| *** ferest has quit IRC | 00:28 | |
| *** ferest has joined #puppet-openstack | 00:28 | |
| *** ferest has quit IRC | 00:33 | |
| *** fedexo has quit IRC | 00:45 | |
| *** fedexo has joined #puppet-openstack | 00:49 | |
| *** gildub has joined #puppet-openstack | 00:53 | |
| *** markvoelker has quit IRC | 01:18 | |
| *** ilbot3 has quit IRC | 01:44 | |
| *** ilbot3 has joined #puppet-openstack | 01:48 | |
| *** markvoelker has joined #puppet-openstack | 02:34 | |
| *** markvoelker has quit IRC | 02:39 | |
| *** fedexo has quit IRC | 03:10 | |
| *** fedexo has joined #puppet-openstack | 03:12 | |
| *** chem has quit IRC | 04:00 | |
| *** markvoelker has joined #puppet-openstack | 04:23 | |
| *** markvoelker has quit IRC | 04:28 | |
| *** sks has joined #puppet-openstack | 04:35 | |
| *** rushiagr_away is now known as rushiagr | 04:52 | |
| *** jtomasek has joined #puppet-openstack | 05:37 | |
| *** markvoelker has joined #puppet-openstack | 06:12 | |
| *** markvoelker has quit IRC | 06:16 | |
| *** fedexo has quit IRC | 06:35 | |
| *** rhefner has quit IRC | 06:43 | |
| *** tfz has quit IRC | 06:52 | |
| *** dgurtner has joined #puppet-openstack | 07:04 | |
| *** spredzy|afk is now known as spredzy | 07:07 | |
| *** panda has quit IRC | 07:21 | |
| *** panda has joined #puppet-openstack | 07:21 | |
| *** tfz has joined #puppet-openstack | 07:25 | |
| *** gfidente has joined #puppet-openstack | 07:26 | |
| *** gfidente has quit IRC | 07:26 | |
| *** gfidente has joined #puppet-openstack | 07:26 | |
| *** gildub has quit IRC | 07:27 | |
| *** _nick has quit IRC | 07:33 | |
| *** _nick has joined #puppet-openstack | 07:35 | |
| *** jistr has joined #puppet-openstack | 07:39 | |
| *** paramite has joined #puppet-openstack | 07:40 | |
| *** imcsk8_ is now known as imcsk8|zZz | 07:50 | |
| *** jpena has joined #puppet-openstack | 07:52 | |
| *** markvoelker has joined #puppet-openstack | 08:01 | |
| *** markvoelker has quit IRC | 08:06 | |
| *** kkirkpatrick has quit IRC | 08:08 | |
| *** paramite is now known as paramite|afk | 08:35 | |
| *** paramite|afk is now known as paramite | 08:41 | |
| *** cdent has joined #puppet-openstack | 08:56 | |
| *** jpena has quit IRC | 09:06 | |
| *** jpena has joined #puppet-openstack | 09:20 | |
| *** paramite is now known as paramite|afk | 09:32 | |
| *** markvoelker has joined #puppet-openstack | 09:49 | |
| *** markvoelker has quit IRC | 09:54 | |
| *** kkirkpatrick has joined #puppet-openstack | 10:07 | |
| *** jpena has quit IRC | 10:11 | |
| *** jpena has joined #puppet-openstack | 10:14 | |
| *** blentz has quit IRC | 10:15 | |
| *** blentz has joined #puppet-openstack | 10:15 | |
| *** openstack has quit IRC | 10:20 | |
| *** openstack has joined #puppet-openstack | 10:27 | |
| *** jpena has quit IRC | 10:34 | |
| *** danieru has joined #puppet-openstack | 10:42 | |
| *** paramite|afk is now known as paramite | 10:43 | |
| *** danieru_ has quit IRC | 10:43 | |
| *** bogdando has quit IRC | 10:43 | |
| *** mordred has quit IRC | 10:43 | |
| *** mordred has joined #puppet-openstack | 10:44 | |
| *** bogdando has joined #puppet-openstack | 10:45 | |
| *** jpena has joined #puppet-openstack | 10:49 | |
| *** sgolovatiuk has quit IRC | 10:58 | |
| *** rcallaw__ has joined #puppet-openstack | 11:15 | |
| *** rcallawa_ has quit IRC | 11:17 | |
| *** arnaud_orange has joined #puppet-openstack | 11:18 | |
| *** ducttape_ has joined #puppet-openstack | 11:21 | |
| *** panda has quit IRC | 11:21 | |
| *** panda has joined #puppet-openstack | 11:22 | |
| *** rhefner has joined #puppet-openstack | 11:28 | |
| *** ducttape_ has quit IRC | 11:29 | |
| *** markvoelker has joined #puppet-openstack | 11:35 | |
| *** markvoelker has quit IRC | 11:40 | |
| *** ducttape_ has joined #puppet-openstack | 11:46 | |
| *** ducttape_ has quit IRC | 11:48 | |
| *** rushiagr is now known as rushiagr_away | 11:57 | |
| *** rcallaw__ has quit IRC | 11:58 | |
| *** dprince has joined #puppet-openstack | 11:58 | |
| *** jpena is now known as jpena|lunch | 11:59 | |
| *** saju_m has joined #puppet-openstack | 12:04 | |
| *** paramite is now known as paramite|afk | 12:06 | |
| *** paramite|afk is now known as paramite | 12:11 | |
| *** jayg|g0n3 is now known as jayg | 12:13 | |
| *** morazi has quit IRC | 12:20 | |
| *** morazi has joined #puppet-openstack | 12:20 | |
| openstackgerrit | Dan Prince proposed stackforge/puppet-tripleo: Use node IP lists for HA Proxy ipaddresses https://review.openstack.org/186775 | 12:22 |
|---|---|---|
| openstackgerrit | Dan Prince proposed stackforge/puppet-tripleo: Configure virtual IPs for split out networks https://review.openstack.org/186240 | 12:22 |
