Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2013-08-13

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

All times shown according to UTC.

Time Nick Message
00:03 otherwiseguy joined #puppet-openstack
00:53 blkperl bodepd: gah... zones can't be strigns
00:53 blkperl that was not clear by that stupid err msg
01:30 bodepd blkperl: feel free to add validation to ensure no one gets bit by it in the future
01:39 xingchao joined #puppet-openstack
02:08 openstackgerrit A change was merged to stackforge/puppet-nova: Remove nova_volume  https://review.openstack.org/41170
02:09 openstackgerrit A change was merged to stackforge/puppet-openstack: Added force_config_drive param for compute  https://review.openstack.org/41173
02:29 xingchao_ joined #puppet-openstack
02:57 xingchao joined #puppet-openstack
05:47 openstackgerrit William Van Hevelingen proposed a change to stackforge/puppet-openstack: Pass swift variables to keystone in openstack::all  https://review.openstack.org/41577
06:07 morazi joined #puppet-openstack
06:22 openstackgerrit William Van Hevelingen proposed a change to stackforge/puppet-swift: Validate that swift zones are integers.  https://review.openstack.org/41607
06:23 openstackgerrit William Van Hevelingen proposed a change to stackforge/puppet-swift: Validate that swift zones are integers.  https://review.openstack.org/41607
07:07 puppet-razor joined #puppet-openstack
07:11 puppet-razor left #puppet-openstack
07:23 EmilienM good morning here
07:30 bodepd EmilienM: good morning.
07:31 EmilienM the challenge of today for me: send a first patchset of nova-cells support in puppet-nova
07:33 * bodepd reads about cells
07:33 EmilienM bodepd: http://docs.openstack.org/trunk/openstack-compute/admin/content/ch_cells.html
07:33 bodepd I'm reading the blueprint atm
07:35 * EmilienM can feel that won't be easy
07:36 bodepd EmilienM: I'm working on a replacement for puppet-openstack
07:37 bodepd EmilienM: I would be curious about your opinion of it when it gets a bit more settled
07:37 EmilienM bodepd: nice, could you add me as reviewer when it's done ?
07:37 bodepd Sure, it's more in teh review stage.
07:37 bodepd I'm also not sure if it will become puppet-openstack or something else.
07:37 bodepd but it's intended to be flexible enough to be used by everyone (unlike puppet-openstack)
07:38 bodepd it uses hiera and assumes Puppet 3.2
07:38 bodepd as it gets a little closer, I would be interested in:
07:38 bodepd 1. if it can replace whatver composition layer you built
07:38 bodepd 2. if it works for the multi-cell mode
07:38 bodepd model
07:38 bodepd 3. What you think of the hiera based external data stuff
07:39 EmilienM bodepd: imho, it's going to be hard to manage both puppet-openstack + regular modules (puppet-nova, etc)
07:40 EmilienM bodepd: with puppet-openstack, I think there won't be a lot of people who will use multi-cell
07:41 EmilienM for the 3) I don't have thoughts yet
07:42 EmilienM bodepd: iirc, puppet-openstack is more for testing, discovering, right ?
07:42 bodepd EmilienM: that is what is wound up becoming b/c it was too hard to make it flexible enough for everyone
07:42 bodepd EmilienM: I would have liked it to become what everyone uses, but it did not happen
07:43 bodepd EmilienM: I'm glad your a doubter, if I can convince you, I can convince everyone :)
07:44 EmilienM bodepd: :)
07:44 EmilienM bodepd: from my side, I won't make any patch in puppet-openstack
07:44 EmilienM bodepd: 1) I have no time for this module
07:44 EmilienM bodepd: 2) we don't use it in production
07:44 EmilienM bodepd: 3) again, imho, it's useless to maintain twice same things.
