Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2013-10-16

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

All times shown according to UTC.

Time Nick Message
00:00 mgagne iwi: at least it's working now =)
00:00 iwi mgagne: btw - there seems to be a small problem in modules dependencies when installing puppetlabs/openstack
00:01 iwi some underlying modules require puppetlabs-concat where swift requires rippenar-concat
00:01 mgagne iwi: swift still requires rippenar-concat?
00:01 iwi yes
00:01 mgagne :O
00:02 iwi u want a bug report for that ?
00:02 mgagne iwi: fixing it right now
00:02 mgagne iwi: just a matter of cherry-picking the change from master to stable/grizzly
00:03 iwi mgagne: i guess that i can say "thanks for that on behalf of the community" ;)
00:03 mgagne =)
00:04 iwi mgagne: are you testing those modules against havana ?
00:04 mgagne iwi: unfortunately, not yet. I did try once a week-end. =)
00:04 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-swift: Remove puppetlabs-create_resources dependency.  https://review.openstack.org/51982
00:05 iwi mgagne: ;)
00:05 dmsimard joined #puppet-openstack
00:07 openstackgerrit Mathieu Gagné proposed a change to stackforge/puppet-swift: concat module was moved from ripienaar to puppetlabs  https://review.openstack.org/51983
00:15 ryanycoleman joined #puppet-openstack
00:18 ryanycoleman joined #puppet-openstack
00:26 xingcha__ joined #puppet-openstack
00:53 openstackgerrit A change was merged to stackforge/puppet-swift: Remove puppetlabs-create_resources dependency.  https://review.openstack.org/51982
00:53 openstackgerrit A change was merged to stackforge/puppet-swift: concat module was moved from ripienaar to puppetlabs  https://review.openstack.org/51983
00:54 bodepd_ iwi: mgagne already merged
00:55 mgagne =)
00:55 bodepd_ mgagne: I don't have the patience to look into that keystone bug atm
00:55 bodepd_ I'm sure it's an easy fix
00:55 bodepd_ and I remember someone else complaining about it recently
00:56 bodepd_ basically, it need to make sure it deletes the right url
00:56 bodepd_ darnit, not I'm interested in the root cause..
00:56 bodepd_ looking now...
00:59 bodepd_ mgagne: at a glance at the code, I don't see it
00:59 mgagne bodepd_: my bug explains the source of the problem
00:59 bodepd_ I read it, I may be just a bit slow atm
01:00 bodepd_ updates fail to delete the old endpoint because the service_id is cached
01:00 mgagne bodepd_: you have to update 2 or more URLs for the problem to occur
01:00 bodepd_ I don't understand that
01:00 mgagne bodepd_: yes
01:00 bodepd_ ah, b/c it calls create
01:00 mgagne bodepd_: when you update an URL, the endpoint has to be deleted
01:00 mgagne bodepd_: and has to be recreated
01:01 mgagne bodepd_: when updating the 2nd URL, it tries (again) to delete the endpoint which doesn't exist anymore
01:01 mgagne bodepd_: and then create a new (duplicated) one
01:02 bodepd_ ah.
01:02 bodepd_ gotcha
01:02 bodepd_ I was somehow missing that
01:02 mgagne bodepd_: I have to leave now, feel free to work on something else if you have to. I don't need it to be fixed for tonight.