| openstackgerrit | Dan Prince proposed stackforge/puppet-tripleo: Add interface_for_ip function https://review.openstack.org/186776 | 12:22 |
| *** markvoelker has joined #puppet-openstack | 12:34 | |
| *** sks has quit IRC | 12:38 | |
| *** paramite is now known as paramite|afk | 12:38 | |
| *** ferest has joined #puppet-openstack | 12:40 | |
| *** saju_m has quit IRC | 12:43 | |
| *** saju_m has joined #puppet-openstack | 12:43 | |
| *** paramite|afk is now known as paramite | 12:45 | |
| *** rcallawa_ has joined #puppet-openstack | 12:46 | |
| *** ferest has quit IRC | 12:46 | |
| *** ducttape_ has joined #puppet-openstack | 12:58 | |
| *** jpena|lunch is now known as jpena | 13:01 | |
| *** ducttape_ has quit IRC | 13:14 | |
| *** chem has joined #puppet-openstack | 13:46 | |
| *** jpena has quit IRC | 13:48 | |
| *** jpena has joined #puppet-openstack | 13:49 | |
| *** RichardRaseley has joined #puppet-openstack | 13:55 | |
| *** kkirkpatrick has quit IRC | 13:58 | |
| RichardRaseley | mgagne: Have you had a chance to review my most recent comments on https://review.openstack.org/#/c/186109/6 ? I am hoping we can bring this issue to an amicable close, or at least outline what will be required to do so, today. | 14:03 |
| *** paramite is now known as paramite|afk | 14:04 | |
| *** jpena has quit IRC | 14:05 | |
| *** sgolovatiuk has joined #puppet-openstack | 14:05 | |
| *** RichardRaseley has quit IRC | 14:05 | |
| *** paramite|afk is now known as paramite | 14:06 | |
| *** tshtilma has joined #puppet-openstack | 14:13 | |
| *** jpena has joined #puppet-openstack | 14:19 | |
| *** kkirkpatrick has joined #puppet-openstack | 14:24 | |
| *** vinay__ has joined #puppet-openstack | 14:37 | |
| *** imcsk8|zZz is now known as imcsk8 | 14:42 | |
| *** mdorman has joined #puppet-openstack | 14:46 | |
| *** tshtilma has quit IRC | 14:47 | |
| *** tfz has quit IRC | 14:56 | |
| *** dfisher has joined #puppet-openstack | 15:01 | |
| *** ducttape_ has joined #puppet-openstack | 15:01 | |
| *** stevemar has joined #puppet-openstack | 15:04 | |
| *** dprince has quit IRC | 15:13 | |
| *** saju_m has quit IRC | 15:15 | |
| *** vrao has quit IRC | 15:18 | |
| *** vrao has joined #puppet-openstack | 15:19 | |
| *** panda has quit IRC | 15:21 | |
| *** panda has joined #puppet-openstack | 15:22 | |
| *** arnaud_orange has quit IRC | 15:34 | |
| *** tfz has joined #puppet-openstack | 15:44 | |
| *** kkirkpatrick has quit IRC | 15:57 | |
| *** vrao has quit IRC | 15:57 | |
| *** vrao has joined #puppet-openstack | 15:58 | |
| *** RichardRaseley has joined #puppet-openstack | 15:58 | |
| *** kkirkpatrick has joined #puppet-openstack | 15:59 | |
| *** cwolferh has joined #puppet-openstack | 16:07 | |
| *** jistr has quit IRC | 16:11 | |
| *** rcallawa_ has quit IRC | 16:13 | |
| *** paramite has quit IRC | 16:17 | |
| *** rcallawa_ has joined #puppet-openstack | 16:23 | |
| *** richm has joined #puppet-openstack | 16:24 | |
| *** beddari_ has joined #puppet-openstack | 16:29 | |
| *** beddari_ has quit IRC | 16:32 | |
| *** beddari_ has joined #puppet-openstack | 16:34 | |
| *** dgurtner has quit IRC | 16:35 | |
| *** dprince has joined #puppet-openstack | 16:38 | |
| *** beddari has quit IRC | 16:39 | |
| *** beddari_ is now known as beddari | 16:39 | |
| *** gfidente has quit IRC | 16:48 | |
| *** gfidente has joined #puppet-openstack | 16:48 | |
| *** gfidente has quit IRC | 16:48 | |
| *** gfidente has joined #puppet-openstack | 16:48 | |
| *** kkirkpatrick has quit IRC | 16:50 | |
| *** kkirkpatrick has joined #puppet-openstack | 16:51 | |
| *** rcallawa_ has quit IRC | 16:56 | |
| *** jpena has quit IRC | 16:58 | |
| *** rcallawa_ has joined #puppet-openstack | 17:02 | |
| *** beddari1 has joined #puppet-openstack | 17:06 | |
| *** xarses has quit IRC | 17:25 | |
| *** aimon has joined #puppet-openstack | 17:25 | |
| *** _ody has quit IRC | 17:39 | |
| *** _ody has joined #puppet-openstack | 17:40 | |
| *** _ody has quit IRC | 17:42 | |
| *** _ody has joined #puppet-openstack | 17:42 | |
| *** _ody has quit IRC | 17:50 | |
| *** _ody has joined #puppet-openstack | 17:50 | |
| mgagne | RichardRaseley: I replied to your comment in https://review.openstack.org/#/c/186109/ | 17:50 |
| RichardRaseley | mgagne: Thank you, reviewing now. | 17:50 |
| *** _ody has quit IRC | 17:54 | |
| *** _ody has joined #puppet-openstack | 17:54 | |
| RichardRaseley | mgagne: Thank you for your response, I am very happy that we're able to move forward and look forward to working with you more in the future. I am in agreement with your suggestions and will make the appropriate modifications. Just to make sure that I understand fully, this means I would be abandoning the two changes I've proposed to puppet-nova and puppet-swift and 'resurrect' (if that's the term) the relevant patches w | 17:54 |
