[01:00] <X-Rob> https://bugs.launchpad.net/bugs/1409548
[01:00] <mup> Bug #1409548: Multicast Address isn't randomized <hacluster (Juju Charms Collection):New> <https://launchpad.net/bugs/1409548>
[01:00] <X-Rob> Anyone wanna have a chat about that one?
[01:11] <X-Rob> jamespage: dosaboy
[01:11] <X-Rob> jamespage: dosaboy, apparently the hacluster stuff is your baby. https://bugs.launchpad.net/bugs/1409548
[01:11] <mup> Bug #1409548: Multicast Address isn't randomized <hacluster (Juju Charms Collection):New> <https://launchpad.net/bugs/1409548>
[01:17] <rick_h_> X-Rob: I know jamespage is in the UK so he's beddy-by atm I'm sure
[01:17] <rick_h_> X-Rob: an email to the list or the like might be your best bet tbh
[01:18] <X-Rob> rick_h_: It'll be easier to chat when they're awake
[01:18] <rick_h_> X-Rob: k, just heads up on delayed response at the moment
[01:18] <X-Rob> rick_h_: thanks, yeah
[01:19] <X-Rob> It's one of those 'I don't understand if this is amazingly clever, or fundamentally broken' things
[01:21] <rick_h_> if james is involved I'd vote clever :)
[01:21] <X-Rob> The issue is, there's so few bugs about this. I can't understand how it could ever work.
[01:22] <X-Rob> *Sigh*
[01:22] <X-Rob> Anyway, I'll eb around for another 8 hours or so, if I'm lucky one of them will emerge 8)
[09:20] <jamespage> gnuoy, https://code.launchpad.net/~james-page/charms/trusty/ceph-radosgw/embedded-webserver/+merge/246394
[09:20] <jamespage> just a little one pre-freeze
[09:21] <jamespage> (thats running in serverstack now)
[09:21] <jamespage> but I'm guessing that may impact on your HA stuff
[09:21] <gnuoy> jamespage, does this mean you might be up for reviewing my ceph-radosgw ha branch later today ...
[09:21] <jamespage> gnuoy, +1 yes
[09:21] <gnuoy> thanks, just finishing off some test runs
[09:22] <jamespage> gnuoy, it would be nice to get radosgw into a 'we can support this state' for 15.01
[09:22] <gnuoy> jamespage, I think ssl will still be absent
[09:22] <jamespage> gnuoy, I can live with that
[09:22] <gnuoy> kk
[09:22] <jamespage> functional + HA is good for me
[10:03] <gnuoy> jamespage, it doesn't look like there is anything to stop a user enabling the embedded_webserver on an early version of radosgw that doesn't support it
[10:03] <jamespage> gnuoy, its supported on all supported ceph versions
[10:04] <gnuoy> tip top
[10:05] <gnuoy> jamespage, approved
[13:46] <rick_h_> marcoceppi: how often to charm tests run? we're looking to make sure our new gui release yesterday fixed our issues
[13:46] <marcoceppi> rick_h_: they run whenever an update is made to a charm for promulgated charms
[13:47] <marcoceppi> rick_h_: http://juju-ci.vapour.ws:8080/job/charm-bundle-test/
[13:47] <rick_h_> marcoceppi: ok, yesterday we released trusty-17 for the gui and looking here http://reports.vapour.ws/charm-tests but not seeing it?
[13:47] <marcoceppi> rick_h_: I can kick off a test if it helps
[13:48] <rick_h_> marcoceppi: np, I just want to be proactive and make sure we got it and if not we get it fixed asap
[13:48] <marcoceppi> rick_h_: it might be in a blacklist?
[13:48] <rick_h_> looks like the Jan 14th build stopped/died
[13:48] <marcoceppi> not sure, tvansteenburg and the QA team maintain this
[13:48] <marcoceppi> yeah, exceeded public IP addresses for a cloud
[13:48] <rick_h_> ok, will be patient or bug others. This helps thanks
[13:48] <tvansteenburgh> i'm lookinc
[13:48] <tvansteenburgh> looking
[13:54] <marcoceppi> tvansteenburgh: can you also kick off a test for gui when you got a min
[13:54] <tvansteenburgh> marcoceppi: done
[13:54] <rick_h_> ty guys
[14:02] <marcoceppi> tvansteenburgh rick_h_: I think I see a problem with the way tests are structured for gui http://juju-ci.vapour.ws:8080/job/charm-bundle-test/10924/console
[14:02] <marcoceppi> good news, unit tests pass ;)
[14:03] <rick_h_> marcoceppi: yea, we're sure of that every release but making them pass in your infrastructure is our new goal
[14:03] <marcoceppi> rick_h_: +1
[14:03] <rick_h_> ok, so this looks like the other notes in the tooling frankban brought up in that original bug.