01:04 bodepd_ I posted what I think the fix is in the ticket
01:04 bodepd_ it's a little dirty, but I think it's the best fix
01:06 bodepd_ ignore the first proposed solution, it was obviously wrong
01:14 ryanycoleman joined #puppet-openstack
01:20 ryanycoleman joined #puppet-openstack
01:39 xingchao joined #puppet-openstack
01:47 _ilbot joined #puppet-openstack
01:47 Topic for #puppet-openstack is now Place to collaborate on Puppet/OpenStack tools: logs at http://irclog.perlgeek.de/puppet-openstack/today
01:49 badiane_ka joined #puppet-openstack
01:52 e1mer joined #puppet-openstack
01:56 tnoor joined #puppet-openstack
02:02 xingchao joined #puppet-openstack
02:03 dguitarbite_ joined #puppet-openstack
02:11 xingcha__ joined #puppet-openstack
03:48 e1mer joined #puppet-openstack
03:54 tnoor joined #puppet-openstack
04:41 bcrochet joined #puppet-openstack
06:15 otherwiseguy joined #puppet-openstack
07:01 marun joined #puppet-openstack
07:10 openstackgerrit Florian Haas proposed a change to stackforge/puppet-openstack: repo.pp: Support havana  https://review.openstack.org/52012
07:13 francois1 joined #puppet-openstack
07:19 openstackgerrit A change was merged to stackforge/puppet-neutron: Scrap nonsense provider setting requirements  https://review.openstack.org/51603
07:38 openstackgerrit Florian Haas proposed a change to stackforge/puppet-openstack: repo.pp: Support havana  https://review.openstack.org/52012
07:49 qba73 joined #puppet-openstack
07:50 dafter joined #puppet-openstack
07:51 tvb|afk joined #puppet-openstack
08:13 derekh joined #puppet-openstack
08:54 qba73 joined #puppet-openstack
09:06 marun joined #puppet-openstack
09:18 openstackgerrit Florian Haas proposed a change to stackforge/puppet-neutron: Set OVS/enable_tunneling in the plugin, not only the agent  https://review.openstack.org/52041
09:25 openstackgerrit Florian Haas proposed a change to stackforge/puppet-neutron: Set OVS/enable_tunneling in the plugin, not only the agent  https://review.openstack.org/52041
09:29 openstackgerrit Florian Haas proposed a change to stackforge/puppet-neutron: Set OVS/enable_tunneling in the plugin, not only the agent  https://review.openstack.org/52041
09:34 mmagr joined #puppet-openstack
09:36 dafter joined #puppet-openstack
09:58 michchap_ joined #puppet-openstack
10:05 otherwiseguy joined #puppet-openstack
10:33 e1mer joined #puppet-openstack
10:35 michchap joined #puppet-openstack
11:06 e1mer joined #puppet-openstack
12:24 prad joined #puppet-openstack
12:29 grog500 joined #puppet-openstack
12:43 dprince joined #puppet-openstack
13:08 grog500 joined #puppet-openstack
13:54 e-vad maybe this is a question better suited to #puppet, but, do y'all have anything that will enforce the style guide automagically, say in vim or sublime?
14:09 dprince joined #puppet-openstack
14:10 dmsimard joined #puppet-openstack
14:38 rcrit joined #puppet-openstack
15:43 badiane_ka joined #puppet-openstack
16:01 badiane_ka joined #puppet-openstack
16:07 xingchao joined #puppet-openstack
16:08 ryanycoleman joined #puppet-openstack
16:11 fvollero xingchao: ni hao! :)
16:12 fvollero fellas, there's a way to change a variable content based on the release version of the package?
16:17 ari_ joined #puppet-openstack
16:22 ari__ joined #puppet-openstack
16:24 xingchao fvollero: ni hao :)
16:26 xingchao fvollero: how about write custom facter to get the package version and use case statement to do that ?
16:27 ryanycoleman joined #puppet-openstack
16:29 fvollero xingchao: yeah :) i was writing this in fact :)
16:29 ari_ joined #puppet-openstack
16:29 ryanycol_ joined #puppet-openstack
16:30 fvollero xingchao: xiexie
16:30 xingchao fvollero: bu keqi :)
16:31 fvollero xingchao: wow, but bu keqi can also mean rude, or i'm wrong?
