=== zz_CyberJacob is now known as CyberJacob [07:24] * arturt is away: I'm busy === CyberJacob is now known as zz_CyberJacob === Spads_ is now known as Spads [11:19] apuimedo, it was excellent! Next time, I hope you can make it :) [11:20] lazypower: I hope so too ;-) [11:20] lazypower: so who is going to the OSt Summit? [11:20] jamespage will surely be there [11:20] I will yes [11:21] some presentation on nova-lxd, jamespage? [11:22] apuimedo, I submitted talks, but they did not get chosen [11:22] apuimedo, but I expect I'll still get an opporunity... [11:25] jamespage: lightning talks? [11:33] maybe [12:42] gnuoy, did you have any opiniuon on https://code.launchpad.net/~james-page/charms/trusty/neutron-api/wily/+merge/271917 ? [12:44] jamespage, it's ok [12:45] I can't say I'm over the moon about abusing the interpreter to install packages [12:46] gnuoy, its horrid [12:46] but probably required [12:46] I'm open to suggestions [12:47] jamespage, I wonder whether it's clear to actually use the install hook for this. Remove the install symlink, replace with a bash script and then call the old renamed install function [12:47] s/clear/clearer/ [12:48] gnuoy, that might be a bit neater yes [12:49] so splitting any install related activities from interpreter resolution [12:49] yeah [12:51] gnuoy, so something like - https://code.launchpad.net/~james-page/charms/trusty/neutron-api/wily/+merge/271917 [12:52] jamespage, exactly (although line 25 look wrong) [12:52] gnuoy, hmm - really? [12:53] jamespage, which method in ./hooks/neutron_api_hooks.py is going to be called? [12:53] I think the basename will remain install so that might dtrt [12:53] oh, ok [12:53] I stand corrected [12:54] just testing that now [12:57] gnuoy, nope that does not work as I thought === scuttle|afk is now known as scuttlemonkey [13:10] gnuoy, https://code.launchpad.net/~james-page/charms/trusty/neutron-api/wily/+merge/271917 [13:10] revised [13:11] and tested :-) [13:13] jamespage, approved. Not sure if you want to wait on osci [13:23] gnuoy, jamespage - on that ^ n-api test, precise-icehouse timed out (45 minutes), stuck @: neutron-openvswitch/0 maintenance executing (config-changed) installing charm software; not sure if that was an infra hiccup or indicative of a real issue. [13:23] beisner, did not touch thatbit honest guv [13:42] gnuoy, jamespage - a re-run of that is also stuck in the same state, on precise-icehouse. [13:42] this one, caught in the act if you want to poke at it. [13:52] jamespage, gnuoy - the hang happens when this happens: [13:52] 00:12:56.675 2015-09-22 13:19:34 Adding relation neutron-api:neutron-plugin-api <-> neutron-openvswitch:neutron-plugin-api [14:07] There is a better way to do this: juju debug-log -n 100 | grep --line-buffered my-service-name | awk '{$2=$3=$4=$5=$6=""; print $0}' [14:07] ?? [14:45] redelmann: wow that's quite a string of commands [14:45] there's a way to filter debug-log, but I still haven't quite figured it out [15:09] jamespage, observing some new wily deploy woes, assume that is the py2 affect? ie. 2015-09-22 15:08:31 INFO install ImportError: No module named apt_pkg [15:10] beisner, yes - I just stuck up MP for all OpenStack charms that resolved that problem [15:10] beisner, wily will lose py2 on the cloud-image, its already loosing some implicit depends which charms assume are installed [15:11] jamespage, right-o. thanks, just wanted to make sure i wasn't bumping into something else. [15:11] fun times! [15:11] * beisner re-coffees [15:19] Hello Team, [15:46] jamespage: do you have a bug for adding things like apt_pkg or py2 back into cloud images by way of juju? [15:48] bundletester from pypi seems broken for me. 0.5.4/0.5.3 fail to start, and any of the 0.5 series seems to get stuck in a recursive __getattr__ lookup. [15:48] is there somewhere else I should be looking for bundle tester [15:48] ? [16:03] marcoceppi, nope [16:03] jamespage: do you think it's worth having juju do as part of it's cloud init? [16:04] marcoceppi, I think we'll put in a install wrapper for this cycle [16:43] Hello Team, I am charming Platform RTM using juju framework, when i am following the silent installation procedure. i was asked give input as (Yes/No in GUI) to configure database for rsyslog-mysql with dbconfig-common. As per my requirment i will have to give input as "NO". I have tried to use pipe and expect in order to avoid tat prompt but unfortunately it doesnt worked out. Please advise how to avoid configuring rsyslog-mysql p [17:11] Could someone help me on my query regarding prompt? [17:19] Prabakaran: hmm, how did you give hte requirement? this is during hte install step? I'm assuming it's the package install? [17:19] Prabakaran: http://www.microhowto.info/howto/perform_an_unattended_installation_of_a_debian_package.html walks you through helping to provide commands at install time of packages that might help? [17:22] yes it is a package install along with my product , Here for database i have used seperate charm so that we dont need to configure configure database for rsyslog-mysql with dbconfig-common. That is why i have to give No in the GUI prompt [17:27] Thank you rick_h ..i wil refer the link which you had sent.. [17:27] Prabakaran: ah makes sense. Yes, I think you'll need to get the config passed during the install step like that doc mentions. Good luck! [17:28] ya k i wil try.. Thanks for ur help === natefinch is now known as natefinch-afk [21:47] Hi everyone. Trying to deploy nfs charm on Azure I see that the agent state on the machine for a very short period becomes dead (right after VM provisioning) and the nfs service status becomes unknown. Has anyone seen this before? === zz_CyberJacob is now known as CyberJacob === CyberJacob is now known as Guest44232