Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2015-05-29

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

All times shown according to UTC.

Time Nick Message
00:05 Hunner bundle++
00:14 aimon joined #puppet-openstack
00:31 aimon joined #puppet-openstack
01:00 xarses joined #puppet-openstack
01:13 markvoelker joined #puppet-openstack
01:13 ducttape_ joined #puppet-openstack
01:18 nibalizer spredzy|afk: ping on https://github.com/puppetlabs/puppetlabs-stdlib/pull/444
01:43 kkirkpatrick joined #puppet-openstack
01:46 openstackgerrit Gilles Dubreuil proposed stackforge/puppet-keystone: Restructures authentication for resource providers  https://review.openstack.org/181299
02:07 openstackgerrit Gilles Dubreuil proposed stackforge/puppet-keystone: Restructures authentication for resource providers  https://review.openstack.org/181299
02:08 ducttape_ joined #puppet-openstack
02:24 gildub joined #puppet-openstack
02:34 ducttape_ joined #puppet-openstack
02:38 openstackgerrit Cody Herriges proposed stackforge/puppet-neutron: Make tests pass on Puppet 4.x  https://review.openstack.org/184646
02:39 ducttape_ joined #puppet-openstack
02:59 ducttape_ joined #puppet-openstack
03:02 markvoelker joined #puppet-openstack
03:08 ducttape_ joined #puppet-openstack
03:18 kkirkpatrick joined #puppet-openstack
03:59 openstackgerrit Matt Fischer proposed stackforge/puppet-keystone: Add support for Fernet Tokens  https://review.openstack.org/185441
04:02 markvoelker joined #puppet-openstack
04:07 _ody joined #puppet-openstack
04:29 sks joined #puppet-openstack
04:31 fedexo joined #puppet-openstack
04:34 ducttape_ joined #puppet-openstack
04:45 jerryz joined #puppet-openstack
05:00 gildub joined #puppet-openstack
05:14 aimon joined #puppet-openstack
05:52 markvoelker joined #puppet-openstack
06:13 jtomasek joined #puppet-openstack
06:16 aimon joined #puppet-openstack
06:40 gfidente joined #puppet-openstack
06:48 sks joined #puppet-openstack
06:56 arnaud_orange joined #puppet-openstack
07:21 beddari joined #puppet-openstack
07:31 paramite joined #puppet-openstack
07:36 jpena joined #puppet-openstack
07:38 jistr joined #puppet-openstack
07:40 markvoelker joined #puppet-openstack
08:17 ducttape_ joined #puppet-openstack
08:17 aimon joined #puppet-openstack
08:23 derekh joined #puppet-openstack
08:24 tfz joined #puppet-openstack
08:58 ducttape_ joined #puppet-openstack
09:14 paramite joined #puppet-openstack
09:29 markvoelker joined #puppet-openstack
09:34 paramite joined #puppet-openstack
09:46 openstackgerrit sac proposed stackforge/puppet-manila: Add share driver options.  https://review.openstack.org/186699
09:59 ducttape_ joined #puppet-openstack
10:19 aimon joined #puppet-openstack
11:01 ducttape_ joined #puppet-openstack
11:12 wds__ joined #puppet-openstack
11:12 wds___ joined #puppet-openstack
11:18 markvoelker joined #puppet-openstack
11:21 panda joined #puppet-openstack
11:28 maximov joined #puppet-openstack
11:28 mwhahaha joined #puppet-openstack
11:39 cdelatte joined #puppet-openstack
11:59 markvoelker joined #puppet-openstack
12:01 dprince joined #puppet-openstack
12:02 rcallawa_ joined #puppet-openstack
12:02 ducttape_ joined #puppet-openstack
12:20 ducttape_ joined #puppet-openstack
12:32 morazi joined #puppet-openstack
13:24 openstackgerrit sac proposed stackforge/puppet-manila: Add share driver options.  https://review.openstack.org/186699
13:26 kkirkpatrick joined #puppet-openstack
13:29 gsilvis joined #puppet-openstack
13:34 dfisher joined #puppet-openstack
13:36 ducttape_ joined #puppet-openstack
13:59 ferest_ joined #puppet-openstack
14:03 aimon joined #puppet-openstack
14:15 gsilvis joined #puppet-openstack
14:23 rcallawa_ joined #puppet-openstack
14:25 dfisher i can't remember - were there plans to do another Juno release of the modules?
