[00:20] hi again [00:42] CI has failed to build for me 3 times in a row: "35 repositories updated; 2 failed" [00:42] any thoughts? [00:47] godeps: cannot update "/var/lib/jenkins/workspace/github-merge-juju/tmp.uAmw1jccO3/RELEASE/src/bitbucket.org/kardianos/service": package bitbucket.org/kardianos/service: https://api.bitbucket.org/1.0/repositories/kardianos/service: 503 Service Unavailable [00:47] bitbucket is down [00:47] davecheney: ah === ericsnow_ is now known as ericsnow [00:51] * thumper needs a casual (╯°□°)╯︵ ┻━┻ [02:21] * perrito666 hangs the call and immediately falls asleep on the desk [02:21] see some of you tomorrow, cheers [02:42] wwitzel3: did you look at Open in api/apiclient.go? [02:43] wwitzel3: that's where the cert stuff happens on the client side [02:46] wwitzel3: past that there isn't a whole lot you can plug into [03:18] ericsnow: looking now [03:22] ericsnow: yeah I started there but wasn't able to actually figure anything out, as far as I can tell, in that file, on line 231, we are already doing that exact tls config for the websocket. [03:23] wwitzel3: maybe it's a websockets vs. normal HTTP thing then (as backups is using direct HTTP requests for download) [03:24] ericsnow: maybe, yeah, I will dig in to more in the morning, I'm going to take some medicine and try to get some extra zzz's tonight. [03:24] ericsnow: appreciate you taking a peak at it though [03:24] wwitzel3: yeah, I'm EOD soon too [03:24] wwitzel3: np [03:55] davecheney: care to leave a ship it? http://reviews.vapour.ws/r/133 === thumper is now known as thumper-afk [04:30] thumper-afk: looking [04:30] done === uru_ is now known as urulama [05:39] davecheney: ta [07:50] morning [07:51] morning TheMue [08:05] TheMue: you're gone tomorrow, right? I'm surprised to not see Dimiter here yet. [08:05] fwereade: I'd like to chat with you about possible TXN designs. [08:06] jam2, heyhey [08:06] jam2: no, have been here yesterday, had my hangout with dimiter and missed you [08:06] TheMue: yeah, sorry I missed the hangout, ended up getting distracted and missed the ping [08:06] jam2, sgtm -- juju-sapphire? [08:07] fwereade: I'm in a cafe, but we can certainly give it a shot [08:07] jam2: I made my final changes to my comming PR in the evening and right now doing the last checks before proposal *yay* [08:25] jam2, hey [08:26] jam2, TheMue, http://reviews.vapour.ws/r/123/ - if you can have a look [09:02] jam2, did I fall over or did you? === jam21 is now known as jam1 [09:05] fwereade: I'm at a cafe that gives "free wifi" but only in 1 hour blocks [09:05] so I just switched back to tethering my phone [09:05] but I don't trust the 3G for a hangout === jam21 is now known as jam1 [09:06] fwereade: anyway, I think I've gotten enough to at least progress from here, thanks [09:09] jam1, cool [09:13] jam1, fwereade, hey, I really like to land this branch, if you can have a look and say it's fine - http://reviews.vapour.ws/r/123/ [09:14] dimitern: will take a look now while my suite is running in the background ;) [09:15] TheMue, cheers! [09:24] dimitern, ah yeah sorry I got some of the way through the second page there yesterday [09:24] fwereade, np, was it looking ok? [09:24] dimitern, yeah, maybe a couple of minor quibbles [09:24] dimitern, just going through the rest now [09:25] fwereade, great, thanks! [09:26] fwereade, re doing all ports in a single api call in finalize.. if it's ok with you, now once we have versioning in the uniter api, it will be easier to add a FinalizeHookContext api call that does all in one [09:27] dimitern, yeah, agreed, we don't want that yet [09:27] but as a follow-up, once the rest of the port ranges PRs land [09:27] fwereade, cheers [09:28] dimitern, that's LGTM, but please bear all the quibbles in mind -- I think they do need work, but they don't have to be that PR [09:28] dimitern, the v0/v1 split is worth landing even if all the details aren't pefect [09:29] fwereade, sure, will do [09:32] fwereade, re "FWIW, I would be really happy to see this removed. I guess it's a derail right now, but it's completely pointless -- part of an early abortive metering thing." - can you expand on this one a bit please? === rogpeppe is now known as rogpeppe1 === rogpeppe1 is now known as rogpeppe [09:35] fwereade, specifically, what's wrong with calling ParseUserTag at the end of ServiceOwner? [09:41] dimitern, sorry, it's about ServiceOwner itself [09:43] fwereade, so you're saying we should get rid of it at some point? [09:43] dimitern, yeah [09:43] fwereade, that's fine, what I did is just fix it to be consistent with other api calls (permission checks, bulk ops, etc.) [09:44] that definitely doesn't have to be in this one [09:44] dimitern, it's just something I mention whenever I see it [09:44] fwereade, right :) [09:44] TheMue, thanks for the review as well [09:45] dimitern: yw ;) [09:46] dimitern: just proposing mine now [10:00] jam2: I have to say while I'm happy with the way we found to handle API versioning the original approach based on the EnvironmentCapabilities as dimiter