07:45 bodepd EmilienM: I know. most people don't use it, and I think that is a huge missed oppurtunity for calloboration
07:45 EmilienM bodepd: but I don't mind to make support on it, we need to educate the community to use other modules
07:45 EmilienM bodepd: you know, it's like DevStack
07:45 bodepd EmilienM: I want to kill puppet-openstack, call it a failure, and try to build something that everyone can actually use
07:45 EmilienM bodepd: +1
07:46 bodepd the thing I'm working on is very different:
07:46 bodepd https://github.com/bodepd/kickstack/blob/master/manifests/cinder/all.pp
07:46 EmilienM bodepd: for that, we need to write a documentation, and release some manifests
07:46 bodepd a little snippet
07:46 bodepd (it's mostly not ready)
07:46 EmilienM yeah I know kickstack, it's really cool
07:46 bodepd but the idea, is to make it so that you can grab the bits you want, compose them how you wnt
07:46 bodepd EmilienM: this kind of was kickstack
07:46 xingchao bodepd: Hi~ do you mean we will build a tool like devstack or packstack?
07:47 bodepd xingchao: yes
07:47 bodepd not devstack
07:47 bodepd and for packstack
07:47 bodepd I would say that is two seperate things
07:47 bodepd the thing I am working on is a replacement to the templated top level manifests
07:47 bodepd (they don't use puppet-openstack either)
07:48 xingchao wow, but it's a tool make people feel easy to use?
07:48 bodepd I would see the installer being somethign that sits on top of this
07:48 * blkperl thinks he is starting to understand how kickstart is differnet than puppet-openstack now :)
07:48 EmilienM we should use puppet-{nova,keystone,cinder,...} directly
07:48 bodepd blkperl: it's all hiera'ed out
07:48 blkperl I see that :)
07:48 EmilienM kickstack use them ?
07:48 bodepd use what?
07:48 EmilienM the modules
07:48 bodepd yes
07:48 blkperl I hiera'd out my wrappers around puppet-openstack
07:48 EmilienM that's very cool
07:48 EmilienM that what we miss
07:48 EmilienM a top level stuff
07:49 bodepd blkperl: did you see dev_env?
07:49 blkperl no?
07:49 blkperl what is it?
07:49 bodepd blkperl: https://github.com/stackforge/puppet-openstack_dev_env/blob/master/manifests/site.pp#L13
07:49 EmilienM bodepd: do we need to run more tests locally (more than rake) since mgagne added some stuffs in gate ?
07:50 bodepd blkperl: it's what I was using for testing, but I am not maintaining it atm
07:50 bodepd bodepd: it should be rake lint
07:50 EmilienM ok
07:50 bodepd (and it should be fast :) )
07:50 EmilienM bodepd: did you run it on puppet-nova ?
07:51 bodepd rake line?
07:51 bodepd lint?
07:51 bodepd no. I just just guessing.
07:51 bodepd what do you see when you do: rake -T ?
07:51 bodepd EmilienM: ^
07:52 bodepd blkperl: would you be interested in trying out my fork of kickstack
07:52 bodepd blkperl: it absolutely does not work
07:52 EmilienM xps  puppet-nova git:(cells-support) ✗ rake -T lint
07:52 EmilienM rake lint  # Check puppet manifests with puppet-lint
07:52 bodepd blkperl: has no tets, and has never been run before.
07:52 bodepd blkperl: but other than that, it should be fine
07:52 blkperl bodepd: haha your selling this so well :)
07:52 bodepd blkperl: it will be better :)
07:53 bodepd I promise
07:53 EmilienM bodepd: are you a sale guy by the way ? ahah
07:53 bodepd EmilienM: just a realist
07:54 bodepd blkperl: It will probably have tests and start working soon
07:55 blkperl :)
07:56 blkperl bodepd: so how do you deal with removing swift exported resources after they enter puppetdb? I had to write some sql to drop from puppetdb... which kinda sucks
08:00 bodepd blkperl: yeah, that sucks
08:00 bodepd blkperl: I have some functions that make everythign as inactive in puppedb
08:01 bodepd blkperl: and wipe out certificates
08:01 bodepd blkperl: it's a real pain
08:03 Gwyxx joined #puppet-openstack
08:04 xingchao blkperl: bodepd: I suffered too, so it's not suitable to use exported resource in managing swift ring, currently it just could add nodes or devices, but can't reduce easily
08:05 bodepd xingchao: ah.