14:34 ducttape_ joined #puppet-openstack
14:36 openstackgerrit Richard Raseley proposed stackforge/puppet-nova: Enable nova service by default  https://review.openstack.org/184656
14:37 kkirkpatrick joined #puppet-openstack
14:37 openstackgerrit Dan Prince proposed stackforge/puppet-tripleo: Configure virtual IPs for split out networks  https://review.openstack.org/186240
14:37 openstackgerrit Dan Prince proposed stackforge/puppet-tripleo: Use node IP lists for HA Proxy ipaddresses  https://review.openstack.org/186775
14:38 openstackgerrit Dan Prince proposed stackforge/puppet-tripleo: Add interface_for_ip function  https://review.openstack.org/186776
14:39 chem joined #puppet-openstack
14:39 stevemar joined #puppet-openstack
14:47 jistr joined #puppet-openstack
15:18 openstackgerrit sac proposed stackforge/puppet-manila: Add share driver options.  https://review.openstack.org/186699
15:22 panda joined #puppet-openstack
15:25 marvin999 joined #puppet-openstack
15:27 marvin999 joined #puppet-openstack
15:30 mdorman joined #puppet-openstack
15:45 tfz joined #puppet-openstack
15:51 cwolferh joined #puppet-openstack
16:06 jistr joined #puppet-openstack
16:14 rwsu joined #puppet-openstack
16:18 RichardRaseley joined #puppet-openstack
16:20 openstackgerrit Richard Raseley proposed stackforge/puppet-nova: Improve Endpoint Configurability  https://review.openstack.org/186109
16:32 xarses joined #puppet-openstack
16:41 jerryz joined #puppet-openstack
16:47 dfisher ok, dumb question time!
16:48 ducttape_ joined #puppet-openstack
16:48 mdorman no such thing
16:48 dfisher in Neutron's params.pp file, various fields are set to false.  in other params.pp files, it's undef
16:48 dfisher oh, you just wait.  I can astound you with my idiocy.
16:48 dfisher is there a difference between false and undef?
16:49 dfisher http://docs.puppetlabs.com/puppet/3.8/reference/lang_datatypes.html#undef  — says that when used as a boolean, undef is false … i assume that's what these are used for?
16:52 clayton false is defined :)
16:56 dfisher :)
16:56 mdorman yeah there is a difference
16:57 RichardR_ joined #puppet-openstack
16:58 mdorman i mean it all depends on the context.  there are probably certain params that are intended to default explicitly to false, vs. being undef/not present which then actually get set to some other default value… and/or removed from neutron.conf and therefore neutron uses whatever default it has for it.
16:58 dfisher ok.  will keep my eye out on this one.
16:58 dfisher neutron for us is … hard.
16:58 dfisher none of the existing linux neutron drivers work on Solaris.
16:59 dfisher so, we have our own SDN.
16:59 dfisher which doesn't work on Linux.
16:59 dfisher naturally.
16:59 dfisher ODL will save us all … or so I'm told.
17:00 ferest joined #puppet-openstack
17:01 mdorman hah
17:01 mdorman is there openvswitch for solaris?
17:02 dfisher nope.
17:02 mdorman fun
17:02 dfisher there's some work in that space but it's not ready.
17:02 dfisher and i'll likely be client-only
17:02 dfisher it'll*
17:02 dfisher also, mdorman:  thanks for the explanation of undev vs. false.
17:02 dfisher undef.
17:03 dfisher stupid fingers.  type what I mean!
17:03 barthalion I know that feeling
17:03 barthalion they disobey me as well
17:03 RichardR_ dfisher: I was also told these things about ODL. =]
17:03 mdorman dfisher:  np
17:03 dfisher I can barely spell ODL let alone understand it.
17:04 RichardR_ dfisher: +1
17:05 dfisher ok, so .. another fun question.
17:05 dfisher metadata.json has dependencies on puppet modules that flat out won't work on Solaris
17:05 dfisher nova has
17:05 dfisher { "name": "duritong/sysctl", "version_requirement": ">=0.0.1 <1.0.0" },
17:05 dfisher once I get all this stuff working on Solaris
17:05 dfisher and the required upstream stuff done.
17:06 dfisher what do I do about that file?   looking up metadata.json docs didn't say anything about optional dependencies
17:06 dfisher i mean, i guess if they get downloaded, they sit there and do nothing on Solaris so .. no harm
17:07 * RichardR_ looking at why that is a requirement.
17:07 dfisher or -    { "name": "stackforge/vswitch", "version_requirement": ">=1.0.0 <2.0.0" }, for neutron
17:07 Fin1te joined #puppet-openstack
17:07 dfisher swift:  -    {"name":"puppetlabs/xinetd","version_requirement":">=1.0.1 <2.0.0"},
17:08 dfisher etc.
