Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2015-09-10

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

All times shown according to UTC.

Time Nick Message
00:28 yogesh-pc joined #puppet-openstack
00:30 yogesh-pc_ joined #puppet-openstack
00:38 gildub joined #puppet-openstack
00:51 openstackgerrit Ivan Chavero proposed openstack/puppet-neutron: Split SR-IOV configuration file into driver and agent pieces  https://review.openstack.org/221991
00:55 ducttape_ joined #puppet-openstack
00:58 markvoelker joined #puppet-openstack
00:58 linkedinyou joined #puppet-openstack
01:02 ducttape_ joined #puppet-openstack
01:04 tiswanso joined #puppet-openstack
01:05 xarses joined #puppet-openstack
01:07 tiswanso joined #puppet-openstack
01:21 ducttape_ joined #puppet-openstack
01:24 xingchao joined #puppet-openstack
01:30 gildub michchap, would you mind reviewing the following ones? https://review.openstack.org/221119 https://review.openstack.org/221120 and https://review.openstack.org/221120
01:30 michchap gildub: sure.
01:31 michchap gildub: last two are the same link
01:31 gildub michchap, sorry https://review.openstack.org/221121
01:31 gildub michchap, hi - thanks!
01:32 gildub michchap, those are to address flaws with managing domains in keystone V3
01:33 gildub michchap, and they might land in stable/kilo too, ultimately
01:33 EmilienM gildub: why?
01:33 EmilienM gildub: which bug are we fixing?
01:33 gildub EmilienM, the domains
01:33 EmilienM regarding https://review.openstack.org/221567
01:33 EmilienM I don't see what is solving https://review.openstack.org/221567 in stable/kilo
01:34 EmilienM michchap: o/
01:34 gildub EmilienM, I didn't mention 221567
01:34 EmilienM but I do :)
01:35 EmilienM michchap: I'm not conservative in backports, I just want to make sure we're backward compatible
01:35 EmilienM just saying
01:35 gildub EmilienM, ah sorry I thought you're asking about the other ones
01:35 EmilienM everything that fix bugs needs to be backported
01:35 gildub EmilienM, if  221567 is not backported then we won't be able to backport anything else
01:35 EmilienM only if it's backward compat
01:36 richm we will be able to backport without 221567, but it will be much harder
01:36 richm == error/bug prone
01:36 gildub EmilienM, yeah ^
01:37 richm so there is risk either way
01:37 EmilienM richm: if it's harder, let's backport what we need
01:37 EmilienM to make it easier (ie error prone)
01:37 gildub EmilienM, understood the theory, but I think we're in a grey area here, the domain was still experimental when we moved from kilo to liberty, now unless we don't want to have better domain stuff in kilo then...
01:38 gildub EmilienM, well and it's still experimental.
01:38 EmilienM we added domains support in stable/kilo. If there are some critical bugs, let's fix them
01:38 EmilienM like "not able to change the default domain id" is to me critical
01:38 linkedinyou joined #puppet-openstack
01:38 EmilienM so I vote for backporting the domain bugfix to stable/kilo
01:39 EmilienM but not the new features
01:39 gildub EmilienM, ok, btw the initial idea of 221567 was to help with v2 issues too
01:40 EmilienM gildub: good, I'll +2 it
01:40 EmilienM gildub: once chem will update his -1
01:40 richm gildub: The issue that chem found - do we need to fix that in master too?
01:41 richm or was that some artifact of the cherry-pick?
01:41 EmilienM a backport is a cherry pic
01:41 EmilienM so do not modify it
01:42 EmilienM except if code conflict
01:45 EmilienM if something's wrong in master, let's fix master
01:45 gildub richm, we can definitely add two more cases for testing #ssl? but that can be done later, not blocker
01:47 ilbot3 joined #puppet-openstack
01:47 Topic for #puppet-openstack is now Puppet/OpenStack: Weekly meeting (every Tuesday 15:00 UTC): https://wiki.openstack.org/wiki/Meetings/PuppetOpenStack - Logs http://eavesdrop.openstack.org/irclogs/%23puppet-openstack/ - On going efforts: https://trello.com/puppetopenstack
01:49 gildub richm, yes I missed that one but same thing, not a blocker, in master later
01:50 richm It is ok because the only places where get_section is actually used, it always uses 'DEFAULT' as the group/section name.
01:51 richm but if someone tries to use get_section with a different group/section name, they will have to fix get_section first
01:52 gildub richm, oh! When I said I missed that one, I meant line 40
01:52 gildub richm, no I'm looking at line 89...
01:54 gildub richm, yes that's a bug but we're not hitting it because all the calls to get_section are made using 'DEFAULT' for group.
01:55 richm right - so I'm ok with it
01:55 mesamoo joined #puppet-openstack
01:56 gildub richm, ok. It needs a fix for future use, I'll make a fix in master
01:57 richm so you just need to get chem to +1
01:59 gildub richm, yes thanks
02:12 gildub richm, Do you think we've to keep both --domain and '<name>::<domain>' approaches since they are redundant?
02:13 gildub richm, would-it make sense to keep --domain only?
02:14 gildub richm, I'm thinking about future providers too, well besides the trust since we don't have a choice with this one
02:15 richm gildub: having a domain parameter to keystone resources will only make it confusing
02:16 richm gildub: we will not allow specifying a name without a domain in the future
02:16 richm gildub: so the domain parameter will be useless
02:16 gildub richm, I know that, what I'm talking about is between using --domain or ::
02:16 richm using --domain how?
02:17 richm we have to use --domain in openstack command lines?
02:17 richm OSC doesn't understand "user::domain"
02:17 gildub richm, I mean as a resource parameter (forget the --)
02:17 richm gildub: Right, that's what I mean - we should get rid of the domain resource parameter
02:18 richm as soon as possible
02:18 gildub richm, yes, that's what I thought
02:19 gildub richm, but when possible (unlike trust) if a domain parameter can be used why not use it, it makes things more natural
02:19 gildub richm, for instance  keystone_tenant { 'service': domain => 'domain', ... }
02:20 gildub richm, the domain being mandatory
02:21 richm gildub: What will Keystone_tenant.self.instances name that resource?
02:22 gildub richm, ah yeah, forgot about that one.
