Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2017-01-19

| Channels | #puppet-openstack index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:05 openstackgerrit Oliver Walsh proposed openstack/puppet-nova: WIP: Fix cell_v2 setup  https://review.openstack.org/422248
00:11 acormier joined #puppet-openstack
00:52 openstackgerrit Alex Schultz proposed openstack/puppet-nova: Implement a proper cell_v2 setup  https://review.openstack.org/422248
00:52 mwhahaha owalsh-afk: -^
00:58 owalsh-afk mwhahaha: cool..
01:04 owalsh-afk mwhahaha: concern I have with the cell provider is that we really should require a transport_url and database_connection and grep for both to check if a cell exists
01:06 mwhahaha owalsh-afk: we can work on improving it but that IMO that's the break out of peices for this process so they can reuse it
01:07 mwhahaha owalsh-afk: if we change the nova::cell_v2::cell to a provider we can do what your proposing
01:07 mwhahaha Very simply
01:07 owalsh-afk mwhahaha: also ensure: absent won't work I think
01:08 owalsh-afk no delete (what that implies is being dicussed I believe) and it would require the cell uuid I expect
01:08 mwhahaha So I just wanted to get the structure right so we can swap out the prices as necessary
01:09 owalsh-afk cool, I'd comment to keep track of what's needed but gerrit says 503... it's a sign I think. I'll take a closer look tomorrow
01:09 mwhahaha With a provider we can get the uuid no problem. To me the cells resource is wip until I can get an env to work with
01:14 ducttape_ joined #puppet-openstack
01:25 owalsh-afk cool, gerrit is back... I've just brain-dumped before I forget the details. Thanks mwhahaha... I really gotta go now
01:29 EmilienM back
01:29 EmilienM and alive
01:29 EmilienM https://review.openstack.org/#/c/422248/ omg
01:49 sofie joined #puppet-openstack
01:59 markvoelker joined #puppet-openstack
02:09 ducttape_ joined #puppet-openstack
02:33 fc___ joined #puppet-openstack
02:35 ducttape_ joined #puppet-openstack
02:36 partner_ joined #puppet-openstack
02:41 jayg|g0n` joined #puppet-openstack
02:41 mtreinish_ joined #puppet-openstack
02:41 tlbr_ joined #puppet-openstack
02:42 mjb joined #puppet-openstack
02:48 ilbot3 joined #puppet-openstack
02:48 Topic for #puppet-openstack is now CI status: green - Puppet OpenStack modules http://docs.openstack.org/deve​loper/puppet-openstack-guide/
02:50 alee joined #puppet-openstack
03:10 markvoelker joined #puppet-openstack
03:10 saneax-_-|AFK joined #puppet-openstack
03:32 ducttape_ joined #puppet-openstack
04:33 agrebennikov_ joined #puppet-openstack
04:42 nkinder joined #puppet-openstack
04:44 furlongm_ joined #puppet-openstack
05:31 udesale joined #puppet-openstack
05:58 skramaja joined #puppet-openstack
06:03 dgurtner joined #puppet-openstack
06:14 openstackgerrit John Fulton proposed openstack/puppet-ceph: Adding support to force zapping of ceph OSDs and journals  https://review.openstack.org/420992
06:16 agurenko joined #puppet-openstack
06:20 acormier joined #puppet-openstack
06:22 xarses joined #puppet-openstack
06:30 dgurtner joined #puppet-openstack
06:30 dgurtner joined #puppet-openstack
06:45 acormier joined #puppet-openstack
07:00 acormier_ joined #puppet-openstack
07:02 acormier_ joined #puppet-openstack
07:07 dgurtner joined #puppet-openstack
07:07 dgurtner joined #puppet-openstack
07:20 hjensas joined #puppet-openstack
07:20 hjensas joined #puppet-openstack
07:58 rasca joined #puppet-openstack
08:01 yprokule joined #puppet-openstack
08:13 tesseract joined #puppet-openstack
08:22 owalsh-afk joined #puppet-openstack
08:23 ccamacho joined #puppet-openstack
08:42 gfidente joined #puppet-openstack
08:42 gfidente joined #puppet-openstack
08:42 owalsh joined #puppet-openstack
08:46 jpich joined #puppet-openstack
08:56 openstackgerrit joined #puppet-openstack
08:56 openstackgerrit Arnaud Morin proposed openstack/puppet-keystone: Make email parameter optional in roles::admin  https://review.openstack.org/422474
