[11:54] hi frankban, i see you're working on bug 1103204. are you able to duplicate it? [11:54] <_mup_> Bug #1103204: Removing a relation sometimes fails < https://launchpad.net/bugs/1103204 > [11:55] bac, yes [11:55] good [12:27] bac: could you please take a look at https://codereview.appspot.com/7199043 ? [12:28] frankban: ok. [12:30] frankban: it is odd you had to go to such lengths to reproduce the bug. it frequently happened right away for me after reloading the page. [12:30] bac: in uistage or locally? [12:31] locally [12:31] frankban: approved [12:32] bac: thanks. [12:32] with that i'm going back to bed. perhaps i'll feel better after lunch. [13:42] benji: are you around? [13:43] frankban: I am. What's up? [13:43] benji: do you have time for a review? https://codereview.appspot.com/7199043 [13:43] sure, I'll look now [13:43] benji: cool, thanks [13:44] my pleasure [13:54] frankban: I got distracted for a minute, but the review is done now. [13:58] benji: thanks, re the test, agreed. what do you think about creating another card for it? it would be easier to add missing test after your current card is landed. [13:59] frankban: makes sense; in fact, I may do that card too (unless you really want it) [14:03] benji: cool, creating a card, please feel free to tackle it if that means just a reasonably small addition to your branch [14:11] benji: added card for bug #1103477 [14:11] <_mup_> Bug #1103477: Add test for relation removal < https://launchpad.net/bugs/1103477 > [14:11] thanks [14:23] bacm bcsaller frankban teknico, as I said in my email, please let me know if you are interested in working on the Go Juju changes. It will help our planning discussion. Some indication of degree of interest would be somewhat helpful a well. Let's say a scale of 1 to 10, where 5 represents ambivalence, 1 represents an abhorrence verging on the irrational, and 10 represents ecstasy and bliss at the idea. :-) [14:23] bac, that is [14:23] benji, thanks for your email. a degree of interest indicator would be nice but not necessary [14:24] gary_poster, oh, I thought it wasn't needed, at least on my part [14:24] gary_poster: for Go? [14:24] gary_poster, let's turn it to eleven! ;-) [14:24] teknico, ok :-) [14:24] benji, yes [14:24] gary_poster: 8.5 [14:25] benji, cool thanks [15:30] Meeting today? [15:32] benji, frankban teknico - meeting? [15:33] Makyo: sounds like fun [15:33] Makyo: yep [15:33] Makyo, yep [15:33] + bac [15:33] gary_poster might even join us [15:33] Makyo: bac isn't feeling well [15:34] benji, Ah, thanks. [15:43] bcsaller: something that might make the run-the-tests-and-scrape-the-result script easier: http://pypi.python.org/pypi/xvfbwrapper [15:43] bcsaller: in particular the "Example: Headless Selenium WebDriver Tests" section [15:44] benji: thats how its done in the charm test I think [15:44] it might be, I suggested it but am not sure what they settled on [15:51] sorry, meeting benji :-) [16:00] hola! [16:00] http://paste.ubuntu.com/1563427/ is my problem of the day :) [16:11] :-/ [16:21] therve: could you please paste the full charm.log? [16:21] therve, your world is broken :-) [16:22] frankban, where's that? [16:23] teknico, I know! there is most certainly something wrong with me [16:23] therve, most likely not :-) [16:24] inside the juju gui instance, it should be in /var/lib/juju/units/juju-gui-0/charm.log, or something similar... [16:24] therve: ^^^ [16:24] I don't have such a thing [16:27] I have 'juju-gui-0.1.5' instead of 'build-prod' in the /var/lib/juju/units/juju-gui-0/charm/juju-gui/ directory [16:33] therve, uhm, now our world is broken too :-P [16:33] therve: is the provider lxc? [16:33] frankban, indeed [16:34] it's nice that someone is able to actually deploy on lxc :-) [16:35] teknico, it's been rather challenging :) [16:50] therve, related topic. I'm in the process of describing the details of what Landscape is providing the GUI in annotations so that I can work with the UX folks to figure out the user stories and presentation. Could you comment on/correct this? http://pastebin.ubuntu.com/1563586/ [16:51] therve: do you have the file /var/log/juju/unit-juju-gui-0.log or similar (inside the container)? [16:52] frankban, new instance: http://paste.ubuntu.com/1563593/ [16:53] gary_poster, so [16:53] gary_poster, right now, it's unit, not machine [16:53] gary_poster, I think there are a couple of others annotations, at a technical level [16:53] gary_poster, ie, a global link to resolve security and reboot alerts [16:54] and probably a root_url to landscape [17:00] therve, so if we get a security alert on a unit, we will go to a Landscape link that is not specific for a given unit [17:00] ? [17:00] to resolve it [17:00] and same for reboot [17:05] gary_poster, yeah unfortunately we don't have one per unit [17:06] gary_poster, although, considering all the machines are the same, at least the whole service ought to get the alert [17:06] ok cool therve. And you also have a Landscape link to the machines that represent the current environment, right? [17:06] gary_poster, correct [17:06] cool [17:07] therve, "all the machines are the same": really? Juju allows you deploy machines with a given set of constraints, and then add machines later to the same service after having changed constraints [17:07] and even over time I wonder if new machines might be