| RichardRaseley | hich map to the query you provdied? | 17:54 |
| mgagne | RichardRaseley: My main concern is that original authorship and history are "lost in translation" when code are literally copied/paste in new change instead of restored. | 17:56 |
| mgagne | RichardRaseley: I propose 2 solutions with existing changes: add co-authored-by or abandon and integrate the work in the previously abandoned changes and add co-authored-by too. | 17:57 |
| mgagne | RichardRaseley: or any other reasonable proposition | 17:58 |
| *** rcallawa_ has quit IRC | 17:58 | |
| mgagne | RichardRaseley: as for the others changes, I understand that it might not be that easy to rebase them against master. I propose myself to restore them and rebase them against master. | 17:59 |
| mgagne | RichardRaseley: and anyone will be welcome to work on them. | 17:59 |
| *** alex_bh has joined #puppet-openstack | 18:00 | |
| *** xarses has joined #puppet-openstack | 18:02 | |
| RichardRaseley | OK, I understand. I am in agreement. I will abandon my two changes and wait for you to 'resurrect' and rebase the currently abandoned ones (maybe start with nova and swift?). I will ensure appropriate usage of the co-authored-by tag(s) on those resurrected changes. | 18:06 |
| *** kkirkpatrick has quit IRC | 18:07 | |
| *** rcallawa_ has joined #puppet-openstack | 18:08 | |
| *** kkirkpatrick has joined #puppet-openstack | 18:22 | |
| *** alex_bh has quit IRC | 18:29 | |
| *** aimon_ has joined #puppet-openstack | 18:58 | |
| *** aimon has quit IRC | 19:00 | |
| openstackgerrit | Mathieu Gagné proposed stackforge/puppet-nova: Introduce public_url, internal_url and admin_url (and v3/ec2) https://review.openstack.org/70458 | 19:03 |
| mgagne | RichardRaseley: here is the previous abandoned change rebased and adapted against master | 19:03 |
| mgagne | RichardRaseley: all previous unit tests are still passing and have been moved under "when providing deprecated endpoint parameters" | 19:04 |
| RichardRaseley | mgagne: Thank you, I see it. | 19:12 |
| RichardRaseley | mgagne: So, now it seems like we're at the stage of determining the correct implementation. | 19:13 |
| RichardRaseley | or rather, the preferred implementation. | 19:14 |
| mgagne | sure | 19:14 |
| mgagne | RichardRaseley: I'm very concerned with backward compatibility which I think I managed to address in my implementation. I have yet to receive a compelling reason to not use inline_template other than "please don't use it" or "it's not operator friendly" | 19:15 |
| mgagne | RichardRaseley: the idea is to detect *ANY* uses of the deprecated parameters and warm the user to not use them anymore. that's why they are now all undef by default and previous default values are restored if any values are passed to deprecated parameters. | 19:17 |
| *** cdent has quit IRC | 19:17 | |
| RichardRaseley | mgagne: So am I to infer that you do view 'operator friendliness' and 'readability' as a compelling factor? I am just trying to better understand what goals you view as most important within this context. | 19:17 |
| mgagne | RichardRaseley: so if the user *only* overrides the public_protocol (and nothing else), we need to detect it and ask him to migrate to the new parameters. | 19:17 |
| RichardRaseley | OK, I understand. | 19:18 |
| mgagne | RichardRaseley: I'm not the one that mentionned "operator friendliness" but this concern was raised | 19:18 |
| mgagne | RichardRaseley: I'm an "operator" but some say I'm more than that so I guess my opinion is a little biased | 19:19 |
| RichardRaseley | mgagne: Correct, but you said 'I have yet to receive a compelling reason not to use...', which implied to meet that the reasons you enumerated afterwards were considered 'compelling'. I just wanted to make sure I understood fully. | 19:19 |
| RichardRaseley | not to use other than... * | 19:19 |
| RichardRaseley | OK | 19:19 |
| mgagne | RichardRaseley: sorry, let me reword that: I have yet to received objective reasons to not use it. | 19:20 |
| RichardRaseley | OK, I understand your position. | 19:20 |
| mgagne | all puppetlabs folks I talked to said "don't use it" but couldn't provide a technical reason | 19:21 |
| *** panda has quit IRC | 19:21 | |
| mgagne | it feels like inline_template was a function people didn't want to introduce (for some reason) but had to due to customer's pressure or something like that. | 19:22 |
| *** panda has joined #puppet-openstack | 19:22 | |
| mgagne | unless you tell me it spawns a whole ruby process just to compile a template | 19:22 |
| RichardRaseley | So, if we are to merge our respective considerations, it sounds like the amount to (A) ensuring that *any* use of a deprecated parameters is handled appropriately (as you outlined above) and (B) ensuring that the solution is easily parseable and readable by operators. | 19:23 |