02:23 gildub richm, ok, never mind then
02:31 openstackgerrit Adam Vinsh proposed openstack/puppet-swift: WIP Manage swift with swiftinit service provider  https://review.openstack.org/203220
02:33 xingchao joined #puppet-openstack
02:35 yogesh-pc_ joined #puppet-openstack
02:40 openstackgerrit Gilles Dubreuil proposed openstack/puppet-keystone: Providers tenant/user/user_role: domain check/name  https://review.openstack.org/218059
02:54 openstackgerrit Gilles Dubreuil proposed openstack/puppet-keystone: Fixes get_section for future use  https://review.openstack.org/222017
02:55 markvoelker joined #puppet-openstack
02:57 llua^_^AznSmile- joined #puppet-openstack
02:58 openstackgerrit Adam Vinsh proposed openstack/puppet-swift: WIP Manage swift with swiftinit service provider  https://review.openstack.org/203220
03:09 Vinsh gates broken again on centos
03:09 Vinsh 2015-09-10 03:08:37.965 | Error: Execution of '/usr/bin/yum -d 0 -e 0 -y list xinetd' returned 1: One of the configured repositories failed (RDO trunk),
03:09 EmilienM Vinsh: ok
03:09 EmilienM I'm working on it
03:09 Vinsh You rock :)
03:10 EmilienM Vinsh: logs?
03:10 Vinsh https://jenkins06.openstack.org/job/gate-puppet-swift-puppet-beaker-rspec-dsvm-centos7/46/console
03:10 EmilienM ok
03:11 Vinsh maybe it works through it this time...
03:11 EmilienM Vinsh: it's random issue
03:11 Vinsh fail before it was: http://logs.openstack.org/20/203220/32/check/gate-puppet-swift-puppet-beaker-rspec-upgrade-dsvm-centos7/ccf1963/console.html
03:11 EmilienM sometimes rdo repo are down
03:11 EmilienM I was about to change the repo
03:11 EmilienM we are going to to stop using delorean
03:11 EmilienM I got an email from RDO guys, told me we should switch
03:12 Vinsh what is it you add to a review to re-trigger? is it "recheck nobug" ?
03:12 Vinsh ah good to do then
03:12 EmilienM recheck is enough
03:12 EmilienM Vinsh: see https://etherpad.openstack.org/p/puppet-liberty-blocker
03:13 EmilienM I have to update repos
03:13 EmilienM I do it now
03:13 Vinsh I had added a "recheck" to this https://review.openstack.org/#/c/203220/  but didn't notice the old results drop off and trigger a new check
03:14 Vinsh maybe it does so without dropping existing results first
03:14 EmilienM Vinsh: zuul has not finished
03:14 Vinsh ah that one tripleo job that takes forever. :)
03:14 EmilienM no
03:14 EmilienM it's experimental
03:14 EmilienM so they don't count like others
03:15 Vinsh Ok, well I'll try the "recheck" comment again some time and see if it triggers.
03:15 EmilienM yeah
03:16 gildub EmilienM, were are going to get the RPMs if not from delorean?
03:16 EmilienM SECRET
03:16 EmilienM sorry I'm tired
03:16 EmilienM http://cbs.centos.org/repos/cloud7-openstack-liberty-testing/x86_64/os/
03:16 EmilienM http://cbs.centos.org/repos/cloud7-openstack-common-testing/x86_64/os/
03:17 gildub EmilienM, no worries mate!
03:17 EmilienM the "new repos" lol
03:17 Vinsh Yeah, sleep time.  TTYL.
03:17 EmilienM Vinsh: o/
03:17 gildub EmilienM, cool
03:26 EmilienM I might have missed something
03:26 EmilienM http://trunk.rdoproject.org/centos7/current/
03:26 EmilienM it's empty
03:27 gildub Interesting...
03:30 EmilienM oh I got it
03:30 EmilienM so we still need delorean
03:30 EmilienM but we can drop epel & rdo kilo
03:31 gildub EmilienM, looks like something went wrong during the build in  http://trunk.rdoproject.org/centos7/current/
03:33 gildub EmilienM, not following you, what about all the epel dependent packages then. And what's the point of cbs.centos.org...liberty-testing if we need delorean which contains the lastest builds anyway?
03:33 EmilienM I don't know
03:34 EmilienM I get an emal from alan
03:34 EmilienM apevec
03:34 gildub Ah
03:35 gildub EmilienM, Would it be possible to forward it to the list?
03:35 EmilienM gildub: yeah
03:35 gildub EmilienM, great, thanks.
03:36 EmilienM gildub: done
03:40 gildub EmilienM, I can see it, merci!
03:41 EmilienM :)
03:42 gildub This one should actually be in RDO list too
03:43 openstackgerrit Emilien Macchi proposed openstack/puppet-nova: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222025
03:43 EmilienM let' see if it works
03:48 EmilienM http://lists.openstack.org/pipermail/openstack-dev/2015-September/074040.html
03:48 EmilienM it's the most hilarious mail I've never read.
