[09:17] <cjohnston> mornin... /15
[09:44] <AskUbuntu> Unable to access MAAS nodes | http://askubuntu.com/q/473350
[10:35] <ejat> ERROR bootstrap failed: cannot start bootstrap instance: could not create hosted service: POST request failed: BadRequest - The hosted service is not valid. (http code 400: Bad Request)
[10:35] <ejat> itry to bootstrap in azure
[10:35] <ejat> anyone can help me ?
[10:36] <ejat> $ juju status -e azure
[10:36] <ejat> ERROR Unable to connect to environment "azure".
[10:36] <ejat> Please check your credentials or use 'juju bootstrap' to create a new environment.
[10:36] <ejat> Error details:
[10:36] <ejat> environment is not bootstrapped
[14:11] <jcastro> anyone see this? http://manage.jujucharms.com/~fgimenez/precise/cakephp
[14:18] <lazyPower> jcastro: i saw it in the rev queue yesterday. Looks interesting
[14:21] <dvr077> \quit
[14:43] <dpb1> Hi -- how do get this into the charm store? bzr+ssh://bazaar.launchpad.net/~landscape-charmers/charms/trusty/landscape-client/trunk/
[14:44] <dpb1> I want to be able to do juju deploy cs:trusty/landscape-client and have it work
[14:44] <khuss> maas is failing to power on sever. I am using Dell PowerEdge. Any ideas on where to look for why this is happening
[14:45] <dpb1> Right now I get: $ juju deploy cs:trusty/landscape-client
[14:45] <dpb1> ERROR charm not found: cs:trusty/landscape-client
[14:47] <dpb1> khuss: look in the /var/log/maas/celery* logs
[14:48] <dpb1> khuss: iirc
[14:48] <khuss> i don't see anything in the celery logs
[14:48] <khuss> iirc?
[14:49] <khuss> the celery log shows power on request
[14:49] <khuss> but the server is not powering on
[14:49] <dpb1> iirc = if i remember correctly
[14:49] <khuss> got it
[14:50] <dpb1> khuss: good.  and what did you configure for power settings in the page for this server?
[14:50] <khuss> IPMI2
[14:50] <khuss> the power off is working
[14:50] <khuss> but not power on
[14:50] <lazyPower> dpb1: you don't have a trusty branch promulgated. There's only a precise branch afaict
[14:51] <dpb1> lazyPower: how do I get that done?  I have to submit for review?
[14:51] <lazyPower> dpb1: has anyone reviewed landscape against trusty? I know that when I looked at the precise charm, it was g2g after some minor tweaks. if you just pushed to the landscape-charmers branch of trusty, and it wasn't promulgated into teh store API - it wont show up in the store.
[14:52] <lazyPower> dpb1: yep. A charmer will need to look at it and ack the branch then promulgate it.
[14:52] <dpb1> khuss: sec, let me check
[14:52] <khuss> dbp1: ok
[14:54] <dpb1> khuss: my next step would be to try 'ipmipower' with the right options.  The script is just a shell script and it's in /etc/maas/templates/power/ipmi.template  You can see the commmands that it sends and try them by hand.
[14:55] <dpb1> lazyPower: ok, let me put that up then.  We know of no changes necessary for trusty and have been using it there for a while now
[14:56] <khuss> dbp: sure let me check ipmipower
[15:00] <dpb1> lazyPower: ok, put up the bug: https://bugs.launchpad.net/charms/+bug/1324163  thanks!
[15:00] <_mup_> Bug #1324163: Please promulgate landscape-client to trusty <Juju Charms Collection:Fix Committed by davidpbritton> <https://launchpad.net/bugs/1324163>
[15:47] <khuss> is it possible to have maas server and juju bootstrap server on the same machine
[15:48] <khuss> i guess i don't understand why we need a separate juju instance for deploying services
[15:54] <lazyPower> khuss: seperation of concerns. You can achieve what you're looking for if you do run a manual environment. but that quickly gets cumbersome if you're managing > 10 machines.
[15:54] <dpb1> khuss: no, you cannot.  but you can deploy a service "--to 0" which will put that service on the bootstrap node.
[15:59] <khuss> dpb1: lets say we have 10 nodes in the MAAS. how to tell juju to use the first node as the bootstrap node
[16:00] <dpb1> khuss: look at 'juju help constraints'.  You can even do things like "tags=foo", which will allow fine-grained placement, if you need it.
[16:01] <khuss> can we tag a machine from MAAS and then ask juju to use that machine for bootstrapping
[16:01] <dpb1> yup
[16:01] <dpb1> works
[16:02] <khuss> dpb1: i don't see an option to tag a node from MAAS GUI
[16:02] <dpb1> khuss: let me check where it's at
[16:07] <dpb1> khuss: my mistake.  tagging maas nodes is only available through the api (maas-cli package).   'maas <profile> tag -h' will give you more details.
