Perl 6 - the future is here, just unevenly distributed

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

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

All times shown according to UTC.

Time Nick Message
00:02 gildub joined #puppet-openstack
00:06 aimon joined #puppet-openstack
00:17 mgagne lets say I don't care about directory environment in Puppet, how can I have a similar behavior to the previous one? https://docs.puppetlabs.com/puppet/3.6/r​eference/release_notes.html#deprecation-​config-file-environments-and-the-global-​manifestmodulepathconfigversion-settings
00:24 nibalizer mgagne: you can just use normal [environmentname] blocks or do the moduledir=/etc/puppet/$environment/modules thing
00:24 mgagne nibalizer: isn't it deprecated?
00:24 mgagne nibalizer: I'm using Puppet 3.7
00:24 xingchao joined #puppet-openstack
00:25 nibalizer deprecated doesn't mean it is gone
00:25 nibalizer just that its not reccomended any more
00:25 mgagne nibalizer: what's the point of deprecating if not to suggest users to move away from a feature?
00:26 nibalizer i guess i misunderstood the question
00:26 nibalizer i thought you were asking if you could do it in the old way
00:26 mgagne nibalizer: yes, that's the question but ...
00:26 nibalizer are you asking for help setting up directory environments to work the way the old system did/
00:27 mgagne nibalizer: Now I have a gazillion warnings at the console. the idea is that I do not use the directory environment thing. It imposes a view/opinion on how I should manage environments and it doesn't fit with my way of doing things.
00:28 mgagne nibalizer: If I could have ONE directory for all my modules and manifests, that would be great. without having to create 3-4 mostly empty folders each time I wish to introduce an environment
00:30 mgagne I found basemodulepath which looks to do what I want. will have to try that one out. but my reading indicates that if an environment is not defined (I suppose it means creating a directory), catalog compilation will fail which is not cool.
00:32 nibalizer ya, i tend to agree that directory environments seem to be opinionated for not that much value
00:38 rcallawa joined #puppet-openstack
01:04 markvoelker joined #puppet-openstack
01:25 ericpeterson joined #puppet-openstack
01:25 rmoe joined #puppet-openstack
01:27 NathanT joined #puppet-openstack
01:36 xarses joined #puppet-openstack
01:49 ericpeterson joined #puppet-openstack
02:05 ericpeterson joined #puppet-openstack
02:06 rcallawa_ joined #puppet-openstack
02:07 rcallaw__ joined #puppet-openstack
02:26 openstackgerrit Matt Fischer proposed stackforge/puppet-keystone: Enable setting the revoke/token driver.  https://review.openstack.org/160402
02:26 markvoelker joined #puppet-openstack
02:47 openstackgerrit craig delatte proposed stackforge/puppet-cinder: added HP 3par iscsi backend module  https://review.openstack.org/155035
02:49 openstackgerrit craig delatte proposed stackforge/puppet-cinder: added HP 3par iscsi backend module  https://review.openstack.org/155035
03:03 gildub joined #puppet-openstack
03:23 badiane_ka joined #puppet-openstack
03:41 topshare joined #puppet-openstack
04:35 ericpeterson joined #puppet-openstack
05:00 markvoelker joined #puppet-openstack
05:30 markvoelker joined #puppet-openstack
05:42 xingchao_ joined #puppet-openstack
06:03 topshare joined #puppet-openstack
06:06 topshare joined #puppet-openstack
06:31 markvoelker joined #puppet-openstack
06:44 topshare joined #puppet-openstack
06:48 gildub joined #puppet-openstack
06:52 imcsk8 joined #puppet-openstack
06:58 topshare joined #puppet-openstack
07:05 sbfox joined #puppet-openstack
07:21 sanjayu joined #puppet-openstack
07:22 openstackgerrit joined #puppet-openstack
07:31 rissmann joined #puppet-openstack
07:32 markvoelker joined #puppet-openstack
07:50 ByteSore hi all
07:50 ByteSore how can i enable the neutron-l3-agent for the role compute easily?