08:05 bodepd xingchao: we could create that tooling somehow
08:05 bodepd xingchao: I always wanted to add mcollective in the mix to help orchestrate swift tasks
08:05 bodepd xingchao: but I've suffered from lack of time, and lack of real-life use cases
08:06 blkperl or the devs could fix puppetdb to properly support removing nodes :)
08:08 blkperl http://projects.puppetlabs.com/issues/14608
08:08 xingchao bodepd: yes, we could improve puppet-swift :)
08:09 xingchao bodepd: You mean adding mco in managing swift tasks, could you give me some example ?
08:11 EmilienM bodepd: how to fix the lint error ? >
08:11 EmilienM Evaluating manifests/keystone/auth.pp
08:11 EmilienM ERROR: neutron::keystone::auth not in autoload module layout on line 1
08:12 bodepd xingchao: I don't have a really good example. I wanted to take some time to think about it.
08:13 bodepd xingchao: somethign like: mcollective runs and checks the current state of the disks
08:13 bodepd xingchao: in the case of failure or an absent disk, it can mark things as expired in puppet db
08:13 bodepd (sorry, it's called INACTIVE)
08:14 bodepd blkperl: ah, I have some code that does that.
08:14 bodepd hold on...
08:14 blkperl bodepd: I have a script that blows away everything... no worries
08:15 blkperl I suppose I should add it the Exported resource chapter in pro-puppet.....
08:15 bodepd I have a script, where you can run queries
08:15 bodepd and use those queries to select the thigns to desctivate
08:16 blkperl bodepd: https://gist.github.com/blkperl/6218966
08:16 blkperl does it look like that?
08:19 bodepd https://gist.github.com/bodepd/6218998
08:20 bodepd blkperl: that desctivates everything, but you can be more choosey
08:27 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-glance: Clean all warnings for puppet-lint gate  https://review.openstack.org/41621
09:27 toMeloos_ joined #puppet-openstack
09:30 toMeloos_ joined #puppet-openstack
12:52 dmsimard joined #puppet-openstack
12:56 prad_ joined #puppet-openstack
12:59 openstackgerrit Francois Deppierraz proposed a change to stackforge/puppet-cinder: Creates an empty init script override file if necessary  https://review.openstack.org/41275
13:07 dprince joined #puppet-openstack
13:16 prad_ joined #puppet-openstack
13:20 morazi joined #puppet-openstack
13:22 mjblack joined #puppet-openstack
13:24 mjblack joined #puppet-openstack
13:39 dmsimard1 joined #puppet-openstack
14:17 otherwiseguy joined #puppet-openstack
14:30 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-ceilometer: Clean all warnings for puppet-lint gate  https://review.openstack.org/41699
14:30 sileht 28
14:34 openstackgerrit A change was merged to stackforge/puppet-ceilometer: Use Puppet boolean for better consistency  https://review.openstack.org/41184
14:38 EmilienM fc__: hey :)
14:59 technolo-g joined #puppet-openstack
15:05 EmilienM mgagne: Hello
15:05 mgagne EmilienM: hello
15:05 EmilienM mgagne: I started to clean modules for lint
15:05 mgagne EmilienM: cool :D
15:08 mgagne EmilienM: unless we care about adding a full documentation to all classes and defines, a # before the class/define definition is enough to satisfy lint
15:08 EmilienM mgagne: ack
15:13 openstackgerrit A change was merged to stackforge/puppet-ceilometer: Clean all warnings for puppet-lint gate  https://review.openstack.org/41699
15:38 bkero joined #puppet-openstack
16:27 hogepodge joined #puppet-openstack
16:41 britthouser joined #puppet-openstack
16:42 britthouser joined #puppet-openstack
16:51 openstackgerrit Logan McNaughton proposed a change to stackforge/puppet-nova: Fixes some puppet lint warnings  https://review.openstack.org/41732
16:52 openstackgerrit Logan McNaughton proposed a change to stackforge/puppet-nova: Fixes some puppet lint warnings  https://review.openstack.org/41732
16:54 openstackgerrit Logan McNaughton proposed a change to stackforge/puppet-nova: Fixes some puppet lint warnings  https://review.openstack.org/41732