03:53 openstackgerrit Emilien Macchi proposed openstack/puppet-nova: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222025
04:03 EmilienM gildub: look http://trunk.rdoproject.org/centos7/current/rpmbuild.log
04:04 EmilienM it seems like python-oslo-policy can't be installed
04:05 openstackgerrit Emilien Macchi proposed openstack/puppet-keystone: DNM - CI test  https://review.openstack.org/222028
04:05 EmilienM I'm afraid our CI is broken
04:11 EmilienM good night
04:16 gildub EmilienM, yeah I saw that :( - Bonne nuit
04:32 mflobo joined #puppet-openstack
05:31 stamak joined #puppet-openstack
05:51 zhangjn joined #puppet-openstack
05:52 zhangjn joined #puppet-openstack
05:53 jerrygb joined #puppet-openstack
05:56 markvoelker joined #puppet-openstack
06:29 cgfbee joined #puppet-openstack
06:40 arnaud_orange joined #puppet-openstack
06:50 jtomasek joined #puppet-openstack
07:01 Oszkar joined #puppet-openstack
07:11 openstackgerrit Gilles Dubreuil proposed openstack/puppet-openstack-specs: Adds deprecation for using a name without a domain  https://review.openstack.org/219127
07:22 chandankumar joined #puppet-openstack
07:32 jpena joined #puppet-openstack
07:35 stamak joined #puppet-openstack
07:48 arnaud_orange joined #puppet-openstack
07:57 markvoelker joined #puppet-openstack
08:00 bapalm joined #puppet-openstack
08:01 jistr joined #puppet-openstack
08:19 jerrygb joined #puppet-openstack
08:22 Oszkar joined #puppet-openstack
08:27 IBerezovskiy joined #puppet-openstack
08:40 xingchao joined #puppet-openstack
08:45 karume joined #puppet-openstack
08:47 openstackgerrit Lukas Bezdicka proposed openstack/puppet-trove: Remove trove ubuntu package hack  https://review.openstack.org/221273
08:51 derekh joined #puppet-openstack
08:54 karume joined #puppet-openstack
08:56 openstackgerrit Gael Chamoulaud proposed openstack/puppet-designate: Fix rspec 3.x syntax  https://review.openstack.org/222077
08:58 markvoelker joined #puppet-openstack
09:03 openstackgerrit Gael Chamoulaud proposed openstack/puppet-glance: Fix rspec 3.x syntax  https://review.openstack.org/222081
09:08 chem joined #puppet-openstack
09:09 Oszkar joined #puppet-openstack
09:20 jerrygb joined #puppet-openstack
09:23 paramite joined #puppet-openstack
10:01 openstackgerrit Gael Chamoulaud proposed openstack/puppet-horizon: Remove duplicate key for ':servername'  https://review.openstack.org/222100
10:06 openstackgerrit Gael Chamoulaud proposed openstack/puppet-ironic: Fix rspec 3.x syntax  https://review.openstack.org/222103
10:10 Oszkar joined #puppet-openstack
10:13 Oszkar1 joined #puppet-openstack
10:13 openstackgerrit Lukas Bezdicka proposed openstack/puppet-neutron: Install service packages on api nodes  https://review.openstack.org/200877
10:15 openstackgerrit Gael Chamoulaud proposed openstack/puppet-manila: Fix rspec 3.x syntax  https://review.openstack.org/222109
10:19 openstackgerrit Gael Chamoulaud proposed openstack/puppet-monasca: Fix rspec 3.x syntax  https://review.openstack.org/222112
10:27 openstackgerrit Martin Mágr proposed openstack/puppet-nova: Rename default endpoint names  https://review.openstack.org/222120
10:33 openstackgerrit Gael Chamoulaud proposed openstack/puppet-neutron: Fix rspec3 syntax  https://review.openstack.org/222123
10:34 openstackgerrit Gael Chamoulaud proposed openstack/puppet-neutron: Fix rspec 3.x syntax  https://review.openstack.org/222123
10:42 cgfbee joined #puppet-openstack
10:44 openstackgerrit Gael Chamoulaud proposed openstack/puppet-swift: Fix rspec 3.x syntax  https://review.openstack.org/222128
10:59 markvoelker joined #puppet-openstack
11:12 openstackgerrit Lukas Bezdicka proposed openstack/puppet-nova: Remove installation of pm-utils  https://review.openstack.org/222141
11:15 openstackgerrit Lukas Bezdicka proposed openstack/puppet-neutron: Install service packages on api nodes  https://review.openstack.org/200877
11:20 openstackgerrit Denis Egorenko proposed openstack/puppet-sahara: Revert "Fixed /etc/sahara handling"  https://review.openstack.org/222142
11:21 jerrygb joined #puppet-openstack
11:26 openstackgerrit Denis Egorenko proposed openstack/puppet-sahara: Revert "Fix Sahara installation for Ubuntu"  https://review.openstack.org/222144
11:27 openstackgerrit Denis Egorenko proposed openstack/puppet-sahara: Revert "Fix Sahara installation for Ubuntu"  https://review.openstack.org/222144
11:28 openstackgerrit Denis Egorenko proposed openstack/puppet-sahara: Revert "Fix Sahara installation for Ubuntu"  https://review.openstack.org/222144
11:31 TC01 joined #puppet-openstack
11:32 yogesh-pc_ joined #puppet-openstack
11:32 morazi joined #puppet-openstack
11:35 EmilienM good morning folks
11:35 degorenko o/ hi
11:36 openstackgerrit Denis Egorenko proposed openstack/puppet-sahara: Remove POSIX users, groups, and file modes.  https://review.openstack.org/222146
11:39 arnaud_orange joined #puppet-openstack
11:50 tiswanso joined #puppet-openstack
11:50 xingchao joined #puppet-openstack
11:51 openstackgerrit Merged openstack/puppet-keystone: domain name from id lookups return empty  https://review.openstack.org/221119
11:55 markvoelker joined #puppet-openstack
11:59 karume joined #puppet-openstack
12:02 xingchao joined #puppet-openstack
12:02 delattec joined #puppet-openstack
12:03 leanderthal joined #puppet-openstack
12:08 Oszkar joined #puppet-openstack
12:09 Oszkar1 joined #puppet-openstack
12:09 krain_ joined #puppet-openstack
12:12 zhangjn joined #puppet-openstack
12:13 zhangjn joined #puppet-openstack
12:19 dprince joined #puppet-openstack
12:20 dprince joined #puppet-openstack
12:23 ducttape_ joined #puppet-openstack
12:31 bklei joined #puppet-openstack
12:32 karume joined #puppet-openstack
12:37 gfidente joined #puppet-openstack
12:40 bklei joined #puppet-openstack
12:40 jerrygb joined #puppet-openstack
12:40 openstackgerrit Emilien Macchi proposed openstack/puppet-nova: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222025
12:43 bklei_ joined #puppet-openstack
12:45 bklei_ joined #puppet-openstack
12:57 openstackgerrit Emilien Macchi proposed openstack/puppet-nova: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222025
12:58 openstackgerrit Emilien Macchi proposed openstack/puppet-nova: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222025
12:59 openstackgerrit Merged openstack/puppet-keystone: able to use wrong value for default_domain_id  https://review.openstack.org/221120
13:00 richm joined #puppet-openstack
13:02 openstackgerrit Merged openstack/puppet-keystone: able to use wrong resource name  https://review.openstack.org/221121
13:05 karume joined #puppet-openstack
13:06 bklei joined #puppet-openstack
13:10 Vinsh Starting CI... like an old lawnmower ;)
13:19 pradk joined #puppet-openstack
13:19 tiswanso joined #puppet-openstack
13:36 EmilienM Vinsh: today I'm working on new repos
