/srv/irclogs.ubuntu.com/2015/10/22/#juju.txt

bdxthedac, beisner, jamespage: here is my keystone.endpoint table post deploy: http://paste.ubuntu.com/12890703/00:16
bdxthe endpoints ending in .3x are vips, and are noticably absent00:17
bdxthedac: how did your deploy go? same?00:17
thedacturns out I can't do lxc deploys in a nested cloud. I'll have to find another way to test01:14
bdxthedac: ha-bindiface also sets to 'eth0' by default01:21
bdxI going to redeploy with these set appropriately....I'm hoping for a win here01:22
thedacack. let us know01:27
lathiatmarcoceppi: interesting, blew away my environment,m installed again.. got the same EOF errors about the containers etc but it otherwise worked fine.. weird.01:38
bdxthedac: notice my latest comment -> https://bugs.launchpad.net/charms/+source/keystone/+bug/150857503:36
mupBug #1508575: Keystone DB gets all non vip endpoints + openstack service conf files get keystone non vip <ha> <keystone> <openstack> <server> <keystone (Juju Charms Collection):New> <https://launchpad.net/bugs/1508575>03:36
bdxthedac: second to latest03:37
jamespagemarcoceppi, hey - could you add a 16.01 and 16.04 milestones to the charms distro please?10:56
jamespageI need to bump outstanding bugs off 15.10 and can't do that yet as 16.01 does not exist :-)10:58
joechi folks, any juju azure users or experts in the house?12:28
joec.. can anyone tell me if juju supports Azure Resource Groups or can anyone lead me to related documentation?12:29
joechttps://azure.microsoft.com/en-gb/documentation/articles/resource-group-overview/12:30
joecAFAIK juju still uses classic deployment but was wondering if anyone had info regarding support for Resource Groups in Azure https://azure.microsoft.com/en-gb/documentation/articles/resource-manager-deployment-model/12:34
lazypowerjoec, afaik it does not.12:49
lazypowerjoec, however if this is something you need, it would be good to file a bug on to track the conversation12:50
joecthanks @lazypower, for now I'll stick with classic deployment, I'm not sure I actually need Resource Groups functionality yet, still planning -  but will raise a bug on launchpad12:55
=== JoshStrobl is now known as JoshStrobl|Nap
marcoceppijoec: 16.01 >13:40
marcoceppierr, jamespage 16.01 >13:40
marcoceppiugh ?*13:40
jamespagemilestones on the charms distro13:40
jamespageI use them for targetting bugs13:41
marcoceppiyeah, but what's 16.01 ?13:41
jamespagejanuary next year13:41
marcoceppiokay, I get that, but it's not an ubuntu release13:41
jamespageI know - its a charm release13:41
jamespagewe do them every 3 months13:41
marcoceppithe 16.01 isn't a milestone in the distro, it's a series13:42
marcoceppioh wait, I see13:43
jamespagemarcoceppi, you create them here - https://launchpad.net/charms/+milestones13:46
jamespagethey are tied to arbitary dates of our own choosing13:46
marcoceppijamespage: yes, I see that now. I've created the xenial series in the distro, so it's open for charms to be uploaded to13:46
marcoceppiI'll add the milestone now13:46
jamespageta13:47
marcoceppijamespage: good to go13:54
jamespagemarcoceppi, thanks!14:09
=== perrito667 is now known as perrito666
=== Odd_Blok1 is now known as Odd_Bloke
=== jcsackett_ is now known as jcsackett
=== frobware_ is now known as frobware
=== Tristit1a is now known as Tristitia
jcastromthaddon: http://blog.dasroot.net/2015-charming-2-point-oh.html15:54
jcastromthaddon: https://jujucharms.com/docs/devel/authors-charm-composing15:54
=== CyberJacob is now known as zz_CyberJacob
=== JoshStrobl|Nap is now known as JoshStrobl
=== wendar_ is now known as wendar
mattraehi i'm making a charm layer.. should a file in a higher layer override a file in a lower layer?17:28
mattraei'd like to override site.conf in the apache-php layer17:28
marcoceppimattrae: we have a vendoring method for that17:56
=== skay_ is now known as skay
=== skay is now known as Guest69639
mattraemarcoceppi: ahh cool where would i look to see how to use that?18:00
marcoceppimattrae: https://github.com/johnsca/apache-php#usage18:01
=== Guest69639 is now known as skay
mattraemarcoceppi: looks like each site specified in apache.yaml will use the apache config template site.conf. i don't see a way to use a file other than site.conf, or override site.conf in a different layer18:17
marcoceppimattrae: site.conf is put in templates. If your layer has a site.conf file in a templates directory it'll overwrite it18:17
mattraemarcoceppi: great, thats what i expect.. i have a site.conf in my layer but it when i compose, the composed charm still has the site.conf from the apache-php layer18:18
marcoceppimattrae: interesting18:18
marcoceppibcsaller cory_fu ^ ?18:18
cory_fumarcoceppi, mattrae: Just tested on the vanilla layer and I'm able to override site.conf with no problem.18:22
marcoceppimattrae: do you have your layer somewhere?18:22
mattraecool thanks guys for confirming. i don't have my layer uploaded anywhere yet. i'll put it up on github18:23
cory_fumarcoceppi: Has the newest charm-tools been released?18:23
marcoceppicory_fu: no, I'm OTP, but it's going out the door tomorrow18:23
marcoceppierr18:23
marcoceppitoday18:23
cory_fuOk18:23
marcoceppiTODAY!18:23
cory_fumattrae: Can you confirm which version of charm-tools you're using?18:23
mattraecory_fu: i'm using version 1.7.1-0ubuntu4~ubuntu14.04.1~ppa118:24
cory_fuOk, that's the one I tested with18:24
marcoceppicory_fu: I've got a feature request18:28
marcoceppicory_fu: for reactive18:28
cory_fuWhat's that?18:28