16:33 xingchao fvollero:
16:33 xingchao fvollero: bukeqi means your are welcome
16:34 fvollero xingchao: ok :)
16:37 hogepodge joined #puppet-openstack
16:41 hogepodge joined #puppet-openstack
16:49 ari_ joined #puppet-openstack
16:51 dprince joined #puppet-openstack
16:59 ryanycoleman joined #puppet-openstack
17:04 ryanycol_ joined #puppet-openstack
17:08 xingchao joined #puppet-openstack
17:18 tnoor joined #puppet-openstack
17:18 mgagne 3 years ago... http://projects.puppetlabs.com/issues/4145
17:30 ari joined #puppet-openstack
17:52 openstackgerrit A change was merged to stackforge/puppet-nova: Removes unnecessary api-paste.ini configuration  https://review.openstack.org/50915
17:54 ryanycoleman joined #puppet-openstack
18:02 ryanycoleman joined #puppet-openstack
18:07 grog500 joined #puppet-openstack
18:37 bodepd_ fvollero: only with a custom fact, or you could do it in a native type
18:37 bodepd_ e-vad: one of the gates is a lint check
18:37 bodepd_ puppet-lint
18:37 bodepd_ mgagne: can you implement that patch that I recomended in that ticket?
18:37 bodepd_ and test it
18:58 mgagne bodepd_: will test asap
18:58 ryanycoleman joined #puppet-openstack
18:59 openstackgerrit John Griffith proposed a change to stackforge/puppet-cinder: Add solidfire manifest  https://review.openstack.org/50736
19:01 bodepd_ obviously that fix needs to go in all three methods
19:01 bodepd_ using the latest from UCA updates, I do not wind up with a functional metadata service
19:01 bodepd_ (havana)
19:02 bodepd_ anyone run into this? Anyone have the metadata service working with havana?
19:02 mjblack joined #puppet-openstack
19:02 bodepd_ I'm thinking about just enabling config_drive for now to see if I can get around it
19:02 mjblack hmm?
19:03 bodepd_ mjblack: I tried to message you, but I had an autocomplete fail
19:03 mjblack lol
19:03 mjblack I was disconnected, internal network here doesnt seem to like freenode
19:03 bodepd_ have you tested the metadata service?
19:03 mjblack we tried but the problem is it needs l3
19:03 bodepd_ I keep a linode instance just for persistent irc :)
19:03 mjblack and that cant be done in an HA modal
19:04 mjblack *model
19:04 bodepd_ I may just punt on it for now, and enable config_drive
19:04 mjblack cloud init might need to be upgraded to support it
19:05 mjblack epel repo is a little behind, ubuntu cloud image works though
19:05 mjblack they changed I think in grizzly for it to be a cdrom and an older cloud init did not look for that as a device to check for the metadata
19:05 bodepd_ my failure is that 169.254.169.254 is not resolvable from the instances
19:06 mjblack are you doing l3 agent?
19:06 bodepd_ yeah, it may be an l3 failure.
19:06 bodepd_ I was going for the initial litmus test of, does anyone have it working :)
19:06 bodepd_ it could be a faliure in the l3 service (but I *can* allocate a floating ip)
19:07 mjblack did you enable the metadata proxy?
19:07 mgagne bodepd_: does iptables has the appropriate nating for 169.254.169.254 on the compute node?
19:07 bodepd_ proxy should be enabled.
19:08 bodepd_ hold on, let me check the nat rules
19:08 bodepd_ mgagne: there isn't , but I don't think there needs to be
19:09 bodepd_ b/c I am using ovs/gre
19:10 bodepd_ ip netns exec qrouter-05e54a09-eec1-4895-8531-0bc1ef1c863c iptables -t nat --list
19:10 bodepd_ I can see the metadata service rules in the network namespace
19:10 Kupo24z joined #puppet-openstack
19:10 bodepd_ (feel free to correct my bad terminology usage for this stuff, still learning these bits)
19:10 Kupo24z Hey all, does puppetlabs-openstack have compatibility for allowing migrations? (configuring SSH keys for nova user)
19:10 mgagne bodepd_: I don't use metadata service =)
19:11 mgagne Kupo24z: no, I had to do it myself
19:11 bodepd_ I thought jtopjian added something ages ago in the nova module
19:12 mgagne bodepd_: I think it is for libvirt live-migration support (which I tested today)
19:12 mgagne bodepd_: https://github.com/stackforge​/puppet-nova/commit/9844ebbd
19:13 Kupo24z Can you enable that with localized storage?