08:06 stamak joined #puppet-openstack
08:17 ericpeterson joined #puppet-openstack
08:22 AnxiousGarlic joined #puppet-openstack
08:23 AnxiousGarlic left #puppet-openstack
08:26 topshare_ joined #puppet-openstack
08:33 andriikolesnikov joined #puppet-openstack
08:33 markvoelker joined #puppet-openstack
08:36 moshele joined #puppet-openstack
08:42 ddmitriev joined #puppet-openstack
08:44 BrianShang joined #puppet-openstack
08:44 dgurtner joined #puppet-openstack
08:46 jistr joined #puppet-openstack
08:49 jpena joined #puppet-openstack
08:57 xingchao_ joined #puppet-openstack
08:58 openstackgerrit Yanis Guenane proposed stackforge/puppet-cinder: Allow one to configure multiple backend.  https://review.openstack.org/160369
09:35 markvoelker joined #puppet-openstack
09:37 dgurtner joined #puppet-openstack
09:54 xingchao_ joined #puppet-openstack
10:01 ddmitriev joined #puppet-openstack
10:35 markvoelker joined #puppet-openstack
10:40 dgurtner joined #puppet-openstack
10:40 dgurtner joined #puppet-openstack
11:36 markvoelker joined #puppet-openstack
11:49 cdent joined #puppet-openstack
11:54 xingcha__ joined #puppet-openstack
12:03 dgurtner joined #puppet-openstack
12:08 EmilienM morning!
12:25 social joined #puppet-openstack
12:37 moshele EmilienM: hi
12:38 markvoelker joined #puppet-openstack
12:57 andriikolesnikov joined #puppet-openstack
13:00 rcallawa joined #puppet-openstack
13:00 rcallawa joined #puppet-openstack
13:04 markvoelker joined #puppet-openstack
13:10 topshare joined #puppet-openstack
13:21 social I need to annoy someone with reviews
13:21 social too many in backlog
13:26 dprince joined #puppet-openstack
13:37 openstackgerrit Yanis Guenane proposed stackforge/puppet-cinder: Allow one to configure multiple backend.  https://review.openstack.org/160369
13:38 EmilienM social: i've put some comments on some of them, not sure you've seen them
13:43 social EmilienM: which ones?
13:44 * social checks mail
13:44 social EmilienM: while you were on plane I was fixing them
13:59 rcallawa joined #puppet-openstack
14:02 EmilienM social: cool, I'll look at them today, thanks
14:19 richm joined #puppet-openstack
14:20 EmilienM social: on https://review.openstack.org/#/c/159548/ - would you mind to add a test like I suggested ?
14:22 EmilienM social: also, a question on https://review.openstack.org/#/c/158299/
14:23 tfhartmann joined #puppet-openstack
14:30 openstackgerrit Merged stackforge/puppet-cinder: added HP 3par iscsi backend module  https://review.openstack.org/155035
14:35 EmilienM social: https://review.openstack.org/#/c/106144 is better
14:38 EmilienM mgagne: can you have an eye on https://review.openstack.org/#/c/106144 when you got time?
14:40 social EmilienM: https://review.openstack.org/#/c/159548/ should have test for =False
14:40 EmilienM dachary: I guess Ceph CI still has issue with Vagrant ?