suggested it has a somehow more clear logic. === thumper-afk is now known as thumper [10:18] fwereade: hey, got a few minutes to chat? [10:40] trivial review someone please! [10:43] TheMue, you've got a review [10:43] mgz, looking [10:43] dimitern: just got the notification, thanks [10:45] mgz, LGTM [10:46] dimitern: thanks! [11:41] jam, fwereade, reproposed ports sandboxing against master - http://reviews.vapour.ws/r/141/ please take a look [11:45] * dimitern needs to step out for ~1h [11:48] rogpeppe: when you have a mo [11:49] mgz: LGTM [12:20] jam, was that tokumx build i gave for 1.5 targeted towards a build dir / bin .. rebuilding 2.0.0 atm [12:40] jam, fwereade, review poke :) [12:51] rogpeppe: if you add me to ~godeps-maintainers I'll go ahead and land [12:52] jam, tokumx 2.0 binary https://www.dropbox.com/s/f78gcixuerpbpjw/tokumx-2.0.0-linux-x86_64.tar.gz?dl=0 [12:52] w/ ssl [12:52] mgz: done [12:58] rogpeppe: landed, thanks! [12:58] mgz: definitely nicer error messages, thanks [13:10] jam, ping me if you've got a minute to chat toku, added some comments to the docs [13:14] jam, I don't seem to be able to change teams in githib -- can you? [13:21] fwereade: you should be able to, but I can if needed as well [13:21] fwereade: (things related to the juju team at least) [13:21] mgz, yeah, I thought I could once, didn't seem to be happening today though [13:21] mgz, would you remove vlad from hackers, and add gabriel-samfira, please? [13:22] sure thing. [13:22] I'll also owner you. [13:25] any folks seen this error before [13:25] http://pastebin.ubuntu.com/8479146/ [13:26] is this a out of memory on the provider (maas) or juju bootstrap? [13:26] gabriel has been sent an email, he'll need to accept to be added [13:27] arosales: ...that's a good one [13:27] arosales: I would say it, but I bet maas is misbehaving here as well [13:27] arosales: check the maas logs? [13:33] mgz: ya I see OperationalError: out of memory in maas logs [13:33] * arosales sighs [13:42] natefinch: ericsnow ? [13:46] perrito666: here [13:47] ericsnow: we seem to have the standup now [13:47] instead of later [13:47] :p [13:54] gsamfira: twothings, pr707 was go fmt sad, , and you need to join juju/hackers on github for $$merge$$ to be obeyed [13:54] mgz: yup, fixed and pushed [13:55] mgz: I think I have :) [13:56] okay, should be all set then, let me see [13:58] gsamfira: github seems to be lagging, juju has you as in the hackers team, but your profile doesn't have you in juju [13:59] mgz: eventual consistency ftw :). No rush, its fine [13:59] gsamfira: ah, you need to set your membership of juju to public [13:59] hmm..lemme see [14:00] gsamfira: on the page should see a link [14:00] ahh [14:00] thanks! [14:01] was looking for that miniscule link [14:02] gsamfira: itsgoing [14:02] worked :) [14:02] mgz: thanks! [14:43] I really hate the local power company === kadams54_ is now known as kadams54 === dames is now known as thedac [15:08] fwereade, updated http://reviews.vapour.ws/r/141/diff/1-2/ [15:08] fwereade, it should be good to land now [15:12] natefinch, hey, can you have a look as well? http://reviews.vapour.ws/r/141/ I really need to land this [15:20] brb [15:42] dimitern, posted a quick question/concern, does it make sense? [15:45] fwereade, so, currently at the api level there's no relation given or needed when calling OpenPorts or ClosePorts [15:45] fwereade, but we can add it easily later in the params.EntityPortRange struct [15:45] dimitern, we can always fix up the api later, yeah [15:46] dimitern, it's keeping track at the context level that I'm mostly concerned about [15:46] dimitern, we half keep track and half not? because we have machinePorts that does know about relations, but pending that doesn't [15:47] fwereade, pending does know now, but the RelationTag field of PortRangeInfo is ignored for now [15:47] dimitern, but the keys don't know [15:47] dimitern, as it is that data structure can only hold one info per range [15:48] fwereade, that's correct, but isn't this how it should be - in the context of a relation hook? [15:48] dimitern, don't think so -- just because there's a default relation in play doesn't mean you're barred from changing others [15:49] dimitern, you just need to be explicit about it [15:50] fwereade, I agree, but later we'll add a [-r rel-id] argument to open-port, close-port, and opened-ports [15:50] fwereade, it doesn't make sense to do it now, because it's ignored [15:51] dimitern, mmm, I'd rather have the context-level stuff understand relations but always be using a magic null relation -- rather than have to rewrite that logic when we try to incorporate them [15:52] dimitern, I think it can even be written and tested sanely right now, you just need to pass a hardcoded not-really-a-relation into those methods [15:52] dimitern, then enabling it at the command level becomes trivial, instead of a hidden whoops-rewrite-logic thing [15:54] fwereade, so you're saying pendingPorts should be something like map[X]PortRangeInfo, where X is struct { Ports