[00:04] why is the only place to get the required version of Mongo in an ec2 bucket? Even worse, a binary coming over an unencrypted connection. [00:05] bigjools: we're open to suggestions [00:05] davecheney: I need to understand why it's there before I can make a meaningful suggestion, but PPAs spring to mind [00:06] i can't answer the PPA question [00:06] the reason we user our own version of mongo is we need 2.2.0 [00:07] which isn't in LTS [00:07] and we also need TLS enabled, which isn't in any published version that we could find [00:09] davecheney: where is the source? I am happy to get it in a PPA [00:09] (assuming you meant there's a diff for TLS?) [00:10] not, just a compile option [00:10] ok [00:10] bigjools: lets take this to the juju-dev / canonical-juju mailing list [00:10] gustavo needs to be involved, but is not online in this timezone [00:10] I had better join them :) [00:11] I will stick it in a PPA for my own use anyway, and it can be re-used if people see fit === slank_away is now known as slank [00:41] any juju devs in the house [00:41] trunk is super broken [00:41] i am trying to confirm it isn't a local problem === slank is now known as slank_away [05:59] never mind, it was a problem with my machine [08:08] *: morning [08:12] davecheney, fwereade_, TheMue: morning! [08:12] rogpeppe1: hiya [08:12] rogpeppe1, davecheney, TheMue: heyhey [08:12] fwereade_: hiya 2 [08:12] actually just popping out for a croissant [08:12] bbiab [08:12] * rogpeppe1 learned how to make croissants over new year [08:12] ooh, nice [08:19] rogpeppe1: so we know what we'll get during the next meeting? [08:19] TheMue: only if there's an oven and lots of butter! [08:20] rogpeppe1: i'll try to organize it :) [08:20] rogpeppe1: butter should be no problem, but the oven ... [08:27] hello. [08:27] aram: hiya [08:29] aram: hi [09:21] bigjools: one reason to use a binary rather than a ppa, is because installing mongod starts the service running, rather than just having a binary that gets started and stopped by the test suite. [09:22] (though I imagine the *real* reason is just people not being familiar with PPAs) [09:25] jam: and configuration [09:29] anyway I have one in a PPA that does the trick [10:21] I didn't say where, but it's here: ppa:julian-edwards/mongodb [11:01] rogpeppe1, TheMue: https://codereview.appspot.com/7091043/ is trivial and important; https://codereview.appspot.com/7092044/ is slightly more involved and slightly less likely to break the average user's deployment but still good to have for 1.9.6 [11:02] rogpeppe1, TheMue: https://codereview.appspot.com/7058073/ is strictly less important but still a nice-to-have-agreement-on so I feel ok about progressing with the AUST removal [11:02] fwereade_: so, did you decide that the charm you found using JUJU_REMOTE_UNIT in a relation-broken hook isn't crackful after all? [11:02] fwereade_: *click* [11:02] rogpeppe1, it's crack IMO [11:02] fwereade_: or is this just a backward-compatibility sop? [11:02] rogpeppe1, but python sets it to "" rather than leaving it unset [11:03] rogpeppe1, exactly [11:03] fwereade_: ah, i see, so that charm would've failed to get the svc name anyway [11:03] rogpeppe1, and it doesn't actually use it anyway [11:03] fwereade_: but at least the env var test would succeed [11:03] rogpeppe1, yeah [11:03] rogpeppe1, both suggested ones verified against actual charms on actual instances [11:04] fwereade_: LGTM for that one [11:04] rogpeppe1, cheers [11:27] lunchtime [11:27] fwereade_: first lgtm is in, second one looks good so far too, but i have to continue after lunch [11:27] TheMue, cool, thanks [11:29] rogpeppe1, I think https://codereview.appspot.com/7094043 is trivial [11:30] fwereade_: i'm still trying to absorb 7092044... [11:30] rogpeppe1, no worries -- sorry, I seem to be channelling a firehose this week :/ [11:30] fwereade_: i think you've got two LGTMs for 7094043 