14:41 social EmilienM: sorry I forgot about https://review.openstack.org/#/c/158299/
14:42 EmilienM social: indeed, my bad - +2 on https://review.openstack.org/#/c/159548/
14:43 mschmitt joined #puppet-openstack
14:43 social EmilienM: https://review.openstack.org/#/c/158613/ than nova::compute::ironic needs fixing
14:44 social EmilienM: thus we need to separate it so we could include it in nova::compute::ironic
14:44 social EmilienM: nova compute does not need to have ::ironic as that is on controller but if it's set to provide ironic stuff it needs the packages
14:45 EmilienM social: IMHO, ::ironic whould be needed on compute
14:45 EmilienM on compute, we also have ::ceilomter
14:45 EmilienM ceilometer*
14:45 EmilienM for compute agent
14:46 dgurtner joined #puppet-openstack
14:47 dachary dgurtner: will know more EmilienM
14:48 EmilienM dgurtner: hey, it seems there are still some issue with Vagrant in the CI
14:48 EmilienM dgurtner: my first patch in puppet-ceph is tough lol
14:49 dgurtner EmilienM: sadly it's never really easy with the CI
14:49 dgurtner EmilienM: but to be honest, it seems it genuinely fails
14:55 ericpeterson joined #puppet-openstack
15:10 dgurtner EmilienM: http://paste.ubuntu.com/10500988/
15:10 dgurtner EmilienM: that's a fully run through output of patchset 9 (the current one)
15:11 EmilienM dgurtner: good
15:12 dgurtner EmilienM: well, or not so much, as it showes that there are still further issues until el7 is supported
15:13 EmilienM dgurtner: thx for highlight, I'll have a look today;
15:14 arnaud_orange joined #puppet-openstack
15:14 dmsimard dgurtner, EmilienM: Yeah.. I definitely need to start reading about beaker..
15:17 prad joined #puppet-openstack
15:23 rcallawa joined #puppet-openstack
15:31 openstackgerrit Lukas Bezdicka proposed stackforge/puppet-ironic: Fix rabbitmq support  https://review.openstack.org/158299
15:36 mdorman joined #puppet-openstack
15:41 dontalton joined #puppet-openstack
15:44 openstackgerrit Javier Peña proposed stackforge/puppet-glance: Support identity_uri in provider  https://review.openstack.org/160845
15:48 xingchao joined #puppet-openstack
15:48 badiane_ka joined #puppet-openstack
15:54 rcallawa joined #puppet-openstack
15:55 rcallawa joined #puppet-openstack
16:01 crinkle joined #puppet-openstack
16:05 EmilienM crinkle: jpena: weird, I still have the issue with https://review.openstack.org/#/c/159318
16:06 EmilienM crinkle: jpena http://paste.openstack.org/​show/QFytjh5Obi6gC5qFypnS/
16:08 jpena EmilienM: that's a different issue. The command is failing, so we are not filtering any output
16:08 jpena EmilienM: I've seen this in my tests, now I need to remember the fix
16:08 EmilienM lol
16:08 jpena EmilienM: oh, yes! Which distro?
16:09 EmilienM jpena: F21 / Kilo
16:10 jpena EmilienM: ok, try yum update python-cliff
16:10 jpena EmilienM: I think the version in the RDO Juno repo should do
16:10 EmilienM jpena: I'm using jilo
16:10 EmilienM kilo*
16:10 EmilienM jpena: RDO trunk
16:11 jpena EmilienM: yep, but if you are using the delorean trunk repo you need to have RDO Juno repo available. Not all dependencies are in the current trunk yet
16:11 EmilienM jpena: do you have the .rep ?
16:11 EmilienM .repo*
16:12 jpena EmilienM: https://repos.fedorapeople.org/repos/openstack​/openstack-juno/rdo-release-juno-1.noarch.rpm
16:12 EmilienM jpena: thx
16:13 EmilienM jpena: indeed...
16:13 jpena EmilienM: Just make sure the RDO trunk repo is not called /etc/yum.repos.d/rdo-release.repo or it will be overwritten
16:13 EmilienM jpena: thx !
16:13 jpena EmilienM: np
16:14 EmilienM jpena: http://paste.openstack.org/​show/J4BlzUhTCsaJEFQd9mFR/
16:14 EmilienM richm: imcsk8 : did you also had it?
16:15 openstackgerrit Richard Megginson proposed stackforge/puppet-openstack-specs: Support Keystone v3 API in openstack puppet modules  https://review.openstack.org/150108
16:16 andriikolesnikov joined #puppet-openstack
16:17 richm EmilienM: had it?
