[00:00] <tiger7117> hmm.. yes, i noticed this magic thing on GUI. otherwise command line deploying/configuration is kind a same as those old ages way :)
[00:01] <lazyPower> tiger7117: i'm working on a demo video to cover one of our stories with juju, which is getting moving with a big data stack (Hadoop to be specific) - which can take hours for an expert to manually setup, and scale. Using juju and bundles, you can have your hadoop installation ready to go and scale with your needs in literally 14 minutes.  Keep in mind that Service orchestration is a layer above config management.
[00:02] <lazyPower> We're talking about services talkign to one another out of the box in Service oriented architecture principals. Meaning when you deploy your web app that has a MySQL dependency, it exchanges teh username, password, and database - transparently to you. It eases quite a few of your operational goals by intelligently automating away pain points.
[00:02] <lazyPower> https://www.youtube.com/watch?v=CEfFy6tODrQ - this may help shed some light on what we're talking about
[00:02] <tiger7117> ahan.. yes, its like a make a diagram on paper and magically those moutains/rivers are making itself at the back of that page :)
[00:02] <sarnold> I like that :)
[00:02] <lazyPower> tiger7117: if you want a really good intro talk i gave - i can link you to that as well
[00:02] <lazyPower> its about 40 minutes long
[00:03] <lazyPower> tiger7117: but it doesn't just stop at database credential exchange. There is work beign done to automate your DNS solution(s), Framework Charms (like Openstack, Rails, and CakePHP) that aim to be as generic as possible to support as many deployment scenarios as you will need
[00:04] <tiger7117> talk :).. naa don't have too much spare time for talking . hav to setup Godzilla on papers :P
[00:04] <lazyPower> and the best part about these is they are peer reviewed before the land in the recommended charm store for you to consume/use - which means you get the comfort of knowing you're not using some abandoned devops recipe from many months ago that will no longer work - there's a lot of work going into our end user scenario this cycle to bring automated testing into this
[00:05] <lazyPower> When you asked sarnold what the best part of juju is, its the community.
[00:05] <lazyPower> thats my 2 cents anyway.
[00:05] <lazyPower> Hope that wall of text helped tiger7117. if you have any questions, dont hesitate to ping me directly.
[00:06] <tiger7117> i appreciate you guys, its not easy or small thing, in future it can eat elephants in one bite :)
[00:07] <lazyPower> we're dangerously close to eating buffalo in one bite, next step - elephants.
[00:08] <sarnold> mmm buffalo...
[00:08] <tiger7117> haha two buffalo = 1 elephant :)
[00:09] <tiger7117> so u are near to your target.
[00:09] <sarnold> yes, world domination feels quite close ;)
[00:09] <lazyPower> every cycle we get closer.
[00:10] <tiger7117> well, i also appreciate to one guy which made plugin (JuDo), i also have been tested it, it worked too well with digital Ocean cloud, OD is SSD based too fast. you guys put built in thing of JuDo or it will be good if build for RackSpace
[00:11] <lazyPower> tiger7117: Juju DigitalOcean plugin is a pseudo provider built on top of the manual provider
[00:12] <lazyPower> juju switch manual - and you can add machines/containers from any cloud provider you wish, and orcehstrate them with juju
[00:12] <sarnold> wow
[00:12] <lazyPower> your enlistment phase will be manual, and you'll need to use the CLI to deploy services until machine view lands
[00:13] <tiger7117> yesterday, i tested JuDO and added machines and things had configured in 5 mins
[00:13] <tiger7117> LaxyPower: RackSpace also support to JuJu ?
[00:13] <lazyPower> Not natively - this is why you'd be using the manual provider.
[00:13] <tiger7117> ahan.
[00:13] <sarnold> .. which I'll admit has some niceness to it, it won't surprisingly cost you money :)
[00:14] <lazyPower> there's a few requirements to land a provider that introduce some dependency issues - such as lack of object storage - but i've been following the mailing list and some of those are planning on being lifted in the near future
[00:14] <tiger7117> because of this LXC thing, i tried Judo/Digital Ocean, there it is assigning direct Public IPs to Machines.
