Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2013-08-05

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

All times shown according to UTC.

Time Nick Message
00:26 openstackgerrit Xingchao Yu proposed a change to stackforge/puppet-heat: Add fixtures.yml for puppet-heat  https://review.openstack.org/40100
02:58 otherwiseguy joined #puppet-openstack
03:44 openstackgerrit Daneyon Hansen proposed a change to stackforge/puppet-openstack: Allow ovs_enable_tunneling to be Passed Through  https://review.openstack.org/40146
04:15 openstackgerrit A change was merged to stackforge/puppet-heat: Add fixtures.yml for puppet-heat  https://review.openstack.org/40100
04:23 openstackgerrit A change was merged to stackforge/puppet-openstack: Changed the auth_host for swift test file bug/1208117  https://review.openstack.org/40089
05:05 marun joined #puppet-openstack
06:39 openstackgerrit Xingchao Yu proposed a change to stackforge/puppet-heat: Update allowed_hosts conditional statement  https://review.openstack.org/40165
08:38 openstackgerrit Robert Starmer proposed a change to stackforge/puppet-openstack: Pass external_interface_bridge through controller to l3 agent  https://review.openstack.org/40179
09:36 andi- Is there any particular reason why firewall_driver can not be used with nova-network in the puppet modules?
09:44 openstackgerrit Francesco Vollero proposed a change to stackforge/puppet-heat: Added support for QPID rpc  https://review.openstack.org/40188
09:45 fvollero \o/
10:43 key joined #puppet-openstack
10:44 openstackgerrit Francesco Vollero proposed a change to stackforge/puppet-heat: Setup of the test environment  https://review.openstack.org/40201
11:22 fvollero EmilienM: ping
11:29 EmilienM fvollero: hey
11:49 fvollero EmilienM: bonjour... thanks that you reviewed my two patches
11:49 fvollero EmilienM: Did you also checked up the ones for ceilometer ?
11:49 EmilienM fvollero: oops not yet :)
11:51 fvollero EmilienM: no problem. is because there I also have the tests for qpid.
11:53 fvollero EmilienM: and i have a dubt on how to write test for heat_engine (for qpid) if the parameters are in heat class :/
11:53 fvollero heat init
11:57 EmilienM ok
11:57 EmilienM I'll have a look asap
12:30 red_trela joined #puppet-openstack
12:48 fvollero EmilienM: merci
12:59 dmsimard joined #puppet-openstack
13:05 dmsimard joined #puppet-openstack
13:10 dmsimard joined #puppet-openstack
13:18 dprince joined #puppet-openstack
13:57 prad_ joined #puppet-openstack
14:12 otherwiseguy joined #puppet-openstack
14:27 key4 joined #puppet-openstack
15:17 openstackgerrit A change was merged to stackforge/puppet-ceilometer: Update mysql allowed_hosts conditional statement  https://review.openstack.org/40099
15:25 newptone joined #puppet-openstack
15:31 mgagne joined #puppet-openstack
15:33 openstackgerrit A change was merged to stackforge/puppet-openstack: Update openstack::db::mysql specification  https://review.openstack.org/40055
15:42 newptone_ joined #puppet-openstack
15:47 xingchao joined #puppet-openstack
15:57 mgagne bodepd: ping (let me know when you are there)
16:19 otherwiseguy joined #puppet-openstack
16:32 sbadia joined #puppet-openstack
16:36 fvollero mgagne: tbh, i got the layout from rspec-puppet-init thanks for the feedback gonna fix immediately my commit
16:39 fvollero I am trying to understand what's the best way to test a submodule that will setup a conf file using params from the main module. with pre_condition should be enough ?