| RichardRaseley | mgagne: I do have concerns regarding performance of an in-line template, but haven't been able to substantiate them thus far. | 19:23 |
| mgagne | RichardRaseley: cool, it's the first time I hear about the performance tbh | 19:24 |
| mgagne | RichardRaseley: would it be any different than calling template() ? | 19:24 |
| RichardRaseley | mgagne: I do not know, TBH - but I am going to try to get the answer. | 19:25 |
| RichardRaseley | mgagne: The feedback I got from one of our Principal Engineers for Puppet was "the use of inline_template the way you pointed to is like declaring a local function and calling it - this can be done using the 'with' function and a lambda with future parser/4.x (i.e. if you do not want to leak temporary (local) variables)" | 19:42 |
| RichardRaseley | mgagne: He also mentioned that using ERB "creates a ruby context where every variable in sight becomes an instance variable - this is somewhat expensive". | 19:44 |
| mgagne | RichardRaseley: we have yet to fully support 4.x and we still wish to support v3 | 19:50 |
| mgagne | RichardRaseley: but it's as expensive as template() ? | 19:50 |
| RichardRaseley | Absent the read to the filesystem to open the actual template, yes - it is just acting on the passed string. | 19:50 |
| RichardRaseley | (as I understand it). | 19:50 |
| *** panda is now known as panda|ko | 19:52 | |
| *** finchd has quit IRC | 19:57 | |
| RichardRaseley | mgagne: ^ I am guessing that (fs read) would represent the majority of the overhead from a practical perspective, but you'd still have the overhead of the new Ruby context and all the associated instance variables (I am assuming some or all of those are duplicated from the 'parent' context). | 19:59 |
| *** finchd has joined #puppet-openstack | 19:59 | |
| RichardRaseley | Please take any speculation on my part with a grain of salt. =] | 20:00 |
| mgagne | RichardRaseley: I understand. Is it something we should really be concerned about? | 20:00 |
| mgagne | RichardRaseley: because I guess we could come up with a function instead of a template to deal with it | 20:00 |
| Hunner | hi! I see mentions of inline_template | 20:00 |
| mgagne | Hunner: does it mean we can summon by invoking inline_template ? | 20:01 |
| Hunner | Haha. I probably should highlight that in weechat :) | 20:01 |
| RichardRaseley | mgagne: I kind of doubt that the performance difference would be appreciable. | 20:01 |
| Hunner | Usually people discourage the use of inline_template() because to non-ruby devs it might as well be called magic(), and all sorts of arbitrary code happens and if cargo-culted around should have been in a reusable library in the first place | 20:02 |
| Hunner | I use inline_template all the time when hacking on stuff, but never in a final implementation since it is basically always a "dirty hack" | 20:03 |
| RichardRaseley | mgagne: But, as an operationally-focused individual, I can say say that the readability is a concern for me. I'd like to take a shot at developing a solution which would make use of as many 'native' conditionals as possible. If I can meet your concerns outlined above, does that seem reasonable? | 20:03 |
| Hunner | I would recommend using an actual template if you want custom templated content, or using a custom function if you want ruby in a manifest | 20:03 |
| mgagne | RichardRaseley: sure. As you might already know, I don't want to end up with 56 _real which is (to me) worst than the "unreadability" introduced by inline_template | 20:04 |
| mgagne | Hunner: "dirty hack' is subjective | 20:04 |
| Hunner | (I don't know the context of the discussion; is that for 70458?) | 20:04 |
| RichardRaseley | mgagne: That is a fair concern. I'll take a stab in one of the now reopened changes, and then we'll iterate on it from there. | 20:04 |
| Hunner | mgagne: totally | 20:04 |
| mgagne | Hunner: If I introduced a function named inline_ruby, I'm sure it would be seen as "better" | 20:05 |
| Hunner | mgagne: Eh, whatever you call it, it's just "on priciple" but obviously you can change the priciples | 20:06 |
| mgagne | hehe | 20:06 |
| mwhahaha | forgive me for piping in but for all those inline templates, wouldn't it be better to put in a custom parser function in puppet-openstacklib that would generate all those urls for you? | 20:07 |
| mgagne | Hunner: because even if we introduce a new function to handle the deprecation, we are just moving the "magic" elsewhere in a puppet function which is still ruby code anyway. | 20:07 |
| mwhahaha | something that could be universal and shared across all the openstack modules? | 20:07 |