17:08 openstackgerrit Logan McNaughton proposed a change to stackforge/puppet-nova: Fixes some puppet lint warnings  https://review.openstack.org/41732
17:09 openstackgerrit joined #puppet-openstack
17:18 openstackgerrit Logan McNaughton proposed a change to stackforge/puppet-nova: Fixes some puppet lint warnings  https://review.openstack.org/41732
17:20 openstackgerrit A change was merged to stackforge/puppet-openstack: Expose nova security_group_api parameter  https://review.openstack.org/41112
17:20 openstackgerrit A change was merged to stackforge/puppet-openstack: Expose nova security_group_api parameter  https://review.openstack.org/41117
17:24 openstackgerrit A change was merged to stackforge/puppet-quantum: Add missing quantum-plugin-cisco package resource  https://review.openstack.org/41438
17:46 openstackgerrit A change was merged to stackforge/puppet-cinder: Creates an empty init script override file if necessary  https://review.openstack.org/41275
17:50 openstackgerrit Logan McNaughton proposed a change to stackforge/puppet-nova: Fixes some puppet lint warnings  https://review.openstack.org/41732
17:52 bodepd mgagne: https://review.openstack.org/#/c/41732
17:53 bodepd mgagne: since you are puppet-lint man, I wanted to run that by you before I merge it ;)
17:54 mgagne bodepd: not sure about that warning in nova::migration::libvirt. Either the wording is wrong (IMO) or this could be converted to an if instead
17:55 mgagne bodepd: I'm ok with disable_class_inherits_from_params_class
18:07 mgagne is glance retarded or what?
18:08 mgagne --os-endpoint-type OS_ENDPOINT_TYPE
18:08 mgagne Defaults to env[OS_ENDPOINT_TYPE]
18:08 mgagne every other services use NOVA_ENDPOINT_TYPE
18:25 bodepd mgagne: when you get a second, can you have a look at Dan's patches
18:26 bodepd mgagne: they are all the same, actually, the conversation is in the latest one
18:26 bodepd mgagne: https://review.openstack.org/#/c/40958/
18:26 bodepd mgagne: I really don't like the idea of the public auth endpoint not defaulting to the internal one
19:22 openstackgerrit Daneyon Hansen proposed a change to stackforge/puppet-keystone: Adds Support for Managing Token Driver Backend  https://review.openstack.org/41771
19:25 JonnyNomad joined #puppet-openstack
19:28 openstackgerrit Daneyon Hansen proposed a change to stackforge/puppet-keystone: Adds Support for Managing ithe Token Driver Backend  https://review.openstack.org/41771
19:32 openstackgerrit Daneyon Hansen proposed a change to stackforge/puppet-keystone: Adds Support for Managing the Token Driver Backend  https://review.openstack.org/41771
19:35 otherwiseguy joined #puppet-openstack
19:40 openstackgerrit Daneyon Hansen proposed a change to stackforge/puppet-keystone: Adds Support for Managing the Token Driver Backend  https://review.openstack.org/41771
19:41 openstackgerrit A change was merged to stackforge/puppet-nova: Fixes some puppet lint warnings  https://review.openstack.org/41732
19:42 openstackgerrit A change was merged to stackforge/puppet-glance: Clean all warnings for puppet-lint gate  https://review.openstack.org/41621
19:45 openstackgerrit A change was merged to stackforge/puppet-nova: updated fixtures to correct postgresql link  https://review.openstack.org/41574
19:49 openstackgerrit A change was merged to stackforge/puppet-openstack: Removes un-needed cinder parameter  https://review.openstack.org/41180
20:10 openstackgerrit Daneyon Hansen proposed a change to stackforge/puppet-keystone: Adds Support for Managing the Token Driver Backend  https://review.openstack.org/41771
20:31 openstackgerrit Daneyon Hansen proposed a change to stackforge/puppet-keystone: Adds Support for Managing the Token Driver Backend  https://review.openstack.org/41784
20:47 hogepodge joined #puppet-openstack
21:54 bodepd mgagne: I'm glad you responded to that.
