Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-12-15
14:53:42 artom And they'll get the migration on their next major upgrade like everyone else
14:54:03 artom But people who are affected will have one more step in their .z upgrade path
14:54:07 maciejjozefczyk giblet: thanks too ;)
14:54:09 lyarwood artom: right, the NVR really doesn't change that for us, but it's still a minor release bump change as bauzas highlighted.
14:55:04 artom lyarwood, ah, I see
14:55:30 artom In any case, it's not like our deployment tooling/packaging does 'db sync' on yum update, so I think we're good
14:55:56 lyarwood artom: yup indeed
14:59:28 mriedem so after your red hat meeting, what have you agreed the rest of us must do?
14:59:31 mriedem minor bump?
14:59:51 artom Well, we'd like a interpretative-dance-off
14:59:58 artom But we'll settle for minor bump
15:00:03 mriedem i seem to remember the rdo / rhosp tooling used to do a full db sync on a minor update or something
15:00:28 mriedem idk, but last time we backported something like this and i didn't use the proper semver scheme i got in trouble
15:00:54 mriedem from frank in L3 RHOSP QE
15:01:06 bauzas mriedem: it's just the fact we don't package by numbering with semversioning
15:01:18 bauzas I know, it's sad
15:09:37 mriedem andreykurilin_: seems we should deprecate get_resource_manager_extra_kwargs within novaclient
15:09:52 mriedem used to call hooks to get extra boot kwargs to pass to POST /servers
15:10:10 mriedem which is likely broken since 2.1 in the server
15:10:55 andreykurilin_ mrieden: give me a sec to look what get_resource_manager_extra_kwargs is :)
15:13:24 andreykurilin_ mriedem: actually, not sure that it is broken. it looks like not used at all
15:13:31 openstackgerrit Matt Riedemann proposed openstack/python-novaclient master: WIP: Add support for the 2.57 microversion https://review.openstack.org/528128
15:13:54 mriedem andreykurilin_: i mean, if someone did have a hook and they passed extra arguments in the request body to POST /servers,
15:13:55 mriedem it would fail
15:14:03 mriedem because we have strict schema checking on the request body
15:14:45 mriedem https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/schemas/servers.py#L59
15:16:20 andreykurilin_ mriedem: yes. but get_resource_manager_extra_kwargs iterates over hooks which were added via add_resource_manager_extra_kwargs_hook. And I do not see any calls for that method. Since add_resource_manager_extra_kwargs_hook is documented as CLI helper, It should not be used outside the novaclient and I suppose we can just remove both of these methods without deprecation
15:17:09 openstackgerrit Andrey Kurilin proposed openstack/python-novaclient master: [CLI] Fix token auth type https://review.openstack.org/424256
15:18:17 openstackgerrit Andrey Kurilin proposed openstack/python-novaclient master: [CLI] Fix token auth type https://review.openstack.org/424256
15:19:01 andreykurilin_ mriedem: btw, https://review.openstack.org/#/c/424256 is waiting for your review ;)
15:19:22 mriedem ok
15:19:45 mriedem fun
15:19:45 mriedem btw, single bed at the ptg hotel, it'll be like living in the dorms in college again
15:20:04 andreykurilin_ lol
15:20:47 leakypipes mriedem: errr, no.
15:20:56 mriedem you don't miss those days?
15:21:01 mriedem shower caddy, flip flops?
15:21:14 mriedem rascals stealing your towel
15:21:49 leakypipes mriedem: as someone who went to boarding school with showers with no privacy... no. no, I don't.
15:22:07 mriedem we at least had curtains in the great shower room
15:23:16 leakypipes mriedem: no curtains for me. :(
15:23:31 giblet at some point we lost our dorm room door
15:23:32 mriedem leakypipes: oh well that's terrible
15:23:32 leakypipes mriedem: four spigots in the ceiling.
15:23:45 mriedem so gym class showers
15:23:47 leakypipes mriedem: thanks for the memories.
15:23:53 mriedem which went out of style in the late 80s i think
15:23:59 leakypipes mriedem: only not gym class...
