[02:36] QUESTION: Why am I seeing [Review Queue] emails now? [02:44] jose: oops, did you do enough good work to earn more work? :) [02:44] sarnold: hehe nope, they're on the juju mailing list [02:44] oh :) [02:44] I was just wondering if it had something to do with the new revq that's on the works! [03:03] jose: no that's just a new thing we're doing. we're taking 2-hr time blocks on the review queue and reporting our progress to the list [03:03] tvansteenburgh: oh, it's good to know :) === uru-afk is now known as urulama === CyberJacob|Away is now known as CyberJacob === CyberJacob is now known as CyberJacob|Away [11:48] I have a question about the appropriate way to write a charm configuration. [11:48] I have in mind a very simple charm that I give my account credentials to, and it issues "juju destroy-environment" after a given timeout. [11:49] So that I can deploy the charm when I have just a testing env, and make sure it dies even if I forget about it. [11:50] But I'd like to have it allow a syntax like "destroy after 45 minutes" or something along those lines [11:50] but it seems odd to have a config setting that would "reset" the state every time config-changed runs. [12:32] stokachu: what do you mean it breaks it? [12:40] marcoceppi: there you are. I ran into https://bugs.launchpad.net/charm-tools/+bug/1359170 trying to use charm-tools with trunk juju [12:40] Bug #1359170: juju charm create fails in an unhelpful fashion with juju-1.21 in PATH [12:40] I'm not sure that you're the maintainer, but I thought I'd start with you [12:41] jam: I am one of them, this is concerning, I'm not sure what I can do to fix it [12:41] seems like juju-core is not being so nice [12:41] ? [12:41] marcoceppi: so something is failing in the subprocess starting other things, but its unclear what [12:42] jam: I'll compile from trunk and punch on it a bit [12:42] jam: any idea when 1.21 is expected? [12:42] marcoceppi: so I think we set "command.Env" [12:43] and probably juju-create expects PATH to be in there, but we aren't putting it in? [12:43] no, we are appending to os.Environ() so that should be ok. [12:43] ... [12:43] i mean, we're just using argparse in python [12:43] though the way we are doing that makes me wonder if we are actually mutating os.Environ() in-procses (not that it really matters) [12:43] nothing too crazy [12:43] marcoceppi: you're looking for /usr/bin/charm-create and invoking that [12:43] from juju-chram [12:44] juju-charm [12:44] which then uses setuptools to again look up where main is, etc. [12:44] oh, right, for "plugin mapping" [12:44] let me compile, throw in some debug statements, see what I can find [12:50] marcoceppi: doing "print sys.argv" in juju-charm shows me: [12:50] $ juju charm create -h [12:50] args ['/usr/bin/juju-charm'] [12:50] marcoceppi: vs 1.18 does: [12:50] $ /usr/bin/juju charm create -h [12:50] args ['/usr/bin/juju-charm', 'create', '-h'] [12:50] ah, so that's what is different [12:51] bug in core I guess? [12:52] marcoceppi: yeah [12:52] \o/ less work for me [12:52] jam: thanks for catching that though [12:58] could anyone help possibly with charm debug [12:58] juju debug-hooks wordpress/3 loadbalancer-relation-joined [12:59] I was hoping to see a special prompt with % [12:59] but I just get: [12:59] root@vagrant-local-machine-34:~# [12:59] rm4: did you resolved --retry yet ? [13:00] don't quite follow sorry [13:01] rm4: something needs to trigger the hook executing again, assuming it is in a failed state. juju resolved --retry will trigger this. [13:02] rm4: do that in a different terminal and your debug-hook tmux will open a new tmux window with the special prompt [13:02] vagrant@vagrant-ubuntu-trusty-64:~$ juju resolved --retry wordpress/3 ERROR unit "wordpress/3" is not in an error state [13:03] it all works but I would like to see what is happening [13:03] so I can make a charm [13:04] rm4: In that case, you could remove the unit and add it again and immediately after running add run the debug-hook and hope you were fast enough to catch the hooks (you likely will be). [13:05] rm4: or if you are only interested in that relation, remove and add that relation [13:07] In cloud-images, what's the functional difference supposed to be between the vagrant boxes in http://cloud-images.ubuntu.com/vagrant/trusty/ vs. http://cloud-images.ubuntu.com/vagrant/trusty/current/? [13:08] thanks jrwren [13:08] aisrael: might be good to pin utlemming on that. He's in CO so might not be on for a bit. [13:08] I'll try and come back [13:08] adeuring: do you know the answer to aisrael's question ^ ? [13:09] aisrael, the first one has all the boxes [13:09] the 2nd one only has the latest builds [13:09] the boxes get published every few weeks, so we keep those around too [13:09] what is in /current will always be the latest one [13:10] The reason I'm asking: the image in current/ only sets up one adapter (nat) and doesn't provision juju. [13:10] The larger image bootstraps, but sets up two adapters (nat and hostonly) [13:10] ok, the ones without juju in the filename [13:10] are just the normal vanilla vagrant boxes [13:11] the ones with juju in them