[00:14] <lazyPower> so you might see a native rackspace provider land sooner rather than later - but manual provider will always be around, and always offer a method to build a juju orchestrated deployment
[00:14] <tiger7117> so easy to access things publicaly.
[00:15] <lazyPower> yeah, networking is difficult when you boil it down to the specifics of a datacenter, and how you want your traffic routed - so its no surprise that density with containers isn't a fleshed out story yet.
[00:15] <tiger7117> My Best Wishes are with you Guys.. :)
[00:15] <lazyPower> Thanks tiger7117
[00:15] <sarnold> yeah, I'm hoping hallyn and stgraber can work some magic there :)
[00:16] <tiger7117> i dont like, AWS (EC2 etc) , its kind a insecure.
[00:17] <tiger7117> well, so JuJu is totally depend on Python or else as well ?
[00:19] <sarnold> the juju core depends upon go, ssh, and shell (probably bash specifically) -- charms can be written in shell, or python, or ruby, or even in compiled languages, but you'd probably want to stick with a scripting language since that makes iteration on charm authoring a lot faster
[00:20] <tiger7117> Sarnold: by the way, i can call to Dennis Ritchie /Ken Thompson, Linus Torvalds for this JuJu project :D
[00:21] <tiger7117> hmm..
[00:22] <tiger7117> Ruby didn't get too famous, like now days PHP is..
[00:23] <tiger7117> well, mine 3rd day in JuJu is going to start :) have to configure few more things in it, hope to see you guys again ..
[00:24] <tiger7117> Sarnold, LazyPower: Nice to chat with you both, Good Luck !!
[00:24] <tiger7117> Bye !!
[00:24] <lazyPower> take care tiger7117
[03:48] <veebers> what could cause a local lxc container created using 'juju deploy cs:ubuntu' to give this error when attempting to lxc-start it? "lxc-start: Executing '/sbin/init' with no configuration file may crash the host"
[03:48] <veebers> if I create one just using lxc (sudo lxc-create -t ubuntu -n testing) and try to start that it works
[03:51] <veebers> err, disregard that question. It was a user error :-\
[12:20] <tvansteenburgh1> stub: thanks for the review on the Config stuff, appreciate that!
[12:21] <stub> np
[13:07] <PiKiM> Buenos dias, tengo un inconveniente con juju-gui, lo instale pero por el navegador web me sale este error --> wss://jujugui.prueba.net/ws' failed: Unexpected response code: 400
[13:08] <PiKiM> que puede ser??
[13:10] <frankban> PiKiM: a bad request for the wss connection is weird. So you have the gui at https://jujugui.prueba.net? what kind of provider are you using?
[13:11] <PiKiM> I'm using locally do not have it in the cloud
[13:13] <frankban> PiKiM: local env? lxc?
[13:14] <PiKiM> yes lxc
[13:14] <frankban> PiKiM: so why the browser is trying to connect to wss://jujugui.prueba.net/ws?
[13:19] <PiKiM> I do not know, I have the team here and set up a virtual host to see the interface from any other team, and placed in the url https://juju.prueba.net.ws the logo stays turning and seeing the wss console displays the error: //jujugui.prueba.net/ws' failed: Unexpected response code: 400
[13:21] <frankban> PiKiM: so this seems related to an internal misconfiguration of the network. To exclude something client side, could you check if you see the same after clearing the browser cache (e.g. when browsing the GUI URL in incognito mode)?
[13:23] <PiKiM> yes, every time I see the delete browser cache
[13:23] <frankban> PiKiM: so, deleting the cache does not work, correct?
[13:25] <PiKiM> yes, correct
[13:25] <frankban> PiKiM: what URL are you using for the GUI?
[13:27] <frankban> PiKiM: could you please paste the output of a get request to https://juju.prueba.net/gui-server-info ?
[13:31] <PiKiM> frankban: https://jujugui.prueba.net
[13:33] <frankban> PiKiM: could you please paste the output of a get request to https://juju.prueba.net/gui-server-info ?
[13:34] <PiKiM> can this http://nginx.com/blog/websocket-nginx/ ??
[13:42] <PiKiM> is this what you are asking for?
