Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2014-08-04

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

All times shown according to UTC.

Time Nick Message
00:49 shakamunyi joined #puppet-openstack
00:52 uberj joined #puppet-openstack
01:10 tfz joined #puppet-openstack
01:29 xingchao joined #puppet-openstack
01:35 gildub joined #puppet-openstack
01:48 mmaglana joined #puppet-openstack
01:53 shakamunyi joined #puppet-openstack
01:55 topshare joined #puppet-openstack
02:01 tfz joined #puppet-openstack
02:36 shakamunyi joined #puppet-openstack
03:40 saju_m joined #puppet-openstack
04:19 daneyon joined #puppet-openstack
04:27 daneyon joined #puppet-openstack
04:46 daneyon joined #puppet-openstack
05:35 mmagr joined #puppet-openstack
05:42 topshare joined #puppet-openstack
05:42 marun joined #puppet-openstack
05:45 sputnik13 joined #puppet-openstack
06:19 saju_m joined #puppet-openstack
06:21 bodepd joined #puppet-openstack
06:24 ddmitriev joined #puppet-openstack
06:38 EmilienM good morning o/
06:38 EmilienM (back from 2 weeks off, have a lot to catch up :) )
06:40 sputnik13 joined #puppet-openstack
06:42 topshare joined #puppet-openstack
06:49 beddari try 4 weeks :)
06:54 andi- joined #puppet-openstack
07:03 sputnik13 joined #puppet-openstack
07:21 holser joined #puppet-openstack
07:41 comptona joined #puppet-openstack
07:45 saju_m joined #puppet-openstack
07:48 artem_panchenko joined #puppet-openstack
07:51 openstackgerrit Michael Chapman proposed a change to stackforge/puppet_openstack_builder: Add ceilometer to stacktira  https://review.openstack.org/111667
07:51 artem_panchenko joined #puppet-openstack
07:51 michchap EmilienM: morning :)
07:51 michchap EmilienM: welcome back
08:07 EmilienM michchap: hey, thanks !
08:09 michchap EmilienM: I have a question for you if you are back working already :)
08:09 EmilienM michchap: shoot
08:10 michchap EmilienM: we started deploying everything behind SSL termination, and we are finding that the ceilometer client doesn't support SSL. There is a patch in upstream that fixes it. Are you packaging your own client for people or anything like that?
08:14 openstackgerrit Michael Chapman proposed a change to stackforge/puppet_openstack_builder: Add ceilometer to stacktira  https://review.openstack.org/111667
08:15 EmilienM michchap: we don't
08:15 EmilienM michchap: I reported the bug in ceilo
08:15 artem_panchenko joined #puppet-openstack
08:16 michchap EmilienM: I tested https://review.openstack.org/#/c/68939/ and it's working for me
08:16 EmilienM michchap: awesome, is it going to be backported?
08:17 michchap EmilienM: I have no idea. I was going to ask you what the status of RDO packages might be - I think it's a pretty critical bug and it should be backported and pushed into packages ASAP, but it's not up to me.
08:18 sputnik13 joined #puppet-openstack
08:53 openstackgerrit Semyon Deviatkin proposed a change to stackforge/puppet-keystone: Install dependency python-memcache when token driver memcache.  https://review.openstack.org/105015
09:15 saju_m joined #puppet-openstack
09:47 tdb joined #puppet-openstack
09:54 EmilienM michchap: so you are using HAproxy 1.5 ?
09:56 michchap EmilienM: yep. same as you I think?
09:56 EmilienM michchap: yup
09:57 EmilienM michchap: you can have a look at our puppet module if you need inspiration :P
09:57 michchap I have all the SSL endpoints working already, but haven't made internal comms SSL yet
09:58 EmilienM michchap: except for Spice and cloud-init, everything works fine for me in SSL
09:59 michchap EmilienM: and ceilometer client?