17:09 RichardR_ dfisher: Is 'Solaris Support' captured as a bug report? I am sure we have some people internally which would love to help bring some of that stuff up to snuff.
17:09 dfisher not yet.
17:09 dfisher i asked in Vancouver
17:09 dfisher and was told …. $$$
17:09 dfisher well, for official puppetlabs stuff
17:10 dfisher like … puppetlabs/apache, mysql, rabbitmq, memcached, etc.
17:10 RichardR_ dfisher: Meh. The real answer is more nuanced than that.
17:10 dfisher lack of expertise?
17:11 dfisher so, I'm doing the ohsofun legal work to bring in the puppetlabs-specific modules formally for Solaris.  Part of that work will be sending out pull requests for these modules.
17:12 dfisher shit, I have 25+ resource types/providers for Solaris that I want to give you (and have legal permission to do so) for puppet core.  They lack unittests so I can't give them to you :(
17:12 RichardR_ dfisher: It's a combination of factors that aren't really worth detailing. Part of it is definitely expertise.
17:12 RichardR_ dfisher: Are they in a PR somewhere?
17:12 RichardR_ dfisher: I'd like to see if we can help bring that over the finish line.
17:13 * dfisher goes to look
17:17 openstackgerrit Richard Raseley proposed stackforge/puppet-nova: Improve Endpoint Configurability  https://review.openstack.org/186109
17:28 kkirkpatrick joined #puppet-openstack
17:41 rhefner can I get the final +2 on Arne's NetApp patch here for puppet-manila? https://review.openstack.org/#/c/184804/
17:41 openstackgerrit Dan Prince proposed stackforge/puppet-tripleo: Configure virtual IPs for split out networks  https://review.openstack.org/186240
17:41 openstackgerrit Dan Prince proposed stackforge/puppet-tripleo: Add interface_for_ip function  https://review.openstack.org/186776
17:45 clayton so apparently the neutron provider code doesn't work if you only have auth_uri in your neutron.conf
17:45 clayton it requires auth_host, auth_port, auth_protocol
17:55 RichardR_ clayton: Does this affect one particular provider?
17:58 openstackgerrit Dan Prince proposed stackforge/puppet-tripleo: Configure virtual IPs for split out networks  https://review.openstack.org/186240
17:59 mdorman clayton:   yeah that’s true.  it’s on my list of stuff to fix when we go to kilo.  although obviously you are hitting it first
17:59 clayton I've got a patch for neutron that is almost ready
17:59 mdorman clayton:   iirc it’s a pretty trivial change, yeah
18:00 mdorman cool
18:00 clayton although I dont' see any evidence that the old parameters have been removed
18:00 clayton the old way seems to work just fine
18:00 mdorman iirc they’re deprecated
18:00 clayton yeah, but they were deprecated in juno, I thought they'd be removed in kilo
18:00 mdorman oh i see what you mean
18:00 mdorman yeah iwould’ve thought so too
18:01 clayton it's annoying to get puppet deprecation warnings telling me to move to newer parameters that don't actually work ;)
18:01 mdorman i saw med at  meetup last night…  i told him i’m pretty jealous of how fast you guys move on upgrades
18:01 mdorman true, true
18:01 clayton I spent the whole week of the summit working on kilo automation :)
18:01 mdorman yeah i suppose that helps
18:01 clayton that's almost like two weeks if everyone else is gone :)
18:01 mdorman amazing how much you can get done w/ nobody around to bother you
18:01 mdorman yeah, hah
18:09 Fin1te joined #puppet-openstack
18:11 _ryanv_ joined #puppet-openstack
18:14 openstackgerrit Clayton O'Neill proposed stackforge/puppet-neutron: Providers should use auth_uri if available  https://review.openstack.org/186863
18:24 fedexo joined #puppet-openstack
19:22 panda joined #puppet-openstack
19:32 openstackgerrit Colleen Murphy proposed stackforge/puppet-openstack-specs: Fix RST tests  https://review.openstack.org/186107
19:37 openstackgerrit Matthew Kirby proposed stackforge/puppet-nova: Enable nova service by default  https://review.openstack.org/186887
19:38 openstackgerrit Dan Prince proposed stackforge/puppet-tripleo: Configure virtual IPs for split out networks  https://review.openstack.org/186240
19:38 openstackgerrit Dan Prince proposed stackforge/puppet-tripleo: Add interface_for_ip function  https://review.openstack.org/186776
19:40 Hunner That was one of the summit conversations... about how we deprecate parameters and then don't remove them
19:43 mgagne Hunner: "then don't remove them" I happen to be the one that finally removed them after thoroughly researching each of the version in which each parameter were deprecated.