16:56 openstackgerrit Francesco Vollero proposed a change to stackforge/puppet-heat: Initial setup for the test environment in puppet-heat  https://review.openstack.org/40201
17:21 openstackgerrit Dan Prince proposed a change to stackforge/puppet-quantum: Rename quantum to neutron.  https://review.openstack.org/40252
17:22 openstackgerrit Francesco Vollero proposed a change to stackforge/puppet-heat: Initial setup for the test environment in puppet-heat  https://review.openstack.org/40201
17:29 bodepd mgagne: I'm here
17:29 mgagne bodepd: cool
17:29 bodepd I just merge your mega-lint-patch
17:29 mgagne bodepd: nice :P
17:29 bodepd mgagne: well, I approved it anyways
17:30 openstackgerrit A change was merged to stackforge/puppet-swift: Various Puppet lint fixes  https://review.openstack.org/35219
17:51 dmsimard joined #puppet-openstack
18:02 marun bodepd: ping
18:03 mgagne bodepd: do you mind adding puppet-heat to your list of watched projects? ;)
18:04 marun bodepd: where are we wrt the quantum -> neutron rename?  I still don't see a grizzly branch...
18:04 marun bodepd: Also, I'm backing out your shared network change.  The test is definitely in error, neither the public nor private networks should be shared.
18:05 mgagne marun: Did you see the email on puppet-openstack?
18:06 marun mgagne: ah, looking now
18:06 mgagne marun: dprince is pushing hard to make it happen asap
18:06 marun mgagne: I promised him I'd help but failed to do so :(
18:07 mgagne marun: we aren't looking for blames ;)
18:07 bodepd marun: I was kind of hoping someone else was going to do it
18:07 mgagne bodepd: can odyi be of any help?
18:08 bodepd mgagne:  I thought they were going to do it, but I guess it never got done
18:09 marun mgagne: arg
18:09 marun There needs to be a change happen *now*
18:09 EmilienM hello here
18:09 marun the fact that changes are going in for quantum but targeting neutron features is pretty suspicious.
18:10 marun quantum -> grizzly, neutron -> havana
18:10 EmilienM mgagne: I need a review on https://review.openstack.org/#/c/39836/
18:10 marun i question whether proper testing is being done
18:10 marun spec test are not enough
18:10 bodepd marun: I'm not aware of what neturon specific features have gone in
18:10 marun the firewall patch that was just merged is one
18:10 marun fwaas doesn't exist in grizzly
18:11 marun EmilienM - how are you testing your patches?
18:11 marun EmilienM - Are you actually deploying or just running spec tests?
18:11 EmilienM marun: i'm running tests in local
18:11 EmilienM spec tests
18:12 marun EmilienM - That's not good practice.
18:12 marun EmilienM - The fw changes won't work
18:12 marun EmilienM - Because the module targets quantum, not neutron.
18:12 bodepd marun: then that was a mistake to accept
18:12 EmilienM marun: you're right
18:12 EmilienM marun: but we don't have neutron puppet module
18:13 bodepd marun: the reality, it that I was trying ot balance supportin both versions
18:13 marun EmilienM - So don't submit changes that target them until we do.
18:13 EmilienM marun: when we will have, I think that will be a copy from quantum one
18:13 bodepd and AFAIK, had not submitted havana specific configuration
18:13 marun EmilienM - you're effectively breaking quantum support by doing otherwise.
18:13 marun bodepd: We need to branch asap so this doesn't get worse.
18:13 bodepd I know of one example where we messed up a bit
18:14 EmilienM marun: I don't break anything. If you have an extra flag which is not supported you don't have any error.
18:14 marun EmilienM: Except if someone decides to use it!
18:14 marun EmilienM: All I'm saying - no more neutron changes until the repo is openstack-neutron.
18:15 bodepd it's also a but confusing
18:15 marun EmilienM: Keep them to yourselves.
18:15 bodepd b/c it won't be clear to users if it is supported or not
18:15 marun +1
18:15 EmilienM marun: we work upstream here, I don't commit for myself ;)
18:15 EmilienM but ok I wait
18:15 marun EmilienM: Thank you.
18:15 EmilienM marun: do you have an idea when we will have a repo ?
18:15 EmilienM release is *close*
18:16 EmilienM and some people is already deploying Havana & Neutron ;)
18:16 marun bodepd: Can we cut a new repo and work on getting a grizzly branch created?
18:16 marun bodepd: I have no permissions to do so or I would help.
18:16 EmilienM but I don't understand
18:16 marun At the least the new branch.
18:16 EmilienM why don't we work with branches ?
