| openstackgerrit | Arthur Berezin proposed stackforge/apps-catalog: This adds a Cloudify Orchestrator to the app catalog, which is capable to launch TOSCA Blueprints. https://review.openstack.org/217429 | 00:02 |
|---|---|---|
| kzaitsev_mb | and I'm not really sure what to do about that.. | 00:08 |
| kzaitsev_mb | duh it's 3am here, so I'll go to bed, maybe I'll have better ideas tomorrow =/ | 00:08 |
| openstackgerrit | Arthur Berezin proposed stackforge/apps-catalog: This adds a Cloudify Orchestrator to the app catalog, which is capable to launch TOSCA Blueprints. https://review.openstack.org/217441 | 00:11 |
| openstackgerrit | Arthur Berezin proposed stackforge/apps-catalog: Adding Cloudify Orchestrator to the app catalog which capable to launch TOSCA Blueprints. https://review.openstack.org/217441 | 00:21 |
| *** Berezin has joined #openstack-app-catalog | 00:24 | |
| Berezin | Hi | 00:25 |
| Berezin | I just re-submitter a patch https://review.openstack.org/#/c/217441/ | 00:25 |
| Berezin | I'd be happy for a review | 00:25 |
| kzaitsev_mb | hi Berezin I see you have submitted 2 patches | 00:27 |
| openstackgerrit | Arthur Berezin proposed stackforge/apps-catalog: fixing typo https://review.openstack.org/217445 | 00:27 |
| kzaitsev_mb | I see. you're not familiar with how gerrit works, right? | 00:28 |
| Berezin | yep :) | 00:28 |
| Berezin | how do I submit changes to -1 :) | 00:29 |
| Berezin | ? | 00:29 |
| kzaitsev_mb | ok. so first of all a review can take some time | 00:29 |
| kzaitsev_mb | so do not expect a lightning-fast review. I guess someone, who is core would have to download the image and see if it's ok, before adding the asset to the catalog | 00:30 |
| kzaitsev_mb | 2d — there is no need for 3 commits | 00:30 |
| Berezin | ok | 00:30 |
| docaedo | Berezin: I would suggest abandoning 217445 and 217441, and just update your 217429 commit to include the fixes plus add the image in | 00:30 |
| docaedo | and if you have something in progress (i.e. you want to commit, but are still working on it) you can review your own patch and mark it a -1 WIP | 00:31 |
| kzaitsev_mb | exactly — you should abandon 2d and 3d and just update the 1st commit | 00:31 |
| docaedo | (you can also do .. git review +W or something? kzaitsev_mb do you know how to submit for review but mark WIP at the same time?) | 00:31 |
| kzaitsev_mb | mark with -1 WIP or just mention WIP in the 1st line | 00:32 |
| kzaitsev_mb | docaedo: it's impossible afaik =) | 00:32 |
| docaedo | Berezin: thank you though! it's close, and should be pretty straighforward from here to get the original patch into shape | 00:32 |
| docaedo | kzaitsev_mb: ah ok, I thought maybe there was an option in gerrit-review that let you do that | 00:32 |
| kzaitsev_mb | nope =/ it's rather limited | 00:33 |
| kzaitsev_mb | docaedo: we've found a bug in our KubernetesCluster app — it was actually a simple typo, but the app won't work on a newer murano engine. And there is no way I can think of uploading an update-request. =/ | 00:34 |
| kzaitsev_mb | If i won't come up with some solution for that — may I ask you to update the binary manually tomorrow? | 00:34 |
| kzaitsev_mb | Berezin: do you need any help with editing your commit? | 00:35 |
| docaedo | kzaitsev_mb: yes, I think easiest is just a manual update TBH | 00:35 |
| docaedo | the binary store is just plain bad, and I don't think it's worth trying to come up with some complicated process | 00:36 |
| Berezin | kzaitsev_mb: docaedo I'm working on it now :) | 00:36 |
| Berezin | thanks | 00:36 |
| docaedo | better to agree it's a short term "get it done" thing, that was essentially a mistake, and we'll fix it later | 00:36 |
| docaedo | (but I'm stepping away to dinner and evening with fam - back later most likely!) | 00:36 |
| kzaitsev_mb | yep, hope we'll be able to solve that during M | 00:37 |
| kzaitsev_mb | Berezin: so do you know how to edit your commit? =) | 00:37 |