09:08 rmart04 joined #puppet-openstack
09:12 owalsh joined #puppet-openstack
09:12 owalsh joined #puppet-openstack
09:40 ccamacho joined #puppet-openstack
09:45 openstackgerrit Mykyta Karpin proposed openstack/puppet-keystone: Add default_transport_url option  https://review.openstack.org/422509
09:46 salmankhan joined #puppet-openstack
09:48 derekh joined #puppet-openstack
09:49 owalsh_ joined #puppet-openstack
09:57 paramite joined #puppet-openstack
10:00 nmathew joined #puppet-openstack
10:01 social joined #puppet-openstack
10:30 salmankhan joined #puppet-openstack
10:43 openstackgerrit Mykyta Karpin proposed openstack/puppet-glance: Add default_transport_url option  https://review.openstack.org/422545
11:11 richm joined #puppet-openstack
11:28 owalsh joined #puppet-openstack
11:42 salmankhan1 joined #puppet-openstack
11:45 dgurtner joined #puppet-openstack
11:45 dgurtner joined #puppet-openstack
11:47 salmankhan joined #puppet-openstack
11:57 salmankhan joined #puppet-openstack
12:02 hjensas joined #puppet-openstack
12:02 dgurtner joined #puppet-openstack
12:02 dgurtner joined #puppet-openstack
12:02 EmilienM hello
12:14 dgurtner joined #puppet-openstack
12:14 dgurtner joined #puppet-openstack
12:25 openstackgerrit Steven Hardy proposed openstack/puppet-openstack-integration: Temporarily remove *_quotas from swift pipeline  https://review.openstack.org/422084
12:30 openstackgerrit Harald Jensås proposed openstack/puppet-keystone: Federation mellon - spec test wrong filename + typos  https://review.openstack.org/421173
12:38 iurygregory good morning
13:02 mwhahaha fyi, ceilometer fix passed CI https://review.openstack.org/#/c/421561/
13:02 morazi joined #puppet-openstack
13:12 openstackgerrit Pradeep Kilambi proposed openstack/puppet-ceilometer: Configure compute params in config  https://review.openstack.org/420908
13:13 ansmith joined #puppet-openstack
13:21 hjensas joined #puppet-openstack
13:21 hjensas joined #puppet-openstack
13:24 morazi joined #puppet-openstack
13:29 nmathew joined #puppet-openstack
13:31 agrebennikov_ joined #puppet-openstack
13:44 acormier joined #puppet-openstack
13:44 pradk joined #puppet-openstack
13:56 EmilienM shardy, mwhahaha: I have some thoughts about https://review.openstack.org/#/c/421341
13:56 EmilienM I think we should introspect the pipeline parameter and detect the _ and transform in -
13:56 EmilienM so it's transparent for our users
13:57 mwhahaha that could work
13:57 EmilienM and we need a release note to communicate on the fact they need to update the pipeline
13:57 shardy Yeah that sounds reasonable
13:57 EmilienM because it's clear that https://review.openstack.org/#/c/421341 is not backward compatible
13:57 EmilienM I mean, yes we could fix our CI to make it pass
13:57 EmilienM but our users will break their deployments for sure
13:58 EmilienM shardy: I can work on a patch but it will be later today, I have urgent tasks to finish before. But I can do it before tomorrow morning for sure.
13:58 shardy EmilienM: Ok, if you have time that would be great, if not I can probably do it
13:58 EmilienM I'll do it today
13:58 shardy Ok, sounds good, thanks!
14:01 markvoelker joined #puppet-openstack
14:01 udesale joined #puppet-openstack
14:05 dprince joined #puppet-openstack
14:14 agrebennikov_ joined #puppet-openstack
14:14 mkoderer joined #puppet-openstack
14:15 openstackgerrit Pavel Gluschak (scsnow) proposed openstack/puppet-cinder: Added support of vstorage cinder backend  https://review.openstack.org/419012
14:23 amoralej EmilienM, mwhahaha, https://review.openstack.org/#/c/421196/ if you have a chance
14:23 EmilienM ok
14:30 salmankhan joined #puppet-openstack
14:37 openstackgerrit Merged openstack/puppet-ceilometer: Fix Ubuntu Ocata issues  https://review.openstack.org/421561
14:41 alee EmilienM, mwhahaha - can I get a +W on https://review.openstack.org/#/c/422200/ please ?
