Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2015-03-05

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

All times shown according to UTC.

Time Nick Message
00:03 mfisch I can look tonight
00:16 openstackgerrit Jamie Lennox proposed stackforge/puppet-swift: Use keystonemiddleware instead of client  https://review.openstack.org/161476
00:22 aimon joined #puppet-openstack
00:54 openstackgerrit Jamie Lennox proposed stackforge/puppet-swift: Use keystonemiddleware instead of client  https://review.openstack.org/161476
01:02 Madkiss_ joined #puppet-openstack
01:08 markvoelker joined #puppet-openstack
01:15 topshare joined #puppet-openstack
01:15 markvoelker joined #puppet-openstack
01:24 badiane_ka joined #puppet-openstack
01:53 rmoe joined #puppet-openstack
01:55 xingchao_ joined #puppet-openstack
02:16 markvoelker joined #puppet-openstack
02:19 aarefiev_ joined #puppet-openstack
02:20 abramley_ joined #puppet-openstack
02:20 rwsu_ joined #puppet-openstack
03:17 markvoelker joined #puppet-openstack
03:28 ericpeterson joined #puppet-openstack
03:38 xarses joined #puppet-openstack
04:19 markvoelker joined #puppet-openstack
04:33 mfisch joined #puppet-openstack
05:19 markvoelker joined #puppet-openstack
05:20 zerda joined #puppet-openstack
06:22 markvoelker joined #puppet-openstack
06:42 dgurtner joined #puppet-openstack
06:51 Krast joined #puppet-openstack
06:54 ericpeterson joined #puppet-openstack
06:56 stamak joined #puppet-openstack
07:15 moshele joined #puppet-openstack
07:22 markvoelker joined #puppet-openstack
07:30 rissmann joined #puppet-openstack
07:36 openstackgerrit joined #puppet-openstack
07:57 dgurtner joined #puppet-openstack
07:57 dgurtner joined #puppet-openstack
08:04 openstackgerrit Gael Chamoulaud proposed stackforge/puppet-manila: Add puppet-lint-param-docs plugins to puppet-lint  https://review.openstack.org/161140
08:19 tfz joined #puppet-openstack
08:25 markvoelker joined #puppet-openstack
08:33 stamak joined #puppet-openstack
08:49 jistr joined #puppet-openstack
08:57 jpena joined #puppet-openstack
09:15 mmagr joined #puppet-openstack
09:20 moshele can core reviews review this https://review.openstack.org/#/c/158123/
09:26 markvoelker joined #puppet-openstack
09:27 michchap moshele: is there any doc for that config option? I can see the commit here: https://github.com/openstack/cinder/commit/ffdfd0f3bd6d1f72feaefc2b12efdb4ed8b9bc72
09:29 moshele michchap: what kind of doc do you mean?
09:31 michchap moshele: I mean that as a reviewer in order for me to check that a patch is correct, I consult the openstack documentation. I can't see any reference to that config option in the docs.openstack.org so I had to go searching for the commit where it was added to confirm what the new option actually does.
09:34 ddmitriev joined #puppet-openstack
09:40 arnaud_orange left #puppet-openstack
09:40 arnaud_orange joined #puppet-openstack
10:02 igajsin joined #puppet-openstack
10:27 markvoelker joined #puppet-openstack
10:33 saju_m joined #puppet-openstack
10:52 zerda joined #puppet-openstack
11:05 moshele joined #puppet-openstack
11:28 markvoelker joined #puppet-openstack
11:40 andriikolesnikov joined #puppet-openstack
11:47 ddmitriev joined #puppet-openstack
11:55 cdent joined #puppet-openstack
11:55 saju_m joined #puppet-openstack
12:04 EmilienM good morning
12:11 dgurtner joined #puppet-openstack
12:12 moshele @EmilienM: hi
12:13 moshele EmilienM: hi
12:29 markvoelker joined #puppet-openstack
12:38 markvoelker joined #puppet-openstack
13:01 dprince joined #puppet-openstack
13:14 morazi joined #puppet-openstack
13:31 rcallawa joined #puppet-openstack
13:40 EmilienM any chance to have a review on https://review.openstack.org/#/c/151324/ ?