| Hunner | mgagne: Looking at 70458 I would name a function something like `construct_url($public_protocol, $public_address, $compute_port, $compute_version)` though it's kind of nice the way you have it. | 20:07 |
| mgagne | mwhahaha: sure, I think we are enough with interest in this change to come up with a reasonable implementation | 20:07 |
| mgagne | Hunner: there is no default value provided in those parameters. we will have to provide them somehow | 20:08 |
| richm | do custom functions and inline_template have the same execution context on the puppet master? | 20:08 |
| Hunner | Ooo, yeah | 20:08 |
| Hunner | And doing something like passing the variables and all the defaults would be kind of awkward | 20:08 |
| mwhahaha | compute_url(host, port, version) | 20:08 |
| Hunner | richm: Almost | 20:09 |
| mgagne | Hunner: we could pass an hash instead which default values and values provided by the user | 20:09 |
| Hunner | richm: functions need to use scope.lookupvar iirc | 20:09 |
| Hunner | richm: Where templates can use instance variables | 20:09 |
| Hunner | But functions that look up arbitrary variables are discorage because spaghetti code | 20:09 |
| richm | ok - so a custom function can be equivalent in both scope and functionality to inline_template code | 20:09 |
| mgagne | Hunner: true, I would prefer an hash | 20:09 |
| Hunner | It's my one big gripe about templates too, but oh well | 20:10 |
| mgagne | Hunner: but I think some modules used to have multiple "default values" | 20:10 |
| Hunner | richm: But in the case of a function that looks up a bunch of things, I'd take the current inline_template format over that because at least the inline template keeps the code local | 20:10 |
| mgagne | but maybe it's gone now | 20:11 |
| richm | hmm - ok | 20:11 |
| mgagne | https://github.com/stackforge/puppet-neutron/blob/master/manifests/keystone/auth.pp#L89-L93 | 20:11 |
| mgagne | like that one | 20:11 |
| richm | yeah - that pattern is extremely common in puppet openstack code | 20:13 |
| Hunner | mgagne: That's like pick() isn't it? | 20:13 |
| mgagne | Hunner: come to think of it, yes | 20:13 |
| Hunner | mgagne: oh oh oh | 20:14 |
| mgagne | lets read our minds | 20:14 |
| Hunner | `construct_url(pick($public_protocol,'http'),...)` | 20:14 |
| mgagne | already done I guess | 20:14 |
| Hunner | Wait, but then why do you even need construct_url if it's just string concatination? | 20:14 |
| mgagne | Hunner: can we pass an hash to a function? I would prefer explicit over implicit | 20:15 |
| Hunner | Hmm, because you can't do "${pick('foo','bar')}" I guess | 20:15 |
| mgagne | Hunner: here is some static part to it | 20:15 |
| Hunner | mgagne: because ordering of arguments is bad? | 20:15 |
| mgagne | Hunner: yea, it's awful after 2-3 parameters | 20:16 |
| mgagne | Hunner: and because https://github.com/stackforge/puppet-openstacklib/commit/0536a214f46a7125fc634d359ae22a0cb6f35267 | 20:17 |
| mgagne | Hunner: is there a sprintf in Puppet? | 20:19 |
| Hunner | $url_hash = { 'public_protocol' => pick($public_protocol,'http'), 'public_address' => pick($public_address,'127.0.0.1'), 'compute_port' => pick($compute_port, '8774'), 'compute_version' => pick($compute_version,'v2') } ; $url = "${url_hash['public_protocol']}://${url_hash['public_address']}:${url_hash['compute_port']}/${url_hash['compute_version']}/${tenant_id}s" | 20:19 |
| Hunner | eww | 20:19 |
| mgagne | I see where this is going | 20:19 |
| Hunner | https://gist.github.com/hunner/8f2d1a7134e363e0f30b | 20:20 |
| Hunner | mgagne: Yep https://docs.puppetlabs.com/references/latest/function.html#sprintf | 20:21 |
| mgagne | See the documentation for the Kernel::sprintf function in Ruby for all the details. | 20:21 |
| mgagne | yep, helps a lot =) | 20:21 |
| Hunner | Anyway, yeah a function for that could be used if you always format the urls the same | 20:22 |
| Hunner | with compute port and version, etc | 20:22 |
| mgagne | indeed | 20:22 |
| mgagne | I think we are on to something | 20:22 |
| *** jayg is now known as jayg|g0n3 | 20:25 | |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-glance: Correctly munge glance_image is_public property https://review.openstack.org/187317 | 20:29 |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-glance: Fix is_public munge https://review.openstack.org/187318 | 20:32 |
| *** stevemar is now known as stevedore | 20:37 | |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-glance: Correct docs on format for glance::policy data https://review.openstack.org/187320 | 20:37 |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-glance: Add openstack tag to glance packages https://review.openstack.org/187322 | 20:38 |
