Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2013-08-07

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

All times shown according to UTC.

Time Nick Message
00:17 openstackgerrit A change was merged to stackforge/puppet-heat: Update allowed_hosts conditional statement  https://review.openstack.org/40165
00:23 openstackgerrit A change was merged to stackforge/puppet-openstack: Removes Unneeded ovs_local_ip Error Message  https://review.openstack.org/39966
00:44 digicloseup left #puppet-openstack
02:10 prad joined #puppet-openstack
02:58 openstackgerrit Pradeep Kilambi proposed a change to stackforge/puppet-quantum: Update QUANTUM_PLUGIN_CONFIG for cisco plugin.  https://review.openstack.org/40491
04:40 openstackgerrit Robert Starmer proposed a change to stackforge/puppet-openstack: Add missing parameters to enable provider networks in all.pp  https://review.openstack.org/40543
04:58 openstackgerrit Terry Wilson proposed a change to stackforge/puppet-nova: Quantum to Neutron rename  https://review.openstack.org/40544
05:02 openstackgerrit Terry Wilson proposed a change to stackforge/puppet-openstack: Quantum to Neutron rename  https://review.openstack.org/40547
05:53 marun Is Chris Hoge here?
06:35 marun joined #puppet-openstack
07:35 tavasti joined #puppet-openstack
07:40 derekh joined #puppet-openstack
07:55 xingchao joined #puppet-openstack
08:05 EmilienM xingchao: thx for review :) I hope I was not too cold in the last review message. In fact, I just want to be clear in my point of view :)
08:15 xingchao EmilienM:  Your explanation in the last msg gives me a clear view. Here is hot, so don't worry about me :)
08:16 EmilienM cool :)
08:18 derekh joined #puppet-openstack
08:18 red_trela left #puppet-openstack
09:43 fvollero EmilienM: I am changing the puppet-heat pp files (and in api-cloudwatch.pp i saw something that caught my eye. Why there's heat_api_config and  heat_api_cloudwatch_config ?
09:44 EmilienM fvollero: https://review.openstack.org/#/c/39980/
09:44 EmilienM I fixed that
09:44 EmilienM but still under review :P
09:45 EmilienM fvollero: I'm stuck with API, since I found a bug : https://bugs.launchpad.net/heat/+bug/1209141
09:45 * fvollero scared
09:45 EmilienM and I think we should move verbose & debug in init.pp
09:46 EmilienM fvollero: actually, move all the common stuff in init.pp, and let specific stuffs in api.pp, api-*.pp
09:46 fvollero EmilienM: but can be also possible that someone would like to install separated machines just for api
09:46 EmilienM agree ?
09:46 fvollero EmilienM: But it was what I did in my patch
09:46 EmilienM sure, init.pp is always used
09:47 EmilienM which patch ?
09:47 fvollero EmilienM: Having the rabbit_ and qpid_ in init and access from there.
09:47 * EmilienM looks
09:47 fvollero EmilienM: Now i'm moving all those things in the classes they belong (as mgagne pointed)
09:48 EmilienM fvollero: in your patch, it's only for qpid
09:48 EmilienM no debug & verbose stuff
09:48 fvollero EmilienM: Not sure i follow.
09:48 fvollero EmilienM: Ahhhh, of course, because the patch is the qpid-support and nothing else
09:48 EmilienM yep
09:49 EmilienM fvollero: could you give me an example about "moving all those things in the classes" ?
09:49 fvollero EmilienM: so, we can work on clarify this point ?? I will send a new patchset with the rabbit and qpid setup for all the pp subclasses
09:49 EmilienM same idea I was talking about just before ? moving all the stuffs in init.pp and let in other manifests specific ?
09:50 EmilienM yes we can :)
09:50 fvollero EmilienM: Amazing
09:50 EmilienM so IMO, we should merge mine before : https://review.openstack.org/#/c/39980/
09:51 EmilienM and then merge yours with qpid support in init.pp
09:51 EmilienM fvollero: what you should do is to cherry pick my spec file for init, and amend your stuffs for qpid
09:52 EmilienM and by the way, I let a comment on : https://review.openstack.org/#/c/40188/1
09:53 fvollero the tests
09:53 fvollero ?
09:53 EmilienM fvollero: your patch miss tests
09:54 EmilienM fvollero: since I've already start tests in my patch
09:54 EmilienM fvollero: you should take it and add qpid support
09:54 EmilienM fvollero: and also drop ceilometer in engine.pp :P
09:54 fvollero EmilienM: Yes, of course!
09:54 EmilienM s/should/could
09:54 EmilienM :)
09:54 fvollero EmilienM: yeah.
09:54 EmilienM since we are stuck by the API bug > https://bugs.launchpad.net/heat/+bug/1209141
09:55 EmilienM I'm continue the work on tests (for engine)
09:55 EmilienM this afternoon I start the work on neutron module
10:04 fvollero EmilienM: I am missing something on the bug here. The config file isn't service specific ?