18:17 EmilienM master = havana = neutron
18:17 EmilienM stable/grizzly = grizzly = quantum
18:17 marun EmilienM: We don't have a grizzly branch cut yet.
18:17 EmilienM and that's it...
18:17 EmilienM like OpenStack does
18:17 EmilienM marun: but why ?
18:17 mgagne sorry, I merged an havana only feature ^^'
18:18 marun EmilienM: Simply because it hasn't been done.  this is the first release changeover that puppet-openstack has had to do, as far as I know
18:18 marun bodepd: What can I do to have a grizzly branch created?
18:18 marun Preferably at the revision prior to the most recent commit.
18:18 marun (the fw change(
18:19 marun Hopefully there haven't been other havana-specific changes.
18:19 EmilienM IMO, like all OpenStack project we should have same process in versionning
18:19 EmilienM since OpenStack config is not the same for each release...
18:19 EmilienM we could not stay on master all the life
18:19 marun EmilienM - for sure, but this is stackforge not openstack
18:19 EmilienM but anyway, I'm glad I'm not the only one who is surprised that we don't have to merge H features.
18:20 bodepd I created a branch for quantum, not sure if I did it right
18:20 EmilienM marun: I disagree
18:20 marun you can disagree all you want, but you're wrong
18:20 marun there is no ptl
18:20 EmilienM marun: ceilometer was in StackForge
18:20 marun there is not as much organization or rules
18:20 EmilienM marun: and did support versions
18:20 marun *sigh*
18:20 marun like talking to a wall
18:20 EmilienM marun: a lot of projects are on StackForge and follow versions
18:20 marun i'm not saying how it *should* be
18:20 d3u joined #puppet-openstack
18:20 marun I'm saying how it *is*
18:21 EmilienM yeah, I understand :)
18:21 marun If you can't figure out the difference, there's not much I can tell you.
18:21 EmilienM I just want to help & improve the stuff.
18:21 marun In any case, you're as much responsible as I.
18:21 marun If you see a problem - fix!
18:21 marun That's all I'm trying to do here.
18:21 EmilienM marun: again, my job is to deploy openstack. and I have to deploy havana very soon and I can't see puppet module for neutron
18:21 EmilienM marun: that's what I'm doing
18:21 marun so, fix!
18:22 EmilienM I make patchs
18:22 EmilienM as many I can
18:22 marun patches is one small part
18:22 marun making sure we have coherent release - that's also a part
18:22 marun just because it's done for you in other projects doesn't mean to ignore i there
18:22 fvollero marun: yeah, but obviously it depend on the rack of skillset of everyone
18:22 marun i > it
18:22 marun fair enough
18:23 fvollero marun: I mean, I totally agree with you to the extent of the contributor skillset
18:23 EmilienM marun: thx for discussing. I'm quite new here, and I'm glad to understand finally.
18:23 bodepd marun: feel free to direct your anger this way. I've made all of the decisions that got things to where they are
18:24 bodepd marun: I pushed a branch for quasntum
18:24 marun bodepd: awesome!
18:24 marun bodepd: Ok, I have a few fixes to propose that apply to both.
18:25 marun EmilienM: You can now push all you want to master.
18:25 marun bodepd: How are backports going to be managed?
18:25 bodepd marun: I haven't even started reading about it
18:25 bodepd marun: in the same way that openstack does it
18:25 marun bodepd: It's resource intensive on other projects :(
18:25 bodepd I believe that requires patches to both
18:26 fvollero mgagne: Sorry mate, can you reply, when you've time on 40188 ?
18:26 marun bodepd: either the submitter has to know enough to realize a backport is required and submit
18:26 bodepd yeah, I was trying to short but it by having master support both for as long as possible
18:26 marun bodepd: or a dedicated maintainer has to do it...
18:26 marun bodepd: i don't know who is going to have the cycles to do that though
18:26 bodepd marun: as things stand now, that person does not exist
18:27 digicloseup joined #puppet-openstack
18:27 bodepd so, everything goes upstream
18:27 bodepd and people will need to occassionally cherry-pick things as needed
18:27 marun bodeped: maybe whoever is approving needs to identify the need for a backport and require the change author to submit the backport before merge?