14:47 iurygregory alee, waitting for TripleO recheck since it has failures
14:47 alee iurygregory, ah gotcha thanks
14:48 EmilienM shardy: http://docs.openstack.org/dev​eloper/swift/middleware.html
14:48 EmilienM shardy: why do we have _ in upstream doc?
14:50 shardy https://github.com/openstack/swift/blob/​master/etc/proxy-server.conf-sample#L734
14:50 shardy EmilienM: I don't know, that is also inconsistent with the sample conf
14:50 shardy so I guess that needs to be fixed as well
14:50 EmilienM sigh
14:51 EmilienM shardy: what is the real name? _ or - ?
14:51 EmilienM - I guess
14:51 shardy EmilienM: it doesn't matter, they just have to match
14:51 EmilienM oh ok
14:51 shardy but the pipeline is mandatory for puppet-swift if the hard-coded things don't match the service defaults
14:52 EmilienM but since puppet-swift has container_quotas
14:52 EmilienM it means we need to use _
14:52 EmilienM in that case, why not updating tripleo?
14:53 EmilienM shardy: ^
14:53 nkinder joined #puppet-openstack
14:54 dfisher joined #puppet-openstack
14:55 EmilienM why not doing it in swift-proxy service in THT?
14:55 shardy EmilienM: I can, but I was trying to make it consistent with the swift internal defaults
14:55 shardy also, why should operators be forced to modify their configuration on upgrade?
14:56 shardy it seems like this was a backwards incompatible change to start with
14:56 EmilienM well I think it's the other way around
14:56 EmilienM in puppet-swift, we decided to use "_" like it's documented
14:56 EmilienM so users need to define the pipeline with _
14:57 EmilienM operators are not forced to declare this middleware
14:57 EmilienM but if they do, they need to use the _
14:57 shardy EmilienM: it's not consistent in the swift docs either
14:57 EmilienM because that's the way puppet-swift enable it
14:57 EmilienM shardy: see https://review.openstack.org/#/c/395859/
14:58 EmilienM the problem is Thiago used - instead of _
14:58 fultonj joined #puppet-openstack
14:58 EmilienM 7 weeks ago, we should just use _ and our problem is solved imho
14:58 tdasilva sorry
14:58 EmilienM tdasilva: no problem, it's not very intuitive in puppet-swift
14:59 tdasilva just realized the use of - and _ seems very inconsistent even in https://github.com/openstack/swift/blo​b/master/etc/proxy-server.conf-sample
14:59 iurygregory alee, +A ;) TripleO is fine now
14:59 shardy EmilienM: meh, whatever
14:59 EmilienM well
14:59 EmilienM if you look:
14:59 EmilienM https://github.com/openstack/puppet-swift/blob/​master/manifests/proxy/container_quotas.pp#L25
14:59 alee iurygregory, thanks!
14:59 shardy EmilienM: man proxy-server.conf, and the sample config says container-quotas
14:59 iurygregory alee, yw ;)
14:59 shardy so do a bunch of other places in the swift docs
15:00 shardy grep the swift source to see
15:00 EmilienM I added that in 2013
15:00 EmilienM https://github.com/openstack/puppet-swift/com​mit/4157cb9a4141c071337f7e7c1946f32dd0342945
15:00 EmilienM since then, operators have been using "_" to make it work
15:00 EmilienM shardy: I'm afraid renaming _ to - will beak backward compatibility
15:00 shardy Ok, so a bunch of the swift docs, and the sample config is wrong then
15:00 EmilienM tdasilva: can we have some help here?