cory_fubtw, I also just tested with master from charm-tools on GH and overriding site.conf works as expected18:29
marcoceppicory_fu: it would be cool if all I needed to was have an actions.yaml file, and then on creation it'll create action stubs like it does for hooks and then I can do @hook('action-name') and have it all in my reactive thing18:29
cory_fumattrae: http://pastebin.ubuntu.com/12896191/ is the expected behavior (with the name change from "compose" to "build")18:29
marcoceppicory_fu: Although, I guess that's actuall part of the build process, huh?18:31
cory_fumarcoceppi: Yeah, I had planned to add support for actions, but I hadn't gotten around to it yet.  I'm also not 100% sure if we need to consider anything else for that, though.  When an action is called, you don't necessarily want any other reactive handlers triggering, do you?18:31
mattraecory_fu: great thanks :)18:31
marcoceppicory_fu: potentially18:32
marcoceppithough i guess not18:32
marcoceppiI just don't want to have to mess with paths in my action file18:32
marcoceppito get at deps in lib18:32
cory_fumarcoceppi: Right.  Hook handlers obviously won't trigger, and any other handlers that do trigger ought to be safe, so it's probably the right thing to do.18:33
marcoceppicory_fu: so I guess it's like @action as a handler?18:34
cory_fumarcoceppi: Because you also ought to be able to guard an action based on state (e.g., only run an action handler if the DB is available)18:34
marcoceppiright18:34
cory_fumarcoceppi: The only thing I'm not 100% confident in is how surprising it will be for other, non-@hook handlers to be triggered.  I guess we just document it and teach it18:35
marcoceppiyeah18:35
cory_fumarcoceppi: So, there are a couple of parts to that.  We'll need an @action handler, analogous to @hook, in charms.reactive.  We'll also want `charm build` to parse actions.yaml and generate the boilerplate actions/ files18:38
marcoceppicory_fu: yeah18:38
cory_fuI think with that, it will also become more important to fix the issue that @hook (and now @action) can't be combined with @when, et al.18:39
cory_fumarcoceppi: https://github.com/juju/charm-tools/issues/30 and https://github.com/juju-solutions/charms.reactive/issues/1118:49
bcsallerI feel like the semantics around action handling here are quite different, we should talk about this more18:49
cory_fubcsaller: Ok, I am on board with discussing it18:49
marcoceppiditto18:49
=== natefinch is now known as natefinch-afk
=== urulama_ is now known as urulama
dweaverAnyone know how long it should take for the new openstack charms to appear in the charm store, nova-compute charm is still version 29 and should be updating to version 30?19:05
marcoceppidweaver: it can take a few hours19:19
marcoceppiupwards of 2-319:19
dweaverhi marcoceppi , reason I ask is all the others are there except nova-compute and it looks like 5 hours ago it was committed.19:20
marcoceppidweaver: then you'll need to ping rick_h_ to get some insights into ingestion19:21
dweavermarcoceppi, no problem, not urgent from me, just thought it looked a bit odd and there might be a problem, I was just trying a liberty deployment and nova-compute barfed.  I'll wait until tomorrow.19:23
dweavermarcoceppi, thanks anyway and hopefully see you in Belgium in Feb.19:26
marcoceppidweaver: definitely see you in belgium!19:30
=== zz_CyberJacob is now known as CyberJacob
marcoceppiOffice hours starting in a few mins! https://www.youtube.com/watch?v=hQ4nq_6zKJ019:58
lazypower Greetings office hours lurkers o/20:00
bdxhey whats up guys20:03
cory_fuGreetings20:03
bdxI can't make office hours, but wanted to get a collective take on my findings in comment #5 here: https://bugs.launchpad.net/charms/+source/keystone/+bug/150857520:05
mupBug #1508575: Keystone DB gets all non vip endpoints + openstack service conf files get keystone non vip <ha> <keystone> <openstack> <server> <keystone (Juju Charms Collection):New> <https://launchpad.net/bugs/1508575>20:05
bdxwould you guys be knid enough to address it? thanks20:06
bdxkind*20:08
cory_fubdx: Sure thing, we can touch on it20:09
bdxcory_fu: thanks20:10
marcoceppihttp://blog.dasroot.net/2015-charming-2-point-oh.html20:23
marcoceppihttp://interfaces.juju.solutions/20:24
bdxbeisner: thanks21:03
thedacbdx: ok, I just stood up a maas deployed (partial) openstack with keystone and other api services as lxcs. Everything is getting keystone's vip address. I also notice the lxc's have eth0 and not juju-br0 which is only on physical hosts.22:08
thedacbdx: so I am still trying to re-create your env. Is there anyting you are doing to preseeds (curtin or debiain-installer)? Can you confirm you have juju-br *inside* the lxcs?22:09
thedacSo let's try 'juju run --service keystone -- "ifconfig"'22:11
bdxthedac: omp...give me 522:49
thedacthanks22:50
thedacbdx: I need to head out shortly. To answer your questions in the bug we have not broken any conceptual problem. We have not yet identified what is causing the problem in your environment. I guess the next step is to post keystone juju logs for us to analyze.23:16
bdxthedac: I'll update you with these things tonight at some point23:27
=== CyberJacob is now known as zz_CyberJacob
bdxshoot....after an upgrade to 1.24.7 I cant get lxc units to get past allocating23:43
bdxi'll revert to 1.24.6 and report back23:44
bdxok...for some odd reason when I upgraded only juju was upgraded, not juju-core...things are moving along quite nicely now23:50

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!