different than old [17:08] gary_poster, all the machines in a service [17:08] iiuc [17:09] therve, since you can add units over time, that can change things yeah? But anyway, IIUC it is irrelevant: you are already giving us per unit annotations about reboot/security, and then our Landscape reboot/security control is for...the entire environment or every single machine that your Landscape is managing? [17:09] irrelevant to us right now I mean :-) [17:10] gary_poster, we can decide maybe [17:10] because what you have would handle either scenario, homogenous or heterogeneous [17:10] ok, therve, np, it doesn't change the story much on our side [17:11] therve: just deployed juju gui on lxc, worked fine, what's the juju origin in your environment.yaml? [17:11] frankban, ppa [17:15] therve: in the charm directory, juju-gui should be a link. where does it point to? [17:18] frankban, /var/lib/juju/units/juju-gui-1/charm/release/juju-gui-0.1.5 [17:27] Hey Makyo! Welcome back! How was the trip? [17:33] gary_poster, Excellent! Good to be home, though. [17:36] gary_poster, sent an email about annotations [17:41] therve: output of "ll /var/lib/juju/units/juju-gui-1/charm/release/juju-gui-0.1.5/ ? [17:42] frankban, sorry the instance is gone :/ [17:42] therve: np [17:45] new review up: https://codereview.appspot.com/7195047/ [17:46] frankban, it seems to be that the setup_gui logic in the charm is wrong [17:46] ie the first_path_in_dir call doesn't do what it ough to [17:46] (although I'm not sure how this could end up being build-prod I guess) [17:48] therve: in setup_gui we link the uncompressed release dir to juju-gui [17:48] correct [17:49] at this point, juju-gui should contain build-prod, which is the root of the production GUI. [17:51] therve: then, in start_gui, that directory is build_dir, and we overwrite the configuration file: config_js_path = os.path.join(build_dir, 'juju-ui', 'assets', 'config.js') [17:52] frankban, oh right, there is one more level though [17:53] frankban, it's looking for /var/lib/juju/units/juju-gui-0/charm/juju-gui/build_prod/ whereas the path is /var/lib/juju/units/juju-gui-0/charm/juju-gui/$version/build_prod/ [17:55] therve: you mean, $version is juju-gui-0.1.5? [17:55] frankban, yeah [17:56] therve: the directory structure, starting from the charm root, should be: [17:56] frankban, the ln in setup_gui create a link inside JUJU_GUI_DIR, it doesn't replace it [17:58] the -f in ln doesn't remove the drectory [17:59] therve: but at the time setup_gui call ln -s, JUJU_GUI_DIR should not exist [17:59] frankban, ah that may be the issue. It does exists [17:59] so, the problem is, it exists, and should not [18:00] * therve nods [18:00] oh, hum [18:01] frankban, I have a file in my local charm :/ [18:01] let me try again [18:01] therve: how did you deploy the charm when you first encountered this problem? [18:01] ah, ok [18:15] gary_poster: https://codereview.appspot.com/7205044 (from me and Nicola) should address all the comments in the charm review [18:15] awesome frankban, looking [18:15] thanks [18:21] teknico, frankban nice work, thank you. Land as is [18:21] frankban, ok it works after cleaning up my local charm. Sorry for the noise. [18:21] gary_poster, thanks :-) [18:21] :-) [18:21] therve: no problem :-) [18:22] therve, yeah, glad it's ok :-) [18:22] thank you frankban for walking through with him. :-) [18:24] gary_poster: cool, EOD for us, after the second review, if everything's ok, please feel free to land the branch (or have someone do that), and re-propose. I've tried tests and qa, they pass, but a double check would be great [18:24] cool frankban and teknico have a great evening and thanks again === goodspud_ is now known as goodspud [20:14] allgui: ok, I have some crazyness that I'm having trouble debugging, anyone available to pair? [20:17] benji, that's guihelp I think yeah? [20:18] gary_poster: could be, I couldn't remember and couldn't find it written down anywhere [20:18] :-) [20:18] guihelp: ok, I have some crazyness that I'm having trouble debugging, anyone available to pair? [20:18] benji, it's in the kanban card from the discussion. I try to keep notes there before blogs [20:19] it is also in my notes now [20:19] cool [20:42] hazmat, https://code.launchpad.net/~therve/juju/env-annotations-gc/+merge/144583 for your reviewing pleasure [22:02] Makyo: thanks for the review of my branch, would you mind reviewing another one for me? I want to move them as far along as possible so I'm not gumming up the works. https://codereview.appspot.com/7133069 [22:02] benji, on it. [22:02] thanks! [22:04] benji, This is stacked on your other branch, right? [22:04] Makyo: right... oh! I bet you are asking that because the diff is screwed up. [22:04] It's just big, yeah. Where should I be looking, specifically? [22:05] let me see if I can fix it [22:06] * benji turns the big handle on the side of an oily, smokey machine labled "lbox" [22:11] Makyo: I give up. Here is the real diff: http://paste.ubuntu.com/1564396/ and here is the new (but still bloated) review: https://codereview.appspot.com/7206047 [22:11] (why a new review? because I killed the old one in hopes that the -req switch for lbox would actually work) [22:12] benji, Thanks, will take a look [22:12] I appreciate it. [22:16] hazmat: pdr 2? [22:16] robbiew: pdr 2 [22:32] alejandraobregon: ack