13:43 EmilienM dgurtner: I'm still investigating https://review.openstack.org/#/c/156458 and I see vagrant issues in http://paste.ubuntu.com/10508268/
13:48 dgurtner EmilienM: yes, but the run before (against the same patchset) finished without vagrant issues: http://paste.ubuntu.com/10500988/
13:48 EmilienM oh ok
13:49 dgurtner EmilienM: sorry about CI being so crappy right now :(
13:51 EmilienM dgurtner: no problem !
13:52 EmilienM dgurtner: still an issue with librados
14:09 rcallawa joined #puppet-openstack
14:11 openstackgerrit Merged stackforge/puppet-swift: Use keystonemiddleware instead of client  https://review.openstack.org/161476
14:14 richm joined #puppet-openstack
14:15 rcallawa_ joined #puppet-openstack
14:16 rcallaw__ joined #puppet-openstack
14:16 moshele hi can core reviewer review https://review.openstack.org/#/c/158123/ it has already one +2
14:18 rcalla___ joined #puppet-openstack
14:19 IBerezovskiy joined #puppet-openstack
14:32 mschmitt joined #puppet-openstack
14:33 xingchao joined #puppet-openstack
15:00 ericpeterson joined #puppet-openstack
15:09 prad joined #puppet-openstack
15:27 openstackgerrit Merged stackforge/puppet-ceph: Add Puppet 4.x lint checks  https://review.openstack.org/158241
15:29 EmilienM dgurtner: do you think librados should be taken from ceph.com ?
15:29 EmilienM instead of epel
15:36 rcallawa joined #puppet-openstack
15:44 mdorman joined #puppet-openstack
15:48 c0m0 joined #puppet-openstack
15:55 emccormick EmilienM: IF you don't mind an unsolicited opinion, anything to do with ceph should be taken from ceph.com repos and not EPEL :)
15:55 emccormick that includes librados
15:57 EmilienM emccormick: this is a good feedback, thanks. Though i'm still lost about it. dgurtner : thoughts ?
16:00 xarses joined #puppet-openstack
16:17 imcsk8 joined #puppet-openstack
16:22 dgurtner EmilienM: not sure what the current issue is, but you probably won't be able to (and really shouldn't) mix libraries from ceph.com with EPEL
16:22 EmilienM yeah
16:23 EmilienM I think we should pin ceph.com repo
16:23 EmilienM to ensure everything comes from ceph.com, except dependencies (EPEL)
16:26 jpena joined #puppet-openstack
16:26 dgurtner is librados currently coming from epel with you patch?
16:27 dgurtner EmilienM: ^ s/you/your/
16:27 EmilienM dgurtner: I think so
16:27 EmilienM yes
16:29 EmilienM dgurtner: I start to be tired to rebase \o/
16:31 mfisch joined #puppet-openstack
16:32 EmilienM jpena: o/
16:32 EmilienM jpena: reviewing https://review.openstack.org/#/c/161349/1/manifests/params.pp,cm
16:32 dgurtner EmilienM: :D
16:32 mfisch crinkle: our later deploy did hit a network that was removed between the time we can net-list and net-show
16:32 EmilienM jpena: should be test operatingsystem release ?
16:33 jpena EmilienM: that's an issue. We should be checking for OpenStack release actually. RHEL 7 has osp5 and 6 which have the old packaging, for example
16:33 jpena so either we break backwards compatibility, and hope nobody will use the master branch for juno, or we are in bad shape
16:33 EmilienM jpena: it's fair to keep it like this.
