=== veebers2 is now known as veebers === natefinch-afk is now known as natefinch === frankban|afk is now known as frankban [07:37] Hello Juu world! [08:51] admcleod_: Hey! How is it going? [08:52] admcleod_: Got to play a bit with the bigtop kafka offering we have packaging. Did you use Kafka for your ML demo? [11:02] kjackal: good, we havent worked on the demo yet though === dpm is now known as dpm-lunch === dpm-lunch is now known as dpm === sidarta is now known as soliveira === larrymi is now known as larrymi-link [16:10] is there a charm for celery flower? I searched, but am not getting any results on celery https://jujucharms.com/q/celery [16:10] which seems really weird [16:14] skay - not that i'm aware of [16:14] lazyPower: oh, okay. what do people normally use for monitoring celery stuff? [16:15] we've got a django app that uses the old django-celery package, just to use the admin [16:15] I dont think we've got a charm that models a celery monitor. That would be a great contribution [16:15] it's not actively developed, and the celery site recommends FLower [16:15] skay: I use flower [16:15] I think that would get out of the scope of what I'm currently working on [16:15] otherwise it would be a nice thing for me to do [16:15] bdx: do you have a charm for it? [16:16] lazyPower: I was hoping I could be lazy and [16:16] remove a dependency while I'm at it! [16:16] well, a django dependency [16:16] skay: I actually started up a celery charm last week ... I think flower would be a great configurable option [16:17] skay: I would just 'pip install flower' for the time being [16:18] skay: look out in the coming days/weeks for celery [16:18] bdx: thanks [16:18] np [16:24] skay: now that I think of it, I remember what I ran into last week when looking at it [16:25] skay: celery is interacted with by django, it usually in your django codebase that you will import and call to celery [16:26] skay: as such, I figured it was just best to let celery run where the django app runs, and get installed as one of the django app deps [16:26] bdx: aw. we are going to be deploying some non-django services [16:26] hey lazyPower, what's your charmbox docker run --net param? i have --net=bridge, and i thought i did that for reasons, but i can't remember them. curious what you use. [16:26] bdx: yeah, that is what we are doing with our django app [16:26] roadmr: o/ [16:26] kwmonroe - i've since removed that as i'm not using the 'local' provider. [16:26] kwmonroe - however, what that does is it places the container directly on the hosts networking bridge, instead of running behind a service NAT [16:27] heya skay ! [16:27] skay: yea, its just not something that is easily decoupled from the django application ... I've been on a mission to do so for a while, and the more I look at it, the more it just doesn't make sense to do so [16:28] skay: how are your other apps interacting with celery? from their native codebase? [16:28] bdx: unknown yet, we are just starting to talk about moving some things out of our big django app in to smaller apps, some of which will be go [16:29] skay: usually calls to celery will happen from the applications code base [16:29] this is difficult to decouple [16:29] bdx: before I JFDI making the actual service I want to spec out how we'll deploy in just enough detail to make sure I won't shoot us in the feet [16:30] bdx: hmm, so I guess no charm [16:30] or just a setting, like you said, and it can live int he same application container with whatever app [16:30] skay: I've charmed up a bunch of django apps that use celery [16:30] bdx: we have one that does as well [16:31] as I mentioned, I'd like to stop using the django admin for looking at the task queue [16:31] I'm not sure that package is under active development (let me go find the github issue I remember) [16:32] ack lazyPower. that's good to know. it explains why my charmbox network dies whenever i switch networks. methinks natting might keep them alive. [16:32] Hello Team, Is there any interface available @ interface.juju.solutions for apache vhost configuration? [16:32] Yeah, the only reason we did the --net=bridge was for the old local provider integration. Since we are primarily on juju2 these days, we should be able to remove that stanza from the alias and have a good time. [16:33] https://github.com/celery/django-celery/issues/391 [16:33] and some other