Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2013-11-28

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

All times shown according to UTC.

Time Nick Message
00:05 sanman_ joined #puppet-openstack
00:21 openstackgerrit Loic Dachary proposed a change to stackforge/puppet-ceph: ceph::mon create or destroy monitors  https://review.openstack.org/56841
00:21 rharrison dachary, yeah, Puppet Labs created spec_helper stating specifically that it was for projects like rspec-puppet to use so rojek dutifully used it.  Then Puppet Labs made their gem depend on the thing they specifically created it to be a dependency of...
00:22 dachary it's a convenient environment
00:24 rharrison dachary, so then when I went to package it I discovered the dep loop.  I posted bugs to both projects. rojeck responded instantly and the patches are in master.  Puppet Labs has yet to comment on the bug...
00:24 dachary was it a long time ago ?
00:24 rharrison dachary, it was at least the spring or early summer...
00:27 rharrison dachary, https://projects.puppetlabs.com/issues/18913 10 months
00:28 dachary :-)
00:28 rharrison dachary, though some of the folks I've talked to at Puppet Labs say they really want to kill it off entirely and get folks to use the faces hooks.  I wish they'd just say "stop using this thing and use that thing"
00:28 mjblack joined #puppet-openstack
00:28 dachary faces hooks ?
00:29 dachary I'll read about tomorrow, it's insanely late here ;-)
00:29 dachary night !
00:33 xarses joined #puppet-openstack
00:36 rongze joined #puppet-openstack
00:37 xingchao joined #puppet-openstack
00:38 rharrison dachary, unless you're writing something like rspec-puppet or an additional puppet command you really don't need to worry about faces.  Other than they exist.
00:42 mjblack joined #puppet-openstack
00:44 mjblack_ joined #puppet-openstack
00:51 rmoe joined #puppet-openstack
01:14 dmsimard joined #puppet-openstack
01:16 dmsimard1 joined #puppet-openstack
01:26 badiane_ka joined #puppet-openstack
01:41 dmsimard joined #puppet-openstack
01:43 xingchao joined #puppet-openstack
01:45 mjblack joined #puppet-openstack
01:59 rongze joined #puppet-openstack
02:35 rongze joined #puppet-openstack
02:52 mjblack joined #puppet-openstack
02:52 michchap joined #puppet-openstack
03:04 mjblack joined #puppet-openstack
04:34 slowe joined #puppet-openstack
05:08 dmsimard joined #puppet-openstack
05:47 rongze joined #puppet-openstack
06:47 dalgaaf_ joined #puppet-openstack
06:48 rongze joined #puppet-openstack
07:06 dachary rharrison: I'm happy to ignore faces and heading to breakfast :-D
07:48 rongze joined #puppet-openstack
07:53 EmilienM michchap: o/
08:03 giorgiodinapoli joined #puppet-openstack
08:06 bauzas joined #puppet-openstack
08:15 markvoelker joined #puppet-openstack
08:28 mkoderer joined #puppet-openstack
08:49 rongze joined #puppet-openstack
08:49 michchap EmilienM: morning
08:58 EmilienM michchap: i'm requesting a new review: https://review.openstack.org/#/c/58792/
08:59 EmilienM michchap: for your tomorrow if you can :)
08:59 michchap sure
08:59 marun joined #puppet-openstack
09:05 marun joined #puppet-openstack
09:18 derekh joined #puppet-openstack
09:18 marun joined #puppet-openstack
09:20 EmilienM michchap: you will also have ml2 patchsets, fc__ has some recommendations and i forget to unit test one thing
09:23 marun joined #puppet-openstack
09:47 openstackgerrit Loic Dachary proposed a change to stackforge/puppet-ceph: ceph::mon create or destroy monitors  https://review.openstack.org/56841
09:51 openstackgerrit Loic Dachary proposed a change to stackforge/puppet-ceph: ceph::mon create or destroy monitors  https://review.openstack.org/56841
09:51 rongze joined #puppet-openstack
09:56 openstackgerrit Loic Dachary proposed a change to stackforge/puppet-ceph: ceph::mon create or destroy monitors  https://review.openstack.org/56841
09:56 dachary sorry for the spam... getting used to the magic blueprint syntax parsing...
