[01:07] waigani: would you mind have looking at http://reviews.vapour.ws/r/2290/ ? [01:08] waigani: there's no functional change [01:08] waigani: just a tedious rearranging of tests [01:08] waigani: no rush either [01:08] menn0: looking [01:09] menn0: fancy, how'd you get those lines in the description? [01:09] waigani: markdown renders "---" as a horizontal rule [01:09] cool [01:11] menn0: what's the plan? will we have allWatcher and allEnvWatcher or will the latter replace the former? [01:12] waigani: allwatcher stays as is, allenvwatcher gets added [01:13] okay [02:54] davecheney: argl [02:54] davecheney: cross compiling from ppc64el to amd64 is broken [02:54] http://paste.ubuntu.com/11990676/ [03:39] how, how just that one combination ?!? [03:50] davecheney: i guess some kind of grotty compiler bug [03:50] but i'm surprised that 386 works, that's mostly the same backend code [03:51] davecheney: do you know anything about the linker darwin/arm code? [04:44] not a lot [06:48] menn0: ping [06:55] hello! anyone around? [07:06] jam, hey, are you here today or swapping? [07:51] voidspace, hey :) welcome back [07:51] dimitern: hey, hi [07:51] dimitern: I'm now online, have a little office setup here [07:51] dimitern: next step coffee, then machine updates and checking work emails [07:51] voidspace, are you in Romania now? [07:51] dimitern: yep [07:52] dimitern: same timezone as you now [07:52] dimitern: got in last night [07:52] voidspace, cool :) [07:53] voidspace, TheMue, dooferlad, as none of us were working or feeling particularly well on Friday, I'd suggested to combine today's 1:1 with some retro/planning discussions [07:53] +1 [07:56] ok [07:56] TheMue: o/ [07:57] voidspace: followed you travel notes on the net ;) [07:57] your [07:57] heh [07:58] yeah, it's been fun [07:58] not amazingly *relaxing*, but fun [07:58] voidspace, I bet - after driving 3200km in 10 days [07:58] dimitern: +1 [07:59] dimitern: 5000km [08:00] dimitern: 3200 miles... [08:00] dooferlad: o/ [08:00] oh wow [08:00] voidspace: hello! welcome back! [08:00] dimitern: but that total is over three weeks, not ten days [08:00] although *most* of it was in ten days [08:00] voidspace, still, *quite* a lot of driving [08:00] dimitern: heh, so maybe about 3000km in ten days [08:01] yeah [08:01] dooferlad: thanks [08:01] good to be back at work for a rest [08:04] fwereade: you timed out on irc.c.c, I responded with: "yeah, I'm being a bit pedantic I suppose. I have seen rare errors due to this I'm sure, but you're right in that we do this in lots of other places" [08:04] fwereade: and I've landed it now [08:04] axw, ah, sorry [08:06] axw, I guess a Sync would be a bit less racy but it's maybe still fundamentally problematic [08:06] axw, ie just because you know an event has been delivered to a handler does not imply that the handler has finished handling it [08:06] axw, blasted concurrency :) [08:07] fwereade: indeed. I think we'd need to have deeper control over how the syncs are doing. seems to be infrequent enough atm that it's not pressing [08:07] are doing = are being done [08:08] fwereade: speaking of unusual language, maltese looks exotic to me [08:08] axw, yeah, it's got a lot of arabic but in a roman alphabet [08:10] axw, `sqaq tax-xama` is pronounced more-or-less "sa'a t'sharma" [08:10] fwereade: I never would have guessed that [08:10] axw, there's a g->j, j->y, x->sh mapping that covers *most* of the weirdness [08:10] good to know, thanks ;) [08:11] axw, in particular, we'll be sprinting in st julians, or "san giljan", pronounced more--or-less "jilyan" [08:12] axw, I was living there for a couple of weeks before I figured out that they were the sme place [08:12] fwereade: hah :) [08:27] dimitern: ping [08:29] voidspace, pong [08:29] ahh. [08:36] voidspace, you've disconnected before getting my pong I guess [08:36] dimitern: ah yeah [08:36] dimitern: I was seeing if I was still connected [08:36] dimitern: internet problems, but now I have a wired connection, so better [08:37] voidspace, I see [08:37] dimitern: I assumed the lack of a pong meant I was disconnected :-) [08:37] should be good now [08:37] voidspace, I thought 1GBps internet connections are common in RO (I'm told) [08:37] :) [08:37] dimitern: not in this house... [08:37] dimitern: the problem was the wifi not really going this far anyway [08:38] dimitern: not sure how fast the wired connection is (will check) - but "fast enough" I think [08:44] voidspace, speedtest.net [08:44] dimitern: indeed [08:44] dimitern: will wait until dropbox stops syncing [08:45] and my mail client finishes dealing with email [08:46] thunderbird has been frozen for about twenty minutes processing rules for three weeks worth of email! [08:46] :D [08:59] dimitern: just grabbing coffee! [08:59] will be at standup in a minute or two [09:00] voidspace, grab one for me :) [09:05] dimitern: heh, have done [09:05] omw [09:06] dimitern: ah, need to install hangout plugin :-/ [09:06] looks like I've not done a hangout from this laptop [09:07] or maybe it's just that firefox updated [09:07] voidspace, ok, we'll start slowly I guess until you're ready [09:22] rogpeppe1: pong [09:25] menn0: hiya [09:25] menn0: just wondering what the status is on