issues. tests failing with django 1.9, etc. [16:34] kwmonroe - at some point, in the distant future when we have -spare cycles- we should pair and hax on getting it functional with the lxd cloud, and see what that looks like [16:34] sounds like a summit break-out [16:34] see you in septiembre [16:37] skay: yea, looks like you are going to be dealing with a few different issues moving forward if you are trying to be on latest and greatest django/celery/etc/etc ..... if you want these thing to start in your direction .... you have to be the change you want to see [16:38] skay: otherwise, just wait [16:38] bdx: ha, I've got a django 1.6 app and want to get to at least LTS [16:38] bdx: nod [16:39] Prabakaran: i don't see anything related to vhost data currently. it looks like the http interface only passes hostname/port. the apache2 charm handles vhosts during it's config-changed hook (so for example, changing the apache2 port would update the defined vhost config with the new value) [16:40] skay: yea, I just ransacked throught a few django apps bringing them to latest and greatest, it was pretty rough to say the least [16:40] skay: your on the right track, map out the deps for everything, and where the different projects are at (check that the desired feature sets align) [16:41] kwmonroe - iirc the reverse proxy relation sets up a vhost, but you dont get to specify the vhost information. its a templated one shot reverse proxy config [16:41] so for example, if your hosting static assets out of s3, and want to rewrite anything in the /asset url, it doesn't currently support that. [16:41] skay: then just go throught and do a timeline appraisal (expected dates when critical issues/deps will be fixed) [16:42] so there's potential for a contribution there [16:42] skay: this will give you an overhead view of what/when/where/why/how, then you can start putting the pieces together, knowing which ones will play nice, and when, why [16:43] bdx: nod [16:45] yup, +1 lazyPower [16:53] my requirment is to feed this vhost file http://pastebin.ubuntu.com/21160463/ to apache2 charm .. is it possible to do from my reactive code? Is there any interface/layer helps for this scenario? [17:11] Prabakaran: there is not anything in the interface registry to help your scenario, but it should be possible to use your pastebin with apache2 like this: juju set apache2 "vhost_http_template=$(base64 < my_vhost_template.txt)" [17:15] Prabakaran: to set the vhost template from a connected reactive charm, i *think* you might be able to use apache2's vhost-config relation. according to the apache2 readme (https://jujucharms.com/apache2/, section "Using the vhost-config relation"), you could run "relation-set vhosts="foo" from your reactive charm when website.available. [17:17] Thanks kwmonroe, let me try this [17:18] i'm pretty sure you'd need to base64 encode that, i recall the apache2 charm making heavy use of base64 encoding to ensure it preserves linebreaks and what not [17:18] might be worth while to peek in the relationship hook and see what its doing [17:19] kwmonroe - if you have a spare cycle, i'd appreciate some eyeballs on https://github.com/juju-solutions/layer-certificate-authority/pull/1 [17:22] Allah is doing [17:22] sun is not doing Allah is doing [17:22] moon is not doing Allah is doing [17:23] stars are not doing Allah is doing [17:23] planets are not doing Allah is doing [17:23] galaxies are not doing Allah is doing [17:23] oceans are not doing Allah is doing [17:23] mountains are not doing Allah is doing [17:24] trees are not doing Allah is doing [17:24] mom is not doing Allah is doing [17:24] dad is not doing Allah is doing [17:24] boss is not doing Allah is doing [17:24] job is not doing Allah is doing [17:24] dollar is not doing Allah is doing [17:25] degree is not doing Allah is doing [17:25] medicine is not doing Allah is doing [17:25] customers are not doing Allah is doing [17:25] you can not get a job without the permission of allah [17:26] you can not get married without the permission of allah [17:27] nobody can get angry at you without the permission of allah [17:27] light is not doing Allah is doing [17:27] fan is not doing Allah is doing [17:27] businessess are not doing Allah is doing [17:27] america is not doing Allah is doing [17:28] fire can not burn without the permission of allah [17:28] knife can not cut without the permission