10:04 EmilienM fvollero: no, there is only a DEFAULT section
10:05 EmilienM fvollero: we should have for example a "CFN" section, with bind_host, etc, a "CLOUDWATCH" section with bind_host, etc
10:05 fvollero Ohh! Gotcha!
10:06 EmilienM :)
10:13 fvollero EmilienM: Keep wondering about the bug above, can you do an example and see because I thought I understood but i'm not sure
10:13 EmilienM fvollero: you have several API in Heat
10:13 EmilienM 3 actually
10:13 EmilienM currently, they use a wsgy process to run
10:13 fvollero wsgi
10:13 fvollero yepo
10:14 EmilienM in Heat, the binding options are managed by wsgi and each API check binding options to configure themselves
10:14 EmilienM the problem is now you have ONE config file
10:14 EmilienM with ONE bind_host
10:14 EmilienM with ONE bind_port
10:14 EmilienM etc
10:14 EmilienM so it's a big problem.
10:14 EmilienM we should create sections in Heat
10:14 EmilienM since we have dupplicated flags
10:15 fvollero Ok, I got it now. What suroprised me, is that i installed it on fedora, and we have just 1 config file for each service
10:15 fvollero that mean that now, they merged everything
10:16 fvollero ^^this
10:17 EmilienM fvollero: you still have older config files
10:17 EmilienM fvollero: until https://review.openstack.org/#/c/40257/
10:18 fvollero EmilienM: Yeah, but we will not support grizzly anymore? like having a trunk for it
10:18 fvollero win 55
10:19 EmilienM \o
10:19 EmilienM \o/
10:19 fvollero ahahah :)
10:19 EmilienM yeah we will
10:19 EmilienM since we will create a branch :P
10:19 EmilienM it's a pain .. :/
10:19 fvollero yeah, branch not trunk
10:19 EmilienM but mandatory
10:19 fvollero exactly
10:19 EmilienM to make it as clean as possible
10:20 fvollero so, your patch will go @ master but before this we should implement the qpid settings in grizzly as well
10:20 fvollero and branch on grizzly
10:20 fvollero grrr
10:20 fvollero grizzly branch.. .today i'm totally offset
10:21 EmilienM fvollero: no
10:21 EmilienM fvollero: 1) master 2) cherry pick to Grizzly
10:22 fvollero EmilienM: Ack!
10:22 fvollero EmilienM: so I'll start working on your patchset
10:23 EmilienM fvollero: I'm a bit confused by all the patchs on Heat. I think we should start talking with core for merging what we could
10:24 fvollero EmilienM: core heat developers ?
10:25 EmilienM fvollero: no, puppet ones
10:25 fvollero EmilienM: +1
10:25 EmilienM maybe xingchao is still here ?
10:25 fvollero EmilienM: But seems that the 'core' is in US atm :/
10:26 EmilienM actually, I don't know who is awake from core team
10:26 fvollero so, maybe sleeping
10:26 EmilienM \o/
10:26 EmilienM it's not scalable, ahah
10:26 fvollero ahhahahaha
10:26 fvollero and not distributed :)
10:26 EmilienM right
10:26 fvollero geographically distributed
10:55 EmilienM fvollero: maybe could you review https://review.openstack.org/#/c/39980/ to check everything is OK for you
10:56 fvollero EmilienM: I add three comments but i forgot to say that was +1 for me
10:56 EmilienM \o/
10:56 fvollero now i also said it's +1 :)
10:57 EmilienM yeah
10:57 EmilienM fvollero: I hope the bug will be fixed for Havana, otherwise I'll have to revert the patch on config
10:57 EmilienM anyway, it's trivial
10:58 fvollero EmilienM: Yeah, I understand.
10:58 fvollero EmilienM: Luckily is trivial.
10:59 EmilienM fvollero: we need to go ahead if we want to deliver a module ready for H, follow openstack trunk, but also take care of bugs
10:59 fvollero EmilienM: I understand, as well we need to take care of branches like grizzly
10:59 EmilienM yep
11:00 fvollero EmilienM: gonna be an hard job
11:43 prad_ joined #puppet-openstack
11:48 tavasti joined #puppet-openstack
12:29 dprince joined #puppet-openstack
12:39 jtyr joined #puppet-openstack
12:39 jtyr hi guys
12:40 jtyr I would like to try to deploy OpenStack via puppet and I'm not sure which module I should use ...
12:41 jtyr there are modules like puppet-{ceilometer,cinder,glance,horizon,...} and then there is the module puppet-openstack ...
12:42 jtyr please could you advice me which module is the best to use?
12:45 EmilienM jtyr: for testing or for production ?
13:30 xingchao jtyr: maybe you could try packstack or fuel first,they are based on puppet modules and easy for use.
13:38 jtyr EmilienM: We are in the stage of proof of concept now but we would like to bring it into the production soon ...
