=== blahdeblah_ is now known as blahdeblah === natefinch-afk is now known as natefinch === axw_ is now known as axw === yo61_ is now known as yo61 [12:15] how should one implement a peer relation in an interface layer / layered charm? [12:18] icey: create a peer.py file in the interface [12:19] marcoceppi: yeah, I found that, I'm grasping at straws here because I'm getting a nice build error (charmtools.build.tactics: Missing implementation for interface role: provides.py) and I'm working on adding several interfaces (and writing those) at the same time, they all seem to be correct but for this build error -_- [12:20] I'll just keep digging :) === Guest29413 is now known as CyberJacob [13:10] icey: link to your interface? [13:11] marcoceppi: it's not up yet, but I think it may be fixed in charmtools 2.1.3 [13:11] I popped it into charmbox and I can build it there -_- [13:11] icey: 2.1.3 is in pypi, we're working on getting a package built [13:12] eta for package on xenial? doesn't kill me to use it in charmbox for now but nice to know what to expect :) [13:13] icey: Make sure that the relation is listed under "peers" in the metadata.yaml. That error sounds like it's listed under "provides" [13:14] cory_fu: I have 3 provides relations, and a peer relation [13:14] and none of them are existing interfaces [13:14] it was one of the provides having the problem, if I removed it from the layer.yaml, it worked [13:14] and it worksd with the new charmtools [13:15] icey: Also, if it's helpful, you can see an example of a peer relation interface layer here: https://github.com/juju-solutions/interface-namenode-cluster and how it's used in https://github.com/juju-solutions/layer-apache-hadoop-namenode [13:16] thanks cory_fu [13:16] Glad to hear that the issue is fixed already in charmtools, though. :) [13:18] cory_fu: I'm so confused about it, the other 2 provides interfaces I've written are the same except in name and 2.1.2 has no problem with them -_- [13:18] but yeah, no worries since it seems to be fixed in newer versions [13:26] hi guys, I'm having issues bootstrapping on openstack (liberty) with juju-2, juju is trying http://10.96.5.21:5000/v2.0/auth/tokens (my endpoint is http://10.96.5.21:5000/v2.0) and fails with a 404 [13:27] 2016-05-03 13:27:03 DEBUG juju.provider.openstack provider.go:724 authURL: http://10.96.5.21:5000/v2.0 [13:27] 2016-05-03 13:27:03 DEBUG juju.provider.openstack provider.go:685 authentication failed: authentication failed [13:27] caused by: requesting token: Resource at http://10.96.5.21:5000/v2.0/auth/tokens not found [13:27] any tips? That url was grabbed from novarc, and nova commands works just fine [13:47] Is there any documentation for creating a provider in Juju? === fginther` is now known as fginther === disposab1e is now known as disposable [15:13] Hi kwmonroe [15:14] when i do charm proof on the deployable charm , I am gettese [15:14] charm@islrpbeixv665:~/charms/trusty/ibm-db2$ charm proof E: Unknown root metadata field (terms) E: min-juju-version: invalid format, try X.Y.Z [15:14] These are coming in metadata.yaml file when we do charm build from the ibm-base layer [15:15] What to do for these ? Do we need to remove these from deployable charm or from ibm-base layer ? [15:27] Hi everyone [15:28] Someone knows how to "delete" a command launched in juju? [15:28] the machine is off... but I need to stop this command [15:28] I _think_ juju queues up "commands" in mongodb [15:29] and you know how to enter in this mongodb queue? [15:33] suchvenu - its landed in the repository but is pending release - https://github.com/juju/charm-tools/issues/190 I think you're OK to leave it in for now, i'm fairly certain this 2.1 patch will be going out soon. [15:34] axino, you know how to enter in this mongodb queue? [15:34] oh, and i take that back, it hasn't landed its only been filed. [15:35] nevermind me, i defer to kwmonroe :) [15:35] ho you. [15:35] yo yo axino [15:36] D4RKS1D3: connect to your controller and fire up a mongo client ? I don't know about the structure at all, sorry. What command are you trying to cancel ? [15:36] I enter here because i don not find any command to do this [15:43] suchvenu: can you send me the result of the command 'charm version' ? [15:44] suchvenu: I suspect your charm tools version is not current. [15:46] D4RKS1D3: what command do you want to "delete" ? [15:47] charm-tools 2.1.2 [15:48] I need to go out urgentlly, Can you let me know through mail pls [16:03] Sorry for the delay axino I want to the delete an action [16:03] because i put a wrong command [16:03] but the machine is off [16:03] probably if i remove the command of the queue can "save" the state of this machine [16:17] mbruzek that is the latest charm-tools, 2.1.3 is being released, so they have the latest, but these bugs are being patched [16:20] anyone tried the juju beta with azure? [16:20] no yet [16:23] marcoceppi: yes I sent an email, with about the same information. Copied you, please reply if I said something incorrect. [16:28] axino, I am alredy in the database, you know what is the table? [16:37] mbruzek: it's fine [16:37] ejat: I have inthe past, you having issues> [16:37] ? [16:38] now looking into beta 6 ... trying to use juju client on windows [16:38] and looking for documentation for azure credential to be place on credentials.yaml [16:39] ejat: you just need to run juju add-credentials azure [16:39] ejat: you just need to run juju add-credential azure [16:43] marcoceppi, you know what happend when you send a command?, this command is store in the database? [16:44] I do not know what happens when the machine is not alive and you send a command to this service [16:44] Someone knows?, thanks [16:49] marcoceppi: ok thanks ... managed to get all the credential needed using azure cmd line === pmatulis_ is now known as pmatulis [17:22] It should be possible to run bootstrap Juju to a local LXD installation right? [17:38] marlinc: I think that's the default