15:01 EmilienM maybe you know which one we should use
15:01 shardy EmilienM: Ok, fine I'll fix t-h-t to match puppet-swift, and ignore the swift inconsistencies for now
15:01 EmilienM shardy: I would vote for 1) update THT to use "_" in the meantime" 2) create a release note about it and switch to "-" in Pike
15:01 EmilienM shardy: and 3) fix doc/code in Swift to be consistent
15:01 shardy EmilienM: Ok, thanks for the feedback
15:01 EmilienM shardy can do 1), I can do 2) and tdasilva can do 3)
15:02 shardy ack, sounds good, I'll abandon my puppet patches for now
15:02 EmilienM shardy: actually I'll take over them
15:02 EmilienM shardy: you can abandon the poi patches
15:03 EmilienM shardy: but not the puppet-swift . I'll take over and add deprecated warning & release notes
15:03 shardy EmilienM: ack, sounds good, thanks!
15:03 tdasilva EmilienM: I can work with the swift docs to make it more consistent, but I can't guarantee they will be consistent with they way puppet-swift expects
15:03 EmilienM tdasilva: I think what we want here in consistency in Swift, not in puppet-swift
15:03 EmilienM tdasilva: puppet-switch will adapt to what Swift says
15:03 EmilienM but make it loud and clear before
15:04 EmilienM because right now it's confusing
15:04 tdasilva EmilienM: but then it might break backwards compatibility with previous puppet-swift installs, no?
15:04 EmilienM yes
15:04 EmilienM but I'm working on it
15:04 tdasilva oh ok
15:04 EmilienM don't take care of that
15:04 tdasilva sounds good
15:04 tdasilva :)
15:05 udesale joined #puppet-openstack
15:21 acormier joined #puppet-openstack
15:28 markvoelker_ joined #puppet-openstack
15:31 markvoelker joined #puppet-openstack
15:31 jaosorior joined #puppet-openstack
15:35 rajinir joined #puppet-openstack
15:45 openstackgerrit Ricardo Noriega proposed openstack/puppet-neutron: Add support for BGPVPN plugin  https://review.openstack.org/422051
15:47 dprince joined #puppet-openstack
15:53 openstackgerrit Merged openstack/puppet-nova: Set correct ownership for nova keytab  https://review.openstack.org/422200
15:58 openstackgerrit Merged openstack/puppet-vswitch: Add fail_mode parameter to OVS Ports  https://review.openstack.org/421196
16:00 xarses joined #puppet-openstack
16:00 xarses_ joined #puppet-openstack
16:01 shardy joined #puppet-openstack
16:11 ducttape_ joined #puppet-openstack
16:29 fc__ left #puppet-openstack
16:33 vinsh https://github.com/openstack/puppet-openstack-​integration/blob/master/manifests/swift.pp#L43 also related..
16:35 vinsh https://github.com/openstack/puppet-swi​ft/blob/master/manifests/proxy.pp#L255
16:39 vinsh EmilienM: how is https://review.openstack.org/#/c/421341/2 not backward compatible?
16:39 mwhahaha vinsh: for people who have the _ it fails
16:39 vinsh It really doesn't matter what is set using - or _ in https://review.openstack.o​rg/#/c/421341/2/tests/site.pp
16:39 mwhahaha it does
16:40 vinsh proxy.pp will convert it to _
16:40 mwhahaha so the pipeline (which the operator sets) has to match the config
16:40 mwhahaha not in any of the links you posted :D
16:40 vinsh So maybe in proxy.pp we handle account/container_quota as a special case
16:40 mwhahaha so that was the proposed fix, was convert them as necessary
16:41 mwhahaha s/container_quota/container-quota for the pipeline and update the modules
16:41 mwhahaha so if the operator provides the old style in the pipeline arg, it gets fixed
16:41 vinsh Ok, just keeping myself up to speed :)
16:41 vinsh very good
16:42 vinsh the swift example conf files truly are a mish mash as you guys said.
16:43 vinsh In other puppet-swift news.. next week is de-templitize object/account/container. move them to ini configs.
16:44 vinsh also, add the support for swift crypto settings. is on my list.