13:39 jtyr xingchao: I tried packstack but I would like to automate the build across multiple machines ... so I would like to deploy it directly through puppet ...
13:41 prad_ joined #puppet-openstack
13:41 EmilienM jtyr: then puppet-openstack is better for a PoC
13:42 jtyr EmilienM: we have the POC working now but would like to move to production stage
13:42 EmilienM jtyr: so you should use puppet-nova, puppet-* ...
13:42 jtyr the puppet-openstack is ussed by the packstack right?
13:43 xingchao jtyr:  Maybe this doc could help you:     https://github.com/stackforge/puppet​-openstack#beginning-with-openstack
13:44 jtyr EmilienM: this is what I thought ... thanks for confirming it
13:55 dmsimard joined #puppet-openstack
14:06 dmsimard joined #puppet-openstack
14:42 EmilienM mgagne: ping
14:51 digicloseup joined #puppet-openstack
15:02 mgagne jtyr: use puppet-openstack for your PoC and roll out your own puppet-openstack module if you feel it isn't flexible enough or propose patches ;)
15:02 mgagne EmilienM: coffee first :D
15:03 EmilienM mgagne: sure
15:03 gmi joined #puppet-openstack
15:04 mgagne alright
15:05 EmilienM mgagne: with fvollero we sorted out some issues, let me introduce them.
15:06 mgagne EmilienM: BTW, can we install bridge-utils anyway/irregardless of quantum/nova-network?
15:06 EmilienM 1) I found a bug in heat related to config files > https://bugs.launchpad.net/heat/+bug/1209141 - I'm gonna talk about that in the heat meeting tonight
15:06 EmilienM mgagne: we can, but if we use quantum it's useless.
15:07 mgagne EmilienM: if you use linuxbridge, you are lucky =)
15:07 EmilienM 2) I would need some review on https://review.openstack.org/#/c/39980/
15:07 mgagne EmilienM: I don't know Heat yet. About the bind option, can't they be prefixed? Does heat have multiple daemons binding to tcp ports?
15:08 EmilienM mgagne: they have 3 API running on differents prots
15:08 EmilienM ports
15:08 EmilienM but since we have only one config file now,
15:08 EmilienM with a DEFAULT section
15:08 EmilienM we should fix it before.
15:09 mgagne EmilienM: Although it's mostly for legally reason, cinder uses osapi_volume_listen for binding
15:09 EmilienM mgagne: and the good news is that I managed some time in my next sprint to work on puppet-neutron module.
15:10 mgagne cool :)
15:10 EmilienM mgagne: the fact is each API service uses wsgi.py for config
15:10 EmilienM mgagne: we have 2 ways
15:10 EmilienM but I'll talk about that later tonight with heat devs on #openstack-meeting
15:10 mgagne ok, I'll check your change
15:11 EmilienM it could/should be fixed for havana
15:11 EmilienM anyway
15:11 mgagne can we agree that puppet-heat is targeted at Havana?
15:11 EmilienM sure
15:11 EmilienM if somebody needs it, just raise the hand and make a backport.
15:11 mgagne EmilienM: I don't think we should target Grizzly and spend time backporting stuff to a grizzly branch
15:11 EmilienM +1
15:12 xingchao +1, puppet-heat is surely for havana
15:13 EmilienM :)
15:19 xingchao mgagne:  https://review.openstack.org/#/c/40318/ needs some discussion
15:20 mgagne xingchao: which question? =)
15:23 xingchao instead of setting env variables, does it make sense to set alias ?  For example: alias nova='nova --os-tenant-name admin --os-user admin --os-password admin --no-cache'
15:24 fvollero mgagne: yes.
15:24 mgagne xingchao: I think it would be considered a hack instead of a solution =)
15:24 fvollero mgagne: re: it's targeted for havana
15:25 fvollero mgagne: this mean that is completely unuseful my https://review.openstack.org/#/c/40188/ ?
15:26 mgagne fvollero: I feel it should only go in heat::init
15:26 openstackgerrit A change was merged to stackforge/puppet-heat: Update the module for new config file method  https://review.openstack.org/39980
15:28 fvollero mgagne: So, i just need to add the qpid on the init module. Amazing.
15:28 xingchao mgagne:  The question is if we export all variables, it will lead a error when running puppet in the same session, but if only export related keystone variables, other client cli will not work.
15:29 fvollero mgagne: I was going to send my new patch where you was asking to add rabbit and qpid to the submodules :(
15:30 EmilienM fvollero: https://review.openstack.org/#/c/39980/ merged, we can continue the tests & improvements
15:30 fvollero EmilienM: Oh. Ok.
15:31 fvollero EmilienM: let's work on this!
15:31 EmilienM mgagne: actually, in heat::init we should send rabbit, qpid, syslog
15:31 EmilienM mgagne: and after, each heat::api, heat::api-* should have its specifications
15:31 EmilienM right way ?
15:46 mgagne EmilienM: init should contain generic configs used by other services.