for local [17:38] check out this link: https://jujucharms.com/docs/1.25/config-LXC [17:46] Okay julenl :) [18:01] marcoceppi, tvansteenburgh: have you guys had a chance to look at the TLS problem with deployer/python 2.7? [18:03] natefinch: yes, but considering Juju 2.0 is weeks out we're not going to jump on it right away [18:04] marcoceppi: ok, as long as you think it's fixable for 2.0, I'm fine with letting you figure it out :) [18:11] natefinch: yeah, we'll address before 2.0 [18:12] marcoceppi: cool, one less thing I need to worry about :) [18:38] hows it going everyone? Can someone elaborate on, or link me to some docs on the 'shared-db' network space? [18:38] As seen here: https://jujucharms.com/nova-cloud-controller/xenial/0 [18:45] openstack-charmers: could someone please link me to some docs describing the intrinsics of what a 'compute-data' network is, what information/services communicate on this network? How is 'compute-data' network recognized by openstack services? [18:45] as seen here: https://insights.ubuntu.com/2015/11/08/deploying-openstack-on-maas-1-9-with-juju/ [18:59] any neutron mitaka openstack charmers on? === blr_ is now known as blr [21:00] How can I let Juju use a external LXD 'region'? === natefinch is now known as natefinch-afk [21:07] marlinc: you can't at this point in time. It's up for discussion for 16.10 [21:07] :( [21:08] marlinc: sorry :( [21:08] Is it possible to do it manually by creating a 'cloud'? [21:08] Like with OpenStack [21:10] marlinc: no, Juju is missing some code to handle sending some of it's calls across the networking vs locally [21:10] Damn, okay [21:10] Have to SSH in then I guess [21:14] Thanks anyway r [21:14] Thanks anyway rick_h_, hope to see new cool things in 16.10 then [21:15] All of of the tools are amazing already btw [21:15] I wish I had the money to actually properly try out MAAS etc [21:15] marlinc: yea, we're getting planned up for the next cycle and should be good stuff [21:16] marlinc: there's the virtual maas stuff? [21:16] marlinc: I know some folks use that to try it out on one machine [21:17] Yea I did that once, using libvirt [21:18] marlinc: is this what you want? https://insights.ubuntu.com/2015/11/16/juju-and-remote-lxd-host/ [21:19] Not sure julenl, I actually searched and found that as well but I didn't actually read it [21:19] The reason why I didn't read it was because of the talk about environment.yaml [21:20] I'm not sure whether I understand it any more, Juju used to use a environments.yaml file but I guess that's no longer used now? [21:21] as far as I know... it does [21:22] Mm okay? [21:22] accounts.yaml [21:22] bootstrap-config.yaml [21:22] clouds.yaml [21:22] controllers.yaml [21:22] current-controller [21:22] models.yaml [21:22] ssh [21:22] Woops [21:22] Didn't except ls to do that when not in an interactive terminal [22:48] someone knows what happend when you send a command?, this command is store in the database? [22:48] I do not know what happens when the machine is not alive and you send a command to this service [22:48] Thanks in advance [22:56] D4RKS1D3 - Hey, i saw this morning it was suggested to go poking around in the mongo database. I dont know that I agree with that. Its highly discouraged to go poking about in there unless you're familiar with the data schema [22:57] D4RKS1D3 - I do believe that there is a timeout on the action you've queued. [22:57] D4RKS1D3 if you have the action's UUID that it returned, you can query the status of that action [22:58] I do not know the id of the action lazyPower , this id is save in some log? [22:59] D4RKS1D3 you can list all actions run against your controller with: juju action status [22:59] okey, thanks [22:59] for more information see: juju action --help [22:59] juju action statusERROR no actions found [23:00] marlinc - juju 2.0 uses cloud credentials. it supports autoloading through the environment, or an interactive prompt [23:00] That means if i turn on the machine juju do not destroy my machine [23:01] D4RKS1D3 i'm not sure what you mean. but if you sent a juju destroy command, its entirely likely that it will get reaped yes. [23:01] there may be something we can do to help, but we'll need some very detailed information in a bug report to start the process [23:03] D4RKS1D3 : do you have some juju status output, and a small rundown of whats happened? [23:03] Hello, guys. I have a juju 2.0 related question. I have a bundle with multiple charms. I deploy it. I remove one charm and its machine. When I redeploy the bundle, the charm I deleted get spawned, but also new machine for already deploy services. Is this intended ? [23:03] Thanks for helpme lazyPower [23:03] ionutbalutoiu - yes, unless you remove that charm from teh bundle, it will always attempt to reach the state in which the bundle defines [23:04] but, I don't want new machines for already deployed services. This is how juju-deployer works, I think. [23:04] oh its adding additional machines? [23:05] yep. [23:05] i misunderstood, i thought it was only re-adding the machine that was removed. that seems like a bug, we should most definetly get that filed. [23:05] D4RKS1D3 - pastebin the output of juju status for me [23:05] D4RKS1D3 - and run me down what you've done that you're trying to prevent [23:06] or what somethign else did on your behalf :) as i really have no idea [23:07] ionutbalutoiu https://bugs.launchpad.net/juju-core/+filebug - can you file a bug with juju version, the bundle you're deploying, and steps to reproduce? [23:08] @lazyPower yes, preparing the steps. I was looking now, if anything similar was already reported. [23:08] Thanks :) [23:08] of course lazyPower [23:48] lazyPower, I think it was a bundle problem. I'm good now. Juju deploy from 2.0 is behaving just like juju-deployer with bundles. All good :) [23:50] lazyPower, https://bugs.launchpad.net/juju-core/+bug/1577988 Thanks for help us [23:50] Bug #1577988: Revert destroy service when machine is off