[13:42] <PiKiM> curl: (60) SSL certificate problem, verify that the CA cert is OK. Details:
[13:42] <PiKiM> error:14090086:SSL routines:SSL3_GET_SERVER_CERTIFICATE:certificate verify failed
[13:42] <PiKiM> More details here: http://curl.haxx.se/docs/sslcerts.html
[13:42] <PiKiM> curl performs SSL certificate verification by default, using a "bundle"
[13:42] <PiKiM>  of Certificate Authority (CA) public keys (CA certs). If the default
[13:42] <PiKiM>  bundle file isn't adequate, you can specify an alternate file
[13:42] <PiKiM>  using the --cacert option.
[13:42] <PiKiM> If this HTTPS server uses a certificate signed by a CA represented in
[13:42] <PiKiM>  the bundle, the certificate verification probably failed due to a
[13:42] <PiKiM>  problem with the certificate (it might be expired, or the name might
[13:42] <PiKiM>  not match the domain name in the URL).
[13:42] <PiKiM> If you'd like to turn off curl's verification of the certificate, use
[13:42] <PiKiM>  the -k (or --insecure) option.
[13:44] <frankban> PiKiM: I was asking what do you see if you go to  https://juju.prueba.net/gui-server-info with your browser
[13:44] <PiKiM> this
[13:44] <PiKiM> {"uptime": 59872, "deployer": [], "apiversion": "go", "sandbox": false, "version": "0.4.0", "debug": false, "apiurl": "wss://10.0.3.1:17070"}
[13:48] <frankban> PiKiM: that seems ok. So yeah the problem seems to be a misconfiguration of the internal network, and it's likely that the team that set up the infrastructure can help with the wss failure
[13:48] <frankban> PiKiM: as it's not easy to undestand what's going on from here
[13:50] <frankban> PiKiM: perhaps they are using a proxy between the LXC machines and the network you are using, and that proxy is failing to properly redirect secure websocket connections
[13:52] <frankban> PiKiM: last check, what happens if you visit https://jujugui.prueba.net/ws with your browser?
[13:53] <PiKiM> it shows this: Can "Upgrade" only to "WebSocket".
[13:54] <frankban> PiKiM: it doesn't ask for self-signed certs approval, correct?
[13:54] <PiKiM> no, it doesnt
[13:55] <PiKiM> like you said, we are using a virtualhost to redirect the requests to the LXC machines
[13:56] <frankban> PiKiM: so, as I mentioned, if incognito mode does not work, the problem is likely to be on the network configuration. I am sorry I am not able to be more specific.
[15:02] <nuclearbob> I'm having some problems with the juju local provider, I keep getting "connection reset by peer" when the API server is trying to serve RPCs
[15:02] <nuclearbob> and my "juju ssh" sessions get broken pipe pretty quickly
[15:17] <nuclearbob> okay, I figured it out, all of my containers created by the local provider are getting created with the same MAC address
[15:17] <nuclearbob> I think veebers had that problem as well
[15:19] <JoshStrobl> New juju/doc issue: Filed in relation to the Authors Testing doc - https://github.com/juju/docs/issues/149
[15:20] <JoshStrobl> marcoceppi, when you get the time (aside from earlier ping regarding a bug closure), could you add change and/or enhancement labels to it?
[15:56] <JoshStrobl> mbruzek, or marcoceppi: MP for doc to fix some config-vagrant formatting: https://github.com/juju/docs/pull/150
[16:01] <mbruzek> JoshStrobl, Thanks for the contribution.
[16:01] <mbruzek> JoshStrobl, That does look better.
[16:02] <JoshStrobl> mbruzek, \o/
[16:07] <JoshStrobl> mbruzek, looks like you closed it but never merged (either that or it isn't updating for me)
[16:08] <JoshStrobl> there we go :D
[16:08] <mbruzek> JoshStrobl, I fixed that just now.
[16:08] <JoshStrobl> \o/
[16:19] <JoshStrobl> since the docs use, from what I can tell, GFM, would it make more sense to use their table syntax for the keyboard shortcut stuff at https://juju.ubuntu.com/docs/authors-hook-debug.html. It'd mean we'd be able to reduce the vertical space necessary for the page. Though I'm not sure if the parser detects tables and can correctly convert it to HTML tables.