13:36 EmilienM see https://etherpad.openstack.org/p/puppet-liberty-blocker
13:37 Vinsh cbs +1
13:41 jerrygb joined #puppet-openstack
13:41 openstackgerrit Emilien Macchi proposed openstack/puppet-nova: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222025
13:42 openstackgerrit Denis Egorenko proposed openstack/puppet-sahara: Revert "Fixed /etc/sahara handling"  https://review.openstack.org/222142
13:43 zhangjn joined #puppet-openstack
13:52 openstackgerrit Javier Peña proposed openstack/puppet-swift: Add support for swift-object-expirer service  https://review.openstack.org/211549
13:53 xingchao joined #puppet-openstack
13:55 openstackgerrit Adam Vinsh proposed openstack/puppet-swift: WIP Manage swift with swiftinit service provider  https://review.openstack.org/203220
13:56 dfisher joined #puppet-openstack
13:58 zhangjn joined #puppet-openstack
14:00 paramite joined #puppet-openstack
14:01 ducttape_ joined #puppet-openstack
14:02 tiswanso joined #puppet-openstack
14:07 openstackgerrit Merged openstack/puppet-designate: Fix rspec 3.x syntax  https://review.openstack.org/222077
14:13 degorenko spredzy, hey o/ i want ask you about https://review.openstack.org/#/c/219275/ and https://review.openstack.org/#/c/221005/ Should i rebase on your patch? Or until we didn't receive any positive feedback for your patch, i can use old style? :)
14:13 openstackgerrit Merged openstack/puppet-glance: Fix rspec 3.x syntax  https://review.openstack.org/222081
14:15 jerrygb joined #puppet-openstack
14:15 spredzy degorenko, I'd say yes if you don't mind :) it already got a +2 hopefully it won't last before getting merged
14:16 openstackgerrit Merged openstack/puppet-monasca: Fix rspec 3.x syntax  https://review.openstack.org/222112
14:18 degorenko spredzy, i must upset you - your patch has merge conflict
14:18 openstackgerrit Merged openstack/puppet-ironic: Fix rspec 3.x syntax  https://review.openstack.org/222103
14:18 chandankumar joined #puppet-openstack
14:18 EmilienM degorenko: I never saw spredzy upset :P
14:18 EmilienM I tried... :)
14:18 spredzy degorenko, hahah no worries, it happens quite often lately across projects :)
14:18 spredzy EmilienM, ahaha :D
14:19 degorenko EmilienM, spredzy :D
14:19 spredzy degorenko, will rebase later today :)
14:19 openstackgerrit Merged openstack/puppet-manila: Fix rspec 3.x syntax  https://review.openstack.org/222109
14:19 degorenko spredzy, so, i will wait until your patch will be merged, and then i'll update mine :)
14:20 spredzy degorenko, deal ;)
14:25 openstackgerrit Merged openstack/puppet-neutron: Fix rspec 3.x syntax  https://review.openstack.org/222123
14:27 openstackgerrit Merged openstack/puppet-swift: Fix rspec 3.x syntax  https://review.openstack.org/222128
14:34 openstackgerrit Emilien Macchi proposed openstack/puppet-nova: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222025
14:39 openstackgerrit Emilien Macchi proposed openstack/puppet-nova: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222025
14:44 guimaluf joined #puppet-openstack
14:53 IBerezovskiy joined #puppet-openstack
14:53 degorenko joined #puppet-openstack
14:58 openstackgerrit Ivan Chavero proposed openstack/puppet-neutron: Split SR-IOV configuration file into driver and agent pieces  https://review.openstack.org/221991
15:12 paramite joined #puppet-openstack
15:14 yogesh-pc joined #puppet-openstack
15:19 yogesh-pc I would like to replicate the an openstack environment using 2 VM: 1 Compute and 1 Controller Node. If I am successful I will be implementing the same on actual HW nodes. Could someone let me know how I can go about it.
15:48 yogesh-pc_ joined #puppet-openstack
15:49 tiswanso joined #puppet-openstack
15:49 openstackgerrit Emilien Macchi proposed openstack/puppet-nova: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222025
15:52 IBerezovskiy joined #puppet-openstack
15:55 Fin1te joined #puppet-openstack
15:56 xarses joined #puppet-openstack
15:58 angdraug joined #puppet-openstack
16:09 degorenko joined #puppet-openstack
16:15 openstackgerrit Emilien Macchi proposed openstack/puppet-ceilometer: acceptance: bump to Liberty  https://review.openstack.org/222265
16:15 openstackgerrit Denis Egorenko proposed openstack/puppet-sahara: Add new parameters for Sahara  https://review.openstack.org/220090
16:15 openstackgerrit Emilien Macchi proposed openstack/puppet-cinder: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222266
16:15 openstackgerrit Emilien Macchi proposed openstack/puppet-designate: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222269
16:16 openstackgerrit Emilien Macchi proposed openstack/puppet-glance: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222270
16:16 openstackgerrit Emilien Macchi proposed openstack/puppet-heat: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222271
16:16 openstackgerrit Emilien Macchi proposed openstack/puppet-horizon: acceptance: bump to Liberty  https://review.openstack.org/222272
16:16 openstackgerrit Emilien Macchi proposed openstack/puppet-ironic: acceptance: bump to Liberty  https://review.openstack.org/222273
16:16 openstackgerrit Emilien Macchi proposed openstack/puppet-keystone: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222274
16:17 openstackgerrit Emilien Macchi proposed openstack/puppet-manila: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222275
16:17 openstackgerrit Emilien Macchi proposed openstack/puppet-neutron: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222276
16:17 openstackgerrit Emilien Macchi proposed openstack/puppet-sahara: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222277
16:17 openstackgerrit Emilien Macchi proposed openstack/puppet-swift: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222278
16:18 openstackgerrit Emilien Macchi proposed openstack/puppet-trove: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222279
16:30 chem joined #puppet-openstack
16:31 openstackgerrit Denis Egorenko proposed openstack/puppet-sahara: Remove POSIX file modes.  https://review.openstack.org/222146
16:41 openstackgerrit Emilien Macchi proposed openstack/puppet-keystone: wsgi/ubuntu: do not manage keystone.py file  https://review.openstack.org/222286
16:57 openstackgerrit Adam Vinsh proposed openstack/puppet-swift: WIP Manage swift with swiftinit service provider  https://review.openstack.org/203220
16:58 mwhahaha EmilienM: https://review.openstack.org/#/c/219793/ i had already done that keystone fix :D
16:59 EmilienM oh nice
16:59 EmilienM but it fails
16:59 EmilienM no
16:59 EmilienM it fails in integ err
16:59 EmilienM mwhahaha: sorry, for dup, abandoning mine right now
16:59 mwhahaha it's all good
17:00 EmilienM nice catch :P
17:20 dtalmage_ joined #puppet-openstack
17:45 chandankumar joined #puppet-openstack
17:50 richm Does it matter if the output of `puppet resource` does not match what's in a manifest?