09:58 mattymo dachary, do you simply set your local branch to bp/mon so that when you git review, it imports the blueprint automagically?
09:58 dachary I do
09:58 mattymo that's cool
09:59 dachary but... it's overriden by blueprint mentionned in the README it seems
10:00 dachary How do you make it so a review that depends on another review ?
10:00 mattymo rebase your next patchset against this one
10:00 mattymo git review -d 56841, make changes, git commit -a, then git review
10:18 dachary ah cool, thanks mattymo
10:46 mattymo dachary, did you make a BP for ceph::mon?
10:46 mattymo I don't see it
10:47 dachary mattymo: https://wiki.openstack.org/wiki/Pu​ppet-openstack/ceph-blueprint#mon
10:47 mattymo oh wow okay
10:48 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-neutron: Ensure ml2 is set in core_plugin  https://review.openstack.org/58881
10:48 mattymo is a BP mandatory for a change request with new functionality
10:48 mattymo ?
10:49 EmilienM mattymo: not in openstack modules.
10:49 EmilienM mattymo: blueprint are used to discuss design around huge new feature
10:52 rongze joined #puppet-openstack
10:53 marun joined #puppet-openstack
11:01 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-neutron: Ensure ml2 is set in core_plugin  https://review.openstack.org/58881
11:04 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-neutron: Ensure ml2 is set in core_plugin  https://review.openstack.org/58881
11:10 giorgiodinapoli joined #puppet-openstack
11:10 michchap joined #puppet-openstack
11:43 openstackgerrit Loic Dachary proposed a change to stackforge/puppet-ceph: Make ceph::repo ensurable + add support for emperor  https://review.openstack.org/58989
11:48 xingchao joined #puppet-openstack
11:55 mattymo joined #puppet-openstack
11:59 xingchao joined #puppet-openstack
12:00 xingchao_ joined #puppet-openstack
12:04 dachary I tried following instructions from https://wiki.openstack.org/wiki​/Gerrit_Workflow#Add_dependency in order to make https://review.openstack.org/#/c/56841/ depend on https://review.openstack.org/#/c/58989/ . But when I git review ( last line of the page ) it tries to upload the two commits, which is certainly now what I want ;-)
12:11 * dachary reads http://www.mediawiki.org/wiki/Gerrit​/Advanced_usage#Create_a_dependency
12:22 openstackgerrit Loic Dachary proposed a change to stackforge/puppet-ceph: ceph::mon create or destroy monitors  https://review.openstack.org/56841
12:57 rongze joined #puppet-openstack
13:00 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-cinder: Add Ceph as backup backend support  https://review.openstack.org/59006
13:05 dachary michchap: when you're around I would very much like your input on how to deal with https://wiki.openstack.org/wiki/Puppet-openst​ack/ceph-blueprint#I_want_to_try_this_module.2C_heard_of_ceph.2C_want_to_see_it_in_action
13:06 dachary michchap: there are two class { 'ceph::osd': } which does not work. If they are replaced with ceph_osd { ... } then scenario_node_terminus will not work. How would you resolve this ?
13:07 EmilienM dachary: he should be online in 8h :)
13:08 dachary EmilienM: Oo what timezone is this ? ;-)
13:08 EmilienM dachary: sidney
13:08 dachary shoot :-)
13:19 openstackgerrit Loic Dachary proposed a change to stackforge/puppet-ceph: Make ceph::repo and ceph ensurable + add support for emperor  https://review.openstack.org/58989
13:23 EmilienM dachary: what is the status of the module? is it doing the coffee now ?