21:54 bodepd mgagne: I'm even thinking that this is a bug in core.
21:54 mgagne bodepd: it is a bug, there is a note
21:54 bodepd mgagne: IMHO, the change they proposed makes things worse, not better
21:54 bodepd mgagne: link?
21:54 bodepd mgagne: I do love that we are getting involved with master and
21:55 bodepd mgagne: thinking about the deployment implications of changes as they happen!!!
21:55 mgagne bodepd: https://github.com/openstack/python-keystoneclient/commit/734ef6eb83a9d9d49c181018ba63e45dbcfa1c90#L0R365
21:56 mgagne bodepd: the option descriptions were wrong before
21:56 bodepd mgagne: yeah. I understand that this is the motivation for the change
21:56 bodepd mgagne: but I think I disagree with the warnings
21:56 mgagne bodepd: you cannot authenticate using the admin endpoint
21:57 bodepd mgagne: can the service tenant not?
21:57 bodepd mgagne: then how would anything have ever worked?
21:58 mgagne bodepd: maybe it never worked?
21:58 mgagne bodepd: but I do agree that comment does not communicate well that the admin endpoint will be configured if auth_uri is not defined and it's not a "good" value for authentication to work
21:58 bodepd mgagne: I have stood up many services using authtoken
21:58 bodepd mgagne: I still feel like I'm missing something...
21:59 bodepd mgagne: when does admin not work for authentication?
21:59 bodepd mgagne: I thought you could authenticate with admin, but that admin was used for the serviecs catalog
21:59 mgagne bodepd: always?
21:59 * bodepd reads docs
21:59 mgagne bodepd: were you configuring authtoken with puppet or by hand?
21:59 bodepd puppet
22:00 bodepd and I think I am usually only setting auth_host
22:03 bodepd mgagne: take the example of cinder. we were not confiugring authtoken/auth_uri
22:03 bodepd mgagne: and I know for a fact it was working
22:03 mgagne bodepd: I'll have to test. Could be that only user with admin role can authenticate on the admin endpoint.
22:21 EmilienM mgagne: I'm stalling on puppet-nova for cells support. I hope make progress tomorrow
22:21 mgagne EmilienM: =)
22:22 EmilienM mgagne: I have some tests which are failing, is it a problem if I push the commit and set WIP ?
22:22 mgagne EmlienM: no problem with that, I think some people are already levering the CI infra to run more extended tests they can't run locally.
22:23 EmilienM mgagne: actually, I can run them, but they fail and I don't know why.
22:23 EmilienM maybe need I some help at some points
22:23 mgagne EmilienM: submit and we will see
22:23 EmilienM cool
22:24 EmilienM mgagne: tomorrow morning, because I don't like to see "-1" from mgagne when I wake up :P
22:24 mgagne EmilienM: Jenkins will vote first =)
22:25 EmilienM :)
22:25 EmilienM it's late here, cheers :)
22:25 mgagne cya
22:35 blkperl anyone have an example for rspec validate_re from stdlib?