16:48 openstackgerrit Brent Eagles proposed openstack/puppet-neutron: Remove logging_facility, allow_pagination, allow_sorting  https://review.openstack.org/422755
16:50 dmsimard EmilienM, mwhahaha: are you familiar with that tempest failure in newton ? ceilometer.tests.tempest.api.test_telemetry_n​otification_api.TelemetryNotificationAPITest.​test_check_glance_v1_notifications[id-04b10bf​e-a5dc-47af-b22f-0460426bf499,image,smoke]
16:50 dmsimard it rings me a bell but I forget why
16:50 mwhahaha We disabled iy
16:51 mwhahaha Cause it's broken on Ubuntu
16:51 openstackgerrit Harald Jensås proposed openstack/puppet-keystone: Federation openidc - spec test wrong filename + typos  https://review.openstack.org/421179
16:51 mwhahaha I think EmilienM might have also disabled it on centos
16:52 mwhahaha Go look in the exclude regex in poi
16:52 * mwhahaha is on mobile at the moment
16:52 dmsimard mwhahaha: it's excluded in master but not in newton
16:53 mwhahaha Something might have gotten backported
16:53 EmilienM I excluded it in master I think because we try to disable ceilometer tests that interact with glance api v1
16:54 EmilienM we removed glance api v1 in poi
16:56 EmilienM dmsimard: ^
17:02 mwhahaha it's disabled due to packaging things for ubuntu
17:02 mwhahaha probably need to revisit those excludes now that m2 is published
17:02 mwhahaha it's not excluded for centos
17:03 EmilienM ah right
17:04 mwhahaha of course need to fix all the ubuntu issues first
17:04 mwhahaha which reminds me to get back to those today
17:05 openstackgerrit Alex Schultz proposed openstack/puppet-openstack-integration: Update ubuntu excludes for ocata-m2  https://review.openstack.org/422766
17:06 ccamacho joined #puppet-openstack
17:33 openstackgerrit Rajini Ram proposed openstack/puppet-cinder: Support for ScaleIO Cinder backend  https://review.openstack.org/422778
17:37 openstackgerrit Pradeep Kilambi proposed openstack/puppet-ceilometer: Configure compute params in config  https://review.openstack.org/420908
17:43 openstackgerrit Alex Schultz proposed openstack/puppet-aodh: Fix Ubuntu Ocata-m2 issues  https://review.openstack.org/421564
18:00 _ducttape_ joined #puppet-openstack
18:20 derekh joined #puppet-openstack
18:22 openstackgerrit Alex Schultz proposed openstack/puppet-openstack-integration: Update ubuntu excludes for ocata-m2  https://review.openstack.org/422766
18:28 ducttape_ joined #puppet-openstack
18:33 openstackgerrit Pradeep Kilambi proposed openstack/puppet-ceilometer: Configure compute params in config  https://review.openstack.org/420908
18:39 mwhahaha EmilienM: what's the entry point into the beaker tests in CI? is there a shell script that gets run or what
18:39 mwhahaha i'm looking at adding the mirror hosts
18:40 mwhahaha blah i see it's in project-config
18:43 EmilienM mwhahaha: back from lunch. Have you found?
18:43 mwhahaha yea i found it
18:44 mwhahaha EmilienM: so do you think we should move the beaker stuff to use a script from p-o-i so we can do the mirror handling?
18:45 mwhahaha EmilienM: https://github.com/openstack-infra/p​roject-config/blob/master/jenkins/jo​bs/puppet-module-jobs.yaml#L75-L88
18:45 mwhahaha since we do already checkout p-o-i, should we just move those bits into like the run_beaker_acceptance.sh?