16:17 jpena EmilienM: that's the third problem... Check the Keystone logs (probably under /var/log/httpd if you are using packstack master). If you find traces there, you will need to restart httpd. It's a packaging issue, keystone should depend on python-oslo-middleware but it did not
16:17 EmilienM jpena: no
16:17 EmilienM I'm using a VM with puppet-keystone
16:17 EmilienM I'm not using packstack
16:17 jpena EmilienM: then it will be under /var/log/keystone
16:20 jpena EmilienM: you will need to install python-oslo-middleware, then restart keystone
16:20 EmilienM jpena: already installed, thx
16:21 richm EmilienM: ah yes, the novaclient warning problem
16:22 richm I've been doing this in my vm
16:22 richm sed -i '/^warnings.warn/,/^$/d' /usr/lib/python2.7/site-packag​es/novaclient/v1_1/__init__.py
16:22 richm EmilienM: crinkle also has a patch which fixes this problem
16:27 EmilienM jpena: were you able to test keystone v3 ?
16:28 jpena EmilienM: nope, I think someone else was going to do that
16:39 tfhartmann joined #puppet-openstack
16:52 maximov joined #puppet-openstack
17:05 rmoe joined #puppet-openstack
17:14 xarses joined #puppet-openstack
17:18 jpena EmilienM: if you have a sec -> https://review.openstack.org/160464
17:22 sbfox joined #puppet-openstack
17:44 openstackgerrit David Moreau Simard proposed stackforge/puppet-ceph: Increase flexibility with managing client keys in profile  https://review.openstack.org/159950
17:45 dmsimard dachary, xarses - chatted a bit with dgurtner about this already.. would like to have your opinion on the direction in https://review.openstack.org/#/c/159950/
17:45 dmsimard EmilienM if you want to, you can too :)
17:47 dmsimard Basically the idea is to allow the profile to inject and configure additional client keys - it's otherwise pretty unflexible right now
17:49 dachary dmsimard: that sounds interesting
17:49 EmilienM jpena: +2
17:53 dmsimard dachary: I was ending up having to implement something to inject and configure client keys in my composition layer .. so I guess it's better to do it upstream :)
17:56 dachary right
17:59 xingchao joined #puppet-openstack
18:01 sbfox joined #puppet-openstack
18:02 aimon joined #puppet-openstack
18:07 xarses dmsimard: why wouldn't we want to do this?
18:09 dmsimard xarses: No reason, we probably want to. Asking if the general idea is good and if implementation makes sense. I was also wondering, considering the fact that we haven't yet cut a stable release, how to manage deprecation
18:12 xarses burn baby burn?
18:12 xarses =D
18:13 dmsimard I'm all for entirely removing the "previous" way, that'd be far simpler and cleaner
18:14 dmsimard but we have to all agree considering this would break current implementations
18:14 dmsimard we won't have that luxury once we cut a stable release
18:14 dmsimard the alternative is releasing a stable release with a deprecation built-in
18:14 dmsimard which is sort of weird but meh
18:14 dmsimard dachary: thoughts? ^
18:14 xarses hmm
18:15 xarses how do you differ in the monitor from the keys it must have, VS general client keys?
18:15 xarses with your change you would install client.images on the mon?
18:16 dmsimard If you put everything in common, yes - but a proper implementation would probably have a host-based hiera file with different keys in them
18:16 xarses also, for the mon keys which have strict requirements, do we want people to be able to shoot themselves in the foot and maybe send wrong permissions?
18:16 dmsimard e.g, I don't want to install client.admin on nova compute nodes
18:16 xarses actually, one of the operations requires it iirc
18:17 dmsimard I'm leaving the mon key handled the way it was before
18:17 dmsimard Just touching client.admin, the bootstrap keys and whatever other keys
18:17 dmsimard you don't need client.admin on the compute, cinder or glance nodes afaik
18:18 xarses ya, we ended up needing it
18:18 dmsimard That's odd, I haven't needed it for my setup
18:18 xarses lets see if we kept good bug history
18:18 dmsimard For the fuel implementation ?