13:24 dachary ahaha
13:24 michchap dachary: I'm here
13:24 dachary michchap: cool :-)
13:25 EmilienM michchap: w00t
13:25 dachary I can see a zillion ways to deal with the problem but I'd rather go for something you recommend
13:26 michchap dachary: because you can't create two copies of the same class?
13:28 michchap dachary: I asked Dan B about this and he gave me a somewhat cryptic answer. I don't have a solid solution right now. You can pass in a list of hashes, but that's not very elegant.
13:29 dachary hum
13:30 dachary I can settle with passing a list of hash
13:30 michchap The response I got was "The node object only supports classes and not resources"
13:30 dachary michchap: is there an example of such trick in openstack-installer ?
13:30 dachary I mean in modules openstack-installer depends on
13:31 michchap dachary: no, but I've been thinking about it because this exact problem is in the cinder multi-backend patch that I merged but then later regretted.
13:31 dachary ahahah
13:31 michchap where we want to specify multiple backends
13:31 dachary what does xarses  have to say about it ?
13:31 dachary IIRC he authored the patch for multi-backend
13:31 michchap I guess it's a question of how many parameters you have
13:32 michchap I haven't mentioned it to him, we talked about having native types...
13:33 michchap I want to know if I can get around these issues by having native types do things for me
13:33 michchap but I just don't have enough experience with them. I have to sit down and learn it properly.
13:34 michchap The thing is, if you just pass in a list of osd target devices to manage, you make an assumption that all those devices are uniform.
13:35 dachary I'm thinking class ceph::osds { list => [ { disk => '/dev/vdb' ... }, { disk => '/dev/vdc' ... } ] } that relies on a resource ceph_osd { }. The ceph::osds ( note the s ) would only be there for the sake of scenario_node_terminus. People who do not have the constraint would use ceph_osd { } instead.
13:35 michchap is it just pointing at devices, or is there additional config you would reasonably do that isn't going to be constant across them, such as journal locations or something?
13:36 michchap because class ceph::osds { devices => [ /dev/vdb, /dev/vdc, /dev/vdd ] } would work fine as long as they are all treated the same.
13:36 rongze joined #puppet-openstack
13:37 dachary I don't think it would be good or even better from an architectural point of view to assume which options are shared and which are not.
13:37 michchap I don't know enough about common deployments of ceph to know whether that is the vast majority of cases, or if there are enough times when that won't work that it isn't acceptable
13:37 dachary michchap: I think it would be ok in a number of cases. But it would block you otherwise and that's going to generate forks ;-)
13:38 michchap it would only cause forks for those who use the node_terminus, but yeah, I agree that it's not good.
13:40 dachary ceph::osd
13:40 dachary - 1
13:40 dachary - 2
13:40 dachary - 3
13:40 dachary sorry
13:40 dachary ceph::osd::args
13:40 dachary - 1
13:40 dachary - 2
13:40 dachary - 3
13:40 dachary is this going to translate into
13:40 michchap for now, I think just make the defines and pass hashes from an osds class, as you suggested, and I will keep looking for better options
13:40 dachary class { 'ceph::osd': args => [1,2,3] } ?
13:41 dachary My fu level in both yaml & terminus is still low ;-)
13:41 michchap Yeah mine is still only at medium - I am reading all the doc and preparing a screencast as a way of learning it. I hope to have something ready soon to share with you+others.
13:42 dachary michchap: I'll go this route then, thanks for discussing this :-)
13:43 michchap dachary: no problem. You can try to ping Dan - he is back and working on things, just not puppet-openstack.
13:45 dachary oh, I thought he was still on vacation
13:45 dachary bodepd: ping ?
13:45 michchap I think he would be asleep at the moment, maybe try in a few hours
13:49 dachary hum, it's thanksgiving
13:49 michchap ah true...
