Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2013-07-15

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

All times shown according to UTC.

Time Nick Message
00:22 openstackgerrit Michael Chapman proposed a change to stackforge/puppet-quantum: Support Cisco quantum plugin (WORKINPROGRESS)  https://review.openstack.org/36830
00:33 openstackgerrit Michael Chapman proposed a change to stackforge/puppet-quantum: Support Cisco quantum plugin (WORKINPROGRESS)  https://review.openstack.org/36830
00:39 openstackgerrit joined #puppet-openstack
07:31 openstackgerrit Jens-Christian Fischer proposed a change to stackforge/puppet-openstack: Add rbd option to cinder configuration  https://review.openstack.org/36222
08:02 openstackgerrit Jens-Christian Fischer proposed a change to stackforge/puppet-openstack: Add rbd option to cinder configuration  https://review.openstack.org/36222
08:05 giulivo joined #puppet-openstack
08:24 red_trela joined #puppet-openstack
08:25 derekh joined #puppet-openstack
12:02 prad_ joined #puppet-openstack
12:30 dprince joined #puppet-openstack
12:36 openstackgerrit Jens-Christian Fischer proposed a change to stackforge/puppet-keystone: Select keystone endpoint url bases on SSL setting  https://review.openstack.org/37048
13:45 dmsimard joined #puppet-openstack
14:33 derekh joined #puppet-openstack
14:47 openstackgerrit Daneyon Hansen proposed a change to stackforge/puppet-openstack: Add Support for Quantum OVS VLAN Networking  https://review.openstack.org/36913
14:50 michchap joined #puppet-openstack
15:27 michchap joined #puppet-openstack
15:32 mgagne1 joined #puppet-openstack
16:04 michchap joined #puppet-openstack
16:14 michchap_ joined #puppet-openstack
16:16 michchap joined #puppet-openstack
16:19 openstackgerrit Daneyon Hansen proposed a change to stackforge/puppet-openstack: Add support for SQL idle timeout.  https://review.openstack.org/37083
16:20 michchap_ joined #puppet-openstack
16:20 otherwiseguy joined #puppet-openstack
16:28 prad_ joined #puppet-openstack
16:40 mgagne joined #puppet-openstack
16:40 mgagne joined #puppet-openstack
16:53 michchap joined #puppet-openstack
17:30 michchap joined #puppet-openstack
17:41 dontalton joined #puppet-openstack
17:50 blkperl odyi: so are you working on fixes for openstack tests directory  and readme fixes or should I start submitting some PRs to fix some of the things I've found
17:50 blkperl i've got all.pp running from start to finish now
18:09 michchap joined #puppet-openstack
18:28 openstackgerrit A change was merged to stackforge/puppet-horizon: Update local_settings.py to latest version  https://review.openstack.org/35014
18:47 michchap joined #puppet-openstack
19:25 michchap joined #puppet-openstack
19:27 otherwiseguy bodepd: mgagne: We're (finally) very close to getting puppet 2.7 or 3.1 added to RHOS/RDO. With that said, any problem with me submitting a couple of fixes for 2.6 compatibility in puppet-{openstack,tempest} to bridge the gap for the next couple of weeks while packages make their way through testing/approval processes?
19:28 mgagne otherwiseguy: mostly trailing commas?
19:28 otherwiseguy mgagne: trailing commas and it looks like a weird issue where passing a literal array to ensure_packages (instead of a variable containing said array) for some reason results in a syntax error.
19:29 mgagne otherwiseguy: I see
19:29 otherwiseguy Also, are there significant backwards-compatibility issues that you know of for moving to 3.1 as opposed to 2.7?
19:29 mgagne otherwiseguy: tbh, I don't care as long as it does not break anything and does not introduce major changes
19:30 mgagne otherwiseguy: off the top of my head, I'm not aware of any
19:30 JoeJulian Is there a [sub]module, yet, that ensures that a vm is active under openstack?
19:31 otherwiseguy Looks like we may try to do something like 3.1 for the "layered product common" repos for RHOS/RDO and maybe leave EPEL @ 2.7 since backwards-compatibility is a big thing there, but we'd like to have the same version everywhere if it doesn't cause any problems.
19:31 otherwiseguy But updating to something that is already basically EOL seems kind of silly to me.
19:32 JoeJulian otherwiseguy: There's been so many backward compatibility breakages in epel in the last year, I'm not sure sure you should care anymore.