| *** Berezin has quit IRC | 00:40 | |
| *** kzaitsev_mb has quit IRC | 01:35 | |
| *** kzaitsev_mb has joined #openstack-app-catalog | 01:43 | |
| *** Berezin has joined #openstack-app-catalog | 01:49 | |
| *** Berezin has quit IRC | 01:53 | |
| *** kzaitsev_mb has quit IRC | 01:56 | |
| *** Berezin has joined #openstack-app-catalog | 02:06 | |
| openstackgerrit | Arthur Berezin proposed stackforge/apps-catalog: Adding Cloudify Orchestrator to the app catalog. Cloudify Orchstrator is capable of launching TOSCA Blueprints. https://review.openstack.org/217429 | 02:36 |
| openstackgerrit | Arthur Berezin proposed stackforge/apps-catalog: Adding image https://review.openstack.org/217462 | 02:36 |
| *** Berezin has quit IRC | 02:42 | |
| *** Berezin has joined #openstack-app-catalog | 03:07 | |
| *** Berezin has quit IRC | 03:43 | |
| *** kebray has joined #openstack-app-catalog | 04:59 | |
| *** openstackgerrit has quit IRC | 05:16 | |
| *** openstackgerrit has joined #openstack-app-catalog | 05:17 | |
| *** Berezin has joined #openstack-app-catalog | 07:02 | |
| *** Berezin has quit IRC | 07:09 | |
| openstackgerrit | Arthur Berezin proposed stackforge/apps-catalog: Adding Cloudify Orchestrator to the app catalog. Cloudify Orchstrator is capable of launching TOSCA Blueprints. https://review.openstack.org/217429 | 07:12 |
| *** Berezin has joined #openstack-app-catalog | 07:28 | |
| *** kebray has quit IRC | 07:45 | |
| *** Berezin has quit IRC | 07:55 | |
| *** Berezin has joined #openstack-app-catalog | 08:01 | |
| *** Berezin has quit IRC | 08:15 | |
| *** kzaitsev_mb has joined #openstack-app-catalog | 09:27 | |
| *** kzaitsev_mb has quit IRC | 10:07 | |
| *** kzaitsev_mb has joined #openstack-app-catalog | 11:14 | |
| *** kzaitsev1ws has joined #openstack-app-catalog | 11:53 | |
| *** kzaitsev1ws has quit IRC | 11:53 | |
| kzaitsev_mb | hey the patches are being merged =) | 12:23 |
| *** openstackgerrit has quit IRC | 12:31 | |
| *** openstackgerrit has joined #openstack-app-catalog | 12:32 | |
| kfox1111 | yup. :) | 12:52 |
| kfox1111 | docaedo: Oh. Just tried to test out Berezin's patch then realized all my test code is over on the merged assets branch. | 13:06 |
| kfox1111 | how far out do you think that is? | 13:06 |
| *** kebray has joined #openstack-app-catalog | 14:44 | |
| kzaitsev_mb | kfox1111: https://review.openstack.org/217747 | 14:55 |
| *** kebray has quit IRC | 15:07 | |
| *** kebray has joined #openstack-app-catalog | 15:34 | |
| kfox1111 | ah. cool. I'll try that out. | 15:42 |
| *** kebray has quit IRC | 15:58 | |
| docaedo | kfox1111: I'm digging myself out of things and getting caught up today - regarding single assets file, is the only thing we're waiting for the infra update to make the JSON file? Or also need to change things on the web side as well? | 16:27 |
| docaedo | I haven't looked at 216047 but will shortly | 16:28 |
| docaedo | oh man, is it Thursday already?? | 16:53 |
| kfox1111 | just infra. the plugin and website are all fixed up. | 16:53 |
| kzaitsev_mb | docaedo: yep =) | 16:53 |
| kzaitsev_mb | time flys | 16:53 |
| kzaitsev_mb | *flies | 16:54 |
| kzaitsev_mb | whatever | 16:54 |
| docaedo | I can do that infra patch today, that's easy. We need to merge for app-catalog first a patch that includes the full new path (that v1 path) so I don't have to make the change on the puppet side any more complicated than it needs to be | 16:54 |
| docaedo | just got my calendar reminder for the meeting, but not sure we need it today? guess we can have it, will just be slightly disorganized (and possibly short), unless everyone who is going to meet is already here on the channel | 16:54 |
| kfox1111 | we should probably just merge it. then when the assets.json merge hits, it will automatically trigger. | 16:54 |
| docaedo | opinions? | 16:54 |
| kfox1111 | I'm still in juno->kilo funland all day so may have to skip it. | 16:55 |