[14:03] <tvansteenburgh> marcoceppi: yeah, fixing that is on my to-do list
[14:04] <rick_h_> ok, thanks guys. Just wanted to ake sure they ran/etc
[14:13] <stub> marcoceppi: I've got a replacement Cassadra charm that should be ready for next week. It isn't backwards compatible, so I was going to get it slotted in as trusty/cassandra and leave precise/cassandra the old charm for old Cassandra 1.0.
[14:13] <marcoceppi> stub: is there no trusty/cassandra yet then?
[14:13] <stub> marcoceppi: promulgating the precise charm would make speedbumps
[14:14] <marcoceppi> stub: oh, I see
[14:14] <stub> marcoceppi: Not that I can see, no.
[14:14] <marcoceppi> stub: cool, I'll make sure it's not int he review queue
[14:14] <stub> ta
[14:14] <marcoceppi> stub: looking forward to seeing the new charm!
[14:15] <stub> You will be amazed.
[15:19] <jose> tvansteenburgh: hey, morning! have time for a quick pm?
[15:19] <tvansteenburgh> jose, yup
[15:34] <skay> mojo workspace-new is throwing permission denied, though I am not running commands any differently than I have been. and I'm still doing chmod 755 on the container.
[15:35] <skay> it wasn't throwing it yesterday, and I'm not certain what changed
[16:01] <skay> resolution -- I didn't realize I had upgraded from 0.1.2 to 0.1.3 yesterday, and there's an issue at the moment that requires a manual workaround to change perms on /var/lib/lxc for mojo
[17:23] <mwak> o/
[17:23] <marcoceppi> mwak: \o
[17:42]  * skay waves
[17:43] <skay> maybe now that I've been using mojo for a few weeks (monthish) it could be interesting to post to the juju lest and tell people my experiences. but that may be boring. dunno
[17:44] <skay> s/boring/too obscure
[17:46] <marcoceppi> skay: not boring
[17:46] <marcoceppi> please post!
[17:47] <skay> marcoceppi: I'll try to put together something interesting. I was thinking maybe one of you would want to post about it soon?
[17:48] <skay> if you think, I'll wait to post until after
[17:48] <marcoceppi> skay: it's been announced already
[17:48] <skay> oh! I missed it
[17:48] <marcoceppi> https://insights.ubuntu.com/2015/01/13/mojo-juju-service-orchestration-distilled/
[17:51] <skay> marcoceppi: spiffy! I can write something from the point of view of a new user
[18:34] <jose> marcoceppi: hey, is the queue updating?
[18:38] <wayneoutthere> uh... this is my first time to use IRC. Do I just type something here and then someone from that long list on the right might have some ideas/answer?
[18:42] <wayneoutthere> ok anyways, here is my first juju question: i have installed wordpress and owncloud charms.  wordpress displays a kind of virtual IP address in 'juju status' but owncloud does not.  How do I access owncloud via browser?
[18:45] <skay> wayneoutthere: hi, I don't have an answer to your question. Since you are new to irc, I wanted to mention that if you don't get an answer right away, you can stay in the channel. a lot of times people will be "idling" and not immediately around to answer a question
[18:45] <skay> wayneoutthere: what happens in that case, is that someone may reply to you and preface the replay with your nick. Most irc clients will give you a notification when that happens, and then you will know someone is talking to you
[18:46] <wayneoutthere> thanks skay
[18:46] <skay> wayneoutthere: so I encourage you to stick around
[18:46] <wayneoutthere> thanks, I will.  I really want to blow this juju thing up
[18:46] <wayneoutthere> it's great
[18:47] <wayneoutthere> skay how do I preface it so it beeps on your screen like you just did? Thanks for your help BTW
[18:47] <skay> I have been using it for about a month now
[18:47] <skay> wayneoutthere: if you include someone's nick (nickname) in your message, they will usually get a beep
[18:47] <skay> wayneoutthere: so you may be getting a beep now because I am including your nick in my reply
[18:48] <wayneoutthere> oh, so I don't need any codes arund name or hash tags? Just type the name? skay skay did you get two beeps ;)
[18:48] <skay> also, if you want to talk to someone, you can do tab completition. like, type s and k and then tab. it will autocomplete my nick
[18:48] <wayneoutthere> sweet
[18:48] <wayneoutthere> skay:
[18:48] <wayneoutthere> nice.