22:38 mgagne blkperl: Is this example working? https://review.openstack.org/#/c/41607/2/manifests/storage/node.pp
22:39 mgagne blkperl: oh
22:39 blkperl mgagne: bodepd asked me to add an rspec test for it
22:39 mgagne blkperl: hold on
22:39 blkperl so far I have a test but it always returns true :)
22:39 mgagne it should throw a puppet error
22:39 blkperl yeah.....
22:39 blkperl not sure whats up....
22:40 mgagne spec/classes/ceilometer_api_spec.rb:            raise_error(Puppet::Error, /validate_re\(\): "#{auth_admin_prefix}" does not match/) }
22:40 blkperl yeah I have raise_error(Puppet::Error, /validate_re\(\): The $zone value must be a number/)
22:40 blkperl but its not working
22:41 mgagne blkperl: how is it not working?
22:41 blkperl if I change it to a number it works
22:42 blkperl which means its not raising an error....
22:43 mgagne blkperl: you mean a string
22:45 blkperl err within the context of the test if I set zone => 1 it should fail because a puppet error was not raised
22:46 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-horizon: Clean all warnings for puppet-lint gate  https://review.openstack.org/41807
22:49 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-keystone: Clean all warnings for puppet-lint gate  https://review.openstack.org/41808
22:50 openstackgerrit A change was merged to stackforge/puppet-horizon: Clean all warnings for puppet-lint gate  https://review.openstack.org/41807
22:50 openstackgerrit William Van Hevelingen proposed a change to stackforge/puppet-swift: Validate that swift zones are integers.  https://review.openstack.org/41607
22:51 blkperl mgagne: I pushed my test, can you look at it?
22:51 mgagne blkperl: sure, was already testing on my side
22:54 mgagne blkperl: first, raise_error is a helper only available in some modules
22:55 mgagne blkperl: https://github.com/stackforge/puppet-quantum/blob/master/spec/shared_examples.rb
22:55 mgagne blkperl: second, you will have to (re)define params in your context with the "wrong" value
22:55 openstackgerrit A change was merged to stackforge/puppet-keystone: Clean all warnings for puppet-lint gate  https://review.openstack.org/41808
22:58 blkperl mgagne: is let :zone do 'foo' end not enough to redefine?
22:59 mgagne blkperl: rspec expects params to be defined with the actual parameters to test
22:59 mgagne blkperl: :zone won't be used by the test
23:00 mgagne blkperl: hold on, this could be the worst test file =)
23:00 mgagne blkperl: a lot is missing or misorganized
23:00 blkperl mgagne: yeah... I thought it was a bit strange.... :)
23:02 blkperl ok #1 is working... I think my test works now
23:04 blkperl well nope....backtraces of something
23:05 mgagne blkperl: working on it
23:05 blkperl thanks!
23:08 mgagne blkperl: $ is causing problem
23:08 mgagne blperl: I would change the wording to better fit what is already used elsewhere
23:09 mgagne blkperl: like "zone parameter must be an integer"
23:09 blkperl thanks!
23:09 mgagne blkperl: also ...
23:10 mgagne blkperl: https://gist.github.com/mgagne/6226631
23:17 blkperl expected Puppet::Error with "The zone parameter must be an integer", got #<Puppet::Error: The zone parameter must be an integer
23:17 blkperl is that not the same thing?
23:17 blkperl mgagne: ^
23:18 blkperl nvm
23:19 openstackgerrit William Van Hevelingen proposed a change to stackforge/puppet-swift: Validate that swift zones are integers.  https://review.openstack.org/41607
23:20 blkperl mgagne++ thanks for the help
23:21 mgagne np
23:22 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-nova: Clean all warnings for puppet-lint gate  https://review.openstack.org/41811
23:23 xingchao joined #puppet-openstack
23:47 openstackgerrit A change was merged to stackforge/puppet-swift: Validate that swift zones are integers.  https://review.openstack.org/41607
23:49 blkperl woot
23:52 hogepodge joined #puppet-openstack

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