17:50 richm I think not
17:50 richm I would like to not have to support that they match
17:50 richm The issue, once again, is the '::domain' in the resource name
17:51 richm What if someone decides they don't want to do that, and just create a resource like keystone_tenant { 'projectname': domain => 'domain1' } ?
17:52 richm because, why not?  It is completely unambiguous
17:52 richm does it matter if `puppet resource` outputs keystone_tenant { 'projectname::domain1': domain => 'domain1' } ?
17:53 richm I have two problems 1) I don't know how people actually use this stuff 2) Whichever way we do it, it will piss off someone
17:54 EmilienM richm: I guess people will start complaining or giving feedback if something's wrong
17:56 richm For example, if we require that the name is always "name::domain", do we even need a domain property for resources?  Is that even useful anymore?
17:56 richm If I always have to say keystone_tenant { 'projectname::domain1': ...}
17:57 richm What is the domain property useful for?
17:58 richm When I did the initial design, I was thinking that Puppet has facilities to easily change or provide default values for properties specified as parameters to resource or class declarations
17:58 richm For example, node based defaults, hiera integration, etc.
17:59 richm But based on the limited feedback I have had, it seems that people who deploy this stuff don't really care, because the names are parameterized too, so it is just as easy to generate a name with a domain or without
18:00 richm so if we are going to require 'name::domain' everywhere, it is just as easy for deployers to do that as provide domain parameters
18:00 richm which means we don't need the domain properties
18:00 richm which means we should get rid of them asap
18:01 richm which means officially deprecating them from a lot of places
18:01 richm at the same time as we begin the deprecation process for names without '::domain'
18:02 richm and just wait until you see how ugly this is going to make keystone_trust resource naming . . .
18:03 richm or specifying a keystone_group with a lot of users . . .
18:05 imcsk8 richm: from my point of view the domain should be explicitly set
18:05 EmilienM richm: I suggest to ask operators mailing list
18:05 EmilienM and openstack-dev/puppet
18:05 EmilienM so we get attention from people on the field
18:06 richm imcsk8: explicitly set in the name?
18:07 kindjal joined #puppet-openstack
18:07 richm imcsk8: for example, how would this affect packstack, if you have to change everywhere a keystone resource is passed in, to make sure it is passed in with "::domain" added?
18:07 kindjal Setting up keystone with puppet-keystone (kilo) yeilds this error, “openstack service list: error: unrecognized arguments: --long”...
18:07 imcsk8 richm: as parameter: domain  => 'domain1'
18:08 kindjal appears that —long is hard coded, and I think only supported by openstackclient with API version 2… and not version 3… but I can’t figure out how to specify which API version to use.
18:08 kindjal My environment is empty (bootstrapping)
18:08 kindjal anyone else see this?
18:08 admin0 joined #puppet-openstack
18:09 imcsk8 richm: if we have to do that in packstack we could concatenate the name in the python code
18:10 richm imcsk8: packstack relies on using the default names for users and tenants in many places - there isn't even a way in packstack to specify a name for some users and tenants
18:11 imcsk8 richm: exactly we would have to change that. we are actually going to make packstack use keystone v3
18:11 imcsk8 richm: so we'll have to make a lot of changes
18:15 richm imcsk8: many changes either way - add 'domain', 'user_domain', 'project_domain', etc. parameters in many, many places, or change the way names are created
18:15 imcsk8 richm: yes it's gonna be fun and we'll be asking you a lot of questions ;)
18:15 richm If you just have to change the names, then you don't have to add domain parameters to many, many classes and resources
18:15 severion joined #puppet-openstack
18:16 richm however, you have to make sure that every class understands what 'name::domain' means
18:16 richm all of the puppet-keystone code will
18:16 richm but who knows where that name will be used throughout all of the puppet modules?
18:18 richm for example - lets say I have an admin user - let's say this user exists in corp LDAP server, so I create a domain 'admin_domain' for this user
18:19 richm so I pass in 'admin::admin_domain' as the user name in all of my puppet classes/resources
18:19 richm including nagios, let's say
18:20 richm now, nagios isn't technically an openstack project and technically doesn't have to know about Keystone v3 auth and domain scoped tokens
18:20 richm nagios only knows about the corp LDAP and can use 'admin' without a domain.  Who is going to be responsible for stripping off the '::admin_domain' part in order to pass to the puppet nagios classes/resources?
18:22 richm Another question: In puppet, is it the 'name' property that must be unique, or is it the 'title'?
18:23 richm for example, is it legal to have both keystone_tenant { 'one tenant': name => 'tenant', domain => 'domain1'} and keystone_tenant { 'another tenant': name => 'tenant', domain => 'domain2'}
18:25 Fin1te joined #puppet-openstack
18:30 Vinsh Dang, zuul sure is backed up today.
18:32 linkedinyou joined #puppet-openstack
18:34 kindjal Anyone have a sec to help me run puppet-keystone? I get one error if I use API v2.0 and another error if I try API v3… not sure what it wants.
18:34 kindjal v3 doesn’t like —long, and v2 doesn’t like —format
18:34 iurygregory joined #puppet-openstack
18:35 mflobo joined #puppet-openstack
18:41 openstackgerrit Merged openstack/puppet-keystone: Use Ubuntu provided wsgi.py for keystone.wsgi  https://review.openstack.org/219793
18:42 imcsk8 richm:  title has to be unique https://docs.puppetlabs.com/puppet/latest/reference/lang_defined_types.html#resource-uniqueness
18:43 openstackgerrit Iury Gregory Melo Ferreira proposed openstack/puppet-keystone: Support for Keystone as Identity Provider  https://review.openstack.org/208054
18:52 aimon joined #puppet-openstack
18:53 richm kindjal: I'm not sure what you mean by "use API 2.0" and "use API 3" - the puppet-keystone keystone_service provider is hardcoded to use API 3
18:55 angdraug joined #puppet-openstack
18:56 kindjal richm: Execution of '/usr/bin/openstack service list --quiet --format csv --long' returned 2
18:57 kindjal openstack service list: error: unrecognized arguments: --long
18:57 kindjal that happens with an empty environment… playing around on the command line seems to reveal that the —long option is only available in API v2 in openstackclient
18:57 richm kindjal: /usr/bin/openstack --os-identity-api-version 3 service list --quiet --format csv --long
18:58 kindjal ERROR: cliff.app You must provide a username via either --os-username or env[OS_USERNAME]
18:58 richm you have to tell openstack which identity api version you are using - the arguments are different depending on the version
18:58 richm kindjal: env|grep ^OS_
18:58 iurygregory EmilienM, when you have time can you take a look at the acceptance in https://review.openstack.org/#/c/208054/ ^^
18:59 kindjal richm: No results
18:59 kindjal nothing with OS_
18:59 richm kindjal: did you expect /usr/bin/openstack service list --quiet --format csv --long to return a result with no authentication provided?
