[00:38] alexisb: bauer is down again, :( RT 75536 === urulama-afk is now known as urulama [08:35] fwereade, hey [08:36] fwereade, updated http://reviews.vapour.ws/r/141/diff/3-4/ if you can have a look? [08:58] dimitern, cheers [09:07] dimitern, on further inspection today, I'm worried that we really rather need to define how relation ports work at the context level [09:08] dimitern, otherwise when we start using ones that aren't -1 we *will* forget to test it all properly [09:08] dimitern, because anyone seeing the try* funcs will assume that relationId is a meaningful parameter [09:08] fwereade, I was hoping to discuss this in brussels [09:08] dimitern, tell you what [09:09] fwereade, yeah? [09:09] dimitern, take relationId out of the signatures of the try funcs [09:09] dimitern, we'll still need to rewrite them to take it into account properly [09:09] fwereade, and hardcode -1 inside? [09:10] dimitern, but it'll do what we need to today, and we won't be able to fool ourselves into believing the logic is more capable than it is [09:10] dimitern, yeah [09:10] dimitern, the first line can be "relationId := -1" even [09:11] fwereade, ok, will this be sufficient for now? I can also add a TODO comment about refactoring both funcs to take relation ids properly into account [09:11] dimitern, I think so, yeah [09:12] fwereade, cheers! [09:21] dimitern, given that, LGTM with trivials [09:22] fwereade, great, just finished with the changes and re-proposing, thanks again [09:54] fwereade, last step - a small branch that implements the "opened-ports" hook tool to return a list of ports or ranges opened by the unit: http://reviews.vapour.ws/r/144/ [10:15] dimitern, hmm, I'm wondering how that will mesh with (again) per-relation ports [10:15] dimitern, there's an uncomfortable tension between implicit -1 and implicit current relation [10:18] fwereade, we can add [-r relid] argument later perhaps? [10:20] dimitern, yeah, the concern is that all the *other* -r flags assume current relation [10:21] dimitern, so there's a change in behaviour for charms that open/close/opened in a relation context [10:21] dimitern, OTOH it'll be exactly the same problem for open/close, there's nothing special about opened [10:22] fwereade, yeah, I think so [10:22] dimitern, I'm worrying a bit about output format though [10:22] dimitern, is a list sufficient? [10:23] dimitern, I guess it probably is [10:23] dimitern, we can have an --all flag with a map or something [10:25] dimitern, LGTM with --format yaml, --format json [10:25] fwereade, the format flag is common to all commands btw [10:25] dimitern, yeah, I expressed it insanely, but I'm mainly whining about the lack of tests [10:26] dimitern, doesn't need to be super-complex, just make sure they work [10:26] fwereade, sure, np [10:26] fwereade, thanks! [11:54] * perrito666 bags ready [13:34] natefinch: going to the 1:1? [13:34] perrito666: yep [14:00] nate switched to stdup [15:10] natefinch, perrito666 This issue is blocking the cloud installer. We need to understand if this is juju, ,maas, or both, and do we need a backported fix to 1.20, see bug 1376952 [15:10] Bug #1376952: juju machine stuck on 'pending' with no error message: juju 1.20.9.1, maas provider

[15:11] natefinch, perrito666 - mmcc_ is here if you have further questions [15:12] yep, hi - the environment has been torn down but the systems are available to try to reproduce, and if anyone wants access to look themselves, that’d be fine [15:17] stokachu, does machine 1 exist? Can you get a copy of the /var/log/cloud-init-output.log to tell about what happened before juju's agent was started [15:18] mmcc_, ^ [15:19] sinzui: that machine is not up, and has been recommissioned a couple of times. I’ll try to reproduce and see if I can get that log for you [15:19] the only strange thing I see and can somehow understand is that peergrouper complains that it cannot determine the replicaset [15:19] which leads me to know that the state server might be failing while trying to create the replicaset [15:20] natefinch: any input? [15:21] looking [15:22] natefinch: did you by any change got my package? [15:24] perrito666: I haven't gotten anything yet, I was actually going to ask about that. Do you have tracking info on it? [15:24] perrito666: if you ever need to get anything here in a short period of time, I have Amazon Prime, so I can get stuff shipped 2nd day for free [15:25] perrito666: hell, I can get something delivered tomorrow morning for $5, if you really want :) [15:30] natefinch: lol :p I think this is not amazon delivering so It might take a lot longer, its the size of a usb thumb so you can store it until the next sprint if you get it after :p I would say you can use it but it only is useful on thinkpads [15:30] haha [15:31] perrito666: ahh, yeah, I never buy anything that isn't shipped by amazon unless I'm desperate for that very specific thing... because you never know how long some schmoe is going to take to get it to you, and chances are he's going to send it the slowest, cheapest way possible [15:32] natefinch: well its a bit of a specific part so it was only available from that source [15:35] perrito666: yeah, figured. bummer. Hope it makes it here. It's probably being sent via US Post Office, which delivers today and tomorrow, so we might get lucky. [15:36] natefinch: well laptop spare parts (in this case is the bluetooth module) are almost always sent from china, what I dont understand is why someone made this laptop without the bluetooth module in the first palce [15:37] i'm running into an ordering issue using gocheck; i'm trying to use SetUpTest, but that is precluding something else from setting up state... does this sound familiar to anyone? [15:37] I use a usb bluetooth dongle but I keep breaking those against things [15:37] katco: not following [15:38] perrito666: are you familiar with gocheck's SetUpTest functionality? [15:38] perrito666: it basically is a special function that will be run before every test in a suite [15:39] perrito666: the problem is, if i define it, then state in an embedded suite never gets set up [15:39] katco: you have to manually call sub-suite's setup functions too [15:40] remember, the methods on the struct you're defining override the ones on embedded structs, so you have to manually call the ones on the embedded structs [15:41] natefinch: ah that was exactly it [15:41] natefinch: dependency model i didn't fully understand [15:41] katco: yep. happen to uh... a friend of mine. ;) [15:41] haha [15:42] in the perfect world in my brain, nothing is ever that obscure [15:42] yeah, it's really the fault of the magic in gocheck [15:43] i would have thought that gocheck would handle that for you [15:43] since it has all the suites which have been registered [15:43] ok, back to hacking. thanks perrito666, natefinch [15:44] well, it just looks for the top level method and calls it [15:46] ok bbl, going lunch with the wife to compensate for 1 week driking with friends in brussels [15:51] haha sounds fair ;) [15:56] cmars, mongo ++ :-) [15:56] cmars, er. gonzo ++ [15:57] hazmat, fear and loathing in nosql databases [15:57] idk [15:57] fun little side project i've been hacking on [15:58] hazmat, https://github.com/cmars/gonzodb [15:58] cmars, yeah. i saw [15:58] cmars, reminds me of https://github.com/rick446/MongoTools/blob/master/mongotools/mim/mim.py [15:58] mongo in memory .. make unit tests FAST [16:01] hazmat: that would be amazing [16:02] hazmat, that python version is impressive. even embeds JS. nice! [16:02] natefinch, cmars alexisb we should schedule a topic on this for next week [16:03] hazmat, I was not following the discussion, give me a minute and will do [16:06] alexisb, i'll add, just wanted to put it on your radar [16:06] actually we already have an agenda item that matches this topic [16:06] Juju Testing: Better Merge Gating, Faster Testing === Ursinha is now known as Ursinha-afk [17:09] we used MIM at SourceForge for our testing [17:22] sinzui, natefinch, perrito666 - trying to reproduce bug 1376952 just now, the machine came up correctly, so I will just go ahead and use it. if it breaks again, I’ll update the bug. thanks for looking, and let me know if I can enable any extra logging or anything that’d help diagnose it next time [17:22] Bug #1376952: juju machine stuck on 'pending' with no error message: juju 1.20.9.1, maas provider

=== mmcc__ is now known as mmcc_ [17:24] mmcc_: ok, thanks === Ursinha-afk is now known as Ursinha === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha === Ursinha is now known as Ursinha-afk === fuzzy is now known as Ponyo === Ponyo is now known as Fuzai === Fuzai is now known as Ponyo === mmcc__ is now known as mmcc_ === Ursinha-afk is now known as Ursinha === sebas538_ is now known as sebas5384 === mmcc__ is now known as mmcc_