19:32 JoeJulian s/sure sure/so sure/
19:33 otherwiseguy *I* don't care. But whoever is maintaining the EPEL puppet package seems to care. :p
19:33 comptona_ I would prefer that 2.7 compatibility is preserved if it's possible to do so
19:33 comptona my non-hypervisor machines in different roles are using 2.7 and I'd like to be able to use the same puppet version everywhere
20:02 openstackgerrit Daneyon Hansen proposed a change to stackforge/puppet-openstack: Adds Support for Separating Memcached from Horizon  https://review.openstack.org/37117
20:04 michchap joined #puppet-openstack
20:04 openstackgerrit Daneyon Hansen proposed a change to stackforge/puppet-openstack: Adds Support for Separating Memcached from Horizon  https://review.openstack.org/37117
20:33 otherwiseguy joined #puppet-openstack
20:37 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-swift: Add ini_setting based custom types for configs  https://review.openstack.org/35973
20:39 mgagne bodepd: ping
20:41 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-swift: Add ini_setting based custom types for configs  https://review.openstack.org/37125
20:42 michchap joined #puppet-openstack
21:02 odyi blkperl: Haven't got to them yet so by all means send patches..
21:07 openstackgerrit Terry Wilson proposed a change to stackforge/puppet-openstack: Remove trailing comma for Puppet 2.6 compat  https://review.openstack.org/37132
21:08 openstackgerrit Terry Wilson proposed a change to stackforge/puppet-tempest: Remove trailing comma for Puppet 2.6 compat  https://review.openstack.org/37133
21:09 otherwiseguy mgagne: for some reason the previous error I saw with ensure_packages on 2.6 disappeared. So I'll chalk that up to lack of sleep or something. So just comma fixes for 2.6 compat.
21:09 mgagne otherwiseguy: check your ruby version
21:09 otherwiseguy 1.8.7
21:10 mgagne otherwiseguy: Could the comma issue be related to ruby and not puppet itself?
21:10 otherwiseguy mgagne: I don't think so, because the problem went away when I manually installed puppet 2.7
21:10 mgagne :-/
21:11 openstackgerrit Daneyon Hansen proposed a change to stackforge/puppet-nova: Add Support for Rabbit Mirrored (HA) Queues  https://review.openstack.org/36714
21:11 otherwiseguy (for the trailing comma thing). Not sure what the deal was w/ the array literal.
21:12 otherwiseguy In any case, yay for simple patches.
21:12 otherwiseguy And more yay for finally getting close to us moving off of 2.6.
21:12 mgagne :D
21:21 michchap joined #puppet-openstack
21:24 openstackgerrit A change was merged to stackforge/puppet-tempest: Remove trailing comma for Puppet 2.6 compat  https://review.openstack.org/37133
21:29 openstackgerrit A change was merged to stackforge/puppet-openstack: Remove trailing comma for Puppet 2.6 compat  https://review.openstack.org/37132
21:33 bodepd mgagne: pong
21:34 mgagne bodepd: forgot what I wanted to ask =(
21:34 bodepd mgagne: I would appreciate if you could respond to that email thread
21:34 bodepd (but you already knew that)
21:35 mgagne bodepd: I saw the thread, I don't have anything to propose :-/
21:35 mgagne bodepd: but I do know it'a a problem
21:35 bodepd just comment. I know you have opinions
21:35 bodepd I also don't know what the best solution is
21:35 bodepd in my mind, it's the number one thing that has to be solved
21:36 bodepd maybe, we have to totally rewrite the openstack module to use roles/profildes?
21:36 bodepd mgagne: anyways, we will let you know once we decide so that you can enforce that everyone is consistent :)
21:36 mgagne bodepd: I think that people expect the module to answer all needs/combinaison without having to write a single manifest
21:37 bodepd Cisco wants the module to support N number of deployment scenarios (that are in the process of being defined)
21:37 bodepd and I want to ensure that we solve the general issue of plugability as we support those deployments
21:37 bodepd and we/they are willing to write the code :)
21:38 bodepd no question about that
21:38 mgagne bodepd: problem is that you are not alone and that RedHat will want to support N scenarios too
21:38 bodepd absolutely agree.