18:59 richm kindjal: or were you just looking for it to provide a usage error message?
18:59 kindjal well, I’ve tried with OS_TOKEN and user creds
19:00 kindjal and it fails for different reasons...
19:00 kindjal then I found this: https://bugs.launchpad.net/puppet-keystone/+bug/1488018
19:00 openstack Launchpad bug 1488018 in puppet-keystone "Does not work if v3 API variables are set in the environment" [Undecided,Incomplete]
19:00 kindjal within there is a comment, “To be able to bootstrap Keystone installation properly, I would not recommend to have an openrc or any openstack variables preset."
19:00 kindjal I’m expecting it to read auth_token from /etc/keystone/keystone.conf, I think.
19:00 richm kindjal: no, it won't do that
19:01 derekh joined #puppet-openstack
19:01 richm you have to specify OS_TOKEN= or openstack --os-token "value"
19:01 kindjal and OS_URL too then?
19:01 richm yes
19:02 kindjal but not OS_IDENTITY_API_VERSION?
19:02 richm # OS_IDENTITY_API_VERSION=3 OS_TOKEN="value" OS_URL=http://localhost:35357/v3 openstack service list --long
19:03 kindjal with those set:
19:03 kindjal openstack service list: error: unrecognized arguments: --long
19:03 richm kindjal: openstack --version
19:03 kindjal openstack 0.3.0
19:03 kindjal hmm
19:03 kindjal python-openstackclient | 1.0.3-0ubuntu1~cloud0 | http://ubuntu-cloud.archive.canonical.com/ubuntu/ trusty-updates/kilo/main amd64 Packages
19:03 kindjal not sure how I got the old version...
19:04 kindjal apt-cache policy shows 1.0.3 available...
19:04 richm yeah - 0.3.0 is not going to work for many reasons, this being only one of them
19:04 kindjal python-openstackclient is already the newest version.
19:04 kindjal despite 1.0.3 being a candidate
19:04 kindjal hmm
19:04 kindjal ok
19:04 kindjal thanks!
19:12 ducttape_ joined #puppet-openstack
19:26 Fin1te joined #puppet-openstack
19:28 xingchao joined #puppet-openstack
19:47 openstackgerrit Emilien Macchi proposed openstack/puppet-ceilometer: acceptance: bump to Liberty  https://review.openstack.org/222265
19:49 openstackgerrit Emilien Macchi proposed openstack/puppet-heat: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222271
19:51 yogesh-pc joined #puppet-openstack
19:54 kindjal richm: I got openstackclient 1.0.3 installed, and there’s still an error...
19:54 kindjal with OS_ variables set...
19:54 kindjal /usr/bin/openstack domain list --quiet --format csv
19:54 kindjal ERROR: openstack Invalid command 'domain list --format'
20:00 kindjal if I set API v3 via OS_ then I get: openstack endpoint list: error: unrecognized arguments: --long
20:01 kindjal if I unset all OS_ then I think it’s working!
20:05 linkedinyou|2 joined #puppet-openstack
20:15 iurygregory hey richm, when configure keystone with mellon , i need to modify another file beyond keystone.conf and keystone-apache.conf (apache directory) ?
20:16 openstackgerrit Merged openstack/puppet-neutron: Configure Big Switch ML2 plugin  https://review.openstack.org/214686
20:21 richm iurygregory: looking
20:22 iurygregory ok
20:24 richm iurygregory: not directly - mellon provides a script /usr/libexec/mod_auth_mellon/mellon_create_metadata.sh which creates the SP metadata file, and the cert and key files
20:24 richm https://github.com/richm/rdo-vm-factory/blob/master/rdo-federation-setup/vm-post-cloud-init-rdo.sh#L129
20:25 richm then it uses those files in the keystone-apache.conf
20:25 richm https://github.com/richm/rdo-vm-factory/blob/master/rdo-federation-setup/vm-post-cloud-init-rdo.sh#L129
20:26 iurygregory hummm, for shibboleth we need to update the shibolleth2.xml file and attribute-map.xml
20:30 jerrygb joined #puppet-openstack
20:33 yogesh-pc joined #puppet-openstack
20:36 richm iurygregory: ok - that's one of those things that will be different in each saml implementation
20:38 iurygregory yeah, i'm just wondering which puppet-module can i use to modify shibboleth file, since puppet-keystone will only change keystone conf files
20:49 richm iurygregory: hmm - don't know - did web search find anything?
20:49 richm iurygregory: you might be the first :P
20:54 iurygregory richm, i've found this (https://github.com/aethylred/puppet-shibboleth) XD, so the user who wants keystone as SP using shibboleth will need to use keystone::federation::saml::shibboleth to configure the keystone and use shibboleth from this repo..
20:55 iurygregory to apply changes to shibboleth files
20:56 richm iurygregory: ok - not sure how puppet-shibboleth should be handled - it's not really a puppet-openstack module - sounds like something that should be on puppet forge
20:57 EmilienM clayton: where do you run containers? on a VM?
20:57 EmilienM dprince ^
20:57 clayton on baremetal
20:57 clayton for deploying openstack itself
20:57 dprince clayton: where are your Dockerfiles?
20:57 dprince clayton: for say... openstack service containers?
20:57 clayton right, we run heat and designate in containers right now
20:58 iurygregory thanks richm
20:58 EmilienM clayton: have you investigated magnum?
20:58 rhallisey joined #puppet-openstack
20:58 EmilienM or using Puppet to manage containers is good for you?
20:58 Slower joined #puppet-openstack
20:59 clayton I'm guessing magnum wouldn't make sense unless we were doing tripleo
20:59 EmilienM lol
20:59 clayton but no, I haven't looked into it hardly at all
20:59 EmilienM it's funny because I'm proposing to use Magnum in TripleO
20:59 EmilienM but until now, we don't think it's the right approach for us
21:00 clayton The little I've read, Magnum sounds interesting
21:00 EmilienM clayton: how do you manage configuration files with containers?
