| === brandon is now known as Guest24204 | ||
| Guest24204 | is anyone familiar with setting up node-app on AWS. I am hitting a wall with the config.yaml for some reason | 00:05 |
|---|---|---|
| sarnold | Guest24204: any chance you can pastebin the errors you're getting? | 00:11 |
| sarnold | (I'm not familiar with the charm, and only slightly familiar with node, but you never know :) | 00:11 |
| Guest24204 | :) | 00:11 |
| === alexlist` is now known as alexlist | ||
| Guest24204 | Description of issue: I am trying to set the config file first by --config had issues with that so tried to upload my own charm my own --repository and have issue there too. Here is what I get : | 00:15 |
| Guest24204 | with --config: juju deploy --config config.yaml node-app mongonode-app | 00:18 |
| Guest24204 | error response | 00:18 |
| Guest24204 | error: no settings found for "mongonode-app" | 00:18 |
| Guest24204 | have the config file under http://github.com/TheMindCompany/mongonode-app but as mongonode-app.yaml | 00:19 |
| Guest24204 | then with --repository: juju deploy --repository ~/charms local:node-app mongonode-app -v | 00:20 |
| Guest24204 | error: cannot get latest charm revision: charm not found in "/home/brandon/charms": local:precise/node-app | 00:20 |
| Guest24204 | i followed mark mim's tutorial at for the --repository at http://markmims.com/cloud/2011/09/07/node-mongo-ubuntu.html | 00:22 |
| Guest24204 | my issue is simple getting the up to aws. If i set it up with "juju deploy node-app mongonode-app it will launch the instance but with no app files. | 00:23 |
| Guest24204 | hope that was thorough enough :) and thank you for at least attempting to help me in advance | 00:24 |
| sarnold | Guest24204: hrm, the charm README suggests it is looking for a config/config.js .. yours has "name": "mongonode", but you've called the app "mongonode-app" on the command line and the config.yaml ... | 00:29 |
| sarnold | Guest24204: if that's not it, maybe try changing "app_name: " to "app_name: mongonode-app" in the config.yaml... | 00:30 |
| Guest24204 | that is for the mongo module in node.js. I do have this setup and use it in my node server file | 00:31 |
| Guest24204 | you think I should remove the sub properties type: string | 00:33 |
| Guest24204 | default: mongonode-app | 00:33 |
| Guest24204 | description: Application Name or just add that too? Well trying both :) | 00:33 |
| Guest24204 | nope neither way worked :x | 00:35 |
| Guest24204 | <racking brain here < bang - bang - banging head against cement wall | 00:36 |
| sarnold | Guest24204: maybe? (sorry...) | 00:38 |
| Guest24204 | I know. Thank you for trying though. | 00:39 |
| === Guest24204 is now known as weblife | ||
| sarnold | weblife: there's a fair number of questions about juju on askubuntu.com -- perhaps you can find some inspiration there? | 00:40 |
| weblife | been looking everywhere not many people have used juju with node.js. Plenty on MongoDB. Can't find any related error issues either :( I am going to ask Mark Mims in hopes he might know. | 00:44 |
| sarnold | m_3: ^^^ guest24204 / weblife has some node charm questions I'm useless with.. :) | 00:45 |
| === defunctzombie is now known as defunctzombie_zz | ||
| hazmat | thumper, is local with restart supposed to restart mongo and machine agent? | 02:07 |
| hazmat | hmm looks like it should | 02:17 |
| thumper | hazmat: yes it does | 02:32 |
| === defunctzombie_zz is now known as defunctzombie | ||
| weblife | I posted my issue in a more detailed format at http://askubuntu.com/q/324672/88546 . Anyone think they can help me out with launching a node-app instance on AWS? | 02:53 |
| AskUbuntu | Juju node-app charm on AWS | http://askubuntu.com/q/324672 | 02:55 |
| weblife | Thats cool. Didn't know there was a bot in here for that stuff. | 02:59 |
| weblife | yeahhhhhhh!!!!! It was damn white spaces. | 04:59 |
| === CyberJacob|Away is now known as CyberJacob | ||