21:38 bodepd hence the thread
21:38 bodepd to ensure we design things in a way where we can support our configurations:
21:38 bodepd 1. without making the simplest case more complicated
21:39 bodepd 2. with a pattern that can be generally applied to solve the redhat deployment cases
21:39 bodepd mgagne: the answer maybe that openstack is only for the simplest case and we write a comple ENC that just composes the classes that we need
21:39 bodepd from the non-openstack modules
21:40 bodepd I am totally ok with that as a solution, but it concerns me b/c then people are only sharing 70% of code and not 90%
21:40 bodepd so, I am going to try to define a workable model
21:40 bodepd until I really feel it's impossible
21:41 JoeJulian bodepd: I'm thinking of making a type/provider for a openstack host (unless one already exists that I'm not finding). Any thoughts?
21:41 JoeJulian host = vm
21:41 bodepd another possibility is that I keep working on Cisco/puppet-coi and eventually it replaces puppet-openstack
21:41 bodepd (using roles/profiles and hiera)
21:42 bodepd JoeJulian:two ways to do this.
21:42 bodepd CLI or API
21:42 bodepd everyting so far has ben CLI and we should move to an API mode (probably using ruby-fog)
21:42 bodepd other than that, it needs to be done
21:42 JoeJulian api it is, then.
21:43 bodepd in fact, one of the cisco plugin models we are working on deploys it's own VM as a router
21:43 bodepd JoeJulian: I woudl suggest this is a new module
21:43 bodepd and that eventually all of the keystone/glance/quantum native types are refactored to follow whatever patterns you decide upon
21:43 bodepd JoeJulian: also, you may want to look at thisL
21:43 bodepd https://github.com/bodepd/cloudstack_resources
21:44 bodepd it's basically the same thing except for cloudstack
21:44 JoeJulian That's going to make it way too easy....
21:56 openstackgerrit Daneyon Hansen proposed a change to stackforge/puppet-openstack: Adds support for Nova Memcached  https://review.openstack.org/37143
21:57 mgagne bodepd: .
21:59 michchap joined #puppet-openstack
22:13 bodepd mgagne: not sure what that dot means in the context
22:14 mgagne bodepd: =)
22:14 mgagne bodepd: I don't feel comfortable reviewing all those changes that are just passthrough for underlying classes
22:15 mgagne bodepd: and where nothing is done (no logic) about it in openstack::*
22:15 bodepd mgagne: you mean stuff in openstack?
22:15 mgagne bodepd: yha
22:16 bodepd wait. you mean like the nova memcached patch?
22:16 bodepd or the subsequent patches that will have to forward it through the openstack classes
22:16 mgagne yes, all those patches adding passthrough
22:16 bodepd once it's been added to the underlying class, it has to be passed through to be sed
22:16 bodepd used
22:17 mgagne bodepd: because at that rhythm, it won't be long that controller will have all parameters from all underlying classes
22:18 bodepd mgagne: I'm not sure what to do about that.
22:18 bodepd mgagne: maybe we have to stop using the openstack class.
22:18 bodepd mgagne: and instead build our own top level model
22:19 bodepd (we have one, but it is mostly a hiera layer plugged on top of the openstack classes
22:19 mgagne bodepd: people should be able to compose their own roles using more basic blocks
22:20 bodepd the do the openstack classes have any reason for existing?
22:20 mgagne bodepd: I thought it was an example of what could be done with puppet-nova, puppet-cinder, etc.
22:21 bodepd I would like it to be more
22:21 bodepd (part of the underlying theme of that email)
22:21 bodepd and agree it can
22:22 bodepd s/can/can't/
22:22 bodepd given it's current design
22:23 mgagne bodepd: it's partially why I don't feel comfortable reviewing stuff that I see no future with
22:23 bodepd mgagne: so you think it will either die or be replaced?
22:23 mgagne bodepd: yes
22:24 mgagne bodepd: it's like that huge do_stuff() method with 485 parameters trying to cover each case
22:38 michchap joined #puppet-openstack
22:43 bodepd mgagne: what if we just start implmenting the second proposal and move plugin code out of the openstack module?
22:44 bodepd mgagne: and only leave a single 'default' plugin there
22:44 bodepd mgagne: can that save the model
22:44 bodepd ?
22:44 mgagne bodepd: I'll have to read the proposition first
22:45 bodepd mgagne: https://gist.github.com/bodepd/6004172
22:45 bodepd mgagne:  I am going to work on cleaning up that explanation a little
22:49 mgagne bodepd: I'm sorry, I don't feel I can contribute with ideas or opinions today
22:50 bodepd mgagne: no worries. feel free to ping me when it's convenient. I really value your opinion on usage patterns, etc..
23:15 michchap joined #puppet-openstack
23:55 michchap joined #puppet-openstack

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