18:18 xarses yes
18:19 dmsimard ok let me know.. cause it's definitely working for me with just the client.cinder key on the compute node and the client.glance key on the glance controller
18:19 dmsimard Not using fuel though :p
18:20 xarses ya, ok
18:20 tfhartmann joined #puppet-openstack
18:20 xarses ah, looks like we worked around it
18:21 dmsimard What was the bug ?
18:21 xarses i'd guess it was accessing the cinder volume since this is the permissions of the nova-compute's key
18:21 xarses "mon 'allow r' osd 'allow class-read object_prefix rbd_children, allow rwx pool=${cinder_pool}, allow rx pool=${glance_pool}, allow rwx pool=${compute_pool}'",
18:25 xarses ok, stale information on my end
18:26 xarses ok, back to the keys you moved permissions to in ceph::profiles::params::client_keys
18:28 dmsimard yeah, instead of deduplicating params like admin_key_mode.. what's next, admin_key_owner, admin_key_group ? Reminds me of the way the keystone endpoints were built from 5 different params :)
18:28 xarses do we want the keys that where part of the monitor (besides mon) client.admin client.bootstrap-osd client.bootstrap-mds to have all of these setting accesible and possibly wrong, and then rely on them later to only find that they are broken?
18:29 xarses afk
18:30 dmsimard xarses: I'm okay with that, as long as we provide documentation (ex: that examples file). Otherwise the ceph::key define provides sensible defaults already (ex: root:root, 0600)
18:34 bodepd sbadia: sorry for the slow response, let me see if I can get it done :)
18:40 bodepd sbadia: I added you , crinkle , and openstackinfra, adding EmilienM didn't work for some reason
18:40 bodepd sbadia: let me know if you need anything else
18:40 xarses dmsimard: sorry, by permissions i meant the caps, not the file permissions
18:40 pasquier-s joined #puppet-openstack
18:40 EmilienM bodepd: maybe because I'm not registred
18:41 EmilienM bodepd: thx, what's up ?
18:42 mschmitt joined #puppet-openstack
18:59 EmilienM richm: what are current blockers for https://review.openstack.org/#/c/146926 ?
18:59 EmilienM python-openstackclient?
19:04 richm EmilienM: yes
19:05 EmilienM richm: so if we merge your patch, we break master on ubuntu
19:05 richm EmilienM: yes
19:05 EmilienM cool
19:05 EmilienM good to know
19:06 richm EmilienM: and, given the fact that none of this will actually work with multiple domains in all of the services until we get keystonemiddleware 1.3 or later,
19:06 richm not to mention the fact that this will also require updates and new versions of all of the puppet-* packages for openstack modules,
19:07 richm perhaps we should just hold off on the review until closer to kilo
19:07 crinkle so we're not going to attempt to backport all this to juno any more?
19:08 EmilienM I don't think so
19:08 richm crinkle: I think it's just not possible given the effort required in other areas to make it useable
19:08 crinkle okay, I'm fine with this
19:09 richm so, I can either abandon the review, or just keep rebasing it and updating it
19:10 crinkle so since we're now targeting kilo, what if I port all the openstackclient changes in keystone into openstacklib and we can work on the v3 support on master in openstacklib?
19:10 EmilienM crinkle: +1
19:10 richm That's fine with me - are we ready to say that work on master is no longer easily backport-able to juno?
19:14 crinkle I think that's okay...
19:14 crinkle we'd have to make bugfixes changes in the keystone providers directly against stable/juno
19:15 richm ok
19:16 richm crinkle: Are you proposing to start on https://trello.com/c/htpmuhZd/12-move-​openstack-rb-into-puppet-openstacklib ?
19:17 crinkle yes
19:17 crinkle I can do that this week
19:18 EmilienM crinkle: cool, make sure you assign the card to you
19:19 crinkle should I maybe send a ml post to propose/announce this? since it'll affect workflows when contributing to the keystone module
19:19 richm yes
19:19 richm please
19:19 crinkle mmk
19:19 badiane_ka joined #puppet-openstack
19:20 openstackgerrit Richard Megginson proposed stackforge/puppet-keystone: support for keystone v3 api  https://review.openstack.org/146926
19:20 richm EmilienM: ^^^ - will that suffice?