18:27 marun bodepd: we're going to lose thing otherwise.
18:27 marun things
18:28 marun EmilienM: Are all the options supported in the following change supported in grizzly?
18:28 marun https://review.openstack.org/#/c/39836/
18:29 EmilienM marun: yes
18:29 EmilienM I'm using them right no
18:29 EmilienM now*
18:29 EmilienM marun: but the patch with quota & firewall flag: no. Only H
18:29 marun EmilienM: Ok.  Please submit the l3 change to the grizzly branch as well.
18:30 EmilienM marun: sure I will.
18:30 EmilienM it's time to diner here, I come back later. Cheers
18:30 fvollero EmilienM: 8.30pm seems very european ;)
18:31 EmilienM fvollero: french ;)
18:31 fvollero EmilienM: Italien aussi ;)
18:37 odyi joined #puppet-openstack
18:44 openstackgerrit Maru Newby proposed a change to stackforge/puppet-openstack: Revert provisioning of shared networks.  https://review.openstack.org/40264
18:45 EmilienM marun: ^
18:46 EmilienM marun: and now, I suggest a patch based on "sed" ;-) but IMO we should ask some review from quantum-core.
18:46 marun Emilien: I'm not sure of the context?
18:46 marun (i'm neutron core, btw)
18:47 EmilienM I'm not ;)
18:47 EmilienM but I know quantum is not existing anymore.
18:47 marun dprince has submitted a patch that does the rename
18:47 marun please review: https://review.openstack.org/#/c/40252/
18:48 EmilienM marun: so IMO we should apply the patch to master branch
18:48 marun EmilienM: agreed
18:48 marun EmilienM: I think it's already targeted there
18:49 marun Testing would be much apprecaited
18:50 EmilienM marun: I can't see your name : https://wiki.openstack.org/wiki/Neutron/Teams neither in the blueprints
18:50 EmilienM are you core in neutron ?
18:50 marun I'm there as 'maru'
18:51 EmilienM marun: so we should definitly test it, and asap
18:51 EmilienM why not in master
18:51 marun I'm not sure what you mean?
18:51 marun The patch in question is targeting master already
18:52 EmilienM marun: I would like to test it, in the same way as other modules.
18:52 marun sure.  so check out the branch from the review
18:53 marun EmilienM: do you know how to do that?
18:53 EmilienM marun: actually, I would test it on a single VM before, applying the patch in local and setup all the neutron stuff on my VM
18:54 EmilienM to test if it's actually working
18:54 marun bodepd: I'm confused - I see grizzly and stable/grizzly.  Which should I target?  And can you delete the other?
18:54 EmilienM stable/grizzly
18:55 EmilienM if we want to follow upstream
18:55 mgagne if grizzly branch exists, can we remove it?
18:55 marun mgagne: If you have permission, please do!
18:55 bodepd mgagne: I created it by accident
18:55 mgagne just now?
18:55 bodepd mgagne: and failed to remove it
18:55 bodepd like 15 minutes ago
18:55 marun EmilienM: I suggest using openstack::provision to do the testing
18:55 mgagne I'll see if I can delete it, I'll ask infra otherwise
18:56 bodepd please try
18:56 bodepd I already asked in infa...
18:56 mgagne ok, I'll wait for them then
18:57 bodepd they recommended that I check the UI
18:58 mgagne https://review.openstack.org/#/admin​/projects/stackforge/puppet-quantum,branches
18:58 marun bodepd: gate jobs failing, any idea why? http://logs.openstack.org/64/40264/1/gate/gate-pup​pet-openstack-puppet-unit-3.0/76afed7/console.html
18:59 bodepd says syntax error in the manifest
18:59 bodepd was that patch not just a git revert?
19:00 marun bodepd: Not just a revert, I left the spec test addition in but modified it.
19:00 marun bodepd: I ran the spec tests without error :(
19:00 marun will try again
19:00 bodepd extra comma
19:01 bodepd in the spec tests
19:01 EmilienM marun: do you have the URL of openstack::provision please ?