10:00 EmilienM indeed
10:00 EmilienM https://github.com/enovance/openstack-yaml-infra​-3nodes/blob/master/data/type.yaml.tmpl#L82-L84
10:00 EmilienM michchap: I use HTTP
10:01 michchap EmilienM: :(
10:02 EmilienM michchap: there is a LOT of pending patches in puppet modules. I try to make some, but I would need some help from other core
10:02 michchap EmilienM: I am slowly making progress on them
10:03 EmilienM michchap: I'm happy to see progress on https://review.openstack.org/#/c/68939/ tho
10:07 sbadia joined #puppet-openstack
10:07 michchap EmilienM: I spent a lot of time working on the vswitch patch
10:07 michchap lots of testing
10:08 EmilienM michchap: oh, puppet-vswitch?
10:08 michchap EmilienM: yeah there is a large refactor that preserves interface address
10:08 EmilienM michchap: let me know if you need reviews
10:09 michchap EmilienM: we were using keep_ip so we could have network node and public API services on the same machine
10:09 michchap EmilienM: but this patch is much neater
10:09 michchap EmilienM: test it out on your redhat builds: https://review.openstack.org/#/c/94504/
10:13 EmilienM michchap: ok, I'll
10:29 xingchao joined #puppet-openstack
10:52 britthouser joined #puppet-openstack
11:16 michchap EmilienM: i'm also seeing something that doesn't look quite right. We are setting keystone::admin_endpoint to $address:$port/v2.0/ but this doesn't work with the proposed upstream patch for ceilometer client. Reading the doc for admin_endpoint it's supposed to be just the base, not include the version string. This is a big problem because we use it in the native types so it's hard to change :(
11:25 EmilienM michchap: indeed. I'm curious I did not see that before
11:26 michchap EmilienM: I am trying to determine why the other clients don't seem to notice that it's not correct. If you have a puppet-deployed cloud handy, try doing a curl -i -X GET http://openstack.aptira.com:5000/ -H "Accept: application/json" -H "User-Agent: python-keystoneclient"
11:26 michchap you can see in the links sections it's not right.
11:29 openstackgerrit A change was merged to stackforge/puppet-designate: Hide secrets from puppet logs  https://review.openstack.org/106529
11:30 michchap if we remove the admin_endpoint code from the keystone provider so we can set it correctly, then we have to rely on the bind host and bind port, which are not right in the case where we use a load balancer
11:33 EmilienM michchap: right
11:33 EmilienM michchap: and ssl termination
11:33 EmilienM but I think that's what you meant
11:33 michchap I don't see the keystone version being stored in keystone.conf anywhere so we would have to assume v2.0 as well
11:34 michchap well the bind host might be the local address but the endpoint should point at the vip
11:34 michchap and yeah...no protocol info anywhere
11:34 michchap any ideas?
11:34 rharrison joined #puppet-openstack
11:39 michchap I think "${admin_endpoint}/v2.0" is probably the best we can do right now.
11:39 sputnik13 joined #puppet-openstack
11:40 michchap because the endpoint will have the protocol in there.
11:41 EmilienM michchap: I agree with this one
11:41 EmilienM it's not breaking anything anyway
12:03 mtl1 joined #puppet-openstack
12:04 EmilienM michchap: I'm looking at all patches to move to openstacklib::mysql. ie: https://review.openstack.org/#/c/111073
12:04 EmilienM sbadia: michchap: if you have time to have a look also
12:09 uberj_ joined #puppet-openstack
12:11 toabctl_ joined #puppet-openstack
12:12 tdb_ joined #puppet-openstack
12:21 ianw joined #puppet-openstack
12:23 mattymo joined #puppet-openstack
12:27 michchap EmilienM: because I already have a module called openstacklib I need some time before I can test those changes
12:27 michchap EmilienM: because I have to move everything
12:28 EmilienM michchap: I see
12:34 rharrison joined #puppet-openstack
12:36 dprince joined #puppet-openstack
12:37 tfhartmann joined #puppet-openstack
12:41 openstackgerrit A change was merged to stackforge/puppet-keystone: keystone_user_role supports email as username  https://review.openstack.org/108677
12:49 rmoe joined #puppet-openstack
12:53 morazi joined #puppet-openstack
13:05 andriikolesnikov joined #puppet-openstack
13:14 Mr_D joined #puppet-openstack
13:31 ric` joined #puppet-openstack
13:32 rharrison joined #puppet-openstack
13:32 bodepd joined #puppet-openstack
13:43 openstackgerrit Michael Chapman proposed a change to stackforge/puppet-keystone: Fix usage of admin_endpoint  https://review.openstack.org/111729
13:44 michchap EmilienM: ^^ I added a warning, but it isn't backwards compatible.