21:01 clayton we mount /etc/<service> and /var/log/<service> into the container for now
21:01 clayton the former being a read only mount
21:01 EmilienM clayton: so for you, it's not possible to run 2 heat-api containers if they have different configurations
21:02 clayton no, I still don't really understand why anyone would want to do that
21:02 EmilienM clayton: I'm was just wondering
21:02 gfidente joined #puppet-openstack
21:02 EmilienM dprince, rhallisey, Slower: this approach is interested, we might consider it too
21:02 EmilienM interesting*
21:03 clayton the glue code we use between the openstack puppet modules and our containers is here - https://github.com/twc-openstack/puppet-os_docker
21:03 dprince EmilienM: that is very similar to what we are doing
21:03 dprince EmilienM: we create our config files w/ puppet (in an /etc tree)
21:03 clayton our dockerfiles aren't public right now, but they're not anything fancy.  we basically just build a venv inside the container with a frozen requirementx.txt
21:04 EmilienM dprince: puppet-tripleo is tricking a lot of stuff though
21:04 EmilienM we might need to get rid of this
21:04 rhallisey sounds a lot like what we do
21:04 clayton We've looked at kolla, but right now there are some features missing that we need (like being able to specify a frozen requirements.txt)
21:04 dprince EmilienM: I'm open to ideas
21:04 rhallisey clayton, if you need anything implemented make a bp
21:05 rhallisey and I'll make it happen
21:05 EmilienM Alleluia
21:05 rhallisey for kolla that is
21:05 clayton I know, I talked to sdake about it already, I just don't have time right now
21:05 rhallisey clayton, kk
21:06 dprince EmilienM, clayton : this rhallisey work on the external config interface (that we plan on using with the initial TripleO work) https://review.openstack.org/#/c/216874/
21:06 EmilienM dprince: we could investigate the 'tag' thing
21:06 dprince EmilienM: by all means, yes :)
21:06 clayton and honestly, I think kolla looks like a good project, but it has a lot of flexibility we don't need right now, so switching from our home grown stuff isnt' very high priority.
21:06 dprince EmilienM: I planned on doing it to get you off my case :)
21:06 EmilienM lol
21:06 dprince EmilienM: but if you beat me to it I wouldn't be unhappy
21:07 EmilienM ahah
21:07 EmilienM oh dear
21:07 EmilienM clayton: during the midcycle, could you push more patches for tags?
21:08 clayton @EmilienM: that's the plan, given time
21:09 EmilienM for now, dprince is noop'ing the resources
21:09 EmilienM see https://review.openstack.org/#/c/209030/
21:09 openstackgerrit Gilles Dubreuil proposed openstack/puppet-keystone: Fixes get_section for future use  https://review.openstack.org/222017
21:10 EmilienM dprince: how do you create DB then?
21:10 openstackgerrit Gilles Dubreuil proposed openstack/puppet-keystone: Keystone V3 support with domain checking  https://review.openstack.org/218044
21:10 dprince EmilienM: the nooping is fine grained
21:10 dprince EmilienM: you could simply allow Exec resources
21:10 dprince EmilienM: but for the compute role we don't run DB syncs
21:10 dprince EmilienM: so we are good to go without Exec's
21:11 rhallisey dprince, kolla as a build in flag to run db_sync
21:11 openstackgerrit Gilles Dubreuil proposed openstack/puppet-keystone: Providers tenant/user/user_role: domain check/name  https://review.openstack.org/218059
21:11 rhallisey I think it's bootstrap or something
21:11 dprince rhallisey: right, and we haven't decided where/what will run the db_syncs yet
21:11 EmilienM dprince: I'm more interested by controllers, more challenging
21:11 rhallisey right
21:11 dprince EmilienM: agree
21:11 EmilienM and well, do roles still mean something?
21:11 EmilienM since we are going to run containers
21:12 gildub joined #puppet-openstack
21:12 dprince EmilienM: we want better "composability" for roles
21:12 kindjal joined #puppet-openstack
21:12 dprince EmilienM: but for now they are still fairly rigid. And they may be meaningful because there could be different hardware underneath
21:12 dprince EmilienM: i.e. you've got attached storage in some places, but not others
21:13 dprince EmilienM: better nics on some nodes, etc
21:13 gildub EmilienM, richm, now that we got that far, can we please finish cleaning the domain, the domains bugs got merged and I need to have the domain default in so we can:
21:13 EmilienM dprince: yeah, containers would help to break roles model
21:13 gildub EmilienM, richm, 1.backport 2.have a strong base for now (trust) and later (group)
21:14 dprince EmilienM: moving the stateless containers around would likely be easy. The others... not so much
21:14 EmilienM dprince: right.. pacemaker, etc :/
21:15 kindjal_ joined #puppet-openstack
21:31 aimon joined #puppet-openstack
21:41 chem joined #puppet-openstack
21:44 openstackgerrit Emilien Macchi proposed openstack/puppet-swift: acceptance: switch RDO to CBS repositories  https://review.openstack.org/222278
21:51 EmilienM gildub: ping https://review.openstack.org/#/c/218059/40/spec/acceptance/keystone_wsgi_apache_spec.rb,cm
21:51 EmilienM gildub: you're dropping 'domain' parameter resource
21:51 EmilienM I wonder why
21:51 derekh joined #puppet-openstack
21:57 EmilienM richm, gildub: see my comment on https://review.openstack.org/#/c/218059
21:58 EmilienM yogesh-pc: you should look puppet-nova
21:59 yogesh-pc EmilienM: thanks.
22:00 yogesh-pc EmilienM: can we place neutron within the controller host machine?
22:00 EmilienM you can place whatever you like
22:00 EmilienM our modules are flexible enough
22:11 ducttape_ joined #puppet-openstack
22:11 yogesh-pc So is there a manifest for the controller and compute node setups that I can use ?
22:13 yogesh-pc I want to setup a quick openstack cloud as a proof of concept,.. pretty new to openstack
22:17 EmilienM yogesh-pc: I can give you a single node manifest
22:17 EmilienM yogesh-pc: we're currently working to have more documentation
22:17 EmilienM and this is something we will provide very soon, probably next week or so
22:18 yogesh-pc Awesome.. Can I get a single node to start with... that would be really useful
22:18 EmilienM if you want a single node manifest, look: https://github.com/openstack/puppet-openstack-integration/blob/master/fixtures/scenario001.pp#L17-L237
22:18 yogesh-pc I am using vagrant to test the setup for now
22:18 EmilienM tested on ubuntu trusty & centos
22:18 EmilienM yogesh-pc: me too
22:18 EmilienM the manifest works fine, out of the box with master branch of modules
22:19 yogesh-pc for my final setup, I am thinking of deploying 2 nodes: one controller with all the services and a compute node that will host the VMs
22:20 yogesh-pc so for one node how does it install the hypervisor?