| kfox1111 | but a lot of progress. it looks like all the horizon patches are merging now. :) | 16:55 |
| kfox1111 | so the plugin should be standalone by liberty. | 16:55 |
| kfox1111 | murano support in the plugin for liberty is looking more iffy. | 16:55 |
| docaedo | +1 on merging 216047, I'll take a close look and test myself, if all goes well I'll merge and work on the infra patch to fix as well | 16:56 |
| kfox1111 | kzaitsev_mb has a patch that will fix form passing, so basic functionality may work by then. not sure. | 16:56 |
| docaedo | I propose we skip the meeting (though I'll go on the channel and announce it, and check to see if some random people planned to show up wanting to chat) | 16:56 |
| kfox1111 | hmm.... do we have manuall access to the web server? | 16:56 |
| kfox1111 | without the infra patch in place though, assets.json won't be built initially and the website will fail until its built... | 16:57 |
| docaedo | oh - I did not realize the website is now dependent on the JSON file | 16:58 |
| docaedo | (because I have not looked at that patch at all) | 16:58 |
| kzaitsev_mb | well, it feels like we should wait for infra then | 16:58 |
| kfox1111 | yeah. :/ | 16:59 |
| docaedo | I can add as part of that patch the JSON file that will be created by the infra patch easily, so will kind of be a one-off manual step (which would be necessary to actually test that patch anyway) | 17:00 |
| kzaitsev_mb | shall we go to #openstack-meeting-3 then? | 17:00 |
| docaedo | sure | 17:00 |
| kzaitsev_mb | ok, now, that I see https://review.openstack.org/#/c/205882/1/manifests/init.pp I get, why we have to wait | 17:09 |
| kzaitsev_mb | btw. kfox1111 you can use Depends-On in you patch | 17:09 |
| kzaitsev_mb | you can add "Depends-On: Ib8211b1f0a00e2d9098c7c0eb0624781fdb212d3" line to https://review.openstack.org/#/c/216331/ | 17:10 |
| docaedo | I would rather go the other way though - I can make the slight change to the app-catalog patch (manually create the assets.json file) and merge it, | 17:10 |
| docaedo | otherwise the puppet side gets more complicated (i.e. creating a new directory after checking out the repo, but before creating the JSON) | 17:11 |
| docaedo | and history tells me I'll spend some time explaining why that is necessary | 17:11 |
| docaedo | and then I'll find I chose the less favorable puppet method for creating a directory | 17:11 |
| docaedo | and I'll change it and resubmit | 17:11 |
| docaedo | and then a new person will ask why we are creating a JSON file when we have a YAML | 17:11 |
| docaedo | and then... | 17:12 |
| docaedo | and then... | 17:12 |
| docaedo | :) | 17:12 |
| kzaitsev_mb | =))) | 17:12 |
| kzaitsev_mb | oh docaedo | 17:12 |
| docaedo | BUT if you guys both prefer we handle the creation of the directory via infra patch, I'll concede and work on that | 17:12 |
| docaedo | kzaitsev_mb: yes? | 17:12 |
| kzaitsev_mb | May I ask you to upload a binary for murano app? | 17:13 |
| docaedo | absolutely, you can share details here or if I need to fetch it from somewhere confidential you can msg me directly | 17:13 |
| kzaitsev_mb | https://www.dropbox.com/s/0ne4ggsjfhdzbow/io.murano.apps.docker.kubernetes.KubernetesCluster.zip?dl=1 to /apps/io.murano.apps.docker.kubernetes.KubernetesCluster.zip | 17:13 |
| docaedo | cool, I'm on it | 17:13 |
| kzaitsev_mb | thnx | 17:14 |
| kzaitsev_mb | I'm going to be afk for some time now, will get back in an hour or so. | 17:17 |
| kzaitsev_mb | And I do not really have a stong opinion about infra thing, at least not yet ) | 17:18 |
| docaedo | OK - RAX says the file is uploaded, but could be a little while before it updates all over their CDN | 17:18 |
| docaedo | I'll put up a patch that does what I propose to 216047 and if there's an objection, we can always revert my patchset :) | 17:20 |
| *** kzaitsev_mb has quit IRC | 17:30 | |
| kfox1111 | docaedo: there's a race condition that way. wehn we remove the assets.json back out of git, then if the server ever gets blown away/put back, then the dir is no longer there. | 17:40 |