19:01 bodepd line 35
19:01 bodepd marun: ^^^
19:01 marun bodepd: arg
19:01 marun bodepd: I swear I ran the tests.  #fail
19:01 bodepd yeah, whether or not you can have trailing commas in ruby
19:01 bodepd depends on whether the has is enclosed in {} (IIRC)
19:01 bodepd s/has/hash/
19:02 marun bodepd: ah.  that seems to be biting me a lot, one day maybe I'll learn.
19:02 bodepd it's lame
19:02 marun old (python) habits die hard - trailing commas are always ok
19:02 bodepd Puppet has it's shares of where they are and aren't allowed
19:02 bodepd I think they should be always ok
19:02 marun bodepd: so is a best practice never to put trailing commas?
19:03 bodepd in Ruby yes :)
19:03 bodepd in Puppet you should put trailing commas
19:03 marun but sometimes they aren't allowed?
19:03 bodepd (but they are not supported in class parameter definitions until 2.7.x)
19:03 marun i'm a simple creature.  the more rules i have to remember the less i get done
19:03 marun so commas everywhere except if one works at redhat which uses < 2.7
19:03 marun gotcha
19:04 marun :)
19:04 bodepd yep. and you guys have to chase those down and remove em like crazy
19:04 marun I literally just ran the damn spec tests and it gave no error :(
19:04 mgagne marun: ruby version?
19:04 bodepd what version of ruby?
19:04 marun ah, puppet 3.2?
19:04 marun ruby 2.0
19:04 marun :(
19:04 marun version mismatch
19:05 bodepd I thkn the tests are running 1.9.3
19:05 bodepd that is a ruby error
19:05 dprince joined #puppet-openstack
19:05 mgagne /usr/bin/ruby1.8
19:06 marun ok, mental not
19:06 marun only run tests in my vm
19:06 marun the host os has shiny versions of everything
19:06 openstackgerrit Maru Newby proposed a change to stackforge/puppet-openstack: Revert provisioning of shared networks.  https://review.openstack.org/40264
19:08 marun bodepd, mgagne: I need to ensure we can continue to use puppet-openstack's stable/grizzly, so is it ok if I backport the provisioning script?
19:08 bodepd sure
19:08 EmilienM marun: oh I see the module, my bad
19:08 bodepd I can use a resource default at top scope until I can sort out why it's failing
19:09 EmilienM marun: so to be sure, what you suggest is replacing quantum by neutron in http://goo.gl/MynM2U and test a deployment, right ?
19:09 marun EmilienM: Nope
19:09 marun EmilienM: I suggest using this: https://github.com/stackforge/puppet-open​stack/blob/master/manifests/provision.pp
19:10 marun EmilienM: Oh, wait
19:10 EmilienM but it's not enough imo
19:10 marun EmilienM: I'm no help, sorry.
19:10 marun EmilienM: I don't think it's enough to provision just quantum/neutron
19:10 EmilienM no worries, I think the cleanest way would be to try the real module
19:10 marun EmilienM: It's necessary to deploy everything at once, ala packstack
19:11 marun (or another tool that deploys everything from stackforge/puppet-* modules)
19:11 mgagne bodepd: didn't find a way to delete a branch
19:11 EmilienM marun: oh I see. But I work on debian / ubuntu only.
19:11 bodepd yeah.
19:11 bodepd I guess we'll alwyas remember when I made that mistake
19:11 EmilienM marun: yep, I'll try tomorrow with puppet-quantum module
19:11 starmer joined #puppet-openstack
19:11 marun EmilienM: Ok
19:11 bodepd the ::provision stuff won't work as soon as we merge that patch into master
19:11 marun EmilienM: I'll test as well. Hopefully we can get this patch properly tested.