19:20 richm In conjunction with the link to the bp
19:21 EmilienM richm: I would say yes, it's good
19:38 fvollero joined #puppet-openstack
19:43 sbfox joined #puppet-openstack
19:47 NathanT I've been having some issues with the keystone_tenant provider. Every time I try to use it it gives me an error about an endefined method 'downcase' for nil:NilClass. I've been trying to get more information on it and have not been successful. Could anyone point me in the right direction?
19:48 tfhartmann joined #puppet-openstack
19:49 mgagne NathanT: do you have the enabled property defined?
19:50 mgagne NathanT: I'm trying to figure out which variable is nil
19:53 NathanT Yea I am using it specifically throu the keystone::roles::admin class. It sets enabled to true by default.
19:53 NathanT through*
19:57 openstackgerrit Merged stackforge/puppet-nova: Remove require relationship on python  https://review.openstack.org/160464
19:57 mgagne NathanT: ok, the only other places using downcase are within the openstack provider itself
19:58 mgagne NathanT: I'm not familiar yet with the new provider but my understanding is that an openrc file can be used. Are you using one?
19:59 NathanT Yea I am, before I set that up it would fail earlier stating it couldn't do a service list or project list.
20:03 ericpeterson joined #puppet-openstack
20:06 rcallawa joined #puppet-openstack
20:23 mgagne NathanT: can you make sure there is no blank line in the file?
20:23 mgagne NathanT: at first glance, looks like blank lines aren't ignored
20:25 NathanT mgagne: I appreciate your help so far, if you are busy I understand. openrc has no blank lines.
20:25 mgagne NathanT: yes, I'm context switching a lot those days, sorry
20:25 mgagne NathanT: any line numbers in your error message?
20:26 NathanT mgagne: no problem at all, I totally understand that feeling. No line numbers unfortunately. The error just comes out like this: Error: /Stage[main]/Keystone::Roles::​Admin/Keystone_tenant[admin]: Could not evaluate: undefined method `downcase' for nil:NilClass
20:26 NathanT It does the same with any tenant.
20:27 mgagne crinkle: do you know if there is a way to tell puppet to output the whole stacktrace if a provider fails?
20:27 crinkle mgagne: add --trace to puppet agent --test or puppet apply
20:28 mgagne NathanT: ^
20:30 NathanT mgagne: crinkle yea that is a lot more information. /puppet/provider/keystone_tenant/openstack.rb lines 78, 75, 86, and 30. I can post the actual trace that is relevent to the openstack.rb if you like. I normally try not to post multiple lines to an irc though.
20:31 mgagne NathanT: http://paste.openstack.org/
20:31 NathanT mgagne http://paste.openstack.org/show/186800/
20:32 mgagne https://github.com/stackforge/puppet-keystone/b​lob/master/lib/puppet/provider/openstack.rb#L77
20:32 mgagne maybe try to run the same command as Puppet and see the output?
20:33 * mgagne tries to reverse engineer the actual command
20:33 ericpeterson joined #puppet-openstack
20:34 richm openstack project list --long
20:34 richm It's as if there is no Enabled column in the output, or it is something other than True or False
20:35 mgagne openstack project show --format shell admin
20:35 mgagne because it looks to be in the "show|create" block
20:35 ericpete_ joined #puppet-openstack
20:36 richm ?
20:36 richm That's not lib/puppet/provider/keysto​ne_tenant/openstack.rb:78
20:36 crinkle there's a couple of instances of foo.downcase, NathanT could you show the output with --debug and --trace?
20:36 richm in the juno branch, it is this:
20:36 mgagne richm: sorry, I thought you were NathanT
20:36 richm :enabled     => project[:enabled].downcase.chomp == 'true' ? true : false,
20:37 richm project[:enabled] is nil, which is very strange
20:37 mgagne richm: haha, I missed the keystone_tenant part in the file path
20:37 richm which means the Enabled column is not in the output???