[17:16] <tiger7117> Hi..
[17:17] <tiger7117> how can we change the user and password of JuJu Gui user-admin ?
[17:20] <tiger7117> where is my friend LazyPower :) ?
[17:21] <lazyPower> tiger7117: its set at bootstrap. If you don't have an admin-secret: key defined for your environment in ~/.juju/environments.yaml its autogenerated, and located in ~/.juju/environments/<environment>.jenv
[17:22] <lazyPower> i dont think it can be updated post deployment. rick_h_ - is there a config option to update the login for the GUI at present?
[17:23] <JoshStrobl> Is anyone opposed to me using GFM's tables for the tmux shortcuts at https://juju.ubuntu.com/docs/authors-hook-debug.html. I already branched my fork and tested it, it properly parses into a table and looks better than the current doc (in my opinion, of course).
[17:24] <lazyPower> JoshStrobl: screenshot?
[17:24] <tiger7117> yes, i know where to get the password of user-admin password from file ~/.juju/environments/<environment>.jenv
[17:25] <JoshStrobl> lazyPower, disclaimer: I can't get CSS fetching working with the python stuff, so I copied the HTML generated from my page into the authors-hook-debug running on the actual docs
[17:25] <JoshStrobl> lazyPower, I'll get a screenshot
[17:25] <tiger7117> any other confusion is that .. in file <enviroment>.jenv file have user: admin and Longest Password: ………..  it should be user-admin or admin, then where is user-admin username ?
[17:27] <JoshStrobl> lazyPower, actually, just found a parsing issue :\
[17:27] <lazyPower> booo
[17:27] <JoshStrobl> lazyPower, gonna work on it a bit more
[17:27] <lazyPower> ok
[17:27] <lazyPower> tiger7117: not sure what you're asking.
[17:27] <tiger7117> second might be important thing that.. JuJu-GUI Login page don't load on SAFARI browser. it only works perfect in chrome.
[17:30] <lazyPower> tiger7117: its been tested in Chrome and Firefox - but recommended to use chrome for the best experience.
[17:30] <JoshStrobl> lazyPower, had to do with having a | inside the actual table (which separates columns with |) since apparently once of the tmux shortcuts requires |. I changed it in the markdown to &#124; and it worked fine.
[17:30] <tiger7117> i said in above that .. yes, this file (~/.juju/environments/<environment>.jenv) has the auto generated long password for JuJu-GUI user-admin, even in first two lines, i have confusion that in this file (~/.juju/environments/<environment>.jenv) at first line is User: admin and in second line Password: …….. , it looks like that username of JuJu-GUI is Admin except user-admin.
[17:32] <lazyPower> tiger7117: at present there is only a single user for the GUI, and it's the default populated value of user-admin, which corresponds to user: admin in the .jenv
[17:33] <tiger7117> ahan.. if i manually change/edit to .jenv file for this password: then will it change as well or not ?
[17:34] <bac> tiger7117: the GUI uses the 'admin-secret' to login.  if you set it in your enviornments.yaml then it'll be used.  if you don't set it, juju will generate one for you.
[17:35] <bac> tiger7117: if you shutdown the environment and manually edit the .jenv file then the new admin-secret should be used when you bootstrap the new environment
[17:37] <tiger7117> bac: how can i set admin-secret login info in enviroments.yaml file ?
[17:37] <bac> tiger7117: just put it in the appropriate stanza: local, ec2, whatever
[17:37] <bac> admin-secret:  shhhhh123
[17:39] <tiger7117> bac: actually i am using digital ocean plugin (JuDo).. so mine enviroment.yaml file http://pastebin.com/XdBFN452
[17:40] <tiger7117> should i just put admin-secret: mypassword at the end of this enviroment.yaml file ?
[17:41] <bac> tiger7117: put it inside the section for the provider you are using.  in your case, inside the digitalocean section.  note you'll have to remove the corresponding .jenv file after editing environments.yaml so a new one can be regenerated.