15:47 mgagne xingchao: it should work if only one of the authentication method is used.
15:47 mgagne xingchao: no? I used password authentication and it worked like a charm. I don't know about token based authentication.
15:49 EmilienM yep
15:49 xingchao mgagne: If I just use keystone_admin_token and keystone_admin_url, it doesn't work for nova.glance,etc
15:54 fvollero mgagne: more details on https://review.openstack.org/#/c/39689/ ?
15:56 xingchao mgagne:  For example: export SERVICE_TOKEN=admin  export SERVICE_ENDPOINT=http://192.168.1.9:35357/v2.0/   Then I run  nova list, it fails:  ERROR: You must provide a username via either --os-username or env[OS_USERNAME]
15:59 mgagne xingchao: are you telling me there's no way to generate an openrc that won't trow a warning?
15:59 mgagne throw*
16:00 xingchao mgagne: If we export all env variables, it's just a warning, if only export keystone token, it's a error.
16:01 mgagne xingchao: should we therefore reset env variables when running keystone client from puppet?
16:02 mgagne xingchao: About openrc, is OS_SERVICE_TOKEN required? I mean, can the end user do without it?
16:02 xingchao mgagne:  If they want to use keystone cli, it's needed.
16:04 xingchao mgagne: How about just adding notice in openrc ?  If they want use keystone cli ,enable that variable by hand.
16:04 mgagne xingchao: I'm able to use keystone without the token
16:05 mgagne xingchao: but I guess it would cause problem for the initial configuration if done by cli. But I guess you are using puppet anyway for that part and admin user will be created for you by puppet.
16:08 xingchao mgagne: Oh, I just try it, it works without  keystone admin_token
16:09 mgagne xingchao: I therefore suggest making keystone_admin_token optional. Shall the end user wants it, he will be able to do so but he won't be required to configure it.
16:12 xingchao mgagne: Ok, if we use puppet to create users, keystone_admin_token variable could be optional.
16:14 xingchao mgagne: We'd better add a notice in openstack::auth_file
16:14 mgagne xingchao: sure
16:14 mgagne xingchao: I'll update later today
16:14 xingchao mgagne: ok
16:16 marun joined #puppet-openstack
16:18 mgagne xingchao: I think the notice should be documented but should not trigger a puppet warning
16:20 xingchao mgagne: yes, I means to add the notice in doc :)
16:20 dprince joined #puppet-openstack
16:20 mgagne xingchao: doing that
16:28 xingchao mgagne: how to retrigger  Puppet OpenStack CI ?
16:28 mgagne xingchao: I don't have access to this system. Cisco maintains it (bodepd)
16:29 xingchao mgagne: oh, I see
16:34 fvollero EmilienM: ping. no log dir ?
16:37 EmilienM fvollero: not yet :P
16:37 EmilienM we could add it in init
16:38 fvollero EmilienM: Ok, i'll do it later, i'm adding qpid now
16:38 EmilienM cool
16:39 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-keystone: release "Updated README.md for 2.1.0 release"  https://review.openstack.org/40508
16:39 blentz is there a rationale for having the rabbit and qpid params be different? e.g. rabbit_userid vs. qpid_username
16:49 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-swift: Puppet Forge release 2.1.0  https://review.openstack.org/40673
16:57 mgagne blentz: we had that discussion before. The conclusion was that it is unfortunate that OpenStack did not use generic names for those configs but that we should go with it until they fix it.
16:59 blentz mgagne: who is 'they' in this case? who would need to fix it? is it something i can propose patches for?
17:00 mgagne blentz: "they" is a generic term for whoever will propose the use of generic names in OpenStack and do the work to make it happen.
17:01 mgagne blentz: since it has the potential to generate a lot of work and discussions, not many people will want to raise the question.
17:02 mgagne blentz: or nobody saw it as a problem
17:02 blentz mgagne: i see. so class params can't diverge from the underlying config names?
17:03 mgagne blentz: I would say that we try as much as possible to not confuse the end user with names that would diverge from OpenStack ones.
17:04 otherwiseguy joined #puppet-openstack
17:05 blentz mgagne: what i'm looking for is something like this: http://pastebin.com/6QQsXZcd
17:05 blentz but, if i need to poke the openstack folks to change the config values there, i s'pose i can do that, too.
17:06 mgagne blentz: I see. I have to grab diner but will be back soon.
17:13 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-glance: Puppet Forge release 2.1.0  https://review.openstack.org/40678
17:34 starmer joined #puppet-openstack
17:36 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-openstack: Add ability to configure endpoint_type in openrc  https://review.openstack.org/40318
17:39 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-cinder: Puppet Forge release 2.1.0  https://review.openstack.org/40681
17:47 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-nova: Puppet Forge release 2.1.0  https://review.openstack.org/40683
17:54 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-horizon: Puppet Forge release 2.1.0  https://review.openstack.org/40684
18:02 rcrit joined #puppet-openstack
18:03 rcrit I'm looking to extend the current puppet module, to be executed post-install. Is it possible to include theses classes to be used by another module?