20:37 mgagne looks like so
20:38 crinkle I don't know why that would happen, but we could easily mitigate it by checking it project[:enabled] is nil before operating on it
20:39 richm It shouldn't happen
20:40 NathanT crinkle mgagne richm here is the output with --debug and --trace http://paste.openstack.org/show/186803/
20:40 richm Is there some openrc setting which says "omit the Enabled column"?
20:40 richm NathanT: what is the output of /usr/bin/openstack project list --quiet --format csv --long ?
20:43 mfisch hey mgagne can I ask you a rspec question?
20:43 NathanT rchm http://paste.openstack.org/show/186804/
20:43 NathanT richm: *
20:44 richm NathanT: openstack project list --quiet --format csv --long 2> /dev/null
20:44 crinkle it looks like enabled is there
20:44 richm looks like all that other output is causing problems . . .
20:45 NathanT http://paste.openstack.org/show/186805/
20:45 crinkle it should barf trying to parse it as a csv if the warnings were causing issues
20:47 mgagne mfisch: sure
20:47 richm yeah - I don't know why it would get far enough to even try to project[:enabled].downcase
21:00 rcallawa joined #puppet-openstack
21:02 rcallawa joined #puppet-openstack
21:04 openstackgerrit Matt Fischer proposed stackforge/puppet-designate: WIP: Add virtualenv support  https://review.openstack.org/157032
21:12 kskurtveit_ joined #puppet-openstack
21:13 BrianShang_ joined #puppet-openstack
21:21 crinkle NathanT: what version of puppet-keystone are you running, or which commit are you based on?
21:21 crinkle I'm able to reproduce your problem before 24effa1e360b1eacec7fbfdfed339ba6e639f7fd
21:23 NathanT ahhh. I am before that version. I will update it now. Thank you so much. I have been trying to figure this out for the last couple days.
21:32 EmilienM here we go
21:38 xarses fancy
21:44 EmilienM joined #puppet-openstack
21:47 pasquier-s joined #puppet-openstack
21:55 NathanT crinkle thanks again, you're my hero.
22:00 badiane_ka joined #puppet-openstack
22:01 pator joined #puppet-openstack
22:02 openstackgerrit Ryan Bak proposed stackforge/puppet-monasca: Gives monasca-admin user admin permissions  https://review.openstack.org/160996
22:19 blentz joined #puppet-openstack
22:35 openstackgerrit Merged stackforge/puppet-monasca: Gives monasca-admin user admin permissions  https://review.openstack.org/160996
22:52 openstackgerrit joined #puppet-openstack
23:33 mgagne anyone found a way to run multiple rspec tests in parallel? :D
23:34 crinkle haha that would be amazing
23:36 mgagne we have multiple vagrant boxes in our env. (like ~100) and I'm testing catalog compilation for each of them using rspec-puppet. it takes forever: ~15m
23:36 mgagne so I'm looking to optimize the process
23:39 social mgagne: I'd start with how is it with ruby and the "parallel"
23:40 mgagne social: should I spawn a python process instead? =)
23:41 xarses get a vagrant host?
23:41 xarses erm
23:41 xarses get a faster vagrant host?
23:41 mgagne the tests aren't run within vagrant itself
23:41 mgagne they are run on a jenkins slave
23:42 xarses yep
23:42 mgagne we need to get testr for ruby then: https://wiki.openstack.org/wiki/Testr
23:42 mgagne xarses: the problem is that tests are made in serial
23:42 xarses you're catalog takes 15 min to compile, using puppet-master?
23:42 mgagne xarses: compilating 100 catalog takes 15m
23:43 xarses oh, ok, that sounds slightly better
23:43 mgagne xarses: I'm only compiling the catalog, not applying it. the goal is to compile it one before deploy to avoid surprise in prod
23:43 mgagne once*
23:43 mgagne because typos and missing hiera

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