| kfox1111 | the infra patch really should create the dir I think. | 17:41 |
| docaedo | ah I see what you're saying, makes sense | 17:43 |
| docaedo | we have a race condition indeed. I can't test (or even try to merge) the infra patch because it expects assets.yaml to exist | 18:31 |
| docaedo | so we need an interim patch that creates the merged assets.yaml before the infra patch can go in | 18:31 |
| docaedo | so why not have the web site continue to pull from the assets.yaml instead of having the web site need the JSON file? | 18:32 |
| docaedo | kfox1111: what's in app-catalog has to be basically standalone IMO, and the changes (which are right now transitional). What's in 216047 is dependent on external stuff (infra manifest), for the sake of making it read from a JSON file that doesn't exist yet... | 18:40 |
| *** Berezin has joined #openstack-app-catalog | 18:41 | |
| docaedo | sorry that thought was jumbled - I am trying to say while we are transitioning the app-catalog from being just a web site, to something that is useful externally (the horizon plugin), we shouldn't make the website become dependent on something outside the website itself | 18:42 |
| docaedo | because right now, it's easy/obvious how to pull the repo and run it locally for your own development, etc. But if something outside the repo has to create directories and a JSON file in order to make the repo testable, then it's broken without some manual steps/effort | 18:43 |
| docaedo | kfox1111: I know you're in the midst of the upgrade, we can pick this up when you have time again - in the mean time I'll see if I can make something slightly different that would work for me and add review comments | 18:44 |
| docaedo | In other news though, this merged, yay! https://review.openstack.org/206773 | 18:48 |
| *** Berezin has quit IRC | 18:50 | |
| docaedo | I added a comment to 216047 with suggestion for a change that would make it possible to merge this patch :) | 18:56 |
| *** kzaitsev_mb has joined #openstack-app-catalog | 19:02 | |
| kfox1111 | docaedo: could do an intermediate one where in infra we merge all the non merged yaml's together to assets.json for testing. then we merge the apps-catalog one, which will have a working website, then we do another infra. | 19:08 |
| docaedo | but that still means the app-catalog repo stops working as a simple flat site | 19:09 |
| docaedo | my question is - why make the website pull from the JSON that is being generated from the YAML anyway when it works just fine directly from the YAML? | 19:10 |
| kfox1111 | javascript can natively parse json, not yaml. | 19:11 |
| kfox1111 | the runserver test code I added does the generation so that running it locally still works. | 19:11 |
| kfox1111 | I updated the wiki too. | 19:11 |
| docaedo | so the intention is to stop using "jsyaml.safeLoad"? | 19:13 |
| docaedo | I get that the runserver test code makes testing locally work, but I'm trying to understand the value of loading assets from the JSON instead of the YAML, that's the part I still don't see the value of | 19:14 |
| docaedo | (also, this still has us in a race condition with infra patch - until we have a merge with assets.yaml in it, the init.pp will fail tests) | 19:15 |
| docaedo | (but if we can merge a version with assets.yaml that does not break the website, the infra patch will pass tests and work) | 19:16 |
| kfox1111 | I can try and break up the patches... | 19:17 |
| docaedo | why does the web site need to load from a JSON file that doesn't exist rather than a YAML file that does? | 19:17 |
| kfox1111 | yeah... could split that into 2. do everything but removing the yaml code, then do that in a different patch. | 19:18 |
| docaedo | could be split yes, but why load from JSON instead of YAML? Is it really ONLY so we can stop using jsyaml.safeLoad? | 19:19 |
| kfox1111 | 3 reasons. 1, json's smaller because yaml | 19:19 |