network.PortRange, RelationTag names.RelationTag } ? [15:54] dimitern, something like that, yeah [15:55] dimitern, I know it's kinda tedious [15:55] dimitern, but since we're writing it now we may as well get it right, ratherthan reworking it later [15:55] fwereade, ok, fair point [15:55] dimitern, awesome, tyvm [15:56] fwereade, and re your other question [15:56] fwereade, changing the map key should solve that as well I think [15:56] dimitern, yeah, I think so [15:57] fwereade, sweet! thanks, I will change it and repropose [16:07] fwereade, can you LGTM it conditionally, so I can land it after that? [16:07] dimitern, it kinda feels like that's the heart of the CL is the trouble there [16:08] dimitern, give me an ETA on the changes and I'll set an alarm to come and look though [16:09] fwereade, I'll be done in 1h, will ping you [16:43] if anyone has time http://reviews.vapour.ws/r/142/diff [16:55] fwereade, updated - http://reviews.vapour.ws/r/141/diff/2-3/ [16:55] perrito666, "You don't have access to this review request." [16:57] dimitern: sorry I keep forgetting to publish these things [16:57] done [16:58] perrito666, cheers, looking [17:02] perrito666, reviewed [17:03] * dimitern is way past eod now, time to go [17:13] ok, I think I might actually be able to do a little work now === urulama is now known as urulama-afk [17:27] natefinch: if on a server deployed with juju [17:27] I take down eth0 and bring up eth1 on a different network [17:27] will juju agent detect that change? [17:27] in 1.20.9 [17:28] buh [17:28] I figured [17:28] (asking for a friend ;) [17:28] lhaha [17:29] natefinch: so, not really going to happen? [17:29] uh all the networking guys are off by now, so I'm not sure. I'd be pretty sure that juju wouldn't figure it out on its own. [17:29] can I kick juju in the ass to figure it out? [17:29] that you know of [17:29] like restart the agent [17:39] marcoceppi: you can certainly try... I don't know what it'll do, honestly. [18:53] is there a way to specify an environment for any juju command? i.e. juju -e local (blah) [18:55] yes? [18:55] - e local? [18:55] er -e [18:55] that should work on anything that requires an environment [18:55] hm doesn't seem to work for me? "error: flag provided but not defined: -e" [18:56] so you have to specify after the subcmd? [18:56] might have to be after the command, I forget if it cares [18:56] hm. just trying to figure out if there's a way i can make sure i don't hose my actual juju environment while developing :p [18:56] katco: you can always rename the jenv for that environment [18:57] to something not .jenv [18:57] then it's untouchable [18:57] yeah but then i have to flip bits around everytime i want to switch b/t developing and doing something with my env. [18:57] it's a good suggestion, but not ideal =/ [18:58] katco: there's a request in to be able to "lock" an environment so you don't accidentally muck up a pristine environment, but no one has tried tackling it yet [18:58] ah cool [18:58] damn juju developers. [18:58] never get to the important stuff! ;p [18:59] oh hey natefinch where/when you flying out? wondering if i'll catch any canonical ppl on my flight [18:59] well, the clientside part is trivial, but you want the serverside to respect it too... but then older juju clients wouldn't be able to unlock the environment, so what then? Maybe not a big deal. But anyway... it's tricky. [19:00] katco: Saturday 7pm EST BOS-Heathrow-Brussels [19:00] British Airlines IIRC [19:00] ah... i'm in dc [19:00] i mean i have a layover there, then direct to brussels [19:01] oh for the days when lambert intl. airport was actually an international airport lol. [19:01] *nod* This was the best times for me, gives me the most time on Saturday with my family without being a ridiculously extra long flight for no reason [19:01] jej [19:01] hhe [19:01] yeah, i'm flying out sat. night [19:02] might see you at the brussels airport [19:02] get in at like 7am i think [19:02] I get in at 11am.... I wanted to get in at 8am, but it would have meant leaving home at like noon on saturday [19:42] gotta run, the kids are refusing to take a nap and my wife is feeling worse again. [19:53] I can neve understand how being I kid I also refused to nap [19:53] and now I would give my life for a nap [19:53] katco: ericsnow some of you wants to add a review to http://reviews.vapour.ws/r/142/ ? [19:54] perrito666: I'll take a look [19:54] perrito666: i'm trying to wrap up a change, i'll try and have a look when i'm done [19:54] perrito666: thanks for the ping [20:08] is it my perception or git rebase will hit the same conflict every time? [20:29] ericsnow: ta [21:08] FYI, I'm going to be in and out a lot today as I prepare for the sprint next week [21:08] need to be at the airport in 7 hours [21:08] so need to pack etc. [22:01] random network question: anyone know what this arp line means? [22:01] ? (192.168.0.4) at (incomplete) on en1 ifscope [ethernet] [22:01] i'm having LAN routing issues [22:01] EOD a little early [22:01] and i think that's one of the symptoms