[17:45] <tiger7117> bac: i created two machines as well  one for juju-gui and one for another purpose, so after deleting .jenv or editing to this enviroment.yaml file will it effect on those as well because this .jenv file has lot of info regarding uuid/cert/ state servers IPs/bootstrap info etc?
[17:45] <JoshStrobl> lazyPower, in theory, this is what it is supposed to look like. It'd be nice though if the Ubuntu CSS got rid of the disgusting thead color and made each initial column the same width, but nothing I can do about that. http://i.imgur.com/PssugDU.png
[17:47] <bac> tiger7117: the other services will not use the admin-secret.  realize i don't know exactly what you're doing.  it may not be prudent to tear down your environment and bootstrap another.  i assume this is just a test system?
[17:49] <tiger7117> yes, test system but dont want to spend more time again for building machines and setting up things from again .
[17:51] <lazyPower> JoshStrobl: i think that looks fine
[17:52] <lazyPower> s/fine/great
[17:52] <JoshStrobl> lazyPower, cool, I'll merge my local branches, push to GitHub and do a MP then!
[17:55] <tiger7117> Ok, thanks guys, Good Luck !!
[18:07] <JoshStrobl> MP: https://github.com/juju/docs/pull/151
[18:07] <JoshStrobl> gonna go get some food now
[18:37] <Sh3rl0ck> Is there a way to do HA with rados gateway in the JUJU based OpenStack environment
[18:39] <Sh3rl0ck> Seems like the ha-cluster charm does not have any rados gw support
[18:42] <Sh3rl0ck> Has anyone used haproxy charm with rados gw?
[19:43] <hatch> If I wanted to set up a url rewriting with the apache charm what's going to be the best approach? Fork it and deploy a custom version?
[19:49] <lazyPower> hatch: it has base64 encoded vhost support if i recall correctly
[19:49] <lazyPower> so you can write your url rewrites as a vhost, and ship it as a config option instead of forking the charm.
[19:49] <hatch> ahh I missed that in the config
[19:49] <hatch> that would work
[19:50] <hatch> now the question is, will haproxy or apache be a better choice :)
[19:54] <arosales> lazyPower: asanjar: well done on http://youtu.be/f9yTWK7Z9Wg
[20:03] <hatch> lazyPower: great vid! jw why the text got all garbled towards the end....was that my stream or is that how the source is?
[20:06] <jcastro> lazyPower, post that video to the mailing list as well!
[20:22] <lazyPower> hatch: which text?
[20:22] <lazyPower> thanks arosales :)
[20:22] <lazyPower> jcastro: on it!
[20:27] <hatch> lazyPower:  umm towards the end when you were in the console, it got very pixelated, like over processed or something
[20:27] <lazyPower> hatch: there's some issues that stemmed after the youtube upload :(
[20:28] <hatch> lazyPower:  ahh.... here http://youtu.be/f9yTWK7Z9Wg?t=8m50s is that pixelated for you?
[20:28] <lazyPower> the source is good, the youtube publish seems to have ganked a few spots towards the end. i'm seeing what you're asking about, its reading the clipping when i cut from a video source to another.
[20:28] <hatch> ahhh ok ok np I just wanted to point it out
[20:28] <hatch> great vid
[20:28] <lazyPower> yeah, it is a bit. and it replays the same few frames.
[20:28]  * lazyPower curses youtube for mangling his cut
[20:29] <lazyPower> well not bad for a first time pro-amateur cut. All filmed/cut/edited/compiled on ubuntu 14.04 using lightworks, simple screen recorder, audacity and Tracktion 5 for audio remastering.
[20:29] <lazyPower> i bet my next one i nail it.
[20:31] <hatch> lazyPower:  you should do a video, on how you make videos :D
[20:31] <lazyPower> hatch: i'm still pending one on how i DJ on linux :P
[20:31] <hatch> haha
[20:32] <lazyPower> that one is pending me cleaning my office however.
[20:32] <lazyPower> which let me tell ya is HIGH on my priority list </sarcasm>
[20:32] <hatch> lol
[20:32] <hatch> do you have any hardware?