16:35 aarefiev joined #puppet-openstack
16:54 EmilienM jpena: https://review.openstack.org/#/c/161349 reviewed - no -1 but still a comment
16:58 jpena EmilienM: just replied to the comment. I saw there were similar cases for other variables, so it looked like a good idea to keep it consistent
17:05 openstackgerrit Emilien Macchi proposed stackforge/puppet-ceph: Add EL7 support  https://review.openstack.org/156458
17:05 EmilienM jpena: sounds good, +2
17:05 jpena EmilienM: thx ;)
17:26 dprince joined #puppet-openstack
17:30 xingchao joined #puppet-openstack
17:33 EmilienM mfisch: what do you think about https://review.openstack.org/#/c/106144/ (solution #1) or https://review.openstack.org/#/c/159769/ (solution #2)
17:35 mfisch will look later i'm on a call
17:36 aimon joined #puppet-openstack
17:40 EmilienM you or others folks
17:52 rmoe joined #puppet-openstack
17:55 andriikolesnikov joined #puppet-openstack
18:24 prad joined #puppet-openstack
18:51 dontalton joined #puppet-openstack
18:57 dgurtner joined #puppet-openstack
18:57 dgurtner joined #puppet-openstack
19:00 isd joined #puppet-openstack
19:02 dgurtner dmsimard: ping re: CI boxes running as root
19:05 dmsimard dgurtner: Yeah that's me
19:06 dmsimard dgurtner: See the e-mail I sent you ?
19:07 dgurtner dmsimard: sorry. will have a look
19:08 isd Hey -- I'm trying to deploy a controller node with puppet, and it's having a few issues. I've traced one of them to the absence of the interface br-ex. Does anyone know off-hand where in the manifests this is created?
19:11 EmilienM isd: afik, puppet-neutron does not create it for you. For now, you can deal with your own composition layer
19:23 tfhartmann joined #puppet-openstack
19:24 openstackgerrit Colleen Murphy proposed stackforge/puppet-openstacklib: Move openstackclient parent provider to openstacklib  https://review.openstack.org/134843
19:25 openstackgerrit Colleen Murphy proposed stackforge/puppet-keystone: Clean up keystone provider methods and tests  https://review.openstack.org/158105
19:25 openstackgerrit Colleen Murphy proposed stackforge/puppet-keystone: Move openstackclient to openstacklib  https://review.openstack.org/161848
19:29 tfz joined #puppet-openstack
19:29 rcallawa joined #puppet-openstack
19:32 rcallawa joined #puppet-openstack
19:37 EmilienM crinkle: I think you can drop the TODO L1 on https://review.openstack.org/#/c/134843/15/lib/puppet/provider/openstack.rb,cm
19:38 crinkle haha ty
19:38 EmilienM crinkle: same for https://review.openstack.org/#/c/134843/15/lib/puppet/util/openstack.rb,cm
19:38 EmilienM and for tests as well
19:39 openstackgerrit Colleen Murphy proposed stackforge/puppet-openstacklib: Move openstackclient parent provider to openstacklib  https://review.openstack.org/134843
19:39 EmilienM crinkle: +2
20:22 rcallawa joined #puppet-openstack
20:26 openstackgerrit joined #puppet-openstack
20:29 rcallawa_ joined #puppet-openstack
20:48 xarses dachary: dmsimard EmilienM  dgurtner, @ceph, how do you deal with creating pools, activating OSD's and ensuring the pools are usable before starting?
20:49 dmsimard before starting what ? :)
20:49 xarses to use them
20:50 dmsimard Hmmm, in fact, I don't personally use puppet to manage my pools and my crush map
20:50 dmsimard Call me old school but it's something I'd rather not entrust puppet with
20:50 xarses well I'm not worried about the puppet side of it persay
20:50 xarses just the operational aspect
20:50 xarses at the moment
20:52 xarses so when creating a pool it creates pg_num PG's these must be at least initially placed before the pool can be used
20:52 xarses if you start activating a bunch of OSD's right after this, it becomes a mess
20:53 xarses with as few as 40 osd's it could take 30min for the pool to become usable
20:55 xarses and to make it worse, there doesn't seem to be an easy way to test that the pool is ready for data
20:56 dmsimard Hmm, I don't know.. My current CI setup doesn't do a whole lot of testing (yet) but I'm going to improve that very soon so I might bump into that issue soon
20:56 dmsimard aka, I'm going to want to make sure that I can instanciate an Openstack VM backed by a Cinder Ceph volume
20:58 xarses ya, we are seeing this being done when we try to upload images to glance
20:59 dmsimard xarses: Your tests try to upload an image before the cluster is ready ?