13:56 marun joined #puppet-openstack
13:57 openstackgerrit Loic Dachary proposed a change to stackforge/puppet-ceph: ceph::mon create or destroy monitors  https://review.openstack.org/56841
14:02 dmsimard joined #puppet-openstack
14:09 michchap dachary: is all of the config that needs to be passed to the osd daemons going to be specified in the ceph::conf, or do the ceph::osd defines create the ceph.conf entries for the osds?
14:22 dmsimard It might just be me so I'll ask you guys, what do you feel about mandatory parameters defaulted as undef and later evaluated with a fail() if missing, explaining why the parameter is required ?
14:25 michchap dmsimard: I don't have particularly strong feelings about it. It sounds nice. Are you proposing this as a policy going forward, or to re-do all mandatory parameters this way?
14:26 dmsimard michchap: I don't know, I'm kind of torn.. makes the code heavier and seems to me like a permanent —verbose or —debug switch
14:26 dmsimard michchap: On the other hand I can see the benefits as well
14:26 michchap dmsimard: Is there something up for review that does this?
14:27 dmsimard michchap: Yeah, example: https://review.openstack.org​/#/c/58989/2/manifests/init.pp
14:29 dmsimard michchap: Hm, it actually changed since I last looked at it because it's under an ensure => present now
14:29 michchap dmsimard: If a mandatory parameter with no default is omitted, the catalog will fail saying the parameter must be passed
14:30 michchap Error: Must pass mandatory to Class[Test] at /root/site.pp:7 on node precise64
14:30 dmsimard michchap: yeah, I know that - but in this case, parameters that are mandatory to installation of the software are not required for the uninstallation
14:30 michchap dmsimard: so I actually think unless there is some strange reason to require a parameter that is likely to confuse someone, maybe there's not really any need for adding code that fails
14:31 dmsimard so the mandatory parameters are defaulted to undef and only evaluated on ensure => present
14:32 michchap so they aren't really mandatory then
14:32 michchap hmm
14:33 michchap well it's better to explicity fail than to fail silently
14:34 dmsimard yeah, in that case
14:37 michchap because in that patch, the fsid will be set as '', which is probably not what is wanted, so I guess that is a good case for having a fail()
14:38 dmsimard Yup.
14:41 michchap in fact this might be fairly common - where two optional parameters much be enabled together there would be the same thing.
14:42 michchap s/much/must
14:42 dmsimard Yeah, I know about parameters that must be declared together
14:49 dmsimard1 joined #puppet-openstack
14:50 rongze joined #puppet-openstack
15:02 xingchao joined #puppet-openstack
15:07 fvollero xingchao: Ni hao!
15:19 rongze joined #puppet-openstack
15:30 dmsimard1 dachary: ping
15:35 dachary dmsimard: good morning !
15:36 * dachary reading the backlog
15:38 dmsimard dachary: What should we do for the scoping of ceph configuration ?
15:38 dmsimard dachary: Any parameters can be global, global to daemon type or down to hosts
15:39 dmsimard dachary: I see an overlap of authentication_type in ::init and ::mon
15:39 dmsimard dachary: I didn't think we were going to support multiple clusters off the start, this complicates several things
15:40 EmilienM dachary: actually i was wondering of the current status of ceph module. (what can i do with at the moment, etc)
15:41 dachary dmsimard: we're going to support the majority of use cases from the start, otherwise we have no chance of reducing fragmentation.
15:41 dachary EmilienM: you can't use it at the moment.
15:41 dmsimard EmilienM: It's not suitable for production yet
15:41 dachary it's not suitable for anything.
15:41 dmsimard ^
15:42 EmilienM dachary: dmsimard: thank you, good job anyway on the module :)
15:43 dmsimard EmilienM: We're trying hard, it's not easy :)
15:43 EmilienM dmsimard: yeah, a lot of work is done on puppet modules, it's exiting!
15:43 dachary dmsimard: I don't think we need to do anything about the scoping. Only to remember that it's there and avoid taking a route that is vulnerable to scoping.
15:43 dachary avoiding duplicated arguments is good in this regard
15:46 dmsimard dachary: Can you give me the manifest you're using to test your implementation ?