19:13 Kupo24z im using lvm as my nova image type
19:13 mgagne Kupo24z: that => live-migration?
19:13 Kupo24z (which i diddnt see an option to configure in the manifest btw)
19:13 Kupo24z yeah
19:15 mgagne Kupo24z: with Grizzly, you will have to have a shared storage in /var/lib/nova/instances for this to work or use --block-migrate when invoking live-migration
19:15 Kupo24z im using havana rc1
19:16 bodepd_ so,I ran tcpdump on the private router in the network namespace
19:16 mgagne Kupo24z: might be the same "limitation", will have to check the code. I'm checking in nova.virt.libvirt.driver. Look for check_can_live_migrate_source
19:18 bodepd_ great, it does work. but there appears to be a race condition somewhere b/c it does not work when my test instnace gets booted...
19:18 mjblack bodepd_: gre stinks :)
19:18 bodepd_ it's well beyond me to understand where that race condition lies...
19:19 bodepd_ maybe, I'll just add another sleeo 60 band-aid
19:19 bodepd_ s/speeo/sleep/
19:19 mjblack are you creating the namespace when you spin up the VM?
19:19 bodepd_ everything is created when the VM spins up
19:19 bodepd_ neutron doesit
19:20 mjblack right, so it is probably that delay thats doing that
19:20 bodepd_ oh, yes.
19:20 bodepd_ I am manually creating things with quantum commands
19:20 bodepd_ private nets/subnets router public net/subnet
19:20 mjblack I havent seen the namespace created until the first VM is launched
19:21 bodepd_ yeah. I forget which things are created when you launch the first VM vs. which things are created when you run neutron commands
19:21 mjblack if you pre-allocate a port I believe it will create the namespace
19:22 mjblack which you can easily test by creating a new subnet
19:23 hogepodge joined #puppet-openstack
19:27 bauzas joined #puppet-openstack
19:27 Kupo24z does the module have any LBaaS support?
19:29 ryanycoleman joined #puppet-openstack
19:29 mgagne bodepd_: puppetlabs-mysql is undergoing another refactor =)
19:36 dmsimard bodepd_, mgagne: Yup, it really breaks things.
19:39 hogepodge joined #puppet-openstack
19:51 qba73 joined #puppet-openstack
19:52 mjblack joined #puppet-openstack
19:52 mjblack bleh
19:54 mjblack bodepd_: speaking of ceph, is there a current candidate for the stackforge that is being considered?
19:54 ryanycoleman joined #puppet-openstack
19:58 bodepd_ mgagne: I haven't upgraded since 1.x
19:58 bodepd_ mgagne: I can't imagine that we will upgreade anytime soon
19:59 bodepd_ mgagne: I did see that someone was complaining about a failed openstack install b/c of version 2.x being installed
19:59 bodepd_ dmsimard: are you guys using 2.x for mysql?
19:59 bodepd_ dmsimard: mgagne if so, why?
19:59 bodepd_ mjblack: I just sent an email about that
19:59 bodepd_ mjblack: the plan that I propose is to start from scratch
19:59 bodepd_ mjblack: and allow people to suggest classes from other implementatiosn as patches
20:00 dmsimard dmsimard: No, I just happened to do a puppet module install puppetlabs/mysql and saw 2.0 instead of 1.0
20:00 bodepd_ mjblack: I've seen efforts to consolidate modules fail in the past
20:00 dmsimard muahah, I just replied to myself. I'm tired - that was meant for bodepd_
20:00 bodepd_ mjblack: b/c people can't agree what to start with
20:00 mjblack bodepd_: ah ok, yeah I was looking at some I found and I didnt really like them, and I was told they dont work with ceph-deploy
20:00 bodepd_ mjblack: we had a call with the various ceph developers
20:00 bodepd_ mjblack: that dachary (a ceph deve) setyup
20:01 bodepd_ setup
20:01 bodepd_ mjblack: that is what spurred all of this talk
20:01 bodepd_ mjblack: their recommendation was not to use ceph deploy for this
20:01 dmsimard mjblack, bodepd_: Might have to start from scratch regardless, enovance started puppet-ceph as agpl
20:01 dmsimard My understanding is that apache license would work best
20:01 mjblack bodepd_: interesting
20:02 bodepd_ dmsimard: in general, apache license is seen as the least restrictive
20:02 bodepd_ in that it doesn't scare lawyers
20:02 bodepd_ gpl has redistribution restrctions
20:02 bodepd_ which can limit coorporate acceptance which can inhibit use
20:03 bodepd_ dachary: I'm assuming you can live with Apache 2.0 ?