of allah [17:28] rulers are not doing Allah is doing [17:29] governments are not doing Allah is doing [17:29] sleep is not doing Allah is doing [17:29] hunger is not doing Allah is doing [17:30] food does not take away the hunger Allah takes away the hunger [17:30] water does not take away the thirst Allah takes away the thirst [17:30] seeing is not doing Allah is doing [17:31] hearing is not doing Allah is doing [17:31] seasons are not doing Allah is doing [17:31] weather is not doing Allah is doing [17:31] humans are not doing Allah is doing [17:31] animals are not doing Allah is doing [17:32] the best amongst you are those who learn and then teach quran [17:32] one letter read from book of Allah amounts to one good deed and Allah multiplies one good deed ten times [17:35] Well, that was fun [17:35] devops are not doing Juju is doing [17:36] Spamming before I joined? [17:36] phy1729 yep [17:36] Guest_94843 [17:36] was the nick [17:40] I'll stick around for a while if you don't mind, but he doesn't seem to hit the same channel twice. Continure with your magic juju :) [17:40] Thanks phy1729 for the quick response [17:40] appreciate the backup in case it happens again [17:42] TIL: don't run `charm build` in your $HOME [17:43] It doesn't do any damage, but it sure does take a long time to fail. [17:43] lol [17:43] yuuuup [17:43] gotta love that shallow deep copy ;) [17:43] balloons: tell me about this juju snap in the store [17:43] can I whole hog replace the deb? [17:43] jcastro - iirc mattyw created that snap [17:44] juju-nskaggs 2.0-beta13 nskaggs [17:44] it looks brand new to me [17:44] lazyPower, jcastro, I've put no juju snap in any store [17:44] ooo, ok. that was just prior work on the pi then [17:45] disregard my fallable memory [17:45] yeah it's like less than a day old afaict [17:45] lazyPower, I only ever built it for the pi [17:46] lazyPower, I never put it in the store because there wasn't a store in those days ;) [17:59] lazyPower: pr looks fine - want me to merge, or do you have other eyeballs on it? [18:00] go for it [18:00] donezo [18:20] ta kwmonroe [19:07] petevg: we're in a pickle. the tox.ini in bigtop-base (https://github.com/juju-solutions/layer-apache-bigtop-base/pull/28) gets overridden by the base layer tox.ini. [19:07] (https://github.com/juju-solutions/layer-basic/blob/master/tox.ini) [19:08] this is bad petevg, because charms built with the bigtop base layer *will* get the ./tests/unit/* stuff from bigtop-base, but won't have the unit_test_reqs installed, so tox fails on all bigtop charms. [19:08] kwmonroe: That is bad :-/ [19:09] kwmonroe: Do we have a way to exclude directories from being inherited in a layer? [19:10] petevg: i think there's an ignores options.. i saw beisner mention that recently. yo beisner, how does one use ignores in a layer? [19:17] stand down beisner. got it. [19:17] petevg: pretty simple, http://paste.ubuntu.com/21178635/ [19:18] kwmonroe: Cool. Will add it. Thx. :-) [19:20] kwmonroe: Just one quick question. When two layers contain a file, do you know how we handle the conflicts? I would expect layers to be able to overwrite files from a layer that they inherit from. But that would mean that layer-apache-bigtop-base's tox.ini would clobber the one from the base layer. And it sounds like the opposite is happening ... [19:24] i agree with your expectation and observation petevg. and i thought the layer ordering was fixed with https://github.com/juju/charm-tools/issues/218 [19:25] kwmonroe: Interesting. Just to be safe, I think that I'm going to rename "tox.ini" to "tox-unit.ini", and add it to the ignore list. [19:25] We'll still be able to run tests with "tox -c tox-unit.ini" [19:25] ah - cool petevg. smells like a readme update too ;) [19:26] Just working on that now :-) [19:27] petevg: may want to add unit_test_reqs.txt to the ignore list while you're in there.. [19:27] Done :-) [19:29] kwmonroe: Changes pushed. Going to spend some time doing some additional building/testing to make sure that things are okay, though. [19:42] kwmonroe: Everything looks good from my end. I can build charms, and the right stuff gets ignore from the bigtop base layer. lmk if you see any issues on your end. [20:00] roger that petevg - thx! === natefinch is now known as natefinch-afk === scuttlemonkey is now known as scuttle|afk [22:10] ping bdx === wolverin_ is now known as wolverineav