are supposed to bootstrap/gui, etc. with juju [13:11] aisrael: the smaller files are "plain" ubuntu cloud images, while the larger files have juju preinstalled, and the Juju GUI is installed during Vagrant's provisioning [13:11] jcastro: Oh, duh. [13:11] ok, so that's one doc I should look at getting updated [13:11] which one? [13:11] https://juju.ubuntu.com/docs/howto-vagrant-workflow.html [13:12] points to a non-existent box, in current/ [13:12] ah yes [13:12] damn, I fixed the config-vagrant page [13:12] but totally forgot about this one, yeah, fix it in the docs please. [13:12] and I'm still troubleshooting why the networking problem [13:13] https://github.com/juju/docs has the instructions [13:13] they're in markdown, very easy to fix [13:13] excellent, I'll add to my list today [13:13] mbruzek, coreycb: did you guys not review yesterday? [13:14] jcastro What makes you say that? [13:14] jcastro, review what? [13:14] the review queue, you guys were on the calendar yesterday [13:15] jcastro, hmm not familiar with the calendar.. sorry. I do need to start doing more reviews though. [13:15] we each were supposed to pick 2 hours a week to review, then send the results to the list. [13:15] Yes I reviewed the ufw changes for elasticsearch and I DID email the group about it. [13:16] jcastro, I think you have coreycb mixed up with cory_fu [13:17] hah yes I do! sorry about that coreycb [13:17] I meant cory_fu [13:17] jcastro, ahh, np [13:18] oh excellent! [13:18] mbruzek, I now see both of your [Review Queue] posts, that is quite excellent! [13:19] thanks for the positive reinforcement jcastro [13:19] and for checking up on me [13:19] hah [13:19] it is unclear to me if you merged his changes though [13:29] marcoceppi: turns out it only happens if you don't have an environment set [13:29] I *might* be rare in not having a default env [13:29] jam: that's fun! [13:29] or using juju switch [13:29] marcoceppi: I always run "juju command -e env" [13:29] so that I know what environment it is running in [13:32] jamespag`: Hi, mwhudson sent me... I'm having a problem with the cinder charm on a system with only 1 drive. I am trying to have it use a loopback file for storage. The file is created, the loop device (/dev/loop0) is created, but the volume group is never created. === gsamfira1 is now known as gsamfira [14:03] I managed to get charm debug working [14:03] thanks for that [14:04] bit of a different prompt to documentation though [14:04] https://juju.ubuntu.com/docs/authors-hook-debug.html shows ~ [14:05] rm4: ah, yes, debug-hooks shows unit/hook-context now [14:06] root@vagrant-local-machine-34:/var/lib/juju/agents/unit-wordpress-3/charm/hooks# relation-get [14:06] private-address: 10.0.3.181 [14:06] yes I think so [14:07] rm4: that's the prompt you get for debug-hooks? [14:07] it should look like wordpress/3@ [14:35] that was the prompt for debug-hooks yes [14:38] rm4: is that window 0 of tmux or window 1 ? [14:39] 12.04.5 0:- 1:website-relation-departed* [14:39] I think that is 1 [14:49] I was in /home/ubuntu === urulama is now known as uru-afk === roadmr is now known as roadmr_afk === roadmr_afk is now known as roadmr === TheRedMage is now known as MassuguGo === uru-afk is now known as urulama [19:01] marcoceppi: are you and the other charmers on the juju-dev mailing list? [19:02] natefinch: I am [19:02] marcoceppi: ahh, right, I remember you responding to the default-hook thread. [19:03] marcoceppi: you offered to survey the current charms and see which ones used the all-symlink model... do you know when you might get that done? I know it's a big job. [19:16] natefinch: I can do that now, should take about 15 mins [19:18] marcoceppi: that would be fantastic, thank you [19:32] natefinch: here are precise charms in the store, slightly curated: http://paste.ubuntu.com/8100495/ [19:32] natefinch: and the command used to get that, by running it in a directory which `charm getall` was fetched [19:32] for f in `find */hooks/ -maxdepth 1`; do if [ -L $f ]; then echo "$(echo $f | awk -F/ '{print $1}') `readlink $f`"; fi; done | sort | uniq [19:33] charm common file [19:34] well, "charm" "common file" is the output [19:41] marcoceppi: nice [19:44] marcoceppi: so, are there any that symlink only some of the hooks? [19:44] natefinch: yes [19:44] natefinch: most all hooks that say relation_common [19:44] indicates that only some of the hooks of that are symlinked [19:45] I also removed entries that were just sylinks between relation-joined and relation-changed [19:45] since that's a stupid pattern and easy enough to clean up === CyberJacob|Away is now known as CyberJacob [19:45] marcoceppi: heh [19:46] marcoceppi: so how many charms are there total? I'd like to know what this is 62 out of :) [19:47] 162 [19:47] also, there are some dupliates in there [19:47] as some are inconsistently symlinked (./hooks.py vs hooks.py) [19:47] that list needs to be curated a bit more [19:48] also does not include trusty charms [19:59] marcoceppi: thanks, I got rid of dupes, easy enough [20:54] ctlaugh: :( === CyberJacob is now known as CyberJacob|Away === CyberJacob|Away is now known as CyberJacob