| mgagne | so is anyone working on the implementation proposed by Hunner? | 20:41 |
| * Hunner is working on internal code :( | 20:44 | |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-glance: Support identity_uri https://review.openstack.org/187329 | 20:48 |
| openstackgerrit | Richard Raseley proposed stackforge/puppet-nova: Introduce public_url, internal_url and admin_url (and v3/ec2) https://review.openstack.org/70458 | 20:49 |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-glance: Support identity_uri in provider https://review.openstack.org/187330 | 20:49 |
| RichardRaseley | mgagne: ^ OK - here is the first stab at something which I think addresses the concerns which have been outlined thus far. I am sure I have missed one or more things, so I'd appreciate some review. | 20:49 |
| mgagne | RichardRaseley: we were talking about using pick() to still support previous default values | 20:50 |
| mgagne | RichardRaseley: what if the user provides public_protocol but not public_address nor compute_port? | 20:51 |
| mgagne | RichardRaseley: it fails with "You must either define $public_url or $public_protocol and $public_address and $compute_port" which is not the previous behavior | 20:51 |
| RichardRaseley | mgagne: OK, let me look at the previous behavior under that situation. | 20:51 |
| mgagne | I'm working on an implementation suggested by Hunner which uses pick() | 20:52 |
| RichardRaseley | mgagne: OK. Do you intend to squash over the method I was just working on? I'd be great to be able to see them side-by-side and then choose what one looks best to standardize on. | 20:53 |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-glance: Fix catalog compilation when not configuring endpoint https://review.openstack.org/187335 | 20:54 |
| mgagne | RichardRaseley: I don't know how to propose side-by-side. We can either propose 2 patch set, one after the other and diff. Or propose a new change based on the previous one. You can propose a new change by removing the Change-Id from the commit and generating a new one. | 20:55 |
| mgagne | RichardRaseley: do you suggest I create a new one for my new proposition? | 20:56 |
| RichardRaseley | mgagne: Either you or I can do it - I don't have a preference on who. | 20:56 |
| RichardRaseley | mgagne: So, previously we were shipping defaults for those values which would make it such that if one of those 3 required things wasn't specified we would take that. In this model, since our values are 'undef' would you expect a default value to be assigned should one or more of these items not be defined AND public_url is not defined? I think that is what was happening in your inline_template. | 20:57 |
| mgagne | yes, that's what my implementation adressed | 20:57 |
| RichardRaseley | OK, thank you. | 20:57 |
| mgagne | thanks to Hunner, we now have an implementation without inline_template | 20:57 |
| mgagne | Should I post mine over yours so we can diff? | 20:58 |
| mgagne | I guess we can restore a previous patch set if the need exists | 20:58 |
| RichardRaseley | mgagne: Sure. | 20:58 |
| openstackgerrit | Mathieu Gagné proposed stackforge/puppet-nova: Introduce public_url, internal_url and admin_url (and v3/ec2) https://review.openstack.org/70458 | 20:58 |
| mgagne | I didn,t run the tests, damn | 20:58 |
| mgagne | with syntax errors :D | 20:59 |
| RichardRaseley | No worries, I get the gist. | 20:59 |
| openstackgerrit | Clayton O'Neill proposed stackforge/puppet-nova: Run db-sync if nova packages are upgraded https://review.openstack.org/187336 | 21:00 |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-glance: Command parameter to sync the correct Database https://review.openstack.org/187337 | 21:00 |
| RichardRaseley | I think it is much improved, though the conditionals are reversed from the way I would expect. I would think that we would test for what we want first (the existence of 'public_url') and then if that doesn't exist break into the else block with the deprecated parameters. What are your thoughts? | 21:00 |
| RichardRaseley | mgagne: Definitely more readable from my perspective, though. | 21:01 |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-glance: Add $notification_driver parameter to notify::rabbitmq https://review.openstack.org/187340 | 21:01 |
| RichardRaseley | Ah, but we're shipping a default value for public_url so that would always match. I can't quite decide if that seems right. | 21:02 |
| mgagne | the previous url generated was that one, if NO parameter are passed, you still end up with the previous URL we had | 21:03 |
| RichardRaseley | Also, the v2 and v3 suffixes (e.g. /v2/%(tenant_id)s and /v3) are consistent, correct? Perhaps we should push the concatenation of those bits down into the URL construction (vs. being part of the parameter). | 21:03 |