| === brandon is now known as weblife | ||
| === defunctzombie is now known as defunctzombie_zz | ||
| weblife | I just launched a node-app-22 from the charms repo and am getting the following: http://ec2-54-226-121-230.compute-1.amazonaws.com/ --- Any ideas before I get back to it in th morning? | 06:46 |
| weblife | First guess is it isn't loading my git repo. Check that out first I guess. | 06:49 |
| === CyberJacob is now known as CyberJacob|Away | ||
| === mthaddon` is now known as mthaddon | ||
| === teknico1 is now known as teknico | ||
| mectors | Is there any way to import the shift-D generated script with juju-core? | 10:43 |
| marcoceppi | mectors: yes, with a tool called juju-deployer | 11:07 |
| mectors | how do you install juju-deployer with juju-core | 11:07 |
| marcoceppi | mectors: you can also deploy the juju-gui to your environment and drag-n-drop the deployment file to the juju-gui | 11:07 |
| marcoceppi | mectors: you'll need to add a ppa to your system to install it. The final location for the deployer ppa hasn't been locked down yet | 11:09 |
| * marcoceppi finds email post | 11:09 | |
| theJujuGuy | Hi, i try to get juju working with maas. I read the MaaS Juju Quickstart Guide and run into problems. If i try to get the juju status "juju status" i get an error: ERROR Unexpected HTTP 401 trying to GET http://myip:80/MAAS/api/1.0/files/?filename=provider-state&op=get | 12:42 |
| theJujuGuy | is this a common error? any hints? | 12:42 |
| codybum | marcoceppi: Are you online? | 13:00 |
| marcoceppi | codybum: yes | 13:00 |
| marcoceppi | theJujuGuy: Do you have a successful bootstrap? What version of juju are you using (juju version or juju --version)? Do you see at least one node in use in the maas master dashboard? | 13:01 |
| codybum | marcoceppi: Yesterday my nick was<privteStack>. I changed nick to be the same as other ones, given I will likely be around for a while. | 13:02 |
| marcoceppi | codybum: welcome back o/ | 13:02 |
| codybum | marcoceppi: I figured out the problem you were helping me with yesterday. I was user error. Specifically the index.json had the wrong port number. | 13:03 |
| marcoceppi | codybum: Oh no, glad you got that resolved! | 13:03 |
| codybum | marcoceppi: I basically tore apart juju code and debugged until I noticed the constraints didn't match, and it was the port. | 13:03 |
| theJujuGuy | marcoceppi: i use juju-verison: 0.7; "juju bootstrap" gave the same error. In my dashboard are 1 "ready" node and 2 allocated to me | 13:04 |
| codybum | marcoceppi: I ran into some other stange things that you might know about. | 13:04 |
| marcoceppi | codybum: yes please! | 13:04 |
| marcoceppi | theJujuGuy: If bootstrap fails then status will also fail, could you run `juju destroy-environment` then run `juju bootstrap -v` and pastebin the output of the bootstrap command? | 13:05 |
| codybum | marcoceppi: It looks like juju, when set to use floating ip, uses the nova API to set addresses. I am using Quantum and unfortunately the floating addressed managed by quantum don't show up under the os_floating_ip api call for nova. | 13:05 |
| marcoceppi | codybum: I think this is a known issue with juju-core, that the openstack provider does not work well with quantum yet. Let me check the bugs for you real quick to confirm | 13:06 |
| codybum | marcoceppi: Basically on bootstrap if floating_ip = true the process fails dude to an openstack error. I would expect most people, especially HP use Quantum. So I wonder if it is something strange on my side. | 13:06 |
| marcoceppi | codybum: Maybe I'm confusing the openstack charms with juju then. Let me double check | 13:07 |
| marcoceppi | codybum: could it be related to this? https://bugs.launchpad.net/juju-core/+bug/1188126 | 13:08 |
| _mup_ | Bug #1188126: Juju unable to interact consistently with an openstack deployment where tenant has multiple networks configured <canonistack> <juju:New> <juju-core:Triaged> <https://launchpad.net/bugs/1188126> | 13:08 |