18:45 openstackgerrit Ricardo Noriega proposed openstack/puppet-neutron: Add support for BGPVPN plugin  https://review.openstack.org/422051
18:46 EmilienM mwhahaha: I would support it
19:01 derekh joined #puppet-openstack
19:04 openstackgerrit Alex Schultz proposed openstack/puppet-openstack-integration: Create beaker test script  https://review.openstack.org/422818
19:12 furlongm_ joined #puppet-openstack
19:15 EmilienM mwhahaha: we'll need to backport it to stable branches
19:15 mwhahaha well i'm seeing if it works first :D
19:15 mwhahaha but yea
19:16 EmilienM mwhahaha: we need to take care of other puppet moduels not in Puppet OpenStack who have beaker jobs
19:16 EmilienM they don't use poi
19:17 mwhahaha but they check it out
19:17 * mwhahaha is aware
19:17 * mwhahaha is only testing at the moment
19:18 EmilienM cool cool
19:18 cdelatte joined #puppet-openstack
19:29 furlongm_ joined #puppet-openstack
19:41 openstackgerrit joined #puppet-openstack
19:41 openstackgerrit Harald Jensås proposed openstack/puppet-keystone: Federation shibboleth - spec test wrong filename + typos  https://review.openstack.org/421218
19:59 hjensas joined #puppet-openstack
19:59 hjensas joined #puppet-openstack
20:07 salmankhan joined #puppet-openstack
20:17 dprince joined #puppet-openstack
20:48 openstackgerrit Ricardo Noriega proposed openstack/puppet-neutron: Add support for BGPVPN plugin  https://review.openstack.org/422051
20:59 _ducttape_ joined #puppet-openstack
21:05 openstackgerrit David Critch proposed openstack/puppet-ceph: Adding support to force zapping of ceph OSDs and journals  https://review.openstack.org/420992
21:09 ccamacho joined #puppet-openstack
21:15 ducttape_ joined #puppet-openstack
21:28 dprince joined #puppet-openstack
21:30 acormier joined #puppet-openstack
21:37 mwhahaha so EmilienM do we want to https://review.openstack.org/#/c/422285/?  Since we haven't released it should we nuke it before it goes out with 9.5.0?
21:39 EmilienM mwhahaha: I guess yes
21:39 EmilienM mwhahaha: I can approve it
21:39 mwhahaha k
21:40 mwhahaha i think it might be best given how cell v2 has gone
21:41 mwhahaha iurygregory, EmilienM: https://review.openstack.org/#/c/421564/ has also passed CI
21:42 EmilienM approved
21:43 EmilienM sounds like poi still not happy
21:43 EmilienM mwhahaha: https://review.openstack.o​rg/#/c/422766/2/run_tests.sh
21:43 EmilienM EXCCLUDES
21:43 mwhahaha yea i thought i fixed that
21:44 mwhahaha i guess i forgot to push
21:44 EmilienM also http://logs.openstack.org/66/422766/2/chec​k/gate-puppet-openstack-integration-4-scen​ario001-tempest-ubuntu-xenial-nv/63d514a/c​onsole.html#_2017-01-19_18_46_19_698264
21:44 EmilienM we need to pip install gabbi on ubuntu
21:44 EmilienM like I did in the past for centos
21:44 openstackgerrit Alex Schultz proposed openstack/puppet-openstack-integration: Update ubuntu excludes for ocata-m2  https://review.openstack.org/422766
21:44 EmilienM I'm sure they didn't package it
21:44 EmilienM let me check
21:44 EmilienM omg they did
21:44 EmilienM http://packages.ubuntu.com/xenial/python-gabbi
21:45 EmilienM zigo did :D
21:45 mwhahaha yea they pulled alot of the upstream stuff in as part of ocata-m2
21:45 EmilienM mwhahaha: we need this package, maybe we can deploy it with puppet-tempest? shrug
21:45 mwhahaha which is partially why we have these problems
21:46 EmilienM scenario004 fails to spawn a VM
21:46 EmilienM http://logs.openstack.org/66/422766/2/check/g​ate-puppet-openstack-integration-4-scenario00​4-tempest-ubuntu-xenial-nv/cab67d1/logs/nova/​nova-compute.txt.gz#_2017-01-19_18_43_47_570
21:47 EmilienM http://logs.openstack.org/66/422766/2/check/g​ate-puppet-openstack-integration-4-scenario00​4-tempest-ubuntu-xenial-nv/cab67d1/logs/glanc​e/glance-api.txt.gz#_2017-01-19_18_49_15_048
21:47 EmilienM sounds like in glance
21:47 EmilienM probably with ceph rgw backend
21:48 EmilienM scenario003, designate is not idempotent http://logs.openstack.org/66/422766/2/che​ck/gate-puppet-openstack-integration-4-sc​enario003-tempest-ubuntu-xenial-nv/2a3e21​0/logs/puppet.txt.gz#_2017-01-19_18_41_24
21:48 mwhahaha probably missing something, we had that in centos for a while
21:49 EmilienM http://logs.openstack.org/66/422766/2/check/gate-​puppet-openstack-integration-4-scenario003-tempes​t-ubuntu-xenial-nv/2a3e210/logs/designate/designa​te-pool-manager.txt.gz#_2017-01-19_18_36_25_578
21:50 EmilienM scenario002 fails to spawn a VM, glance fails to find image http://logs.openstack.org/66/422766/2/check/g​ate-puppet-openstack-integration-4-scenario00​2-tempest-ubuntu-xenial-nv/157d5a9/logs/glanc​e/glance-api.txt.gz#_2017-01-19_18_50_41_986
21:50 EmilienM they do a terrible work in testing
21:50 mwhahaha well to be fair, we run off their testing repo
21:50 mwhahaha but yes
21:50 EmilienM http://logs.openstack.org/66/422766/2/check/g​ate-puppet-openstack-integration-4-scenario00​2-tempest-ubuntu-xenial-nv/157d5a9/logs/nova/​nova-compute.txt.gz#_2017-01-19_18_50_42_299
21:51 EmilienM scenario002 fails on the vif also
21:51 EmilienM et voilà
21:51 EmilienM so now? do we spend 24h to fix thing?