18:04 openstackgerrit William Van Hevelingen proposed a change to stackforge/puppet-keystone: Update modulefile and fixtures to use puppetlabs/inifile  https://review.openstack.org/40687
18:08 openstackgerrit William Van Hevelingen proposed a change to stackforge/puppet-cinder: Update modulefile and fixtures to use puppetlabs/inifile  https://review.openstack.org/40690
18:42 openstackgerrit Maru Newby proposed a change to stackforge/puppet-tempest: Rename quantum to neutron.  https://review.openstack.org/40697
18:43 openstackgerrit nightfly19 proposed a change to stackforge/puppet-keystone: Update modulefile and fixtures to use puppetlabs/inifile  https://review.openstack.org/40698
18:48 openstackgerrit Philip Cohoe proposed a change to stackforge/puppet-glance: Update modulefile and fixtures to use puppetlabs/inifile  https://review.openstack.org/40699
18:51 blkperl joined #puppet-openstack
18:51 * blkperl may have introduced some new people to gerrit/openstack puppet today :)
18:56 EmilienM dprince: it seems we also should make a patch in puppet-nova for neutron support
18:56 EmilienM already done actually > https://review.openstack.org/#/c/40544/
19:06 dprince EmilienM: great
19:14 mgagne bodepd: ping
19:14 mgagne bodepd: https://review.openstack.org/#/c/40690/ => puppet-lint: SUCCESS
19:21 openstackgerrit Terry Wilson proposed a change to stackforge/puppet-horizon: Rename quantum to neutron  https://review.openstack.org/40705
19:43 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-horizon: Puppet Forge release 2.1.0  https://review.openstack.org/40684
19:46 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-keystone: Puppet Forge release 2.1.0  https://review.openstack.org/40508
19:46 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-swift: Puppet Forge release 2.1.0  https://review.openstack.org/40673
19:47 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-glance: Puppet Forge release 2.1.0  https://review.openstack.org/40678
19:48 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-cinder: Puppet Forge release 2.1.0  https://review.openstack.org/40681
19:49 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-swift: Puppet Forge release 2.1.0  https://review.openstack.org/40673
19:50 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-glance: Puppet Forge release 2.1.0  https://review.openstack.org/40678
19:51 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-nova: Puppet Forge release 2.1.0  https://review.openstack.org/40683
19:56 prad_ mgagne: got a question for you.. I need to propose a similar change to https://review.openstack.org/#/c/40491/ for grizzly.. But single grizzly quantum uses multi config vs unified config, I need to modify the patch. So my workflow would be slightly different from what you suggested right? since i can't just cherry pick this to grizzly
19:56 prad_ mgagne: I'll have to create a separate patch of grizzly and do a git review stable/grizzly ?
19:56 bodepd mgagne: can we make it a gating change?
19:57 bodepd prad_: this is where things are going to get annoying
19:57 bodepd ...
19:57 mgagne bodepd: there's still modules failing puppet-lint. I can try fixing them
19:57 mgagne prad_: multi config vs unified?
19:58 prad_ mgagne: yea for plugins
19:58 prad_ mgagne: well cisco plugin i should say
19:58 mgagne prad_: you mean cisco plugin changed the way config files are splited (or not)?
19:58 prad_ bottom-line, i can just cherry pick the havana patch
19:58 prad_ so wonder whats the best workflow here
19:59 prad_ mgagne: yea
19:59 prad_ bleu s/can/cant
20:00 mgagne prad_: Can't cherry pick then. You should therefore propose a new change. I see no problem with that.
20:03 openstackgerrit A change was merged to stackforge/puppet-glance: Update modulefile and fixtures to use puppetlabs/inifile  https://review.openstack.org/40699
20:03 openstackgerrit A change was merged to stackforge/puppet-keystone: Update modulefile and fixtures to use puppetlabs/inifile  https://review.openstack.org/40698
20:04 prad_ mgagne: so the workflow would be git co -b my_grizzly origin/stable/grizzly && commit my change && git review stable/grizzly ?
20:06 mgagne prad_: that should be it. you can use git checkout -t origin/stable/grizzly too
20:07 prad_ ok, thanks!
20:07 mgagne prad_: git review seems to be able to detect the branch you are on
20:09 openstackgerrit A change was merged to stackforge/puppet-keystone: Update modulefile and fixtures to use puppetlabs/inifile  https://review.openstack.org/40687
20:11 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-openstack: Puppet Forge release 2.1.0  https://review.openstack.org/40710
20:13 EmilienM mgagne: neutron module looks working
20:13 EmilienM both on nova & neutron side
20:13 bodepd mgagne: I am looking into it
20:13 bodepd mgagne: the lint failure
20:13 EmilienM which is a good news
20:16 bodepd mgagne: I'm confused by that failure
20:16 bodepd mgagne: it claims that openstack::provision has tabs
20:16 bodepd mgagne: but, upon opening it, I don't see them
20:16 mgagne bodepd: tabs are invisible characters
20:17 bodepd I erased everything in the file and added spaces
20:17 bodepd mgagne: also, I have vim configured to show tabs
20:17 mgagne trolololo
20:18 mgagne bodepd: checking
20:19 mgagne bodepd: it's not about tabs, it's about misalignement
20:19 mgagne bodepd: number of spaces is odd, not even
20:21 openstackgerrit Dan Bode proposed a change to stackforge/puppet-openstack: Split device_endpoint into its own file  https://review.openstack.org/40713
20:22 bodepd mgagne: ah, misread it.