15:47 dachary dmsimard: I can do better than that :-)
15:47 dmsimard Can't seem to get it to work on my end
15:47 dmsimard I guess I can use the mon system spec
15:48 dmsimard dachary: the spec doesn't test cephx ?
15:48 dachary https://review.openstack.org/#/c/5​6841/10/spec/system/ceph_mon_spec.rb and the result of the last run is http://paste.ubuntu.com/6489640/
15:48 dachary dmsimard: ^
15:49 dachary dmsimard: no, it does not.
15:49 dmsimard dachary: That's what I'm trying to make work :)
15:50 dachary how do you mean ? for https://review.openstack.org/#/c/56841/10 ?
15:51 dachary dmsimard: ^
15:52 dachary I'm working on improving the coverage of the integration tests. Does the module look good to you otherwise ?
15:54 dmsimard Well I'm actually testing it before giving my approval
15:54 dmsimard but I can't get ceph::mon with cephx to work
15:55 dachary I'm not too surprised
15:55 dachary no integration tests means likely to not work
15:56 dachary if you want to start by trying this, I suggest you wait until I complete the integration tests
15:57 dmsimard ok i'll let you finish
15:57 dachary I'm worried that once I get it working there will be reviews about the style that force me to redo it completely
15:57 dachary because I don't have much experience in puppet
15:58 dachary for instance it is entirely possible that mgagne tells me that the script I included ( although small enough in my opinion ) is heresy and must be replaced by something else ( and I have no clue how I could do that without complicating things significantly )
15:59 dmsimard dachary: I'm not entirely sure about those bash scripts either tbh
16:00 dachary how would you do it ?
16:00 dachary it's 5 lines but I feel this will raise hell
16:00 dmsimard dachary: I don't have a solution so I haven't commented on them - the only other way that comes to mind is a file resource that is sourced by a template which results in the bash script but it's not exactly clean either
16:01 dachary how is it not clean ?
16:01 dachary that's what I don't see
16:01 dachary mgagne said that what matters is that the return code is meaningfull and the output is debuggable. Both of which are taken care off.
16:02 dachary dmsimard: what do you not like ?
16:02 dmsimard dachary: It's not that I don't like it, I am wondering if there is another way
16:03 dachary A simpler way you mean ?
16:03 dmsimard Yes and no, if there are other ways we have a base for comparison and choose the best option :)
16:04 dachary dalgaaf_: if you have a minute I would very much appreciate a quick review of https://review.openstack.org​/#/c/56841/10/manifests/mon.pp
16:04 dachary dmsimard: I'm not sure I follow :-)
16:05 dalgaaf_ dachary: sorry ... I'm on the train ... have to review it later
16:08 dachary dalgaaf_: no worries, thanks in advance :-) A quick review will do. I hope nothing makes you jump. But I'm ready to amend otherwise :-)
16:11 dmsimard dachary: It seems it's possible to do something like this: exec { 'name': command => template('long_bash_script_here') }
16:11 dmsimard (Not saying it's awesome)
16:12 dachary I'm ok with that, sure. I don't see the upside but if that's what it takes, I'll do it :-)
16:12 dmsimard Trying to see how else it could be done
16:14 dachary dmsimard: what do you think of set -ex ? I tend to like that because it stops as soon as a failure happens and it's verbose so you see what's going on. It's not too verbose either because it's barely 4 commands ( and doing it with puppet would output tthat anyway with -v )
16:29 mgagne joined #puppet-openstack
16:29 mgagne joined #puppet-openstack
16:41 dmsimard Off-topic and this obviously isn't targetted to you guys (You already know this stuff) but I started blogging and my first post is about pupet-openstack :D http://dmsimard.com/2013/11/26/how-​to-contribute-to-puppet-openstack/
16:44 mgagne dmsimard: I should get mine up :-/
16:45 dmsimard mgagne: Whenever I start blogging, I usually end up giving up - i'll try to keep it up this time :P
16:51 dachary Jean-Roger: I think that would be useful to your class mates http://dmsimard.com/2013/11/26/how-​to-contribute-to-puppet-openstack/
16:51 Jean-Roger Let me take a look
16:51 dachary mgagne: hi !