[18:48] <skay> wayneoutthere: no, only one. btw, if you are new to irc and open source projects in general, you could join #openhatch.
[18:49] <wayneoutthere> thanks again for your help. i'm a big helper in the ubuntu vancouver group but never much in the backend like this but getting more and more in here.  thanks skay
[18:49] <skay> wayneoutthere: that is a friendly group who helps introduce people to the open source community and the technologies people use when participating, like irc
[18:49] <skay> wayneoutthere: cheers. see ya around
[18:49] <wayneoutthere> :)
[18:55] <jose> wayneoutthere: hey
[18:55] <jose> wayneoutthere: you need for owncloud to turn from pending state to started state, and then everything will be ready for you on juju status
[18:56] <wayneoutthere> jose: owncloud:     charm: cs:trusty/owncloud-2     exposed: true     life: dying     relations:       db:       - mysql       shared-fs:       - nfs
[18:56] <jose> wayneoutthere: you have destroyed the service, see life: dying
[18:56] <wayneoutthere> oh..sorry it's dying
[18:56] <wayneoutthere> i know
[18:57] <wayneoutthere> but it was the same except for that. I couldn't reach it while living/started
[18:57] <wayneoutthere> and that's another issue jose . it should be destroyed by now. why still 'dying'?
[18:58] <jose> wallyworld: because the relations are still up, you haven't destroyed them
[18:58] <jose> if you destroy the relations, there will be nothing left since it's dead
[18:58] <jose> and you can re-deploy them
[18:58] <jose> juju destroy-relation owncloud mysql; juju destroy-relation owncloud nfs
[18:58] <jose> also, the owncloud-nfs relation is kinda broken atm, and I'm looking forward to fix it in the near future
[19:05] <wayneoutthere> jose thanks jose.  i'll try again.  I noticed someting was a bit weird with NFS maybe that is the problem.
[19:05] <wayneoutthere> can I do it without NFS or is the functionality dependent on it?
[19:10] <wayneoutthere> jose: it's interesting. I did both a 'destroy service' via GUI as well as juju destroy-service owncloud via CLI and both seem to leave it in 'dying' state....
[19:13] <wayneoutthere> jose: nevermind. I think maybe I had to destroy the relationships before it would fully die ... let's seee. sorry for my bother by the way and thanks for your help
[19:18] <jose> wallyworld: sorry, had to step away for a min.
[19:18] <jose> let me know if you have further probs
[19:19] <jose> and you *can* do it without nfs
[19:19] <jose> works fine as a standalone service too
[19:24] <wayneoutthere> jose:  np.   i cannot seem to get rid of nfs, owncloud, or any of their relationships no matter what I do.  Could you tell me: if this was working correctly, would I see a 'virtual ip' like the other services? Ilike the dns-name
[19:24] <jose> nope
[19:24] <jose> yep*
[19:25] <jose> yes, you would see an IP
[19:25] <jose> have you tried the commands I gave you?
[19:25] <wayneoutthere> jose:  weird. i have never seen that.  commands? Maybe I missed that one.  I have done the usual ones... let me oh. now i see the commans. sorry let me try that.
[19:26] <jose> np
[19:28] <wayneoutthere> jose: i've done those commands now which I think i have done before. it breaks the relations on one of the units but it doesn't seem to fully clean this out.... very odd.  I'll give it a minmute while it processes that last command set again
[19:29] <jose> ok!
[19:29] <wayneoutthere> jose: i get this for mysql and nfs, but I can't seem to actually destroy the services
[19:29] <wayneoutthere> agent-state-info: 'hook failed: "nfs-relation-broken" for owncloud:shared-fs'
[19:30] <jose> wayneoutthere: juju resolved nfs/0
[19:30] <lazyPower> jose: seems like owncloud is in error state on nfs - so he should resolve owncloud.