| codybum | marcoceppi: That might be related since I do have two networks associated with the tenant. | 13:10 |
| theJujuGuy | marcoceppi: http://goo.gl/xn559h | 13:10 |
| codybum | marcoceppi: I was actually able to bootstrap without using a floating_ip and have a bootstrap node running. | 13:10 |
| marcoceppi | theJujuGuy: Sorry, you'll need to run `juju -v bootstrap`, juju 0.7 is very picky about the order of commandline options. | 13:11 |
| marcoceppi | codybum: awesome, it's a bummer about the quantum issue you're experiencing though. If you think that bug is affecting you, you can subscribe to it an follow it's progress | 13:12 |
| codybum | marcoceppi: I was able to deploy wordpress and mysql charms, and new VMs were created. However, I was never able to successfully add a relation. | 13:12 |
| marcoceppi | codybum: did add-relation throw an error? | 13:13 |
| theJujuGuy | marcoceppi: http://goo.gl/oXZeAk | 13:13 |
| codybum | marcoceppi: That is the question, what logs should I be reading once the system is boostrapped | 13:13 |
| codybum | marcoceppi: on teh bootstrap node /var/log/juju/* contains few logs and nothing that I can see about node deployment | 13:14 |
| marcoceppi | codybum: each unit has it's own unit log. One simple way is to just check `juju status`, if you see a relations: key with a list of the relations added then juju knows about it. If not add-relation failed some how | 13:14 |
| marcoceppi | Otherwise /var/log/juju/unit-*.log on each node will contain the output from the charm hooks | 13:15 |
| marcoceppi | theJujuGuy: Sorry, 401 means it can't authorize. There's a chance that the maas-oauth isn't set correctly | 13:16 |
| marcoceppi | theJujuGuy: check the MAAS master to verify you've got the right key entered in to the environments.yaml file | 13:17 |
| codybum | marcoceppi: Great Tip.. Let me rebuild the environment and come back :) thanks | 13:17 |
| marcoceppi | codybum: also, if that bug is affecting you, feel free to subscribe to it on the right side of the page to get updates for it directly | 13:18 |
| theJujuGuy | marcoceppi: what about the user? is the user identified by the maas-oauth? | 13:27 |
| marcoceppi | theJujuGuy: So, when I did this I used the same user as my local username. So my local user account name is "marco" and I just created a "marco" account in the MAAS master, alloted units to that user, copied that MAAS key to my environments.yaml and used that | 13:29 |
| theJujuGuy | marcoceppi: thank you very much | 13:29 |
| theJujuGuy | marcoceppi: i am a step further | 13:30 |
| marcoceppi | theJujuGuy: Excellent. We really don't have the most stellar maas documentation, so during your process of using MAAS and Juju when you find any points of contention (like this one) if you could just record it and possibly send your experience using MAAS and juju to the mailing list (juju@lists.ubuntu.com) it'll really help us improve our documentation! | 13:31 |
| theJujuGuy | i tried to install juju with the Quick Start: now a get a new error by "juju -v bootstrap" : ERROR SSH authorized/public key not found. http://goo.gl/YojmDj I insert my key to the dashboard-site, also i tried to delete it an lets bootstrap generate the key pair. The error is the same. | 14:06 |
| marcoceppi | theJujuGuy: Do you have an ssh keypair in you ~/.ssh folder? | 14:08 |
| theJujuGuy | marcoceppi: thx again. why i must set my pub-key in the dashboard, when it's need in the folder? | 14:11 |
| marcoceppi | theJujuGuy: I think it's because MAAS doesn't have a mechanism to seed an ssh key during "provisioning", most clouds allow you to push keys to a machine during spin up. So it needs to have the key in maas master as the same in .ssh so juju can drive the machine after provisioning by maas is done | 14:15 |