16:51 Jean-Roger Hello btw
16:52 sbadia joined #puppet-openstack
16:56 hbroyon joined #puppet-openstack
16:56 hbroyon Hello
16:57 Jean-Roger Hi hbroyon, here you go : http://dmsimard.com/2013/11/26/how-​to-contribute-to-puppet-openstack/
16:57 Jean-Roger Thank you dmsimard for the howto
16:57 dmsimard Glad it's appreciated, that's the goal :D
16:58 hbroyon We change it, someone can review and verified please https://review.openstack.org/#/c/56770/
16:58 hbroyon Jean-Roger : Thanks, this will be help us =D
16:59 Jean-Roger My pleasure
17:00 hbroyon You are very interesting by puppet, it s a very good thing =)
17:00 hbroyon We ll try to help other team ^^
17:00 dmsimard hbroyon, dachary: About that review, I haven't commented or approved yet because I'm still not sure about providing code to the user that does not work. I remember dachary commented on that but I forgot what he said :(
17:03 hbroyon dmsimard : Yes of course, it is important that everything is for the best in a project like this one. If there is still a problem, we will change it wrong ^ ^
17:04 dmsimard hbroyon: I'll let other people weigh in on the review, that's why I haven't commented yet
17:04 dmsimard hbroyon: But I'm okay with providing the use cases in a separate file as such
17:05 hbroyon Good, thanks ^^
17:05 Jean-Roger Sure, that makes things more clear
17:06 dachary hbroyon: looking
17:06 dachary hbroyon: I voted for it already ;-)
17:07 dachary dmsimard: what I said is that these are not examples, they are use cases that illustrate how things could be done. It's by no mean an example.
17:07 hbroyon Yes we saw , thanks ^^
17:08 dmsimard dachary: How things should be done as we develop the module ?
17:08 dachary dmsimard: the goal is to roughly show how it would go. The details do not matter.
17:09 dachary dmsimard: it is mostly for the benefit of people looking at the module for the first time. So that they can see the use case they have in mind was considered. And that glance at what it would mean to implement it.
17:10 dmsimard dachary: Why provide code then?
17:10 dachary dmsimard: because it would be both difficult to write and difficult to read otherwise.
17:11 dachary it's like the code you see in the slides of a presentation
17:11 dachary you won't expect it to be a 100% working example and it won't occur to you to resent the author if copy / paste does not apply exactly to what you are trying to do ;-)
17:12 dachary I'm not saying the code *must* be wrong
17:12 bodepd dachary: pong
17:12 dachary I'm saying it does not matter much if it is, as long as the structure is good enough :-)
17:12 dachary s/good enough/close enough to the real thing/
17:12 dachary bodepd: hey !
17:13 dmsimard dachary: Then it needs to be clear to set expectations for the users that the code snippets do not work/are not expected to work
17:13 dmsimard dachary: that it is "pseudo code"
17:13 dachary We can add a reminder at the beginning of the usecases file to that end.
17:16 dachary bodepd: I was wondering how to approach the fact that ceph::osd needs to be used multiple times on the same machine. class { 'ceph::osd': ... } won't work. And ceph_osd {} won't allow for node_terminus. Do you have a suggestion ? ( the backlog has a conversation with michchap if you're willing to browse )
17:18 dachary hbroyon: did you get that ? ( what dmsimard said about usecases )
17:20 hbroyon dachary : I undestand
17:22 dachary hbroyon: cool
17:22 dmsimard dachary: the current puppet-ceph implementation uses ceph::osd and ceph::osd::device specifically for the devices
17:22 hbroyon :)
17:22 dmsimard dachary: You're trying to implement everything directly in ceph::osd ?