13:46 EmilienM michchap: will review in minutes
13:59 thartmann joined #puppet-openstack
14:00 openstackgerrit A change was merged to stackforge/puppet-nova: Add support for os_region_name  https://review.openstack.org/111006
14:02 michchap EmilienM: 110561 and 111019 are opposites of each other. I don't quite understand the motivation for removing the path from an exec - is it for people installing from source to /opt or something?
14:02 EmilienM michchap: indeed
14:04 openstackgerrit A change was merged to stackforge/puppet-horizon: Enable support for secure cookies  https://review.openstack.org/111013
14:06 EmilienM michchap: have a short meeting now. Look after
14:06 michchap EmilienM: no problem. catch you tomorrow
14:06 EmilienM michchap: indeed, it's late for you
14:12 berendt joined #puppet-openstack
14:16 Mr_D joined #puppet-openstack
14:22 daneyon joined #puppet-openstack
14:23 daneyon joined #puppet-openstack
14:36 hogepodge joined #puppet-openstack
14:40 hogepodge joined #puppet-openstack
14:46 mfisch mgagne: can you look at +2ing this since EmilienM won't +2 his own work (understandably): https://review.openstack.org/#/c/111019/
14:47 hogepodge Good morning!
14:47 hogepodge merged mfisch
14:47 mfisch morning!
14:48 mfisch hogepodge: you're on early
14:48 hogepodge 8am meeting
14:48 mfisch ugh
14:48 hogepodge not that I don’t want to be in bed
14:48 EmilienM hogepodge: hey
14:48 crinkle morning
14:48 mfisch I get to drop 3 forks this morning, so excited
14:49 * EmilienM wonders if we have a meeting /agenda for today
14:50 openstackgerrit A change was merged to stackforge/puppet-cinder: cinder-type: delete PATH in exec  https://review.openstack.org/111019
14:50 Mr_D joined #puppet-openstack
14:51 hogepodge EmilienM no agenda, but you’re welcome to suggest anything.
14:51 Mr_D left #puppet-openstack
14:51 hogepodge EmilienM Welcome back!
14:53 EmilienM hogepodge: thanks
14:53 EmilienM hogepodge: a lot of puppet patches in progress, isn't it?
14:55 hogepodge There are a bunch with -1 workflows
14:55 hogepodge Looks like a few need to be rebased
14:55 hogepodge crinkle just sent a bunch of to use the common mysql code \o/
14:56 crinkle ya feedback welcome :)
14:56 prad joined #puppet-openstack
14:56 prad joined #puppet-openstack
14:57 prad joined #puppet-openstack
14:57 crinkle I'm also looking for feedback on https://review.openstack.org/#/c/107179/
14:57 EmilienM crinkle: awesome
14:58 crinkle the problem I was facing when I mentioned it didn't quite work in the comments was that the glance db wasn't syncing correctly
14:58 crinkle but that code doesn't handle syncing, the other modules do
14:58 crinkle so I think the warning should be handled differently
14:59 mdorman joined #puppet-openstack
14:59 Mr_D joined #puppet-openstack
15:01 hogepodge crinkle I think it still belongs in lib since it’s a un underlying problem with the postgres infrastructure (particularly the python migrate modules) rather than with the db_sync commandst themselves.
15:01 hogepodge plus it keeps the warning code in one place, improving maintainability. Opinions EmilienM or mfisch?
15:03 crinkle okay, I should probably at least make the reasoning and possible mitigations more clear in the message
15:04 hogepodge manually upgrading the python libs would be sufficient?
15:04 mfisch sorry I dont know enough to have an opinion
15:04 crinkle you can't really upgrade them because they require python > 2.6
15:04 crinkle i mean it's not that easy to do via yum at least
15:05 hogepodge Right, it’s not easy in general.