20:03 dachary I can live with that
20:03 bodepd_ dachary: given that is what everything else uses
20:03 bodepd_ dachary: this makes using code from gpl projects dificult
20:03 bodepd_ does someone want to volunteer to add it to stackforge?
20:04 bodepd_ I would just add the upstream as something with just a license file, and maybe a modulefile
20:04 bodepd_ we should add all of the puppet content as patches
20:05 bodepd_ If no one gets around to that, I can take a stab at doing it tomorrow
20:06 dachary I volunteer, the process is well explained
20:27 EmilienM bodepd_: who is going to be core reviewer ?
20:30 bodepd_ as of right now, I would say:
20:30 bodepd_ the current core reviewers
20:30 bodepd_ and that we should nominate some additional folks
20:31 bodepd_ EmilienM: ^^^
20:31 EmilienM make sense
20:31 bodepd_ this is one tricky thing about the current process
20:31 bodepd_ the granularity for reviews is across all projects
20:31 bodepd_ I think we will just have to trust people
20:31 bodepd_ that people know what projects they should be reviewing
20:31 digicloseup joined #puppet-openstack
20:31 EmilienM imo, there is two things to consider: skills on puppet and skills on ceph
20:32 bodepd_ and demonstrating understanding of development process
20:32 EmilienM good skills on ceph are rare
20:32 bodepd_ demonstrated
20:32 EmilienM sure
20:32 bodepd_ I'm happy to reveiw Puppet-side of things, but I should probably not be merging code
20:32 digicloseup joined #puppet-openstack
20:33 hogepodge joined #puppet-openstack
20:33 mgagne bodepd_, EmilienM: we can create a new group for that purpose if required
20:33 bodepd_ I would consider: you, francois, dmsimard (although he is slightly unknown in the comunity), and dachary
20:33 bodepd_ and am open to recommendations for other folks ;)
20:34 dmsimard I'm hyped about this - thanks for citing me :D
20:34 bodepd_ it that an all french group :)
20:34 bodepd_ s/it/is/
20:34 dmsimard If anything, mgagne is sitting right next to me and can guide me with all his wisdom
20:34 bodepd_ no, a mostly french group
20:34 EmilienM i would love to see a good synergy with cisco / mirantis guys also, since they made great job on different implementations
20:34 bodepd_ perhaps we should send out a note to the mailing list
20:35 bodepd_ from cisco: it's Don Talton
20:35 EmilienM great
20:35 dmsimard EmilienM, think Sebastien Han would have the time for something like this ?
20:35 bodepd_ mgagne: I think creating a separate review group is probably best
20:35 dmsimard He seems like the most knowledgeable person I know around Ceph
20:36 bodepd_ mgagne: its' either that or trust people to know what they should be merging
20:36 bodepd_ mgagne: what do you think?
20:36 EmilienM dmsimard: i think yes, but only if he agrees
20:36 mgagne bodepd_: I would prefer a dedicated group as I expect more people to join puppet-ceph than puppet-openstack
20:36 bodepd_ mgagne: funny :)
20:36 dmsimard EmilienM: Want to poke him, see what he thinks ? He can even weigh in on the discussion.
20:37 bodepd_ is he in here?
20:37 EmilienM mgagne: +1
20:37 dmsimard bodepd_: He's on #ceph
20:37 bodepd_ you mentioned you are collaborating with some folks from DT?