19:12 EmilienM marun: if both of use have an havana  openstack running with neutron plugin, I think we could merge the patch for rename :)
19:12 EmilienM and have a clean master branch
19:12 marun +1
19:15 dmsimard1 joined #puppet-openstack
19:16 mgagne bodepd:
19:16 mgagne $ git push origin :refs/heads/quantum
19:16 mgagne remote: Processing changes: refs: 1, done
19:16 mgagne To ssh://mgagne@review.openstack.org:29418/s​tackforge/puppet-quantum.git
19:16 mgagne ! [remote rejected] quantum (can not delete references)
19:16 mgagne error: failed to push some refs to 'ssh://mgagne@review.openstack.org:29418/st​ackforge/puppet-quantum.git'
19:21 openstackgerrit A change was merged to stackforge/puppet-openstack: Revert provisioning of shared networks.  https://review.openstack.org/40264
19:22 dmsimard joined #puppet-openstack
19:28 bodepd mgagne: I saw that earlier
19:28 bodepd I have no idea how to get rid of that branch
19:30 bodepd fungi is going to get rid of it for me
19:30 mgagne infra should be able to do it
19:54 mgagne bodepd: I was thinking about fixing (still) puppet-lint
19:54 mgagne and with the quantum rename, we will have no choice but to grep Modulefile for the actual modulename
20:22 otherwiseguy joined #puppet-openstack
20:28 mgagne bodepd: ping
20:50 marun bodepd: what was the name of the failing test that the 'shared' patch was targeting?
20:57 key joined #puppet-openstack
21:03 openstackgerrit Daneyon Hansen proposed a change to stackforge/puppet-openstack: Fix Swift Keystone Authentication Endpoints  https://review.openstack.org/40287
21:15 bodepd marun: I assume you sorted this out? b/c it is resolved.
21:16 marun bodepd: Hmmm, not me.  Happy to hear it's fixed.
21:18 bodepd marun: oh, sorry, I misread your msg
21:19 marun bodepd: Have you re-run the test without share = true in the manifest and things just worked?
21:20 bodepd my tempest tests fail for sure with that set
21:20 bodepd the integration tests that I am using do not use openstack::provision
21:20 bodepd I'm trying to port those tests to use tempest, but I keep getting distracted before I can finish
21:21 marun ok
21:21 marun I'm working on allowing tempest to be configured for testing all the renaming patches
21:22 marun (need to follow with patches to puppet-{openstack,tempest})
21:23 bodepd tempest/tests/compute/serve​rs/test_server_addresses.py
21:23 bodepd that is the test that was failing
21:23 bodepd the ones from that that run as a part of the smoke test
21:26 EmilienM marun: testing seriously neutron migration could be on my official schedule for the next sprint. I'll let you know tomorrow. If it's the case, I'll perform a full testing and submit a patch
21:27 marun bodepd: thanks, will see about fixing that upstream
21:27 marun EmilienM: How would you do the migration?
21:29 marun EmilienM: I've been thinking about the issue and it's a bit of a chicken-and-egg problem.
21:29 EmilienM marun: we have a CI here. I'm planning to setup a full OpenStack running on Havana with neutron using puppet
21:29 marun Uh, sure
21:29 EmilienM marun: I have one problem
21:29 marun But there needs to be migration automation to work in CI
21:29 EmilienM marun: I can only perform the tests on Debian / Ubuntu
21:30 marun why do you need to write your own tests?
21:30 marun that's what tempest is for
21:30 EmilienM actually, we already have the jobs in CI which build OpenStack with Quantum & Puppet
21:30 EmilienM but for Grizzly
21:30 EmilienM I have to clone the job for Havana and create a new puppet moule
21:30 EmilienM module*
21:30 EmilienM marun: I don't write test, we use tempest
21:31 marun ah, you mean run them, gotcha
21:31 EmilienM I just use CI for spawning server
21:31 EmilienM yep, sorry. I should have left laptop, I can't speak english ATM :P
21:31 marun no, my fault.  'perform' is closer to 'run' than 'write'
21:32 EmilienM marun: see you later, cheers
21:32 marun EmilienM: cheers!
21:48 starmer joined #puppet-openstack
22:31 starmer joined #puppet-openstack
22:48 xingchao joined #puppet-openstack
22:56 openstackgerrit Maru Newby proposed a change to stackforge/puppet-openstack: Disable dhcp on the provisioned public subnet.  https://review.openstack.org/40313
23:46 xingchao joined #puppet-openstack
23:58 otherwiseguy joined #puppet-openstack

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