15:05 hogepodge the dependency on the python-libvirt libraries will bite you when you try to go to 2.7 (speaking from experience)
15:05 crinkle you could avoid the puppet error by not doing the sync at all, or doing it manually somehow
15:05 hogepodge You can use an upgrade in a virtual environment with manually installed 2.7 libraries
15:06 hogepodge The sync has to happen at least once to set up the database schema
15:06 crinkle yeah
15:06 hogepodge As CentOS 7 becomes more widely used this will be less of an issue
15:07 crinkle :)
15:07 crinkle i'll have a new patch shortly
15:08 sputnik13 joined #puppet-openstack
15:13 mjblack joined #puppet-openstack
15:15 mjblack_ joined #puppet-openstack
15:16 mfisch hogepodge: if we ever look to moving this meeting, this is our team's daily stand-up, so I'd vote for 30 mins later so you can sleep more too
15:18 xarses mfisch: sleep? what's that?
15:21 hogepodge mfisch yeah, my dog has usually dragged me out of bed by 5:30 anyways. She’s not feeling well today though, so we slept in.
15:21 hogepodge but we can consider moving it back. It’s as early as possible in consideration of our australian friends
15:22 mfisch hogepodge: 5:30 thats worse than kids
15:27 mmaglana joined #puppet-openstack
15:38 daneyon joined #puppet-openstack
15:47 daneyon joined #puppet-openstack
15:49 xarses joined #puppet-openstack
15:51 rwsu joined #puppet-openstack
15:57 rmoe joined #puppet-openstack
16:02 aimon joined #puppet-openstack
16:03 openstackgerrit_ joined #puppet-openstack
16:21 ikke-t joined #puppet-openstack
16:22 hogepodge joined #puppet-openstack
16:22 openstackgerrit A change was merged to stackforge/puppet-ceph: Refactor to remove inheritance in profiles  https://review.openstack.org/109989
16:23 mmaglana joined #puppet-openstack
16:32 saju_m joined #puppet-openstack
16:50 saju_m joined #puppet-openstack
16:57 mmaglana joined #puppet-openstack
17:00 openstackgerrit A change was merged to stackforge/puppet-glance: Hide secrets from puppet logs  https://review.openstack.org/99294
17:03 mgagne How can I make this work? service { 'whatever': ensure => stopped } if the package needed to have the service is not installed? I see this bug but no workaround: https://tickets.puppetlabs.com/browse/PUP-2188
17:07 openstackgerrit A change was merged to stackforge/puppet-trove: First commit in Stackforge  https://review.openstack.org/104788
17:09 mgagne hogepodge: ^
17:18 hogepodge magane nice
17:18 hogepodge More projects to follow
17:24 mgagne hogepodge: didn't do anything for trove, just got added to reviewers ^^'
17:26 mgagne hogepodge: my ^ was more about the puppet bug ^^'
17:29 shakamunyi joined #puppet-openstack
17:30 andriikolesnikov joined #puppet-openstack
17:34 mgagne hogepodge: Don't you think this kind of settings should be moved into its own manifest? https://review.openstack.org/#/c/106055/2
17:34 mgagne hogepodge: ceilometer::logging for example
17:34 hogepodge mgagne you mean its own class?
17:34 mgagne hogepodge: yes, that's what I meant
17:35 hogepodge yeah, that makes sense
17:35 mgagne hogepodge: but then people will expect to have logging settings configured right off the bat just be including ceilometer::init...
17:36 hogepodge I guess there was already a log parameter there, which is why I +2ed it
17:36 hogepodge It’s acutually probably best to just encourage conf
17:36 hogepodge that’s what it’s there for
17:36 mgagne hogepodge: I understand. it's not the first time someone proposed adding advanced logging settings
17:36 mgagne hogepodge: so I'm looking for a good pattern we can reuse across our modules without overloading the init class
17:37 hogepodge How common is the code? Custom resource in lib or extras?