already, no? [11:31] rogpeppe1, that's 1043 [11:31] fwereade_: oh yeah. too darn similar, thouse numbers [11:31] rogpeppe1, yeah, most unhelpful :) [11:49] fwereade_: all but https://codereview.appspot.com/7058073/ reviewed [11:49] rogpeppe1, awesome, thanks [11:50] fwereade_: i'm gonna leave that for a little bit if you don't mind, i'm not quite up to it currently, and i'm wanting to push the agent stuff forward [11:50] rogpeppe1, yeah, np, it's definitely lower priority [12:00] bloody hell, cmd/jujud tests pass with flags removed. that took a while. [12:05] me cheers at rogpeppe1 [12:05] * fwereade_ cheers at rogpeppe1 [12:06] fwereade_: i'm afraid the CL might be a bit of a monster though. there's so much circularity, it's difficult to break up. [12:06] * fwereade_ braces himself ;) [12:06] fwereade_: i'm fairly happy with the way it's turning out though. definitely feels like an improvement [12:07] rogpeppe1, awesome [12:07] fwereade_: then again, i've though that before and... :-| [12:07] fwereade_, rogpeppe1 - can you have a look https://codereview.appspot.com/7073060/ [12:07] rogpeppe1, heh, yeah [12:07] dimitern, *click* [12:12] dimitern: looking [12:51] dimitern: reviewed [12:51] rogpeppe1: tyvm [12:51] fwereade_: I saw yours as well - thanks [12:51] dimitern, np, yw [12:54] fwereade_: ping [12:55] rogpeppe1, pong [12:55] fwereade_: just want to confirm something that i think is wrong currently [12:55] fwereade_: in deployer/simple.go, DeployUnit doesn't seem to pass the --data-dir argument to the unit agent [12:56] fwereade_: am i missing something, or is it only working by luck [12:56] ? [12:56] rogpeppe1, ha, does it not? it probably should, let me check [12:57] rogpeppe1, yeah, it's crack, fix please :) [12:57] fwereade_: will be fixed with all the other changes [12:57] rogpeppe1, <3 [13:03] rogpeppe1, have comments on https://codereview.appspot.com/7094043/ and https://codereview.appspot.com/7092044/ -- but finish your branch first ;) [13:06] fwereade_: replied to one === rogpeppe1 is now known as rogpeppe [13:08] rogpeppe, replied to that :) [13:11] fwereade_: LGTM [13:12] rogpeppe, do you think it qualifies as trivial? [13:12] fwereade_: not quite. after all, i had to think about it quite carefully. better wait for another pair of eyes. [13:13] rogpeppe, true enough -- thanks :) [13:13] fwereade_: another question around deployer/simple.go: we pass the --log-file flag in, but we set upstart.Conf.Out to the same file - is there any point in doing the former when we've got the latter? [13:14] rogpeppe, meh, probably not [13:14] rogpeppe, dropping it will make life simpler, won't it [13:14] fwereade_: i'd like to lose as many flags as possible [13:14] rogpeppe, +1 [13:15] fwereade_: the only dubious one left is --debug [13:15] rogpeppe, hmm [13:15] rogpeppe, I think I'd like to leave it in, but turn off the state subpackage debug spam [13:16] fwereade_: at some point in the future it would be nice to be able to enable/disable it dynamically. [13:16] rogpeppe, was just composing that message ;p [13:16] fwereade_: at which point we don't really want it as a flag [13:16] fwereade_: or maybe we do, for the initial debug state [13:17] rogpeppe, I think starting in that mode and then dropping to a quieter level once that's detected is saner [13:17] fwereade_: yeah, probably [13:24] TheMue, thanks for the review -- did you see the followup with only a single delaying hook check? I think it strikes a decent balance. [13:25] fwereade_: will take a look [13:26] fwereade_: ah, ok, see the change. like it. [13:34] TheMue, cool, thanks [13:34] fwereade_: but now i'm fighting through the larger one. ;) [13:46] lunching === slank_away is now known as slank === slank is now known as slank_away === slank_away is now known as slank [14:03] rogpeppe, fwereade_- when you have time, please - https://codereview.appspot.com/7073060 [14:03] i fixed suggestions & answered