15:24:03 mriedem heh
15:24:12 mriedem leakypipes: well, good effort out there today
15:24:20 leakypipes lmao
15:25:18 andreykurilin_ mriedem: returning to hooks. may I propose a patch? or you will do it yourself, so Vek and me will able to merge it? :)
15:25:47 leakypipes andreykurilin_: ty for changing the subject :)
15:25:58 andreykurilin_ ha
15:27:07 andreykurilin_ leakypipes: sorry, but such topic cannot be changed so quickly. Do you have any photos from your boarding school ? heh
15:27:59 cdent leakypipes: was that brit, yankee, or something else boarding school?
15:28:00 leakypipes andreykurilin_: yes :) you can see every room I lived in over four years in the movie: http://www.imdb.com/title/tt0097165/
15:28:12 leakypipes cdent: yank
15:28:39 cdent thank you, that helps to distinguish the style of horrible abuse you experience
15:28:59 leakypipes heh
15:31:41 leakypipes andreykurilin_: btw, I'm not kidding. that really is the school I went to.
15:31:57 leakypipes andreykurilin_: it's not called the same name in the movie, but that's the school.
15:32:23 leakypipes andreykurilin_: in fact, three of my teachers from high school play parts in the movie :)
15:32:33 leakypipes I arrived the year after it was released.
15:33:20 cfriesen when shelving, do we preserve the contents of the ephemeral disk(s)? the code looks like it only snapshots the main image, but ephemeral disks are supposed to last the life of the server.
15:33:46 mriedem andreykurilin_: comments / questions in https://review.openstack.org/#/c/424256/ but looks pretty good
15:34:45 mriedem cfriesen: i doubt it
15:34:49 cdent aw, leakypipes, you are a youth
15:34:54 mriedem we snapshot the image and then delete the guest
15:34:56 leakypipes heh
15:35:23 mriedem andreykurilin_: go ahead on the hooks thing, i didn't dig into that very deep, just noticed it while doing something else
15:42:44 andreykurilin_ leakypipes: wow. cool
15:46:56 leakypipes figleaf: you can thank mriedem for that.
15:47:29 mriedem maciejjozefczyk: giblet: questions inline https://review.openstack.org/#/c/520024/
15:47:36 mriedem and you probably want leakypipes to look at ^ as well
15:48:16 vivsoni__ mriedem: Hi
15:48:49 mriedem hi
15:49:08 maciejjozefczyk mriedem: ok
15:50:14 vivsoni__ mriedem: the patch proposed by you to fix for tempest issue(connector empty to terminate_connection)
15:51:01 yyangs Hi
15:51:02 mriedem vivsoni__: yeah?
15:51:18 vivsoni__ mridem: is restored right ?
15:51:21 mriedem yes
15:51:24 vivsoni__ ok
15:51:37 vivsoni__ so will wait for smcginnis for review then
15:51:46 mriedem there were some concerns about it potentially breaking force volume detach, but i think i've handled that in the latest revision
15:52:01 vivsoni__ yes i have seen that
15:52:25 mriedem i am, however, no expert on the delicate intricacies of the cinder volume manager
15:52:38 mriedem plus, i just wanted to say "delicate intricacies"
15:53:30 vivsoni__ in case if Sean disagree, then i need to handle that in our code
15:54:00 vivsoni__ by just returning 'None' in case of empty connector ... right ?
15:54:58 yyangs we have some noisy neighbours and we are trying to use cpu_quota and cpu_period on flavor to limit the cpu cycles by those vms. It is fine in single vcpu servers. But for multi vcpus server, we are a little confused about the cpu_quota.
15:56:02 yyangs In our tests, we set cpu_period to 200ms, for 2 vcpus server, if we want to allow 50% cpu, should we set to 200ms? or 100ms?
15:57:02 leakypipes yyangs: are you referring to Nova or Kubernetes?
15:57:14 yyangs nova
15:57:42 yyangs leakypipes: we are using kvm
15:58:24 yyangs leakypipes: I though nova also relies on cgroup, right?
15:58:31 leakypipes yyangs: no.
15:59:02 leakypipes yyangs: well, maybe libvirt is configuring cgroups for the VM behind the scene, yeah
15:59:07 smcginnis vivsoni__: Sorry, distracted with some other things. But you still need to handle it in your code, even after mriedem's delicate bits.

Earlier   Later