17:22 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-glance: Add Cinder backend to image store  https://review.openstack.org/59080
17:24 hbroyon Sorry, I need to go .
17:24 hbroyon I wish you a good evening
17:28 dachary dmsimard: I'm trying to satisfy the requirements of node_terminus
17:28 dmsimard dachary: I guess I need to read on that
17:31 LAnthony joined #puppet-openstack
17:38 bodepd dachary: I think I can answer without browsing :)
17:38 bodepd dachary: this is why create resource exists
17:39 bodepd dachary: have the thing be defined in a defined resource type
17:39 dachary ok
17:39 bodepd dachary: and then wrap it in a class that accepts a hash of the osds to instantiate
17:40 dachary a list of hashes, each hash being the args of the defined resource type ?
17:40 bodepd dachary: u
17:40 bodepd dachary: https://gist.githxb.com/bodepd/7695685
17:41 bodepd dachary: that shoudl be clear :)
17:42 * dachary thinking
17:44 dachary bodepd: https://gist.github.com/bod​epd/7695685#comment-960087 ?
17:44 dachary modulo format fuckup
17:46 bodepd nah, created resources takes the hash
17:46 bodepd of title => params
17:46 bodepd have a look at the docs for it
17:46 dachary ok
17:46 bodepd there kind of is no for loop in Puppet
17:46 bodepd well there is with the experimental parser starting in 3.2.x
17:46 bodepd but I don't think modules should depend on it just yet
17:47 * dachary reading http://docs.puppetlabs.com/references​/latest/function.html#createresources
17:47 dachary bodepd: crystal clear, thanks :-)
17:49 dachary michchap: will be happy too, I think
18:44 rongze joined #puppet-openstack
19:02 tnoor1 joined #puppet-openstack
19:05 openstackgerrit Anthony Leprêtre proposed a change to stackforge/puppet-ceph: Add "I want to operate a production cluster" in the file "USECASES.md"  https://review.openstack.org/59091
19:06 tnoor2 joined #puppet-openstack
19:43 rongze_ joined #puppet-openstack
19:51 Jean-Roger dachary can you explain the -1 : https://review.openstack.org/#/c/59091/ please ?
19:51 rongze joined #puppet-openstack
19:52 dachary what is it exactly that you don't understand in the comment ?
19:52 dachary Jean-Roger: ^
19:52 Jean-Roger Ah, didn't see the comment, my bad
19:55 Jean-Roger Can you be more specific about these white spaces ?
20:00 dachary you must get rid of trailing white space Jean-Roger
20:00 dachary that's the red in https://review.openstack.org/#/c/59091/1/USECASES.md,unified ;-)
20:01 Jean-Roger I see
20:01 Jean-Roger I'll do it, thanks
20:01 dachary you're welcome
20:15 bauzas joined #puppet-openstack
20:44 LAnthony left #puppet-openstack
20:59 xarses dachary: michchap: straighten up the discussion from a while ago?
21:00 dachary xarses: executive summary : it's all good, sleep well, all is under control :-)
21:03 xarses its  only 1300 here, holiday today
21:28 michchap_ joined #puppet-openstack
21:41 dmsimard joined #puppet-openstack
21:47 dmsimard1 joined #puppet-openstack
22:00 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-glance: Add Cinder backend to image store  https://review.openstack.org/59080
22:08 EmilienM michchap_: hello !
22:09 dachary How can I use a temporary file in a module ?
22:10 dachary something like file { '/tmp/a': ensure => present } and later file { '/tmp/a': ensure => absent }
22:14 EmilienM michchap_: when you are around, i have some review for you, you have a look at gerrit
22:26 michchap joined #puppet-openstack
23:29 rmoe joined #puppet-openstack
23:45 rmoe joined #puppet-openstack
23:58 openstackgerrit Loic Dachary proposed a change to stackforge/puppet-ceph: ceph::mon create or destroy monitors  https://review.openstack.org/56841

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