21:52 mwhahaha maybe tomorrow :D
21:53 EmilienM mfisch, _ody, anyone core volunteer to help?
22:07 openstackgerrit David Critch proposed openstack/puppet-ceph: Adding support to force zapping of ceph OSDs and journals  https://review.openstack.org/420992
22:21 ccamacho joined #puppet-openstack
22:23 openstackgerrit Merged openstack/puppet-nova: Revert "Add cell_v2 simple_cell_setup"  https://review.openstack.org/422285
22:25 spredzy joined #puppet-openstack
22:26 openstackgerrit David Critch proposed openstack/puppet-ceph: Adding support to force zapping of ceph OSDs and journals  https://review.openstack.org/420992
22:45 openstackgerrit Merged openstack/puppet-aodh: Fix Ubuntu Ocata-m2 issues  https://review.openstack.org/421564
22:58 owalsh joined #puppet-openstack
23:00 openstackgerrit David Critch proposed openstack/puppet-ceph: Adding support to force zapping of ceph OSDs and journals  https://review.openstack.org/420992
23:25 owalsh mwhahaha: ping, do you have a few minutes to spare?
23:25 mwhahaha owalsh: yea, what's up
23:25 mwhahaha owalsh: let me pull up that review, it's been one of those days :D
23:26 owalsh :-)
23:27 mwhahaha owalsh: so i'm ok with not having delete support initially
23:29 owalsh mwhahaha: I've been thinking this through a bit more.. hear me out...
23:29 mwhahaha sure
23:30 owalsh we use nova::cell_v2::cell must required transport_url & database_connection. This type manages the optional cells.
23:30 owalsh the default cell is not optional
23:31 mwhahaha right so that's https://review.openstack.org/#/c/422​248/3/manifests/cell_v2/simple_setup.pp@50
23:31 owalsh yes
23:31 owalsh also we add an optional name param
23:32 mwhahaha well due to the nature of puppet resources we'll always have a name
23:32 owalsh it will have a title, but transport_url & database_connection tuple is the unique identifier
23:33 mwhahaha only if managed outside of puppet, correct
23:33 owalsh no. a cell is another cloud basically, the db & transport url is how the api talks to that cloud
23:33 owalsh name is just for humans
23:34 mwhahaha yea i get that
23:35 owalsh I think that's a valid model for the additonal cells for a multi-cell setup
23:36 owalsh now, for ocata a single cell setup is required... to do this we run:
23:37 owalsh api_db sync (as we need an api db to add the cells to)
23:37 owalsh map_cell0
23:37 owalsh create_cell (no args, consider this "map_default_cell")
23:38 owalsh db sync (which now syncs the default cell)
23:39 mwhahaha ok yea we've got that
23:39 mwhahaha so what's the issue you're running into?
23:40 owalsh point being we don't use the new cell type in the simple_setup class here https://review.openstack.org/#/c/422​248/3/manifests/cell_v2/simple_setup.pp
23:41 owalsh that's *just* for non-default cells
23:41 mwhahaha so we have exactly what you described if you don't pass any transport or database connection into nova::cell_v2::simple_setup
23:41 mwhahaha the only different is your default cell gets named cell1
23:42 owalsh we *must* pass transport or database connection - that's the only way to differentiate two cells
23:42 mwhahaha i get what you're saying, and it's exactly what we have in that code
23:42 mwhahaha the point is to reuse the same bit of logic for both the default and future cells
23:42 owalsh we are using name
23:42 mwhahaha and?