19:44 Hunner \o/
19:44 Hunner I think it was a passing comment related to master supporting 2 versions of openstack at once. Perhaps no longer relevant :)
19:44 mgagne Hunner: it's true that we don't have any systematic task related to deprecated parameter removal. However it's false to say that we never removed them and still have garbage from grizzly.
19:45 mgagne Hunner: https://wiki.openstack.org/wiki/Puppet/Deprecated_Parameters
19:45 Hunner Nice
19:46 mgagne Hunner: what's unfortunate is that it's a manual process and the page is impossible to find without google
19:46 mgagne Hunner: and I think we should add a 4th column: Removed in version X
19:47 Hunner Hmm. I wonder if we could piggy-back off of lint and write and auto-deprecator thing
19:47 Hunner Like, any parameter followed by #deprecated could be detected by lint and it spits out a report of deprecated params
19:47 mgagne Hunner: I was thinking the same. I would really like for something similar to exist
19:47 mgagne Hunner: can you run functions in parameter declaration?
19:48 Hunner You can have them as the defaults, yeah
19:48 mgagne Hunner: class foo ($bar = baz('whatever')) { ... } ?
19:48 Hunner Well, in puppet 3. I doubt it's changed in 4 though
19:48 Hunner Yeah
19:48 mgagne ok, I have no idea where I'm going with this question but it's cool to know
19:49 mgagne maybe have a function that can deal with deprecation better than a if else and warning()
19:49 Hunner I was thinking like `class foo ($bar = deprecated(undef)) { ... }`
19:49 Hunner But functionss in rspec-puppet are meh because they don't affect the catalog output
19:49 mgagne Hunner: not sure about the purpose of undef but yea
19:49 Hunner And rspec-puppet doesn't yet detect when functions are run
19:49 Hunner undef would just be the default value
19:49 Hunner Oh wait
19:50 mgagne Hunner: we don't care about how, only the result
19:50 Hunner If someone ever declared $bar then it would not call the function
19:50 Hunner So yeah, not what I wanted
19:50 mgagne Hunner: I compile with deprecated parameter: value is set. I compile with non-deprecated parameter: value is set.
19:50 mgagne Hunner: it won't?
19:51 mgagne Hunner: then it's only a matter of checking if the value is undef or not.
19:51 Hunner `class foo ($bar = deprecated(undef)) { ... } ; class { 'foo': bar => 'override', }`
19:51 mgagne Hunner: which we already do but in an undiscoverable fashion
19:52 Hunner You could do `class foo ($bar = undef) { deprecated('bar',$bar) ; ... }` and raise a warning if not undef
19:52 mgagne Hunner: maybe just have a sort of super boosted epic pick($new_param, $old_param, 'deprecation warning') function which throws warning if the 2nd value is used?
19:52 mgagne Hunner: stop reading my mind :D
19:53 mgagne Hunner: the idea so to have a single function that bundle the warning, deprecation handling and assign the "real" value
19:53 Hunner At least it would give a consistent deprecation warning by using a function instead of everyone doing it differently
19:53 Hunner Oh, but yeah, the "new" parameter might need to be mentioned
19:53 mgagne Hunner: yep, we had a similar idea in Atlanta or Paris I think. I unfortunately didn't do any follow up =(
19:53 Hunner Like deprecated_for('bar','newbar',$bar)
19:54 Hunner Hmm. I like your pick idea
19:54 mgagne Hunner: hmmm could we have a function that just accept the variable name (not the value) and tell ruby to lookup the variable in the scope?
19:54 Hunner One problem is that if you do `deprecated($bar)` then the value of $bar is passed, but deprecated() doesn't know what the name of the parameter is
19:55 Hunner oh right
19:55 cwolferh joined #puppet-openstack
19:55 mgagne Hunner: like: $foo_real = deprecated('foo', 'bar')
19:55 Hunner You're reading my mind now ;)
19:55 mgagne haha
19:55 Hunner So in that case $foo is the deprecated parameter, and $bar is the real one?
19:55 mgagne I'm just not sure how to handle multiple deprecated parameters which got replaced by a single parameter
19:55 Hunner or vice versa?
19:56 mgagne since I was about to suggest having a 3rd optional parameter to override the default warning or something similar, I would put the deprecated name last
19:56 Hunner $newvar = deprecated("${one}${two}", $newvar) ?
19:56 mgagne the other way?