20:37 EmilienM dmsimard: no, you can reach him on IRC #ceph as "leseb"
20:38 dmsimard Yeah, I reached out to them, haven't heard back yet
20:38 dmsimard Dalgaaf said he would volounteer in managing upstream if it would be under the ceph namespace so I think he would be interested
20:38 bodepd_ perhaps we can draft a: call for contributors email on the mailing list and folks can just forward it to potentially interested parties
20:38 mgagne bodepd_: should I send a request for the group puppet-ceph-core to be created?
20:39 bodepd_ mgagne: sure. works for me.
20:39 digicloseup joined #puppet-openstack
20:39 mgagne bodepd_: for the project itself, are we planning on importing an upstream project?
20:39 bodepd_ mgagne: can you coordinate with dmsimard to get the patch for stackforge submitted as well?
20:39 bodepd_ mgagne: do we have to?
20:39 bodepd_ mgagne: if we do, I was thinking just put a
20:39 mgagne bodepd_: I don't know?
20:39 bodepd_ Modulefile and a LICENSE file
20:39 bodepd_ in a repo
20:39 mgagne bodepd_: upstream is not required
20:39 bodepd_ then, I would recommend against an upstream
20:40 mgagne bodepd_: I however expect someone to bootstrap something right after the project creation
20:40 dmsimard mgagne: Not sure we can straight up import upstream, license is agpl
20:40 mgagne dmsimard: true
20:40 mgagne bodepd_: lets create from scratch
20:40 bodepd_ mgagne: did you read that email I sent out :)
20:40 bodepd_ mgagne: perhaps you should :)
20:40 mgagne I read it
20:41 bodepd_ it says that we should start from scratch
20:42 mgagne alright
20:42 EmilienM now, it's a matter of design discussion :)
20:46 mgagne bodepd_: should we require CLA for now?
20:46 bodepd_ mgagne: I'm all about the CLAs
20:46 bodepd_ it's easier to start with them, then to have to add them
20:47 bodepd_ although perhaps they are inappropriate given they are openstack specific
20:47 bodepd_ is there a lawyer in here :)
20:53 mgagne bodepd_: do we want the ircbot to announce new changes and merges on #puppet-openstack?
20:54 bodepd_ mgagne: I will refer to someone else
20:54 EmilienM mgagne: since it's on stackforge near other modules, i would say yes
20:54 bodepd_ mgagne: I plan on moving away from that module
20:54 bodepd_ https://review.openstack.org/#/c/50915/
20:54 bodepd_ as an FYI, should I revert that commit?
20:55 mgagne bodepd_: that module?!
20:55 bodepd_ I misunderstood the last message before I committed
20:55 mgagne bodepd_: I meant the puppet-openstack irc channel, not the project on stackforge
20:55 bodepd_ oh, I 'm confused...
20:55 bodepd_ I vote for yes
20:55 mgagne bodepd_: lets create a new change which would purge old configs.
20:56 bodepd_ mgagne: ok. I was thinking I may have messed up there
20:56 bodepd_ mgagne: so you vote 'no revert'
20:56 mgagne no revert
20:56 mgagne BUT the side-effect is that new configs won't be used as api-paste.ini has more priority
20:57 mgagne bodepd_: else we plan on fixing right now, then maybe we should revert
20:57 mgagne bodepd_: do we have an agreement on Apache Licence?
21:01 bodepd_ mgagne: I don't have time for it this week
21:01 mgagne bodepd_: lets revert for now
21:01 bodepd_ mgagne: I feel that apache license is non-negotiable
21:01 mgagne bodepd_: alright
21:01 bodepd_ is martin in irc?