[19:30] <jose> ah, yep. wayneoutthere: juju resolved owncloud/0 should do
[19:31] <wayneoutthere> is that a command to run?
[19:31] <wayneoutthere> never seen that one
[19:32] <wayneoutthere> jose: I'll just run it anyways because that's just how I am. never fear the button!  here we go
[19:32] <marcoceppi> wayneoutthere: yeah, it will tell juju you acknowledge the error and move on to the next event in the juju queue
[19:32] <jose> yep, without the 'should do'
[19:33] <wayneoutthere> jose: ok lets see
[19:58] <wayneoutthere> jose: sorry about that had phone call.  so i do not have owncloud/0
[19:58] <jose> can you paste your juju status on paste.ubuntu.com and give me the link please?
[19:58] <wayneoutthere> jose: sure. didn't know that's there was such a paste tool cool. just a sec
[20:00] <wayneoutthere> jose: cool! Love the paste site: http://paste.ubuntu.com/9751700/
[20:00] <jose> wayneoutthere: juju resolved mysql/0
[20:01] <wayneoutthere> jose: you are the fastest in the west ;) i'll try
[20:02] <wayneoutthere> jose: wow! I think that workd. that stuff seems to be gone now.
[20:02] <jose> perfect!
[20:02] <wayneoutthere> now I will try doing it again without the nfs
[20:02] <jose> now you can just go ahead and do `juju deploy owncloud`
[20:02] <wayneoutthere> beaut
[20:02] <wayneoutthere> thanks coach
[20:02] <jose> do it standalone. no nfs neither mysql
[20:02] <jose> if you have any questions, here to help
[20:02] <wayneoutthere> oh! no relations?
[20:03] <wayneoutthere> jose: cool! Looks like I got my virtual IP!
[20:03] <wayneoutthere> you da man
[20:03] <jose> \o/
[20:03] <jose> once it says open ports: 80 443 then you're ready to go
[20:03] <wayneoutthere> jose: hey, only thing is that I have to use SSH to SOCKS firefox so probably it will kill this session here in IRC but we'll see
[20:04] <wayneoutthere> let me say thanks in advance and I'll hang around here and learn and grow too
[20:04] <jose> cool :)
[20:04] <wayneoutthere> my goal for this week is to try to make suiteCRM into a charm or something
[20:05] <wayneoutthere> jose: juju is super awesome. so far in my understanding it's a 'free cpanel' and 'one click app install' system with the ability to scale customized.
[20:05] <wayneoutthere> is that a good explanation so far?
[20:05] <jose> could be, it's just that it doesn't actually work with one server, but with one cloud environment
[20:06] <jose> so it will fire machines each time a service is deployed unless specified
[20:06] <jose> but the one-click app install is mostly true :)
[20:13] <wayneoutthere> jose: i'm using it with one server (LXC)
[20:13] <wayneoutthere> seems to be working...
[20:13] <jose> oh, ok
[20:13] <wayneoutthere> but maybe it'snot. Myt port 80 is closed so i need to test in real environment soon.  We're considering hosting small hosting bizz using this as backend (friends family to start)
[20:14] <wayneoutthere> ok here goes test likley i will fae out here hope to find my way back. thanks jose
[20:14] <jose> np, hope it works!
[20:15] <wayneoutthere> rrnwexec is that you pal??? haha
[20:15] <wayneoutthere> jose: awesome. looks like owncloud is now working.  you are appreciated.
[20:15] <jose> yay
[20:16] <wayneoutthere> and looks like i'm still online
[20:20] <dalek57> In the charmhelpers servicemanager docs, there are a lot of mentions of MySQLRelation(). Where does that come from? Does the mysql charm author implement that, or do I? There's a class in the docs that takes a RelationContext. Is that the same?
[20:20] <dalek57> jose: ^^
[20:21] <jose> dalek57: sorry, don't work much with charmhelpers. though marcoceppi could probably give a hand :)
[20:26] <dalek57> marcoceppi: you around?