20:23 bodepd mgagne: I can fix that next
20:23 mgagne bodepd: message is confusing
20:23 bodepd mgagne: or make it part of the patch I already submitted
20:23 mgagne bodepd: yea
20:23 otherwiseguy joined #puppet-openstack
20:24 bodepd mgagne: I'll file a ticket against puppet-lint
20:24 mgagne bodepd: ok
20:24 bodepd that failure does not make any sense for our context
20:28 openstackgerrit Francesco Vollero proposed a change to stackforge/puppet-heat: Added support for QPID in Heat  https://review.openstack.org/40714
20:28 fvollero EmilienM: maybe this is the good one! ^^ :)
20:28 EmilienM fvollero: I hope so :P
20:29 EmilienM fvollero: go on #openstack-meeting follow the heat meeting
20:29 EmilienM maybe could we figure out the bug I found this morning
20:29 bodepd mgagne: I opened this: https://github.com/rodjek/puppet-lint/issues/213
20:29 fvollero EmilienM: sure thing
20:31 openstackgerrit Dan Bode proposed a change to stackforge/puppet-openstack: Fix puppet-lint failures  https://review.openstack.org/40713
20:34 CoilDomain joined #puppet-openstack
20:39 mgagne bodepd: you can left align the array items instead of aligning them with the parameter name.
20:42 openstackgerrit Chris Hoge proposed a change to stackforge/puppet-quantum: Puppet Forge 2.1.1 Release  https://review.openstack.org/40724
20:48 CoilDomain joined #puppet-openstack
20:53 bodepd mgagne: make the comment inline if you think that is a better fix
20:53 bodepd https://review.openstack.org/#/c/40713/
20:53 bodepd otherwise, lint checks pass!!!
20:53 mgagne :D
20:54 bodepd mgagne: I'm surprised it's not failing b/c we have class names with'-'
20:54 mgagne bodepd: we do?!
20:54 bodepd openstack::swift::store-node
20:54 mgagne bodepd: only saw them in puppet-heat
20:54 mgagne bodepd: hmmm
20:54 bodepd nah, one snuck into openstack
20:55 bodepd I'm going to merge what you just approved
20:55 bodepd b/c it's blocking
20:57 dmsimard joined #puppet-openstack
20:57 bodepd mgagne: can you revisit this: https://review.openstack.org/#/c/37277
20:57 bodepd mgagne: his explanation of why that parameter was added there seems reasonable
20:57 mgagne bodepd: at that point, I guess I just don't care
21:00 bodepd mgagne: :(
21:01 dmsimard joined #puppet-openstack
21:03 openstackgerrit A change was merged to stackforge/puppet-openstack: Fix puppet-lint failures  https://review.openstack.org/40713
21:21 dmsimard joined #puppet-openstack
21:33 hogepodge joined #puppet-openstack
21:45 openstackgerrit Maru Newby proposed a change to stackforge/puppet-quantum: Fix router interface addition check.  https://review.openstack.org/40740
22:30 starmer joined #puppet-openstack
22:49 starmer joined #puppet-openstack
22:51 marun bodepd: is chris hoge here?
22:54 marun bodepd: and is there any docs on the release procedure to puppet forge?  I'm not clear on how bugfixes to master make their way into a release.
22:55 blkperl bodepd: so the mysql module changed some class names..... going to submit a PR to change it
22:56 blkperl mysql::python -> mysql::bindings::python
22:56 CoilDomain joined #puppet-openstack
23:00 blkperl oh good they added deprecation classes back
23:01 mgagne awesome
23:02 mgagne I wonder who will remove those classes on new year's day
23:07 openstackgerrit Bradley Rasmussen proposed a change to stackforge/puppet-swift: Update modulefile and fixtures to use puppetlabs/inifile  https://review.openstack.org/40765
23:08 blkperl mgagne: I was wondering that too :)
23:08 bodepd marun: honestly, that is something we are figuring out this week
23:08 mgagne why is it someone different each time this fixture gets updated?
23:08 blkperl hehe
23:08 blkperl I
23:09 bodepd blkperl: I just created branches 0.x for mysql
23:09 blkperl I've been teaching new people
23:09 bodepd and I'm targeting that for now
23:09 blkperl bodepd: ok sounds good
23:09 marun bodped: is chris not on here?