questions [14:12] back [14:12] dimitern: looking [14:13] rogpeppe: 10x === TheMue_ is now known as TheMue [14:33] rogpeppe, how long do you think it's reasonable to wait for a dying subordinate to be removed? [14:33] fwereade_: indefinitely? [14:33] rogpeppe, I guess... [14:33] fwereade_: then it can be force-removed if necessary [14:33] rogpeppe, it's not like complaining is going to make the situation any better, I suppose ;p [14:33] fwereade_: indeed [14:33] rogpeppe, yeah, sgtm [14:35] dimitern: reviewed [14:36] rogpeppe: thank you! [14:39] fwereade_: would you take a look please - https://codereview.appspot.com/7073060 - I tried to answer your questions [14:40] dimitern, I asked a followup just above [14:40] dimitern, I'm not following https://codereview.appspot.com/7073060/diff/1/testservices/novaservice/service_http.go#newcode523 [14:40] dimitern, you got the group name right out of the request [14:40] dimitern, surely it could be handled at any stage? [14:40] fwereade_: you mean I should check for it in advance? [14:41] dimitern, it seems strange to me to create the server and *then* fail when checking the security groups [14:42] fwereade_: yeah, put like this it's indeed strange (even though OS does it similarly), I'll rearrange it [14:42] dimitern, well, if OS creates the server, we should copy that [14:43] dimitern, does it really though? I thought SGs were set at instance create time and unchangeable after that -- it that OS or just EC2? [14:43] fwereade_: looking at both responses and nova source things are not clear, but I'll do it as it makes sense [14:45] dimitern, cheers [14:53] fwereade_: take a look now is it better https://codereview.appspot.com/7073060 [14:55] rogpeppe: as well ^^ (just in case I misunderstood something) [14:57] dimitern: LGTM [14:57] so, telephone support for my brother done, he wants to install ubuntu ;) [14:58] rogpeppe: cheers! [15:03] dimitern, LGTM modulo one set of comments, let me know what you think [15:03] fwereade_: thanks, looking [15:04] fwereade_: yeah, the duplication is ugly, but wasn't sure how to fix it best [15:09] fwereade_: I think I did it better now - https://codereview.appspot.com/7073060 [15:11] dimitern, LGTM [15:11] fwereade_: sweet! 10x [15:11] dimitern, yw :) [15:11] dimitern, https://codereview.appspot.com/7094043/ is pretty small if you have a mo after merging [15:12] fwereade_: sure - just 2 mins and I'm on it [15:15] I'm going to get my batch of rasberry pi's. [15:15] back in an hour or so. [15:16] aram: cool, happy hacking :) [15:17] fwereade_: LGTM [15:17] aram, nice, enjoy [15:17] dimitern, awesome, tyvm [16:21] all live tests pass, phew [16:22] not all at once though - we need to sort out our reliability [16:38] i've just discovered this old CL that i never got another review on [16:38] https://codereview.appspot.com/6963050/ [16:38] please could someone other than william give me another review [16:39] aram, TheMue, jam, mgz: ^ [16:39] rogpeppe: *click* [16:39] TheMue: thanks [16:47] rogpeppe: +1 [16:48] TheMue: thank you! [16:48] rogpeppe: yw, the whole script generation and testing isn't simple. [16:49] TheMue: yeah. i prefer being able to see the generated script in the tests now. at least i can eyeball it for problems, even if we can't run it in the tests. [16:51] rogpeppe: absolutely. i needed some time to understand the generation. now i atleast can see the expected result immediatelly. [16:56] fwereade_, TheMue: here's the beast: https://codereview.appspot.com/7102043 [16:57] rogpeppe: uuuh [16:58] TheMue: sorry for the size - i have no idea how i can break it down [16:59] rogpeppe: no problem, but i'll review it over the weekend. our younger daughter and i will drive to the cinema to watch "life of pi in a few moments. [17:00] TheMue: ah, enjoy. i've heard at least one good report. [17:00] dimitern: if you fancy a challenging review: https://codereview.appspot.com/7102043 :-) [17:00] rogpeppe: yes, and what i've seen so far looked also good. [17:00] TheMue: i enjoyed the book [17:00] rogpeppe: i'm on it\ [17:01] dimitern: thanks! [17:01] so everyone have a nice weekend. [17:01] TheMue: and you. [17:27] rogpeppe: reviewed [17:27] dimitern: you're a lovely man [17:28] rogpeppe: I keep forgetting to use 'm' to reply when there are comments, rather than reply and then publish :) [17:28] dimitern: :-) [17:33] rogpeppe, thanks, I'll take a look [17:34] fwereade_: that would be great. [17:34] fwereade_: it would make my week if i could get it submitted this evening :-) [17:50] rogpeppe, LGTM :) [17:50] fwereade_: really, wa hay! [17:50] rogpeppe, assuming you;ve run it live ;p [17:51] fwereade_: i have [17:51] rogpeppe, then I'm happy :) [17:51] fwereade_: i get no more failures than the ones we're regularly getting in trunk anyway [17:52] rogpeppe, heh [17:52] fwereade_: TestBootstrapWithDefaultSeries is failing every time for me in trunk, live. [17:52] fwereade_: but not if i run it on its own [17:52] rogpeppe, hum, that is bad [17:57] fwereade_: I didn't get that - you replied to my review on https://codereview.appspot.com/7102043/ - are you saying +1 on all my comments? [17:59] dimitern, I was only feeling strongly +1 about the one I badged, the others I guess I'm +0 on but I don't feel strongly enough to actually chime in [18:00] rogpeppe, dimitern: if either of you fancies https://codereview.appspot.com/7094045 it would be cool, but it is getting pretty EODy [18:00] fwereade_: you didn't publish your comments BTW [18:01] fwereade_: I'll take a look at yours [18:01] rogpeppe, dimitern: ah, sorry, I am clearly codereview-impaired today [18:02] but rogpeppe, ignore that one in favour of checking my response to https://codereview.appspot.com/7092044/ [18:07] fwereade_: LGTM - nice and short [18:09] fwereade_: btw I managed to fix my machine finally (I hope) - it needed a good dusting and cleaning - that brought the temp -30 deg. down :) and no longer lags terribly [18:11] dimitern, ha [18:11] dimitern, excellent news anyway :) [18:12] :) yeah [18:13] bringing the temp to -30 deg would be impressive [18:13] though perhaps would make using the laptop a little hard... [18:14] mgz: well, according to sensors - it was running most of the time +80 on both cores, which is high (90 is critical), now it's down to 40-50! [18:15] erk. sounds similar to the issues jamespage had, which also required som pretty heavy dust busting [18:15] graphics card getting used more in newer version probably contributed something to general overheating. [18:16] mgz: it's strange it needed that, because I dusted it like 5-6 months ago [18:16] but i guess even the small amount of gruel accumulated on the fins of the heat sink was enough [18:17] mgz: how's public URLs going? I've run out of things to do - good it's EOD :) [18:20] going to land shortly [18:23] mgz: it seems I missed the daily standup somehow - I was there on time (12 UTC), john and mark were there, but nobody talked to me and after staying there an hour, I left :) [18:23] basically Ian's branches are fine I think, so have just tweaked slightly and will land the two together and send note [18:24] i'm done for the day. have a great weekend, all! [18:24] fwereade_: thanks in particular for that timely review! [18:24] rogpeppe: g'nite and happy weekend [18:26] mgz: cool, so any progress on the versioning discussion/doc? [18:27] jam wrote up a few things [18:28] ^on the standup, we generally don't do fri as it's not a UAE work day, but I should have got on mumble and said hi to you this morning probably as we've generally done that [18:28] was just having 'fun' trying to get canonistack to behave at the time... [18:28] mgz: but we're still not applying anything to the workflow [18:28] mgz: I see [18:28] not yet [18:29] apart from simultanious landing and note to mailing list, which is my plan [18:30] ok