21:00 xarses mons -> poos(images,volumes) -> [computes(pool(images)), ceph-osd(prepare, activate) -> glance image-upload
21:00 xarses s/poos/pools
21:01 xarses sorry, the computes pool is compute
21:01 dmsimard Surely there's a way to tell whether or not a pool is ready
21:01 dmsimard Don't have a cluster to play with *right now* though
21:02 xarses if we ran: mos -> ceph-osd(prepare,activate) -> pools(images,volumes,compute) -> glance image-upload we can reduce alot of the time, but still need a sane way to determine if the pools is ready
21:04 xarses it's difficult using 'ceph pg dump' and could possibly still fall through in conditions that the pool has become degraded, but is still usable
21:04 xarses the ID10T test would be to just make rados write to the pool and keep looping until it's successful
21:04 xarses but that seems so wrong
21:05 xarses I'm going to end up on the ceph ml with this by the sound of it
21:09 tfhartmann joined #puppet-openstack
21:15 xarses dmsimard: here is our current attempt to wait for the cluster https://review.openstack.org/#/c/153338/
21:16 dmsimard xarses: You can always try and check in teuthology and stuff, I find how to do certain things in the tests sometimes :)
21:17 xarses can you point me to that repo?
21:19 dmsimard https://github.com/ceph/teuthology
21:19 openstackgerrit Merged stackforge/puppet-manila: Add Puppet 4.x lint checks  https://review.openstack.org/160356
21:19 xarses dmsimard: thanks
21:19 openstackgerrit Merged stackforge/puppet-manila: Add puppet-lint-param-docs plugins to puppet-lint  https://review.openstack.org/161140
21:19 dmsimard There's stuff here too: https://github.com/ceph/ceph/tree/master/qa
21:20 dmsimard ex, I reverse engineered the REST API with https://github.com/ceph/ceph/blob/master/qa/workunits/rest/test.py :)
21:20 xarses lol
21:21 dmsimard Not saying you'll find what you're looking for in those two but hey, if you do, let me know :)
21:22 crinkle EmilienM: is changing the default value of a parameter not backwards-incompatible? https://review.openstack.org/#/c/159185
21:23 isd left #puppet-openstack
21:23 mfisch dprince: hey I'm working on moving puppet-cinder to the dprince identity_uri way, not my way. Can you take a look when I'm done?
21:24 dprince mfisch: sure, np
21:30 crinkle mfisch: do you want to talk about https://review.openstack.org/#/c/158965/ ?
21:46 xarses dmsimard: i was able to find https://github.com/ceph/teuthology/blob/98e061765177630f4814bf3ad0b511c70f7580cb/teuthology/task/ceph_manager.py#L1102 "wait_for_" commands that are similar to, but not sure they cover well if a pool is ready, just the cluster at a whole.
21:50 mfisch crinkle: I was hoping to find sbadia, but as long as thats not an arbitrary thing to make lint happy I'm fine with it
21:50 crinkle i mean, it is arbitrary
21:51 crinkle linting is arbitrary
21:51 crinkle it should still be done
21:51 crinkle i can try to give background on why they have that lint check though if you want
21:52 mfisch Is it good, bad, or indifferent that we're now locked to version 5?
21:52 crinkle we're locked to openstacklib version 5 in $module version 5
21:52 mfisch makes perfect sense
21:53 crinkle because when openstacklib version 6 comes out it will have backwards incompatible changes that we don't want to guarantee work with $module v5
21:53 mfisch so we want that anyway
21:53 mfisch regardless of lint
21:53 crinkle yes
21:53 mfisch okay thats what I wanted to know
21:53 crinkle okay cool
21:53 crinkle the license thing is sort of dumb but there's also sort of a reason for that too
21:55 mfisch yep
21:55 mfisch +2d
21:56 crinkle cool
21:56 mfisch hey have you guys seen this one before?