| kfox1111 | s spaced out for human readableness. | 19:19 |
| kfox1111 | 2, json parsing is native to the web browser. | 19:20 |
| kfox1111 | 3, we can update/change the schema of the user changable stuff assets.yaml seperatly then what the website/plugins consume assets.json | 19:20 |
| kfox1111 | so, stars for example. | 19:20 |
| docaedo | I don't see any way we include user rating (stars, feedback, etc.) with the current approach, at least no reasonable way. | 19:22 |
| kzaitsev_mb | I'd add, that should we make api dynamic (non-static file) — it would most likely be json | 19:23 |
| kfox1111 | right. | 19:23 |
| kfox1111 | thats why having the api seperate from the content allows us to merge in stuff like stars into the api side. | 19:23 |
| kfox1111 | kzaitsev_mb: yeah. | 19:23 |
| kzaitsev_mb | and this feels like a 1st step there | 19:23 |
| docaedo | this feels to me like serious premature optimization (yaml vs. json size) as well, because we will have to make huge changes in the next few months before the site can handle several hundreds of assets. By then we'll be dynamic and only serving via api whatever the client is asking for | 19:24 |
| kzaitsev_mb | also. maybe it should be just /api/v1/assets? | 19:24 |
| docaedo | I agree this is the first step there, I'm just stuck on breaking the web site as part of this first step | 19:24 |
| kfox1111 | maybe? it leaves us open for assets.something_else if we wanted to support a yaml or something in the future. | 19:24 |
| kfox1111 | I should be able to revert just the yaml load. then do that in a follow up patch. | 19:25 |
| docaedo | definitely would be easy :) my comment has the change you'd need | 19:25 |
| kzaitsev_mb | kfox1111: well, speaking api-wise it should be controlled by Content-Type, I guess | 19:26 |
| docaedo | then we can debate the follow up patch endlessly :) | 19:26 |
| kzaitsev_mb | =)) | 19:26 |
| kfox1111 | yeah, I see some advantage to it being a get to just /v1/assets... | 19:27 |
| kfox1111 | now's the time to decide I guess. | 19:27 |
| docaedo | ON that, I'm mainly ambivalent ... to me this seems like pretty short term anyway, for the sake of the horizon plugin and right now it's really only for converting a yaml to json | 19:29 |
| kfox1111 | looking at the other project's api, /v1/assets would be more consistant. | 19:29 |
| kfox1111 | ok. I'll update it at the same time. | 19:29 |
| docaedo | ah yeah that is true, good point | 19:29 |
| docaedo | I'm fine with that expecting just assets, and will adjust the infra bit accordingly | 19:30 |
| docaedo | so is the URL going to become /v1/assets or /api/v1/assets ? | 19:31 |
| kzaitsev_mb | I'd say /api/v1/assets once again to be consistent with other OS projects APIs | 19:31 |
| kzaitsev_mb | hm | 19:32 |
| kfox1111 | +1 /api/v1/assets | 19:32 |
| docaedo | roger that | 19:32 |
| kzaitsev_mb | to tell the truth that would make us inconsistent as far as I can tell ) | 19:33 |
| kzaitsev_mb | but | 19:33 |
| kzaitsev_mb | http://developer.openstack.org/api-ref-blockstorage-v2.html | 19:34 |
| kzaitsev_mb | but — since this is a web site and accessing /index or something makes sense, I'd still say /api/v1 | 19:35 |
| kfox1111 | yeah. we're alittle odd too in that we're providing a website and api on the same endpoint. | 19:35 |
| kzaitsev_mb | With, say, nova accessing something like /login doesn't make much sence | 19:36 |
| docaedo | yeah I think we can kind of do what we want here since it's basically a pretend API right now, and this won't impact future changes (i.e. if we decide v2 will be served from somewhere else or whatever...) | 19:36 |
| kfox1111 | we can also say, / is for humans to see whats in the catalog, | 19:37 |
| kfox1111 | and /api is for machines to discover what /api/v*'s are supported. | 19:37 |
| docaedo | true | 19:37 |
| docaedo | speaking of discovery, I need to go discover a lunch, will be back in a bit | 19:37 |
