=== natefinch-afk is now known as natefinch [01:57] rick_h_: I don't suppose you're actually online? === kadams54 is now known as kadams54-away === axw_ is now known as axw === urulama__ is now known as urulama === l6unchpad is now known as sky-shiny [10:41] hi guys! where could i find more information on state-port and syslog-port? im setting firewall on juju state-server. a bit more understanding on services exposed on those two ports would be helpful. [11:04] x248aGentĀ£ [12:51] Design question. A client charm is related to the PostgreSQL charm, and needs to use a PostgreSQL extension. eg. it needs to tell the unit to install a package from a PPA. [12:52] Should the client charm be allowed to instruct the PostgreSQL charm to add the PPA and install the package? Or do I need to also have a subordinate charm that does the package installation? [12:53] subordinate seems messy (although trivial, as it would be a 3 line install hook and a pile of boilerplate to keep charm-proof happy) [12:54] But allowing a client service to instruct the db to install arbitrary packages from arbitrary locations? That seems a security problem. [12:55] Well, the third option is to have the operator specify the ppa and package to install in the PostgreSQL service's config. I'm guessing that is the best option. [12:56] (and I can now block the client until the operator has actually done it) [13:33] cory_fu: ping [13:33] Hey [13:33] greetings, can you join me in#juju-gui for a minute? Frankban is asking about some specifics w/ SF that i'm not familiar with [13:40] lazyPower: tell him to join us over here ;) === redelmann is now known as rudi|ding|dnd === rudi|ding|dnd is now known as rudi|deploying|d === rudi|deploying|d is now known as rudi|deploying [15:57] is there a way to setup to which interface on machine n juju ssh n will be connect eg configure ssh proxy on the state server? [16:12] I have a subordinate-charm question, which I suppose I could just code up and see, but it's not clear from the docs.. [16:13] if I have subordinate charms A and B, and A needs some info from B, is there a relation there? [16:17] lamont: both are subordinates? [16:18] presumably attached to the same primary? [16:18] lamont: you can write a relationship that will allow them to exchange data, but you cannot deploy a subordinate to a subordinate. [16:18] yes, same primary, both subord [16:19] lazyPower: so I need to create a relationship for them to share info, just like they're real charms and everything. (DOH) [16:19] that would go a long way to explaining WHY THE CHARM ISN'T DOING WHAT THE AUTHOR THOUGHT [16:19] lamont: correct - adding a relationship/interface to do that data-exchange should be teh asme process as any other charm :) [16:20] lamont: can i ask what you're trying to do so i have more context? [16:20] hp-health thinks it's getting data from nrpe-external-master [16:20] and it's all lies [16:20] (it shows up as blank hostnames in the nagios files that are exported) [16:21] thats fun [16:21] I can see we're going to need to work on your understanding of the definition of "fun" :p [16:21] yeah if you need to proxy information through - i ran into this w/ the docker/flannel-docker stuff [16:21] i'm using flannel-docker to proxy info directly to docker (docker manages the docker config - no exceptions) - so teh subordinate is acting as an informatoin proxy and a payload delivery mechanism [16:22] the mechanics are a bit alien at first glance [16:23] it sounds like you have some of the same requirements for behavior in this particular setup [16:28] hrm... and which is the primary charm there? [16:29] properly, the main charm could proxy the nagios hostname from n-e-m and just populate it into all the other subordinate relations [16:29] or n-e-m could write it to a file :D [16:29] even though that feels wrong [16:30] n-e-m needs a good rewrite from scratch [16:30] lamont: thats highly dependent on context, and which is the manager of the concerns [16:30] http://blog.dasroot.net/2015-charm-encapsulation.html [16:35] * lamont will do some reading [16:43] and, of course, lp:charms/trusty/nrpe is the perfect place to begin this quest [16:55] jw4: you around? [16:56] tvansteenburgh: yes, but OTP for a bit [16:56] tvansteenburgh: feel free to ask and I'll try to multitask :) [16:58] jw4: afaict, the Action.ListAll api does not return info about actions run on subordinates [16:59] jw4: wanted to check with you before filing a bug [17:02] jw4: subordinate actions also don't seem to send any updates to the AllWatcher [17:07] tvansteenburgh: hmm - I don't know of a specific concept of subordinate actions, but I assume you mean actions on subordinate units? [17:08] jw4: yes [17:08] natefinch: heads up I'm around now if you still need anything [17:08] tvansteenburgh: actions and the actions watcher don't know about subordinates, so if you query actions on a unit it doesn't know to look for subordinates [17:09] tvansteenburgh: seems like a viable candidate for a bug report [17:09] rick_h_: was wondering why there's no link to the juju blog on the top of jujucharms.com [17:09] rick_h_: most of core didn't even know there was a blog [17:09] jw4: thanks for confirmation, will file bug [17:09] natefinch: because the blog isn't updated/big enuogh to warrant a link at the top :) and the homepage/community shows recent posts [17:09] tvansteenburgh: cool [17:10] "Latest from the blog" [17:10] rick_h_: fair enough. It looks like it's all external posts... do we have the ability to have our own posts hosted on jujucharms.com? [17:11] natefinch: so this is from the aggregator with things tagged juju I thought? /me dbl check. https://jujucharms.com/community/blog [17:11] natefinch: you'd have to check with the web team where/how the aggregation is pulled in. [17:11] rick_h_: ok, thanks [17:16] gah, I hate exceptions === rudi|deploying is now known as redelmann === kadams54 is now known as kadams54-away [18:03] wwitzel3: hey have you ever seen this happen before? [18:03] http://i.imgur.com/7CGtl8k.png [18:03] wwitzel3: it appears I managed to trap 2 executing hooks at the same time === kadams54 is now known as kadams54-away [18:39] tvansteenburgh: ping [18:41] bac: pong [20:17] what does relation-ids want for an argument? interface name? [20:18] yeah. relation-name [20:18] * lamont goes back to headscratching === natefinch is now known as natefinch-afk [20:31] jw4: fyi https://bugs.launchpad.net/juju-core/+bug/1457205 [20:31] Bug #1457205: Subordinate charm Action data not reported by API [20:31] thanks tvansteenburgh === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away