23:09 bodepd I haven't been very impressed with the release cadence
23:09 bodepd marun: no idea
23:09 bodepd you may want to send him and email.
23:09 marun bodepd: We have a boatload of patches waiting for him to do whatever it is he's trying to do.
23:09 bodepd if one of you guys just wants to run with it, it may be faster
23:10 marun Ok, I'll try to touch base with him.
23:10 bodepd marun: please be patient.
23:10 marun Is he new to this or something?
23:10 bodepd marun: I have been talking with him, and this change effects a lot of thigs
23:10 bodepd marun: yes
23:10 bodepd marun: this is the first time we have had this much traffic and moved versions liek this
23:11 bodepd marun: especially with such a large backwards compat breaking change
23:11 marun him being a bottleneck for this is extremely suboptimal
23:11 marun i would hope for participation by more people going forward
23:13 bodepd blkperl: are those inifile changes your pdxcat guys?
23:13 blkperl bodepd: yea :)
23:13 bodepd blkperl: awesome
23:13 bodepd blkperl: I'm going to merge them, but I'm slightly terrified that PL is going to break APIs in inifile
23:13 blkperl why does quantum have it inifile in fixtures but not in the modulefile?
23:14 bodepd blkperl: they seem to be all about it
23:14 bodepd blkperl: b/c it needs to be updated
23:14 bodepd blkperl: it has not been released to the forge yet
23:14 blkperl ok updating now
23:14 bodepd blkperl: I like how everyone submitted one patch
23:14 bodepd blkperl: learning exercise ;)
23:15 blkperl hopefully I can find more patches for them :)
23:15 bodepd marun: that is all partly my fault. I left PL before he could be trained on any of this stuff
23:15 bodepd marun: he is my replacement
23:15 marun oy
23:15 marun those are big shoes
23:15 bodepd marun: this is one part of the project I suck at
23:16 bodepd marun: multiple branches, backporting fixes
23:16 bodepd marun: it terrifies me to be honest.so much work. so complicated. yet so necessary
23:16 marun why is it terrifying?
23:16 marun i thought you had pretty comprehensive ci running
23:18 bodepd I'm trying to build it out to be a little more comprehensive atm
23:18 bodepd marun: just doubles the amount of bits on disk to keep track of
23:19 bodepd marun: at some point, I need to accept that I can't know where everyline of code is
23:19 bodepd marun: and focus on a chunk of it
23:19 marun i have a pretty small brain, so i reached that point a long long time ago :)
23:19 openstackgerrit A change was merged to stackforge/puppet-swift: Update modulefile and fixtures to use puppetlabs/inifile  https://review.openstack.org/40765
23:19 marun thankfully we're pretty good at chunking
23:21 blkperl bodepd: so is vswitch not on the forge?
23:22 MrRacoon joined #puppet-openstack
23:24 openstackgerrit A change was merged to stackforge/puppet-cinder: Update modulefile and fixtures to use puppetlabs/inifile  https://review.openstack.org/40690
23:25 bodepd blkperl: nah.
23:26 openstackgerrit A change was merged to stackforge/puppet-openstack: Add ability to configure endpoint_type in openrc  https://review.openstack.org/40318
23:26 bodepd blkperl: that is kind of a clusterf*ck
23:26 bodepd blkperl: right now, the best one is from bodepd
23:26 bodepd blkperl: and it has patches that need to go upstream
23:26 bodepd blkperl: want to have a look at getting those upstream and getting it released to the forge?
23:27 bodepd marun: I'm getting a couple of things merged in.
23:27 bodepd marun: the things that I know need to go in pre-release
23:27 blkperl bodepd: is upstream cisco? don't you work there? :)
23:27 bodepd blkperl: nah. it's wherever I forked it from
23:27 bodepd hold on
23:28 marun bodepd: Chris mentioned he might have things done as early as this afternoon.  Do you think that's possible?
23:28 bodepd blkperl: https://github.com/ekarlso/puppet-vswitch/pull/12
23:28 blkperl c6441e69 Modulefile (Dan Bode       2013-03-28 21:47:02 -0700 11) #dependency 'CiscoSystems/vswitch', '>= 0.1.0'
23:28 bodepd marun: no. I don't.
23:28 marun bodepd: what's your guess?
23:29 marun and are there ways additional hands can speed things up?
23:29 bodepd marun: I have no guess. The quickest way to get it done is probably if I did it myself.
23:29 blkperl bodepd: wow that does look pretty bad
23:29 marun And what exactly is he doing?
23:29 bodepd no idea
23:29 bodepd oh, the task
23:29 marun ?
23:29 marun Yeah
23:30 bodepd 1. get the Modulefiles and README's updated
23:30 bodepd for release
23:30 bodepd 2. backport remaining patches to stable/grizzly
23:30 bodepd (not sure how to do that)
23:30 bodepd 3. release to forge
23:30 marun And what is he releasing - grizzly?