[16:07] <dpb1> khuss: I personally like to use constraints like memory/cpu/etc if possible.
[16:10] <khuss> dpb1: thats not very useful when you have homogenous nodes
[16:10] <dpb1> khuss: agreed
[16:11] <dpb1> khuss: I actually have a script that adds a tag of the machine name to each node, which allows you to select whatever you like.
[16:12] <dpb1> khuss: FYI: http://paste.ubuntu.com/7537513/  it's pretty simple
[16:12] <khuss> dpb1: ok thanks
[16:17] <khuss> dpb1: found the problem with the ipmi power on issues. The Dell iDRAC is not reachable from MAAS server. There is a networking problem. Thanks for the tip on using ipmipower command
[16:20] <dpb1> khuss: great.  hit that many times myself. :)
[17:08] <hackedbellini> I'm writing a charm here that connects to a db relation and provides another one. In that other relation hook, I want to be able to do a 'relation-get' for the properties defined in the db relation. Is it possible? AFAIK, if I do a relation-get inside a xxx-relation-changed, it will get those values for the xxx relation, not the db.
[17:11] <lazyPower> hackedbellini: you'll have to cache them somewhere on disk on the host, you cannot receive the values from another relation's context through relation-get.
[17:12] <hackedbellini> lazyPower: hrm, I see. It's exactly that workaround I was trying to avoid =P
[17:14] <lazyPower> hackedbellini: yeah, i'm sorry I don't have better news for you. I think that's more of a scoping thing than anything else. It's reasonable to say that relationship contexts are isolated though
[17:15] <hackedbellini> lazyPower: np :)
[17:16] <lazyPower> hackedbellini: however, if you're using charmhelpers, theres a really nice caching class that just got merged in last week.
[17:17] <lazyPower> so it makes your cache lookups fairly inexpensive
[17:21] <hackedbellini> lazyPower: hrm, seems very nice! Where can I find it? I don't know exactly what is charmhelpers, although the name is familiar somehow
[17:23] <lazyPower> hackedbellini: charmhelpers is a suite of python modules that aid charm creators. Its in the config class, it auto-caches to a dictionary on disk and willc heck there first before pulling data down otw
[17:24] <lazyPower> tvansteenburgh is the man to talk to about the feature. He proposed it lastweek
[17:25] <hackedbellini> lazyPower: very nice! I'll take a look at it, lets see if it solves my problems :)
[17:30] <tvansteenburgh> hackedbellini: this is a good intro to charmhelpers: http://www.youtube.com/watch?v=6kWfLujVwNI
[17:32] <tvansteenburgh> hackedbellini: and here's the Config lazyPower mentioned, with some sample usage: http://bazaar.launchpad.net/~charm-helpers/charm-helpers/devel/view/head:/charmhelpers/core/hookenv.py#L158
[17:33] <hackedbellini> tvansteenburgh: ahhh cool! I'll take a look at the video. And the example seems very simple! So, I can import charmhelpers directly inside the hook, if I understood correctly? No need to change PYTHONPATH or include anything?
[17:34] <tvansteenburgh> hackedbellini: yes, you'll need to get it on your path. the video covers a way to do that from within your hooks module
[17:34] <hackedbellini> tvansteenburgh: nice! Thank you for the links :)
[17:35] <tvansteenburgh> hackedbellini: you're welcome!
[18:35] <mhall119> jcastro: where can I get the latest vagrant packages from?
[18:35] <mhall119> you mentioned the one in the archives is a bit out of date
[18:36] <jcastro> http://www.vagrantup.com/downloads.html
[18:37] <mhall119> no PPA?
[18:37] <jcastro> I don't think so
[18:37] <mhall119> ok
[18:37] <jcastro> you only need it until the next ubuntu release, heh
[18:39] <mhall119> that's what they always say :)
[18:42] <lazyPower> mhall119: if you're on precise, you'll want to get it from upstream
[18:42] <lazyPower> otherwise, teh trusty package is only .1 release behind last i checked.
[18:48] <marcoceppi> mhall119: I tried backporting the trusty vagrant to precise in a ppa, it ended up being a mess
[18:50] <lazyPower> There's a nice ruby dependency chain there that needs some love
[18:57] <mhall119> lazyPower: I'm on trusty, but already installed from upstream
[18:59] <lazyPower> its not a bad idea to use the latest and the greatest. hashicorp is known for making decent products. Their newer stuff with integrated proxy'ing is really nice.
[18:59] <lazyPower> jcastro discovered it shortly after our last charm school and was demo'ing it to the team. Deploy juju-gui, let other people deploy to your VM environment.