[20:32] <hatch> "2 turntables and a microphone"
[20:32] <lazyPower> just a midi controller, hercules instinct
[20:32] <lazyPower> the rest is all software based mixing, and my turtle beach headphones + mic combo
[20:33] <lazyPower> I'm looking at investing in a mic arm, with a studio quality mic so i can get away from this headset whine that i have to remove in post
[20:33] <lazyPower> not sure if its the cans, or if its pulse being silly
[20:33] <hatch> ahh cool
[20:34] <hatch> I have 2 tech 12's from days gone by :)
[20:34] <lazyPower> http://i.ytimg.com/vi/K3sWl3QEzA4/sddefault.jpg
[20:34] <lazyPower> i love this little guy. its portable and sturdy enough to backpack
[20:34] <hatch> yeah that's how things are going nowadays
[20:34] <hatch> have you seen any of the NI stuff?
[20:35] <lazyPower> i dunno man, did you see the UK INFEST gear?
[20:35] <lazyPower> everyone there was using CD/Vinyl based setups
[20:35] <lazyPower> and routed through a laptop for recording
[20:36] <lazyPower> not sure about NI stuff - i haven't gotten that deep into teh tech. I've only been at it for ~ 3 months.
[20:36] <lazyPower> i should do a video "How to setup a pirate radio station with Juju so you can DJ with your friends across the globe"
[20:40] <hatch> haha that would be cool
[20:41] <hatch> yeah I used to be pretty big into it but.....well...time :) priorities....and such
[20:43] <Sh3rl0ck> There is no ha-proxy charm for trusty..only for precise. Is there a work around?
[20:48] <lazyPower> Sh3rl0ck: it deploys on trusty - whats preventing it from making it to the trusty side of teh charm store is automated tests (ie: amulet tests)
[20:48] <lazyPower> Sh3rl0ck: in your local charm repository (presumably ~/charms) if you mkdir -p trusty && cd trusty && charm get cs:precise/ha-proxy
[20:48] <lazyPower> then juju deploy local:trusty/haproxy
[20:49] <lazyPower> you will get the trusty charm you seek - but be aware that doing it this way means you'll have to fetch all updates until the trusty version lands in the store. Deploying a local resource is effectively forking the charm.
[20:51] <Sh3rl0ck> Oh I see...Thanks for the information. So this means I can use precise charm for trusty?
[21:08] <lazyPower> Sh3rl0ck: in most cases yes.
[21:09] <lazyPower> there are some charms that will fail due to outdated dependencies, PPA's, etc. - There's an effort at auditing the store to promote charms to trusty, but its mid-range in the priority queue at the moment. We're looking for community help and re-prioritization of the audit efforts.
[21:40] <mwhudson> hi, is it possible to use juju's openstack provider with an openstack that doesn't have swift?
[22:06] <aisrael> When a new unit is deployed, are the relation hooks re-fired?
[22:13] <marcoceppi> aisrael: yes
[22:13] <marcoceppi> relations are created between services, but they're executed on a per unit basis
[22:13] <marcoceppi> mwhudson: Juju uses the object store for metadata about the environment, so it likely won't
[22:14] <mwhudson> marcoceppi: fair enough
[22:14] <mwhudson> frobware: ^
[22:15] <frobware> mwhudson: right, thanks. so swift seems happily deployed for me today...
[22:19] <thumper> mwhudson: not yet, although we are trying to get to the state where juju manages its own object storage
[22:20] <aisrael> marcoceppi: thanks!
[22:48] <frobware> is there a means for juju status to show what is the floating ip address for openstack launched images? I see 10.10.10.x, but would like to see the floating ip assigned. If I look at the horizon dashboard it is listed and I'm able to ssh using the public addr.
[22:48] <marcoceppi> frobware: interesting, it should show in the public-address field. What version of juju are you using?
[22:49] <frobware> marcoceppi: 1.20.5-trusty-arm64
[22:49] <marcoceppi> frobware: interesting, that sounds like a bug
[22:49] <frobware> ubuntu@mustang01:~$ nova list |grep 10.4
[22:49] <frobware> | 0ca83fe4-3055-44d4-a3b2-c56e267a6b39 | juju-openstack-machine-1 | ACTIVE | -          | Running     | internal10=10.10.10.4, 192.168.2.2 |