| RichardRaseley | mgagne: OK, I think I am with you. | 21:04 |
| mgagne | I'm currently fixing all my typos | 21:04 |
| openstackgerrit | Mathieu Gagné proposed stackforge/puppet-nova: Introduce public_url, internal_url and admin_url (and v3/ec2) https://review.openstack.org/70458 | 21:06 |
| dfisher | yay. keystone and glance modules working on Solaris … | 21:07 |
| dfisher | glance is so … fickle. | 21:07 |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-glance: Create a sync_db boolean for Glance. https://review.openstack.org/187342 | 21:08 |
| openstackgerrit | Mathieu Gagné proposed stackforge/puppet-nova: Introduce public_url, internal_url and admin_url (and v3/ec2) https://review.openstack.org/70458 | 21:09 |
| mgagne | there | 21:09 |
| mfisch | crinkle: do you know if puppet keeps a md5 hash of a file if I'm not managing contents on it? | 21:09 |
| mfisch | to ask the right way, can I trigger a puppet action if an external process updates a given file | 21:10 |
| crinkle | mfisch: no, puppet doesn't keep track of files it's not managing | 21:11 |
| crinkle | there are too many files o na system for that to be possible | 21:11 |
| mfisch | no I would add said file but leave contents => empty | 21:12 |
| mfisch | I guess puppet would then empty my file... | 21:12 |
| mfisch | file { 'foo': } | 21:12 |
| mgagne | https://tickets.puppetlabs.com/browse/PUP-1248 | 21:12 |
| *** rcallaw__ has joined #puppet-openstack | 21:12 | |
| Hunner | mfisch: You can audit => content | 21:13 |
| mgagne | Hunner: ok, you can stop reading my mind now | 21:13 |
| Hunner | and then you can do it | 21:13 |
| mfisch | never heard of this magic | 21:13 |
| mgagne | https://docs.puppetlabs.com/references/latest/metaparameter.html#audit | 21:13 |
| mfisch | aha its not on the main types page (aka the bible) | 21:13 |
| Hunner | mgagne: XD | 21:13 |
| mfisch | this is perfect you guys | 21:13 |
| mfisch | thanks | 21:13 |
| Hunner | Yeah, what mgagne said; it's a metaparameter so works on everything | 21:14 |
| mgagne | I learned something today though | 21:14 |
| mfisch | metaparameters are not on the types page so I often dont see them, seriously | 21:14 |
| mgagne | when you search for a puppet feature and the first result is: XXX doesn't work as expected | 21:14 |
| Hunner | Same with function.html | 21:14 |
| Hunner | Luckly there are far fewer metaparameters, so just memorizing them isn't as hard | 21:15 |
| Hunner | (fewer than functions/types) | 21:15 |
| mfisch | yep | 21:15 |
| *** rcallawa_ has quit IRC | 21:16 | |
| *** rcallaw__ has quit IRC | 21:17 | |
| mfisch | I'm going to have to change the wget module if I want to add this | 21:21 |
| *** xarses has quit IRC | 21:23 | |
| *** xarses has joined #puppet-openstack | 21:41 | |
| * dfisher frowns | 21:47 | |
| dfisher | puppet-heat doesn't account for the various heat packages all having the same name | 21:47 |
| dfisher | does puppet-heat init.pp need a blob like what's in glance's init.pp file where it checks to see if glance_api_package == glance_registry_package? | 21:49 |
| *** dprince has quit IRC | 21:53 | |
| mgagne | is there a compelling reason to a multiple package_ensure for all individual openstack packages? nova-api, nova-conductor, etc. ? | 22:05 |
| *** openstackgerrit has quit IRC | 22:07 | |
| *** openstackgerrit has joined #puppet-openstack | 22:08 | |
| dfisher | mgagne: is that directed to me? | 22:08 |
| mgagne | well, more or less, it's just a question I had in my head for a couple of weeks | 22:08 |
| mgagne | and forgot to ask | 22:08 |
| dfisher | well, i'm not entire sure what package_ensure does … :/ | 22:08 |
| dfisher | oh. i see. | 22:09 |
| mgagne | dfisher: it tries to hide duplicated resource declarations by not redefining it if already defined | 22:09 |
| * dfisher nods | 22:10 | |
| dfisher | i'm getting an Alias error: http://paste.openstack.org/show/254081/ | 22:11 |
| *** fvollero_ has joined #puppet-openstack | 22:23 | |
| *** fvollero has quit IRC | 22:23 | |
| *** aimon_ has quit IRC | 22:25 | |
| *** chem has quit IRC | 22:26 | |
| openstackgerrit | Richard Megginson proposed stackforge/puppet-keystone: support for keystone v3 api - add domains to service_identity https://review.openstack.org/180586 | 22:34 |
| openstackgerrit | Richard Megginson proposed stackforge/puppet-keystone: wip - support for keystone v3 api - the user_role resource https://review.openstack.org/174999 | 22:34 |
| openstackgerrit | Richard Megginson proposed stackforge/puppet-keystone: support for keystone v3 api - v3 domain utility code https://review.openstack.org/178385 | 22:34 |