[20:09] <khuss> i've only one environment setup in juju but the juju status gives the following message
[20:09] <khuss> ERROR Unable to connect to environment "". Please check your credentials or use 'juju bootstrap' to create a new environment.  Error details: environment is not bootstrapped
[20:09] <khuss> it doesn't show the environment name
[20:11] <marcoceppi> khuss: what does juju switch say?
[20:12] <khuss> marcoceppi: juju switch says "maas"
[20:12] <khuss> marcoceppi: which is the enviornment name
[20:12] <khuss> marcoceppi: i did juju bootstrap -e maas
[20:13] <khuss> marcoceppi: it is doing the power cycle of the blade and going thru the install
[20:13] <marcoceppi> khuss: is the node on right now?
[20:13] <marcoceppi> it might still be bootstrapping
[20:13] <marcoceppi> how long ago did you run the bootstrap
[20:13] <marcoceppi> can you ssh in to the machien?
[20:14] <khuss> marcoceppi:  it failed the first time and I just started it again
[20:14] <khuss> marcoceppi: this is what I got at the first time - curl: (7) couldn't connect to host tools from http://localhost:80/MAAS/api/1.0/files/?key=aedd9246-e6a1-11e3-af69-b8ca3a5bc3f8&op=get_by_key downloaded: HTTP 000; time 0.000s; size 0 bytes; speed 0.000 bytes/s ERROR bootstrap failed: rc: 1 Stopping instance... Bootstrap failed, destroying environment
[20:15] <khuss> marcoceppi: i did resync-tools and then just tried the bootstrap
[20:15] <marcoceppi> khuss: your mass might not be configured properly. it shouldn't say localhost for URLs
[20:15] <khuss> marcoceppi: i did that configuration because the cluster controller is on the same machine.. shd I give the IP address instead
[20:16] <marcoceppi> khuss: definitely, you can run sudo dpkg-reconfigure maas-region-controller to set the IP address again
[20:17] <khuss> marcoceppi: do I need to reconfigure the region-controller? it is just on the juju environment file right? What if i change from localhost to the ip of the machine
[20:17] <marcoceppi> khuss: OH, this is in your environments.yaml
[20:18] <khuss> marcoceppi: yes
[20:18] <marcoceppi> that should be fine
[20:18] <marcoceppi> misunderstood
[20:18] <khuss> marcoceppi: let me see how the bootstrap goes this time..  i tried curl on that link (in the error message above) and it didn't work
[20:19] <marcoceppi> khuss: well it'll probably give you a new bucket id
[20:20] <khuss> marcoceppi: failed again. 2014-05-28 20:20:03 INFO juju.cmd supercommand.go:302 running juju-1.18.3-precise-amd64 [gc] 2014-05-28 20:20:03 DEBUG juju.agent agent.go:384 read agent config, format "1.18" 2014-05-28 20:20:03 DEBUG juju.provider.maas environprovider.go:30 opening environment "maas". 2014-05-28 20:20:03 ERROR juju.cmd supercommand.go:305 Get http://localhost:80/MAAS/api/1.0/nodes/?agent_name=b4a4c73b-29e3-4333-8e0b-8f82074896
[20:21] <marcoceppi> khuss: set that to the IP address that both you and the machines can get to
[20:21] <marcoceppi> khuss: once you do that, run `rm ~/.juju/environments/maas.jenv`
[20:22] <marcoceppi> then sync-tools
[20:22] <marcoceppi> then do a bootstrap
[20:22] <khuss> marcoceppi: ok
[20:23] <khuss> marcoceppi: docs http://maas.ubuntu.com/docs/juju-quick-start.html seems to be wrong.. juju ---sync-tools doesnt work.. "juju sync-tools" does
[20:24] <marcoceppi> khuss: thanks for the heads up! I'll file a bug
[20:26] <khuss> marcoceppi: np
[20:38] <khuss> marcoceppi: bootstrap completed without error. however, juju status gives this error - ERROR state/api: websocket.Dial wss://1-10-2.master:17070/: dial tcp: lookup 1-10-2.master: no such host
[20:38] <khuss> marcoceppi: may be a DNS problem?
[20:48] <khuss> marcoceppi:  juju status still doesn't work
[20:51] <khuss> marcoceppi: actually it worked after adding the host to /etc/hosts.. may be there is a better way to use the local DNS server
[20:54] <lazyPower> khuss: have you added the maas master node to your /etc/resolv.conf along with a search declaration for the local domain? in this case it appears to be search 'master'
[20:54] <lazyPower> without quotes
[20:54] <khuss> lazyPower: no i've not .. let me try that
[21:09] <khuss> lazyPower: i added it on the /etc/resolv.conf.. looks like i need to edit some other files in Ubuntu
[21:42] <lazyPower> khuss: it wont persist past a reboot
[21:43] <lazyPower> if it works, then add it to /etc/resolve.conf.d/head (i think) and it should fix you up and persist after a reboot.
[21:54] <khuss> lazyPower: ok. thanks