19:57 Hunner and not $'s
19:57 Hunner Hmm. There might not be a good way to do that
19:57 mgagne go on
19:57 mgagne $new_real = deprecated('new_param', 'old_param', 'it is deprecated yo')
19:57 Hunner perhaps we should stick with the simpler `deprecated('new','old')` and `deprecated('old')`
19:58 Hunner And multi-deprecations will just have to be coded
19:58 mgagne Hunner: you got a point though, what if the parameter is deprecated and has no effect?
19:58 mgagne Hunner: you end up with a single parameter only
19:58 mgagne Hunner: or what just introduce 2 functions for 2 distinct purposes
19:58 Hunner Yeah. I guess that's deprecated('old')
19:58 mgagne removed('old') ?
19:58 Hunner deprecated() and deprecated_for() ?
19:59 mgagne Hunner: lets see the taxonomy used by oslo.config
20:00 mgagne Hunner: hmm python has named parameters
20:02 mgagne Hunner: Find this: https://github.com/openstack/oslo.config/blob/master/oslo_config/tests/test_generator.py#L55 and https://github.com/openstack/oslo.config/blob/master/oslo_config/cfg.py#L610-L615
20:02 mgagne Hunner: deprecated_param and deprecated_for_removal_param? that's long :O
20:06 openstack joined #puppet-openstack
20:06 Hunner Well, we don't really deprecate anything other than params, do we?
20:06 Hunner Err, sometimes classes I guess
20:07 mgagne Hunner: sure, the idea is more or less to be explicit about what we deprecate and also avoid any potential name conflict by using a not too generic name (deprecated)
20:08 mgagne Hunner: or we prefix it with os_ and call it a day
20:08 Hunner mgagne: Well, this would be useful to lots of places, so generic name and stick it in stdlib sounds nice too
20:09 mgagne Hunner: sure. I would like however to incubate the idea in our openstacklib module and see where it goes. our lib moves a lot faster than stdlib IMO
20:09 mgagne Hunner: and we have total control over it =)
20:09 Hunner true
20:21 _ody joined #puppet-openstack
20:30 openstack joined #puppet-openstack
21:00 openstackgerrit joined #puppet-openstack
21:17 dfisher stackforge/puppet-openstack is deprecated, right?
21:22 dfisher but puppetlabs/openstack is not?
21:27 Hunner The former is deprecated, and the latter currently has no one dedicated to it at PL, though it's fate is undecided
21:27 dfisher oh.
21:27 dfisher well, shit.
21:27 dfisher ok.
21:27 Hunner The latter is supposed to be a "here's how to get started with openstack and puppet modules" though it didn't really have enough docs to do so
21:27 Hunner What are you looking for?
21:27 dfisher easier way to do an all-in-one manifest
21:28 Hunner Yeah. Really the os puppet modules are the building blocks, and what you build with them will MOSTLY be up to the user
21:28 dfisher gotcha.
21:28 Hunner And every attempt to build a "reference architecture" has failed
21:28 dfisher whee!
21:29 Hunner spinalstack looks better than puppetlabs-openstack if you're looking for an out-of-the-box experience though
21:29 Hunner http://spinalstack.enovance.com/
21:29 dfisher will look there.
21:31 Hunner If you are looking to get started deploying OS with the modules, it's probably best to start by getting keystone up, then adding the parts on that you want
21:31 Hunner You'll end up with an all-in-one manifest that works for you and you actually understand all the parts of
21:32 openstack joined #puppet-openstack
21:32 dgurtner joined #puppet-openstack
21:32 dgurtner joined #puppet-openstack
21:36 * dfisher nods
21:36 dfisher i just need to actually understand the right way to write something.
21:37 dfisher at this point, I'm tripping over ordering of stuff
22:07 kkirkpatrick joined #puppet-openstack
22:17 _ody One of these days...
22:18 openstackgerrit Matthew Kirby proposed stackforge/puppet-nova: Enable nova service by default  https://review.openstack.org/186933
22:30 sgolovatiuk joined #puppet-openstack
22:59 sgolovatiuk joined #puppet-openstack
23:07 xarses joined #puppet-openstack
23:19 openstackgerrit Matthew Kirby proposed stackforge/puppet-nova: Enable nova service by default  https://review.openstack.org/186955
23:21 crinkle anyone using neutron-lbaas on ubuntu in kilo? I think there's a packaging bug where the package isn't providing one of its conf files, has anyone else run into issues with it?
23:34 openstackgerrit Colleen Murphy proposed stackforge/puppet-neutron: Fix fwaas package name for Ubuntu  https://review.openstack.org/186961
23:35 kevin joined #puppet-openstack
23:51 * crinkle files a bug

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