/nova/api/metadata/handler.py#L75 | 23:29 |
| mriedem | so that InstanceMetadata will be cached | 23:29 |
| *** markvoelker has quit IRC | 23:29 | |
| JayF | cloud-init keeps a static copy of metadata in /var/lib/cloud/instance/[] usually | 23:29 |
| JayF | that could be going into play as well | 23:29 |
| flwang | JayF: ah, thanks for the info, let me check | 23:29 |
| JayF | (at least that's true for configdrive; I haven't used openstack w/metadata service much) | 23:30 |
| flwang | JayF: no, nothing under /var/lib/cloud/ in my cirros instance | 23:30 |
| *** ducttape_ has joined #openstack-nova | 23:32 | |
| gmann | mriedem: oomichi: I see, I added that in my list. Nice catch | 23:35 |
| oomichi | gmann: thanks :-) | 23:35 |
| *** jamielennox|away is now known as jamielennox | 23:36 | |
| gmann | mriedem: oomichi: I think we have same issue with other datetime attribute too like created_at etc | 23:37 |
| *** vtech has quit IRC | 23:37 | |
| gmann | anyways i will check those, agree to make them more strict | 23:37 |
| oomichi | gmann: yeah, maybe there are the same problem on several places | 23:37 |
| *** ducttape_ has quit IRC | 23:37 | |
| oomichi | gmann: awesome, thanks in advance | 23:37 |
| *** mylu has quit IRC | 23:38 | |
| hemnafk | mriedem, https://www.gliffy.com/go/publish/10376541/ | 23:39 |
| oomichi | gmann: updated_at: 5 places | 23:40 |
| gmann | oomichi: yea | 23:40 |
| oomichi | gmann: that is many than I expected | 23:40 |
| mriedem | hemnafk: that's almost too much detail :) nice job | 23:41 |
| hemnafk | yah it's a lot of info | 23:41 |
| *** ZZelle_ has quit IRC | 23:41 | |
| *** Jeffrey4l_ has quit IRC | 23:42 | |
| gmann | oomichi: yea. I think along with that let's check other attribute too if we have loosen up anything | 23:42 |
| *** anush_ has quit IRC | 23:43 | |
| gmann | but that might take time and let fix the datetime one now | 23:43 |
| *** mylu has joined #openstack-nova | 23:45 | |
| *** vilobhmm111 has quit IRC | 23:46 | |
| *** akshai has joined #openstack-nova | 23:48 | |
| *** zenoway has quit IRC | 23:51 | |
| *** mingyu has joined #openstack-nova | 23:51 | |
| *** akshai_ has joined #openstack-nova | 23:54 | |
| *** mingyu has quit IRC | 23:56 | |
| *** Sukhdev_ has joined #openstack-nova | 23:57 | |
| *** akshai has quit IRC | 23:58 | |
| *** markvoelker has joined #openstack-nova | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!