| theJujuGuy | who is mainting the juju quick start from the maas-website? | 14:20 |
| marcoceppi | theJujuGuy: not sure, though we'll probably want to move it to the juju doc | 14:22 |
| marcoceppi | evilnick would know, but he's not online atm | 14:23 |
| weblife | @marcoceppi Doesn't node-app make a copy of the git repo that you set in you config.yaml to a S3 bucket? | 14:57 |
| weblife | or anyone else familiar with the charm. | 14:58 |
| marcoceppi | weblife: no, it simply clones it to the unit | 14:59 |
| marcoceppi | weblife: the charm itself is cached prior to deployment, but that cache won't have the git repo in it. Also, not all cloud providers have s3 or comprable object stores avaiable. So we try to make sure charms rely on as little of the cloud infrastructure as possible | 15:01 |
| weblife | ahh. Okay thank you. I'm getting a 503: No server is available to handle this request. | 15:01 |
| marcoceppi | weblife: at what point of the deployment? | 15:01 |
| weblife | after I have complete the all relationships and exposing. | 15:02 |
| weblife | No server is available to handle this request. | 15:03 |
| weblife | oops | 15:03 |
| weblife | ec2-54-226-121-230.compute-1.amazonaws.com | 15:03 |
| weblife | Any idea why? Before I start my witch hunt. | 15:04 |
| marcoceppi | weblife: what service is this? | 15:07 |
| marcoceppi | node-app? | 15:08 |
| weblife | Sorry using ec2 with aws | 15:08 |
| marcoceppi | weblife: right, but what charm is this? node-app? | 15:08 |
| weblife | revision-22? using the following git repo @ https://github.com/TheMindCompany/mongonode-app | 15:09 |
| weblife | @marcoceppi that what you mean | 15:09 |
| weblife | yes node-app | 15:10 |
| weblife | :) | 15:10 |
| marcoceppi | weblife: Okay, so there's a good chance that the node applciation itself didn't start. 503 means nginx can't connect to the backend. | 15:10 |
| den_sheleh | I'm trying to bind my own charm and logstash-agent. As I can see logstash-agent use juju-info interface to connection, but when I try to declare that one(logstash-agent: interface: juju-info scope: container) in metadata of my charm I got error: "error: charm not found". Is it bug or I'm doing something wrong ? | 15:11 |
| marcoceppi | What you can do is `juju ssh node-app/0` (replace node-app if it's deployed under and alias) and look to see if the process is running, check logs for failures (/var/log/juju/unit-*.log, /var/log/node-app) | 15:11 |
| marcoceppi | weblife: the node-app makes quite a few assumptions, so it might not work with every node-applcaition. If there are changes needed to the charm to make it work with your application then you can patch and run the charm locally or submit it to be merged into the official charm, making it better for everyone :) | 15:12 |
| marcoceppi | den_sheleh: can you pastebin the steps you're taking? subordinates and logstash definintely work last I checked | 15:13 |
| weblife | I am currently launching locally, but if I can fix this I will submit a patch. Thank you for your help. | 15:14 |
| marcoceppi | weblife: np, if you need more help debugging let me know! | 15:15 |
| dannf | would it be possible to get ppa:charmers/charm-helpers populated for raring? | 15:16 |
| marcoceppi | dannf: I dont' see why not, let me take a look | 15:18 |
| * dannf is testing on a maas cluster of hw that wasn't supported in previous releases, charms like wordpress fail when they can't find a compatible tools release | 15:18 | |
| dannf | marcoceppi: cool | 15:18 |
| marcoceppi | dannf: Oh, you're talking about old charm-helpers | 15:18 |
| den_sheleh | marcoceppi: after adding section logstash-agent: interface: juju-info and trying to deploy my local charm I get error described early | 15:20 |
| marcoceppi | den_sheleh: can you paste your metadata.yaml file to paste.ubuntu.com? | 15:20 |
| marcoceppi | dannf: I've kicked of the builders, should build in a few hours | 15:22 |
| dannf | marcoceppi: you rock, thanks! | 15:23 |