23:43 mwhahaha so you're telling me that nova-manage cell_v2 create_cell --name default != nova-manage cell_v2 create_cell
23:44 owalsh I saying nova-mange cell_v2 create_cell should be called nova-manage cell_v2 map_default_cell
23:44 owalsh ideally
23:44 mwhahaha what the hell is map_default_cell
23:44 owalsh that's what we are doing, mapping the default cell to the nova db
23:44 mwhahaha and why is there yet another function to do the same thing
23:45 mwhahaha so i must be missing something because i just updated my env and all i've got is: {create_cell,discover_hosts,list_cells,​map_cell0,map_cell_and_hosts,map_instan​ces,simple_cell_setup,verify_instance}
23:45 mwhahaha what I would like is the initial flow to leverage the same stuff that scaling actions would take
23:45 owalsh yea, it will work but it's not correct. add another cell type with a different name but the same transport_url/db
23:46 mwhahaha but it wouldn't
23:46 mwhahaha so it'd be on the operator to define those when adding new cells
23:48 owalsh yes, so they use the cell type, and tranport/db are mandatory params
23:48 openstackgerrit Pradeep Kilambi proposed openstack/puppet-gnocchi: Remove carbonara package reference  https://review.openstack.org/422918
23:48 mwhahaha owalsh: http://paste.openstack.org/show/595709/
23:49 mwhahaha so we don't actually have to make them mandatory in code, we could enforce that via some a class or something. but we could just document that -^
23:49 mwhahaha so what you're getting stuck on is a hard requirement that isn't really necessary to the entire process
23:49 owalsh why not make them manditory?
23:49 owalsh they are manditory
23:49 mwhahaha so we can leverage them in multiple ways
23:49 owalsh what ways, that pastebin is the only way
23:49 mwhahaha or use name 'default' with some valiadtion
23:50 mwhahaha so that pastebin is how it works now with that code i generated
23:50 mwhahaha and it does what you're saying
23:50 mwhahaha the only thing you're hung up on is making transport/database required for nova::cell_v2::cell
23:50 EmilienM https://review.openstack.org/422918 is a promotion blocker
23:50 mwhahaha and i'm saying that's a nice to have but not necessary to this entire thing
23:50 owalsh how can you check if the cell exists otherwise?
23:51 mwhahaha owalsh: so i'm not interested in handling cases where cells are created outside of puppet
23:51 mwhahaha this is puppet-nova
23:51 mwhahaha the expectation is that you do it via puppet
23:51 mwhahaha now over in #tripleo we could have that discussion but if we follow the same flow as what in puppet-nova it's moot
23:52 mwhahaha we can clearly document in releasenotes/testing that this is the expectations and limitations
23:52 mwhahaha because simple setup is optional, the operator could handle it all externally and we wouldn't care either
23:52 owalsh why, we can model this correctly?
23:53 mwhahaha you're over engineering this
23:53 mwhahaha it is correct, except you know the implementation details of how nova is handling it
23:54 mwhahaha from an operator standpoint the transport_url/databse_connection thing would get caught in a puppet failure
23:54 mwhahaha because the cell create would fail
23:54 owalsh but we know that you can't add a cell without this
23:54 mwhahaha add a cell without what?
23:54 owalsh transport_url/databse_connection
23:55 mwhahaha but you can, it just reads nova.conf
23:55 mwhahaha look i'm trying not to duplicate the code
23:56 mwhahaha now here, let me propose what you're talking about
23:56 mwhahaha gimme 5 mins
23:56 owalsh when it reads from nova.conf it's creating the default cell, not adding a cell
23:56 owalsh it's unfortunate that it's the same command IMO
23:58 mwhahaha but it's not, i think you're getting hung up on these details
23:59 pradk joined #puppet-openstack

| Channels | #puppet-openstack index | Today | | Search | Google Search | Plain-Text | summary