| kfox1111 | hehe. | 19:38 |
| openstackgerrit | Kevin Fox proposed stackforge/apps-catalog: Merge asset files. Add v1 API. https://review.openstack.org/216047 | 19:44 |
| kfox1111 | that one has the changes. | 19:46 |
| docaedo | excellent :) | 20:14 |
| *** reed has quit IRC | 20:16 | |
| openstackgerrit | Merged stackforge/apps-catalog: Merge asset files. Add v1 API. https://review.openstack.org/216047 | 20:16 |
| openstackgerrit | Kevin Fox proposed stackforge/apps-catalog-ui: Update the plugin to use the v1 api. https://review.openstack.org/216331 | 20:17 |
| *** reed has joined #openstack-app-catalog | 20:21 | |
| *** rhagarty__ has quit IRC | 20:23 | |
| *** rhagarty_ has quit IRC | 20:23 | |
| docaedo | Review if you have a moment. Runs properly locally, I think it's good to go: https://review.openstack.org/205882 | 20:26 |
| kfox1111 | looking... | 20:26 |
| kfox1111 | extra space? or is my browser rendering odd? | 20:27 |
| kzaitsev_mb | what extra space? | 20:27 |
| kfox1111 | I'm seeing an extra line, or a wrapped set of spaces after ensure directory in file { "${root_dir}/openstack_catalog/web/api": | 20:28 |
| *** rhagarty has joined #openstack-app-catalog | 20:29 | |
| kfox1111 | hmm... doesn't show up in the unified diff though... maybe just a weird browser bug. | 20:29 |
| kzaitsev_mb | I'll checkout the change to check | 20:29 |
| *** rhagarty_ has joined #openstack-app-catalog | 20:30 | |
| kzaitsev_mb | nah, looks like gerrit is not really suited to review long lines | 20:32 |
| kfox1111 | yeah. I think its fine. | 20:32 |
| *** kebray has joined #openstack-app-catalog | 20:49 | |
| docaedo | OK first draft of app-catalog big tent application can be found here: https://etherpad.openstack.org/p/app-catalog-bigtent | 20:56 |
| docaedo | would love feedback, changes, etc. :) | 20:57 |
| kfox1111 | cool. looking... | 21:02 |
| kfox1111 | are we not an official project now? | 21:04 |
| kfox1111 | so this is a commit message? which repo? | 21:04 |
| docaedo | We are not official, this is a commit message that will include a change to the projects file in the governance repo | 21:07 |
| docaedo | Will share an example link when I get back to a computer | 21:08 |
| *** kebray has quit IRC | 21:09 | |
| kfox1111 | k. was thinking that kind of just happened when it was agreed the big tent was a thing. | 21:18 |
| kfox1111 | anything in stackforge was an official project by definition. | 21:18 |
| docaedo | Nope :) it's complicated for sure .. but basically it's only official when it's here: https://github.com/openstack/governance/blob/master/reference/projects/index.rst | 21:23 |
| docaedo | er, not THERE, bad URL :/ .. | 21:24 |
| docaedo | this is what I meant to paste: https://git.openstack.org/cgit/openstack/governance/tree/reference/projects.yaml | 21:27 |
| kfox1111 | ah. ok. | 21:44 |
| docaedo | kfox1111: any feedback on the big tent thing? I want to submit today if possible. I feel like it's pretty non-controversial, but want to make sure I'm not just charging off on my own :) | 22:27 |
| kfox1111 | sec... looked ok the first time. let me go over it one more time real quick. | 22:31 |
| docaedo | ok thanks, don't mean to rush you, just trying to get it on the agenda for the next TC meeting | 22:33 |
| kfox1111 | Added one tweak, but looks good. | 22:33 |
| kfox1111 | yeah, thats a good thing. :) | 22:33 |
| docaedo | excellent thanks! | 22:33 |
| docaedo | nice addition too, thanks. I was going back and forth on how much I wanted to talk about long term plans. Ultimately figured if I get into roadmap stuff there, it opens the roadmap up for debate via the TC. But that addition indicating there's more than just heat, glance and murano is perfect. | 22:34 |
| docaedo | https://review.openstack.org/#/c/217957/ <-- big tent application | 22:55 |
| *** Berezin has joined #openstack-app-catalog | 23:10 | |
| *** kzaitsev_mb has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!