| den_sheleh | marcoceppi: http://pastebin.com/fYnXLtQW | 15:23 |
| arosales | note Charm School "Infrastructure" is being deferred this week | 15:59 |
| arosales | jcastro, was your OSCON charm school video'ed? | 15:59 |
| jcastro | yes | 16:05 |
| jcastro | we still have to be invoiced though | 16:05 |
| jcastro | they told me "a couple of days" | 16:05 |
| === defunctzombie_zz is now known as defunctzombie | ||
| weblife | @marcoceppi I took your advice and re-installed nodejs. I did from source though and not chris leas ppa, I always seem to have issues with his. It is up and running but I get a 504: Gateway Time-out. Node.js debug gives me connection error: [Error: failed to connect to [localhost:27017]]. I think it may be my app code now because that is the mongodb port setting from config/config.js. Before I mess with my code, would you possi | 16:43 |
| weblife | bly concur with this thought? Also npm says I am missing the repository setting but I have never needed this on my local, is this specific to node-app? | 16:43 |
| === CyberJacob|Away is now known as CyberJacob | ||
| === brandon__ is now known as weblife | ||
| === defunctzombie is now known as defunctzombie_zz | ||
| hazmat | SpamapS, still on sup? | 18:23 |
| SpamapS | hazmat: yeah, but I'd love to find something better at what it does. | 18:24 |
| SpamapS | I'd swear it is getting slower and slower.. :-/ | 18:24 |
| hazmat | SpamapS, there's https://github.com/pazz/alot | 18:25 |
| === defunctzombie_zz is now known as defunctzombie | ||
| SpamapS | hazmat: that might be worth a shot | 18:27 |
| SpamapS | all the notmuch frontends were horrible a year ago | 18:27 |
| weblife | Okay, not liking mongodb-cloud with nodejs. | 20:21 |
| weblife | @marcoceppi in your next screen cast session you think you can cover how node.js, mongodb, and haproxy work and communicate together a little? I have my node app running with mongodb but only in a single instance. I am trying to finish up a tutorial series on how to work with node and mongo on ubuntu and then deploy what you make localaly to aws with my sample app: http://ec2-54-226-87-136.compute-1.amazonaws.com | 20:25 |
| weblife | I want to say that what ever you build is scalable. | 20:26 |
| weblife | but cant at this moment with a single instance. | 20:27 |
| weblife | nevermind you inda do talk about it in yourlast one minus haproxy | 20:38 |
| weblife | @marcoceppi I believe your on the charm-tools team. I installed it earlier it messed with my Juju version from the juju/devel PPA. So I removed it and reinstalled it, works fine now but now charm-tools isn't creating all the files(ie. config.yaml, icon.svg, probably more) | 22:37 |
| weblife | I can deal but thought you should know. Im on 13.04 | 22:43 |
| marcoceppi | weblife: did you install from ppa? that problem was good but not back ported yet | 22:55 |
| marcoceppi | was fixed* | 22:56 |
| * marcoceppi needs to open a back port request | 22:56 | |
| weblife | @marcoceppi Cool, thank you. DId you get my message earlier about possibly using the wget for nodejs.org - and build from source. It would make things more universal. Especially with 13.04 servers. | 23:04 |
| marcoceppi | yes. away from computer again but ping me any time after today and we can talk about the logistics of doing such a thing | 23:05 |
| marcoceppi | it would make a great conduit option weblife | 23:05 |
| marcoceppi | config* | 23:05 |
| ahasenack | hi, any charm reviewers lurking? A simple two-line one: https://code.launchpad.net/~ahasenack/charms/precise/rabbitmq-server/rabbitmq-server-juju-core-compat/+merge/177231 | 23:08 |
| ahasenack | marcoceppi: can you review that? It's simple, I promise | 23:15 |
| ahasenack | same fix already existing in cinder | 23:15 |
| ahasenack | and I just deployed with two units of rabbit, worked | 23:16 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!