17:37 mgagne hogepodge: I'm not suggesting creating a lib
17:38 mgagne hogepodge: just having a common pattern across our modules
17:38 mgagne hogepodge: if it goes in init, so be it. if it goes in *::logging, so be it but I want a clear consensus around it
17:39 mgagne hogepodge: so that if we want to manage logging configs, we do it for ALL modules, not just one
17:42 hogepodge Can we drop a message to the mailing list? I can 0 my review and ask for a spec
17:44 mgagne hogepodge: it could be a general consensus in the ML too
17:45 mgagne hogepodge: but tbh, the more I think about it, might be better to leave them in ::init but make sure it's done across all modules in the same way
17:46 marun joined #puppet-openstack
17:47 hogepodge mgagne I’ll write an e-mail later today with the four options options. config class, params in init across all modules, params in class across all modules, or defined resource in lib/extras
17:48 hogepodge (unless you want to send it)
17:48 mgagne hogepodge: I don't like the defined resource in lib/extras, I don't see the value of wrapping something without much code/logic in it
17:49 hogepodge I don’t like having a lot of repetition
17:50 mgagne hogepodge: how does having *::db::mysql not a repitition?
17:50 mgagne repitition*
17:50 mgagne damn, whatever
17:50 mgagne repetition*
18:04 berendt joined #puppet-openstack
18:24 higgins joined #puppet-openstack
18:25 thumpba joined #puppet-openstack
18:28 hogepodge mgagne https://review.openstack.org/#​/c/111073/1/manifests/db/mysql.pp,cm
18:29 hogepodge mgagne things that are happening in places like that are because of our deprecation policy. But we’re still cutting down code copy and paste. the modules were drifting from one another. Now if we want to add db logic it only goes into one place
18:30 mgagne ok?
18:32 hogepodge guess we’re getting it wrong then?
18:34 mgagne hogepodge: no
18:35 mgagne hogepodge: a mysql database is a mysql database, it's not openstack specific
18:36 mgagne hogepodge: configs might be different between services
19:27 morazi joined #puppet-openstack
19:36 tfz joined #puppet-openstack
19:50 tfz joined #puppet-openstack
20:04 openstackgerrit Ivan Chavero proposed a change to stackforge/puppet-neutron: Fixes plugin.ini error  https://review.openstack.org/111816
20:33 EmilienM mgagne: thx for review
20:49 morazi joined #puppet-openstack
20:57 hogepodge_ joined #puppet-openstack
21:35 hogepodge joined #puppet-openstack
21:45 xarses_ joined #puppet-openstack
21:58 openstackgerrit Colleen Murphy proposed a change to stackforge/puppet-openstacklib: Add db::postgresql to openstacklib  https://review.openstack.org/107179
22:04 openstackgerrit Colleen Murphy proposed a change to stackforge/puppet-openstacklib: Add db::postgresql to openstacklib  https://review.openstack.org/107179
22:06 imcsk8 markvoelker: hello, can you take a look to this patch? https://review.openstack.org/#/c/111816/ thanks!
22:15 snk joined #puppet-openstack
22:17 prad joined #puppet-openstack
22:21 mjblack joined #puppet-openstack
22:35 ByteSore joined #puppet-openstack
22:52 daneyon joined #puppet-openstack
22:57 aimon hogepodge openstack::profile::keystone still requires hiera
22:57 aimon there are two hiera vars in the: users and tenants
22:58 hogepodge aimon indeed
22:58 aimon was this missed or was there a reawson to leave them?
22:58 hogepodge I was using trickery, and hadn’t figured out a way out of it yet.
22:58 hogepodge aimon, so it was a bit of both
22:58 aimon :)
22:58 aimon so what’s the plan ?
23:00 aimon if you want to turn them into config vars I can submit a PR.. or.. you can do it of you have time.. or is there another way?
23:01 hogepodge aimon I think that should be the plan. It should be backwards compatible
23:01 aimon ok.. need extra hands?
23:01 hogepodge I want to get a new release out for it. More hands are always welcome
23:02 aimon ok it is blocking me.. so I am glad to make the change
23:02 aimon I’ll have it to you tomorrow
23:02 aimon or later this evening
23:16 aimon omg no tests.. ok in that case I am testing change now and will ahve a PR in a few
23:37 marun joined #puppet-openstack
23:39 rwsu joined #puppet-openstack
23:41 aimon hogepodge here you go: https://github.com/puppetlabs​/puppetlabs-openstack/pull/82
23:42 aimon btw I’ll have nova_netowrk as an option in the next day or two
23:46 aimon joined #puppet-openstack

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