22:20 openstackgerrit Emilien Macchi proposed openstack/puppet-openstack-integration: switch RDO to CBS repositories  https://review.openstack.org/222385
22:21 EmilienM for the compute you would only need https://github.com/openstack/puppet-openstack-integration/blob/master/fixtures/scenario001.pp#L173-L177
22:21 EmilienM https://github.com/openstack/puppet-openstack-integration/blob/master/fixtures/scenario001.pp#L206-L214
22:21 EmilienM https://github.com/openstack/puppet-openstack-integration/blob/master/fixtures/scenario001.pp#L151-L160
22:21 EmilienM and https://github.com/openstack/puppet-openstack-integration/blob/master/fixtures/scenario001.pp#L226-L231
22:21 EmilienM otherwise it's on the controller
22:24 yogesh-pc So other than the 3 classes rest all will be on the controller?
22:25 yogesh-pc sorry 5 total
22:27 yogesh-pc EmilienM: thanks for that... I will start with the single node. If someone could write a manifest for a two node setup, it will be great!
22:42 kindjal_ joined #puppet-openstack
22:43 aimon_ joined #puppet-openstack
22:44 cdelatte joined #puppet-openstack
22:48 gildub EmilienM, re 218059, looking
22:54 xarses joined #puppet-openstack
22:58 openstackgerrit Merged openstack/puppet-openstack-integration: integration tests: use dstat to monitor system  https://review.openstack.org/218587
23:07 openstackgerrit Gilles Dubreuil proposed openstack/puppet-keystone: Providers tenant/user/user_role: domain check/name  https://review.openstack.org/218059
23:07 gildub EmilienM, domain parameter should go away, because they're bad, as discussed with richm. We probably want to do this sooner than later, in a new patch, while domains are still experimental, otherwise we would have to do a deprecation phase
23:07 EmilienM we introduced it in stable/kilo
23:08 EmilienM AFIK
23:08 gildub EmilienM, yes but experimental
23:08 EmilienM but if we want to drop it OK, but not before M cycle
23:08 EmilienM gildub: we can't just drop it
23:08 EmilienM it's not our policy
23:08 EmilienM gildub: we need to deprecate it
23:08 gildub EmilienM, ok, that's not a blocker anyway
23:08 EmilienM richm: ^ fyi
23:09 EmilienM we don't drop it, we deprecate it
23:09 EmilienM and I want to see acceptance working like before at least for some resources
23:09 EmilienM to prove it works like in stable/kilo
23:09 gildub EmilienM, agreed, patch 218059 doesn't remove them
23:10 EmilienM gildub: ack
23:10 EmilienM please continue to test it
23:10 EmilienM at least for one resource
23:10 EmilienM and add some comment maybe in acceptance to explain the plans
23:10 gildub EmilienM, it's in, waiting for CI result....
23:10 EmilienM it's very useful
23:10 EmilienM gildub: ok :)
23:10 gildub EmilienM, I will add comments
23:17 TC01 joined #puppet-openstack
23:18 delattec joined #puppet-openstack
23:18 gildub EmilienM, also, could you +2 https://review.openstack.org/221567 ? chem has done +1 - That would clear the path (I think for starting process to back-port the domain bugs fixes
23:18 gildub EmilienM, please :)
23:18 EmilienM looking
23:20 EmilienM gildub: can you explain that change? maybe I'm missing something
23:21 EmilienM gildub: nevermind
23:21 gildub EmilienM, the initial driver was to help with Keystone v2/v3 issues (fuel was the first to discover them, among others), now things have evolved, if we don't do that the backport of the domains bugs are going to be much harder
23:21 gildub EmilienM, no worries :)
23:22 EmilienM gildub: +2
23:22 gildub EmilienM, thanks
23:25 gildub michchap, G'Day - Could you please review https://review.openstack.org/221567 ?
23:26 richm EmilienM: gildub: The more I think about it, the more I realize that having domain parameters is a good thing
23:26 richm 1) everywhere else Keystone is used, there are explicit domain parameters
23:27 richm 2) we can still have unique resource titles, even if the names are not unique - so keystone_tenant { 'my tenant': name => 'project', domain => 'domain1'} and keystone_tenant { 'another tenant': name => 'project', domain => 'domain2' } are distinguished by the domain
23:28 richm 3) who cares if puppet resource doesn't return things with the right name/title?  It doesn't now anyway
23:28 michchap gildub: ack
23:28 richm so even having "::" in the name is not necessary
23:29 richm That is - perhaps it is better to just get rid of the "name::domain" thing altogether, and just have people pass in domain parameters everywhere
23:30 gildub richm, I disagree with 3) what about idempotency ?
23:32 gildub richm, unless using description field (when available) for storing an id (or equivalent) to uniquely identify the resource, which is probably bad anyway
23:34 gildub richm, ... as you know,  there is no way to identify uniquely the resource otherwise, unless using domain parameter all times, which is not possible when doing discovery (self.instances)
23:35 gildub richm, and if we choose that road, it means adding extra logic into the providers where they shouldn't have to do so because trying to work around things to simulate some sort of idem-potency.
23:36 gildub richm, therefore I think this would end up in the bad design column
23:38 gildub richm, so actually I disagree with 2 and 3. I don't understand 1, unless you mean outside of puppet (like openstackclient), well they don't have to be idempotent, they have unique IDs available
23:41 gildub richm, also forgot to say, the recommendation from Keystone folks are to always explicitly use the domain, we should not forget that because we had to offer a compatibility bridge for PUppet v2 API
23:45 gildub BTW, for everyone else here, the latter ^ was submitted via list, even though we didn't get lots of feedback, the recommendation, especially from Keystone side, seems pretty clear
23:47 gildub michchap, thanks
23:49 gildub richm, ... because we had to offer compatibilty we ended up with no mandatory domain
23:52 openstackgerrit Emilien Macchi proposed openstack/puppet-openstack-integration: switch RDO to CBS repositories  https://review.openstack.org/222385

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