[20:27] <wayneoutthere> jose: now back to my orignal question.  If someone wants to reach my 'owncloud' service and a different person needs to reach my 'wordpress' service, and they are now both running (as you could see in my paste) how do I configiure thaat?  Do I do it with subdomains,e tc?  If i enterr hostIP i think it will default to
[20:27] <jose> dalek57: probably on a meeting, but he'll surely reply once he's back around
[20:27] <wayneoutthere> sorry that wsa incomplete
[20:28] <wayneoutthere> good enough though
[20:28] <jose> wayneoutthere: unfortunately you wouldn't be able to do that with LXC
[20:28] <wayneoutthere> oh?
[20:28] <wayneoutthere> interesting....
[20:28] <wayneoutthere> so it would only work if someone could VPN in or someting?
[20:28] <jose> if you're working with the local provider it would be pretty difficult and would require manual intervention
[20:28] <jose> it can be done, yes, but requires manual intervention
[20:29] <wayneoutthere> I see. it's not automated which totally defeats the purpose of juju...
[20:29] <jose> exactly
[20:29] <wayneoutthere> so we would need to invest in a bunch of proper storage (or rent from provider) and then delply that way in the other nonLXC ways
[20:29] <jose> yep
[20:29] <jose> have you tried AWS?
[20:29] <wayneoutthere> roger. no i have not
[20:29] <jose> they give you a year completely free
[20:30] <wayneoutthere> what is it?
[20:30] <wayneoutthere> aws...
[20:30] <jose> Amazon Web Services, a cloud provider
[20:30] <wayneoutthere> oh, well I'm working with that dude up there rrnexec on something local maybe so it might happen
[20:30] <wayneoutthere> a year free is good
[20:30] <wayneoutthere> hmm
[20:30] <wayneoutthere> good to know.
[20:32] <jose> let me know if you want me to give you a hand setting up aws
[20:34] <sebas5384> from nothing one of the lxc containers change their IP so now juju is lost
[20:35] <sebas5384> somebody know how to solve this?
[20:35] <wayneoutthere> jose: thanks.  so help me understand the very big general picture of juju.  is it designed for a company to instantly launch web services to their own organization?  Like you would set up all the stuff your compayny would need  as charms and then customize it that way  and login to 'environment' would take place via domain?  how does juju help hosting like the way i was expecting it.  It was  a very easy setup.... sorry for the 
[20:36] <jose> sebas5384: sec, let me grab a doc and answer this other question real quick
[20:36] <sebas5384> i should have to change lxc containers old ip ?
[20:36] <sarnold> wayneoutthere: btw, irc has line-length limits, you hit it at "sorry for the"
[20:36] <sebas5384> jose: sure!  :D
[20:36] <wayneoutthere> jose:  haha.  that was the end.  I'll try to control myself
[20:37] <jose> wayneoutthere: yep, I would say that. so Juju would help you save costs since one engineer can manage all services
[20:37] <jose> and you can use a cloud, which means services are scalable
[20:39] <jose> sebas5384: http://blog.dasroot.net/reconnecting-juju-connectivity.html may be able to help you
[20:39] <jose> I believe that in that .jenv file you'll find the IPs of the instances
[20:40] <sebas5384> jose: but for the lxc instances you say?
[20:40] <jose> yes
[20:40] <jose> just change the name values and that stuff
[20:40] <jose> make sure to check the correct environment.jenv file
[20:41] <sebas5384> jose: I think we are talking about different things
[20:42] <sebas5384> https://gist.github.com/anonymous/ef83f4b3d2e26c30ae1b
[20:43] <jose> read the blog post
[20:43] <jose> in the environments.jenv file there's the info about each unit
[20:43] <jose> also, I do not see any discrepancies on that status
[20:47] <sebas5384> jose: take a look at the list of lxc https://gist.github.com/anonymous/97d4b64e9d963c19c6c3
[20:48] <jose> have you checked the environments.jenv?
[20:48] <sebas5384> jose: wow! oO
[20:48] <sebas5384> i restarted the machine
[20:48] <jose> ?
[20:48] <sebas5384> jose: and the ip go back to normal
[20:49] <jose> there you go, then
[20:49] <sebas5384> yeah, strange hehe
[20:49] <sebas5384> but thanks jose!