| openstackgerrit | Richard Megginson proposed stackforge/puppet-keystone: support for keystone v3 api - keystone and keystone::roles::admin https://review.openstack.org/175060 | 22:34 |
| openstackgerrit | Richard Megginson proposed stackforge/puppet-keystone: wip -support for keystone v3 api - the user resource https://review.openstack.org/174976 | 22:34 |
| openstackgerrit | Richard Megginson proposed stackforge/puppet-keystone: wip - support for keystone v3 api - the project resource https://review.openstack.org/174468 | 22:34 |
| openstackgerrit | Richard Megginson proposed stackforge/puppet-keystone: wip - support for keystone v3 api - beaker tests https://review.openstack.org/178828 | 22:34 |
| openstackgerrit | Richard Megginson proposed stackforge/puppet-keystone: support for keystone v3 api - the domain resource https://review.openstack.org/174439 | 22:34 |
| *** markvoelker_ has joined #puppet-openstack | 22:35 | |
| *** markvoelker has quit IRC | 22:37 | |
| crinkle | anyone want to take a look at this fix for the specs repo? https://review.openstack.org/#/c/186107/ the newly broken tests are blocking a couple of specs | 22:37 |
| richm | crinkle: is it really necessary to have an empty line, followed by a line with ::, followed by another empty line, followed by the preformatted and/or code block? | 22:39 |
| richm | afaik, with rst, you just need ::, followed by an empty line, followed by the preformatted/code block indented by 4 spaces | 22:40 |
| Hunner | mgagne: It looks like lint doesn't like the indentation of pick()s on https://review.openstack.org/#/c/70458/ | 22:41 |
| mgagne | Hunner: I'll take a look at it, forgot to follow up on lint | 22:42 |
| Hunner | I wonder if it just needs to be a multiple of 2 spaces, or if it has to be ONLY two spaces... | 22:42 |
| openstackgerrit | Mathieu Gagné proposed stackforge/puppet-nova: Introduce public_url, internal_url and admin_url (and v3/ec2) https://review.openstack.org/70458 | 22:45 |
| openstackgerrit | Ben Nemec proposed stackforge/puppet-tripleo: Allow use of ssl for public api endpoints https://review.openstack.org/175958 | 22:46 |
| openstackgerrit | Ben Nemec proposed stackforge/puppet-tripleo: Stop keystone admin from listening on public interface https://review.openstack.org/187371 | 22:46 |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-openstack-specs: Fix RST tests https://review.openstack.org/186107 | 22:46 |
| crinkle | richm: ^ thanks for the pointer, updated | 22:46 |
| *** openstackgerrit has quit IRC | 22:51 | |
| *** openstackgerrit has joined #puppet-openstack | 22:52 | |
| *** stevedore is now known as stevemar | 22:55 | |
| *** stevemar is now known as stevedoor | 22:55 | |
| *** aimon has joined #puppet-openstack | 22:57 | |
| *** ducttape_ has quit IRC | 22:58 | |
| *** dfisher has quit IRC | 23:03 | |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-neutron: Raise puppet error, if N-API anavail https://review.openstack.org/187381 | 23:13 |
| *** chem has joined #puppet-openstack | 23:16 | |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-neutron: Correct docs on format for neutron::policy data https://review.openstack.org/187383 | 23:18 |
| *** panda|ko has quit IRC | 23:21 | |
| *** panda|ko has joined #puppet-openstack | 23:22 | |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-neutron: Adding vxlan network type support for neutron ML2 plug-in https://review.openstack.org/187384 | 23:24 |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-neutron: Corrects "ip link set" command https://review.openstack.org/187385 | 23:25 |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-neutron: Fix neutron_network for --router:external setting https://review.openstack.org/187387 | 23:27 |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-neutron: Unescape value in parse_allocation_pool https://review.openstack.org/187388 | 23:29 |
| *** RichardRaseley has quit IRC | 23:30 | |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-neutron: Allow l3_ha to be disabled https://review.openstack.org/187389 | 23:31 |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-neutron: Fix ipv6 support https://review.openstack.org/187390 | 23:31 |
| *** gildub has joined #puppet-openstack | 23:32 | |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-neutron: Add MidoNet plugin manifest https://review.openstack.org/187392 | 23:33 |
| openstackgerrit | Colleen Murphy proposed stackforge/puppet-neutron: stop puppet from breaking neutron https://review.openstack.org/187393 | 23:34 |
| *** _ryanv_ has joined #puppet-openstack | 23:45 | |
| *** _ryanv_ has quit IRC | 23:45 | |
| *** stevedoor is now known as stevemar | 23:47 | |
| *** xarses has quit IRC | 23:52 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!