23:30 bodepd yes. the grizzly release of the puppet modules
23:30 marun I think my main frustration with this is the lack of visibility.
23:31 marun He's going off and doing this and holding everything up and nobody knows where he's at.
23:31 marun That's not compatible with active development in the least.
23:31 marun I understand he's new to this, but it doesn't change the breakage.  Red Hat needs these modules to work so we can test our packaging.
23:32 marun I know it's not your fault, and I'm wanting to help fix things not just complain.
23:32 mgagne marun: we need to own that part otherwise we will be dependent on an external actor
23:33 mgagne marun: own as take the ownership and streamline the process
23:33 marun mgagne: agreed.
23:33 bodepd marun: I agree with that as well.
23:34 bodepd marun: this whole process of creating a release, switchign openstack versions, and starting to backport to maintain 2 branches it totally new to this project.
23:34 mgagne marun: there's a lot of work to be done and stuff are pilling up, creating confusion and frustration
23:34 bodepd marun: any work on what the process should be would be helpful
23:34 mgagne bodepd: and I would add, to puppet modules too
23:34 marun I'm unclear on why it should be more complicated than picking a branch point for each repo and doing all the work on the branch
23:35 marun master can keep going, and backports can be done as needed
23:35 marun but holding up master to do all these updates???
23:35 nibalizer joined #puppet-openstack
23:35 mgagne marun: you usually branch when you are ready to release, I guess it was more or less the case ; or I'm confused
23:36 bodepd mgagne: he had branched, but the branches were not really used
23:36 marun so why does that hold up master?
23:36 bodepd mgagne: so work built up in master that shoudl have also been in the branches
23:36 marun that  means backporting
23:36 marun or picking new branch point
23:37 marun i mean, the core openstack projects have a feature freeze period
23:37 marun maybe that's what's missing
23:37 marun nobody expects to get anything done during that period
23:37 marun and people can be prepared for that eventuality
23:37 mgagne marun: we are still figuring out the release process
23:38 marun fair enough,
23:38 marun has there been any evaluation of how openstack projects do it?
23:38 marun given that we are pretty closely aligned with most of the openstack toolset/process already, that would seem an obvious starting point
23:38 mgagne we are considering it
23:39 openstackgerrit Erik Sutherland proposed a change to stackforge/puppet-quantum: Add dependencies to Modulefile  https://review.openstack.org/40771
23:39 openstackgerrit Dan Bode proposed a change to stackforge/puppet-glance: Fix lint failures  https://review.openstack.org/40772
23:39 mgagne unfortunately, some people (including myself) aren't 100% dedicated to openstack puppet modules.
23:39 marun Look, at the risk of being a dick, I'd like to see either visibility on what Chris is doing or otherwise an evaulation of what's going on so we can know when this crap is going to end.
23:39 marun An open-ended ticket is in nobody's interest.
23:39 blkperl bodepd: ok so the PR from ^ Erik adds stuff to quantums modulefile
23:40 bodepd mgagne: can you look at that patch?
23:40 mgagne blkperl: are you working with Erik and all those people updating inifile fixtures?
23:40 bodepd mgagne: glance puppet-lint failures. we still need to wait for the tests to kick off.
23:40 bodepd marun: I'm working on your issue now.
23:40 blkperl mgagne: yep, teaching moment :)
23:40 bodepd marun: step one is that we have to clean up the failures.
23:40 mgagne blkperl: looks like you round up everyone in the office and giving them a try at gerrit =)
23:41 blkperl mgagne: pretty much :)
23:41 blkperl mgagne: I got the windows team to use github for windows last week, I'm having too much fun :D
23:42 mgagne blkperl: "You in the hallway, come here, I'll show you something." "Ok, where do I put my broom?"
23:42 bodepd mgagne: that glance test failure is a github failure :)
23:42 bodepd mgagne: that almost never happens
23:42 mgagne bodepd: github failure? impossibru
23:42 * blkperl is good at harassing people :)
23:43 bodepd mgagne: are we gating on lint?
23:43 bodepd mgagne: the answer is yes as of today, right?
23:43 mgagne bodepd: not yet, please give me some time to go through my backlog =)
23:44 mgagne bodepd: didn't we set an upper bound on dependency versions?
23:52 bodepd mgagne: ah. not yet.
23:52 bodepd mgagne: I probably just need to do this myself
23:52 bodepd mgagne: but I don't really care about a forge release
23:55 bodepd mgagne: do you think it matters if the Modulefile and README updates hit master?
23:55 mgagne bodepd: how would it impact the module?
23:55 bodepd mgagne: or do you think they could just be cherry-picked/forward ported
23:55 nibalizer blkperl is building an army of open sourcerers
23:56 bodepd mgagne: I'm more concerned about the project's history
23:56 bodepd nibalizer: hey spence
23:56 bodepd bring em in. We need the help!@
23:56 nibalizer bodepd: ohai :)
23:56 nibalizer :D

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