[20:49] <sebas5384> jose++
[20:49] <jose> np
[20:54] <marcoceppi> dalek57: hey
[20:54] <marcoceppi> dalek57: so, that's implemented inside of just a charm now, cory_fu was planning on putting it in the contrib/mysql module
[20:54] <marcoceppi> dalek57: not sure the progress of that, but this is what it looks like
[20:56] <dalek57> marcoceppi: so I'm doing a custom rails deployment and I want it to connect to kafka, which is written in shell. Does it have a KafkaRelation()?
[20:56] <dalek57> the charm is written in shell
[20:56] <marcoceppi> which rails or kafka?
[20:56] <dalek57> the kafka charm is in shell, and I'm writing a rails charm in python
[20:57] <sarnold> ..
[20:57] <cory_fu> dalek57, marcoceppi: The documentation for charmhelpers on pythonhosted.org is a little out of date.  The current CH includes a MysqlRelation implementation in charmhelpers.core.services.helpers
[20:57] <marcoceppi> dalek57: no, so the kafka charm would do what it needs to, you would write the KafkaRelation in the charm you're creating to process the data from it
[20:57] <cory_fu> dalek57: http://bazaar.launchpad.net/~charm-helpers/charm-helpers/devel/view/head:/charmhelpers/core/services/helpers.py#L112
[20:57] <marcoceppi> cory_fu: thanks
[20:58] <marcoceppi> dalek57: you'd write the KafkaRelation, much like the MySQL one to define the parameters and keys needed and it woudl be used in your services framework charm
[20:58] <marcoceppi> dalek57: Kafka doesn't even have to know about it
[20:58] <cory_fu> marcoceppi: What ever happened to removing the default charm create template and making it prompt (as well as renaming the current "python" template to "python-framework")?
[20:58] <marcoceppi> cory_fu: next release of charm-tools
[20:58] <marcoceppi> so, next week?
[20:58] <cory_fu> Cool
[20:59] <marcoceppi> I have to sort out some dependency things
[20:59] <dalek57> cory_fu: thanks for the link
[21:00] <dalek57> marcoceppi: thanks as well. I'll surely be back with more questions
[21:00] <dalek57> :)
[21:00] <cory_fu> dalek57: No problem.  The relation classes are just there to encapsulated the keys that you expect and depend on from a particular relation
[21:01] <cory_fu> dalek57: I'm happy to answer any questions regarding the framework.  :)
[21:03] <tvansteenburgh> cory_fu, dalek57: just pushed latest charmhelpers docs to pythonhosted
[21:03] <cory_fu> tvansteenburgh: Awesome.  Thanks
[21:12] <dalek57> cory_fu: Is a helper for postgres? I would imagine that it would look pretty similar to the mysql one
[21:12] <cory_fu> It would look very similar, yes, but I don't think there is one available yet, so you would need to implement it.
[21:24] <lazyPower> What are the implications of --upload-tools? When you run juju upgrade does it expect you to continue to upload the toolchain for that series?
[21:24] <lazyPower> s/series/version
[21:25] <marcoceppi> lazyPower:  I'm going to assume yes, --upload-tools should not really be used
[21:25] <lazyPower> you *have* to use it on the beta tools because streams cannot find them
[21:25] <marcoceppi> lazyPower: we have betastreams...
[21:25] <lazyPower> however - im' wondering why its not default behavior to fall back to --upload-tools vs making me know about it.
[21:25] <lazyPower> marcoceppi: try bootstrapping a provider with 1.21.beta-r
[21:25] <lazyPower> *beta-4
[21:26] <lazyPower> it will fail 100% of the time due to missing tools - i've been passing --upload tools all week, and i'm curious why it doesnt fall back to that
[21:26] <lazyPower> if there's some kind fo inherent behavior modification to long running environments
[21:26] <marcoceppi> lazyPower: you have to set the stream to devel
[21:26] <marcoceppi> there's docs on this somewhere
[21:27] <lazyPower> i mean its nothing serious - but my lack of knowledge around whats going on other than compiling the stuff and shipping it off to the state server - which is all i know its doing
[21:27] <marcoceppi> agent-stream: devel
[21:27] <marcoceppi> it's in the announce email
[21:27] <lazyPower> I'll ping the list with this question, seems like a good place for it
[21:28] <marcoceppi> true, but utlimtealy the answer is never use --upload-tools unless you're doing devel
[21:33] <wayneoutthere> jose: thanks for your answer earlier. had to have a bite to eat.  i'm thinking that there must be a way to use subdomains and domains to direct WAN traffic to specific juju services.... even if it was manual intervention.  My friend could say "I want a wordpress site"
[21:33] <wayneoutthere> jose: and I could say "ok, give me $10/month and I will install and maintain a word press site for you " and then use juju to deploy it in an instant
[21:33] <jose> wayneoutthere: you could with apache, yes. but as you said, requires manual intervention and defeats the purpose of juju
[21:34] <wayneoutthere> so in that case I would be better using a cpanel style traditional system?