21:03 openstackgerrit Dan Bode proposed a change to stackforge/puppet-nova: Given the current precedence of configuration lookup, this patch should never have been accepted without code that also removes the paste_api configuration.  https://review.openstack.org/52213
21:03 dmsimard EmilienM: No replies when trying to ping Sebastien on #ceph - I asked you, thinking you might have some privileged contact :)
21:04 openstackgerrit Dan Bode proposed a change to stackforge/puppet-nova: Revert "Removes unnecessary api-paste.ini configuration"  https://review.openstack.org/52214
21:05 EmilienM dmsimard: it's 11.05pm
21:05 EmilienM myself i should not being here
21:05 EmilienM but quite exciting discussion around ceph tonight
21:06 dmsimard EmilienM: Oh, I forgot about that - You both are in Paris ?
21:09 mgagne here we go guys: https://review.openstack.org/#/c/52215/
21:14 fc__ joined #puppet-openstack
21:16 otherwiseguy joined #puppet-openstack
21:27 dmsimard Ah, I see Danny (Dalgaaf) picked up on that e-mail I sent him, he's replying to a few posts
21:44 hogepodge can gerrit-bot post to multiple channels? Are there any other channels people would want notification on?
21:48 dmsimard hogepodge: Looking at documentation, it seems like it is possible to post to various channels - http://ci.openstack.org/irc.html#gerritbot
21:49 hogepodge From the standpoint of community modules outside of OpenStack, would there be interest in someone like Puppet Labs hosting a Gerrit for community module development?
21:50 dmsimard I was thinking about that too - can't hide that one of the major factors in creating a puppet-ceph initiative in stackforge is because of the review and other measures in place
21:51 dmsimard And there was a debate about ceph not being an openstack component, yet openvswitch was recently added
21:51 dmsimard The need is to have a common, active repository with rigorous review/test processes in place - for the time being stackforge fulfills that need
21:52 dmsimard So in that respect, there might be a niche to fill, here.
21:53 hogepodge Right. I like Gerrit, and I think I could make a convincing case internally for us to have something similar.
21:53 hogepodge It wouldn't happen overnight, though.
21:57 michchap joined #puppet-openstack
21:59 hogepodge It's something I want though (mgagne)
21:59 hogepodge mgagne: re latest e-mail :-)
21:59 mgagne hogepodge: openstack-infra has a great knowledgeable team able to manage and debug (weird problems) all those tools. I'm not sure PL or ourselves would be able to manage (as well) such infra.
22:00 hogepodge magane: yup, true.
22:00 mgagne hogepodge: I don't want to end up with a half-baked ci infra =)
22:00 hogepodge mgagne: that's why we'd borrow from stackforge :-)
22:04 ari__ joined #puppet-openstack
22:32 e1mer joined #puppet-openstack
22:43 badiane_ka joined #puppet-openstack
22:48 ari joined #puppet-openstack
22:58 dachary mgagne: in your opinion, is it better for a puppet module to rely on a cli or a rest interface when given the choice ?
22:58 odyi CI Puppet Labs does, our testing matrix is fairly large.  Code review not so much and I don't know of any plans in Engineering to take on gerrit or similar.
22:58 mgagne dachary: I learned the hardway that cli isn't that stable
22:59 dachary in the case of Ceph it really is the same code so it would not matter https://github.com/ceph/ceph/blob/​master/src/mon/MonCommands.h#L144 ( see cli,rest )
22:59 dachary matter as in stability
23:00 odyi When I do type/provider work I gravitate to the thing that is more easily parsible.
23:00 dachary ok
23:01 mgagne odyi: you haven't learned yet =)
23:01 mgagne odyi, dachary: in fact, it depends on the stability of the interface (cli vs api)
23:02 odyi APIs are more easily versioned than a cli which means the remote API is more predictable/stable.
23:20 ari joined #puppet-openstack
23:21 dalgaaf joined #puppet-openstack
23:47 ryanycoleman joined #puppet-openstack
23:49 ryanycol_ joined #puppet-openstack
23:54 ari left #puppet-openstack
23:55 tnoor left #puppet-openstack
23:55 michchap joined #puppet-openstack
23:57 openstackgerrit A change was merged to stackforge/puppet-neutron: Fix file permissions on examples/neutron.pp  https://review.openstack.org/51605

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