21:56 mfisch Debug: Failed to load library 'augeas' for feature 'augeas'
21:56 mfisch Debug: Puppet::Type::Augeas::ProviderAugeas: feature augeas is missing
21:57 mfisch Error: Could not find a suitable provider for augeas
21:57 mfisch same node worked yesterday with the same code
22:00 openstackgerrit Merged stackforge/puppet-heat: Fix metadata.json lint issues  https://review.openstack.org/158965
22:05 openstackgerrit Colleen Murphy proposed stackforge/puppet-keystone: Clean up keystone provider methods and tests  https://review.openstack.org/158105
22:14 rcallawa joined #puppet-openstack
22:18 rcallawa joined #puppet-openstack
22:21 rcallawa_ joined #puppet-openstack
22:37 rcallawa joined #puppet-openstack
22:40 imcsk8 crinkle: testing your keystone and openstacklib patches, they seem to work ok
22:41 openstackgerrit Merged stackforge/puppet-nova: Fix Nova Scheduler Typo  https://review.openstack.org/161373
22:43 openstackgerrit Merged stackforge/puppet-neutron: Add memcache_servers parameter to base neutron class  https://review.openstack.org/161458
22:43 rcallawa joined #puppet-openstack
22:49 crinkle imcsk8: \o/
22:54 imcsk8 crinkle: i just got this error on the tests: /usr/share/openstack-puppet/modules/keystone/lib/puppet/provider/keystone.rb:2:in `require': cannot load such file -- puppet/provider/openstack (LoadError)
22:55 mfisch this auth_uri deprecation makes me want to cry
22:56 crinkle imcsk8: you mean in the rspec tests?
22:56 imcsk8 crinkle: yes
22:56 crinkle imcsk8: it won't be able to pass until the patch in puppet-openstacklib is merged
22:56 imcsk8 crinkle: i have applied the patch in my setup
22:57 crinkle imcsk8: it doesn't matter for the rspec tests, you'd additionally need to change the .fixtures.yml to look at the patched openstacklib
22:57 crinkle i tested it by removing the openstacklib line from the repositories section and adding a line for it to the symlinks section
22:58 imcsk8 crinkle: ohhh i haven't done that
22:58 crinkle but i obviously don't want to commit that, so jenkins won't pass until the openstacklib patch is merged
22:58 gildub joined #puppet-openstack
23:01 mdorman joined #puppet-openstack
23:11 crinkle I think I'm confused about what constitutes a backwards-incompatible change and when that is okay.
23:12 crinkle Is changing the default value of a class parameter backwards-incompatible?
23:12 crinkle Is changing the value of a params.pp variable backwards-incompatible?
23:12 crinkle Is it okay to make those changes but not to backport them? or should they be backported in some circumstances?
23:13 crinkle When we change the names of parameters we have a deprecation cycle because it will break people's puppet runs, but I think changing the value of a parameter can have unintended and less obvious consequences on someone's infrastructure
23:15 rcallawa joined #puppet-openstack
23:15 openstackgerrit David Moreau Simard proposed stackforge/puppet-ceph: Increase flexibility with managing client keys in profile  https://review.openstack.org/159950
23:16 mgagne wth... https://github.com/puppetlabs/puppetlabs-concat/blob/master/files/concatfragments.rb#L112
23:16 mgagne none of your business mister puppetlabs-concat
23:20 aimon joined #puppet-openstack
23:20 openstackgerrit David Moreau Simard proposed stackforge/puppet-ceph: Increase flexibility with managing client keys in profile  https://review.openstack.org/159950
23:28 openstackgerrit Merged stackforge/puppet-glance: Support identity_uri in provider  https://review.openstack.org/160845
23:40 rcallawa joined #puppet-openstack
23:42 rcallawa joined #puppet-openstack
23:46 gildub joined #puppet-openstack
23:50 openstackgerrit David Moreau Simard proposed stackforge/puppet-ceph: Increase flexibility with managing client keys in profile  https://review.openstack.org/159950

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