[21:34] <jose> haven't worked with cpanel much?
[21:34] <wayneoutthere> it's just the interface where customers can manage their own hosting.
[21:34] <wayneoutthere> email, domains, subdomains, blah blah
[21:35] <jose> yeah. haven't worked with it much and wouldn't be able to provide insight
[21:35] <wayneoutthere> no problem.
[21:35] <wayneoutthere> i think juju has the potential to replace a good part of cpanel somehow
[21:35] <wayneoutthere> that would be a killer market.
[21:35] <wayneoutthere> hmm. let me look a bit deerp in the web world
[21:35] <wayneoutthere> thanks jose
[21:36] <jose> sure :)
[21:39] <dalek57> cory_fu: I'm confused about how "start" works in charmtools. In order to start my application, I need a connection to postgres. So I put PostgresRelation() in the "required_data" field, then implement that class. Then I make an action called "handle_pg" and put a reference to it in "data_ready". Do I need to separately define "postgresql-relation-* hooks?
[21:42] <wayneoutthere> jose: idea;.  is it possible to have a domain (www.example.com) have that point to dns server which points it to MyServer (normal so far) and then somehow tell it to go to the virtual IP like (10.0.3.254) by means of forward slash directory or subdomain?  c
[21:42] <cory_fu> dalek57: You need to define postgresaql-relation-* hooks (at a minimum -changed) such that it calles manager.manage().  With your PostgresRelation in "required_data", your "handle_pg" action will only be called once the postgres relation is connected and it has sent over the expected keys.  The data for the relation can be gotten either from the PostgresRelation() instance (it's a glorified dict) or via normal hookenv methods
[21:43] <cory_fu> dalek57: If you use the PostgresRelation() instance, you would access the data as pg_rel_instance[pg_rel_instance.name][0]['key']
[21:43] <jose> wayneoutthere: nope, it'd make it crash since 10.* are internal addresses
[21:43] <wayneoutthere> right....
[21:43] <dalek57> cory_fu: What is the best way to see what keys postgresql is going to send over?
[21:44] <wayneoutthere> can you install more than one instance of juju on one server? I'm guessing no... like have one in subdom.example.com and another one in subdom2.example.com?
[21:44] <cory_fu> If you want to use the hookenv methods, keep in mind that other requirements (if you have any) might not be met during the posgres hooks, so you will need to figure out the proper relation ID (which is why the PostgresRelation instance is easier)
[21:45] <cory_fu> dalek57: It should be documented in the postgres charm.
[21:45] <jose> wayneoutthere: nope, that's not possible
[21:45] <cory_fu> In the README, hopefully
[21:45] <wayneoutthere> jose: thanks
[21:50] <wayneoutthere> jose: wait... but if i manually change the port of 10.0.3.254 (internal) which is running service (ie. wordpress) to port 82 (for example) instead of default then I should be able to manually direct taraffic to that service... or with the router.. ?
[21:51] <jose> wayneoutthere: eh, first you'd have to make sure that the LXC address is the one exposed to the world with the public IP
[21:51] <wayneoutthere> yes, it already is
[21:51] <wayneoutthere> oh... you mean the internal?
[21:51] <wayneoutthere> lxc is the 10.0?
[21:51] <jose> exactly
[21:52] <wayneoutthere> oh... eeks
[21:52] <wayneoutthere> thanks
[21:52] <jose> so you'd need to redirect the public ip to your internal ip
[21:52] <jose> and then the internal one to the lxc one
[22:03] <wayneoutthere> rrnwexec: hey coach.  this guy jose has been awesome. thanks for juju irc idea
[22:05] <rrnwexec> jose IS awesome :)
[22:05] <rrnwexec> and the juju folks on this channel are the coolest people I know ;)