the gigawatcher [09:26] rog: the state work is *mostly* done. I was about to put up a WIP PR b/c I want some feedback from fwereade about one aspect where resource usage could be an issue [09:26] rogpeppe1: ^^ [09:26] (forgot to hit tab) [09:27] menn0: cool [09:27] menn0: we'd like to start using it at some point :) [09:28] rogpeppe1: I know. I was asked to pick it up again last week at the sprint [09:28] menn0: cool [09:28] (not that I was at the sprint... just that thumper and rick_h_ asked me to pick it up) [09:28] rogpeppe1: a number of PRs prepping for the change have already landed [09:29] menn0: awesome [09:30] rogpeppe1: i'm reusing much of the existing allwatcher and multiwatcher stuff so it's coming along fairly nicely [09:30] rogpeppe1: the plan is the sneak it in before the 1.25 feature freeze [09:30] menn0: nice [09:30] menn0: looking forward to it [09:31] rogpeppe1: do you need it for the same thing that rick_h_ wants it for? [09:31] menn0: yes [09:31] rogpeppe1: cool cool [09:31] menn0: it's me that wants it, really :) [09:31] rogpeppe1: menn0 one thing we have to be careful of. It came up that we sometimes have allwatcher issues because core itself doens't use it. [09:31] rogpeppe1: menn0 so when we get this we need to make sure we work together on a testing plan that takes the uses we put it to into account. [09:32] rogpeppe1: menn0 and think about ways that changes to core, that might sneak past/not into this new watcher can be caught. [09:32] rick_h_: you're worried about stability of the API right? [09:33] menn0: you might also reach out to sparkiegeek as they're another potential users, though they weren't aware of it so it was new to them at the sprint. [09:33] menn0: yea, or things like the bug in the machine agent stuff, that new things are added, or bugs fixed, but they're not made through to the higher level watchers because they're not directly used/hit in core [09:34] menn0: we talked and I'm supposed to setup a chat with folks from our end, landscape, and core to think of ways of getting our uses of the allwatcher into some sort of test system core can rely on. It's not figured out yet. [09:34] menn0: and this fits the same bill I expect. [09:36] rick_h_: the same concerns certainly apply [09:37] rick_h_: well let me know when that call happens [09:37] menn0: will do, in the meantime as you and rogpeppe1 chat about things something to noddle on in the back of your brains. I think this is going to be a hard problem to figure out and having you two on board <3 [09:39] rick_h_: i'll have a think about what we can do [09:48] menn0: one thing that occurred to me is: will this watcher provide a message when the main JES is being destroyed? [09:50] rogpeppe1: I *think* it will, as long as the API server is up for long enough to deliver the message [09:50] rogpeppe1: when the main env goes to Dying an environment update should be emitted [09:50] menn0: cool [11:00] * dimitern steps out for ~ 1/2h [11:43] fwereade, back [11:50] goo morning all [11:50] lots of goo to you too perrigoo === psivaa is now known as psivaa-lunch [12:22] dooferlad: ping [12:22] anyone around who may be able to give me a hand? [12:58] fwereade: you are not around here by any chance, are you? [12:58] perrito666, yeah [12:58] perrito666, what can I do for you? [12:59] fwereade: I could go for a coffee with croissants, but since that is out of reach i am fixing https://bugs.launchpad.net/juju-core/+bug/1479289 [12:59] Bug #1479289: statushistory uses sequence, fails in multi-env state servers

[13:00] I see that probablyUpdateStatusHistory has a commet from you reagarding the explicit setting of EnvironUUID [13:01] I was wondering if you talk about the particular way we set envuuid all over juju or something else [13:04] mgz: I disable build-revision because I don't a build to collide with proposal of 1.24.4 [13:04] perrito666, it's specific to the Insert [13:05] dimitern: ping [13:05] perrito666, envStateCollection really ought to do its job properly and automatically [13:05] voidspace, pong [13:06] fwereade: agreed [13:06] * perrito666 kills 2 birds with one shot [13:51] katco, rogpeppe1, axw, davecheney, a trivial goamz review anyone? https://github.com/go-amz/amz/pull/59 - much appreciated [13:52] mgz, ^^ [13:52] +2,992 −2,721 [13:52] I'd hate to see a non-trivial change [13:52] ;) [13:53] natefinch, :) well, that's due to the package header comments [13:53] natefinch: those start by changing a couple of universal constants [13:53] natefinch, 99.9% of it is moving code around [13:54] dimitern: I can have a look [13:54] mgz, awesome! ta! [13:54] dimitern: all the package comments will get concatenated together... just put it in one file, and put the "this file contains" separate from the package file. Also you need the copyright headers [13:54] yay my new eye-glasses are ready... just need a moment to drive to the other side of the city :p [13:54] dimitern: LGTM [13:54] s/package file/package comment/ [13:54] rogpeppe1, thanks! [13:55] dimitern: and +1 to natefinch's comment [13:55] natefinch, good point about the headers, will do [13:55] dimitern: you should only have one package-level doc [13:55] not that server.go got them in the first place [13:55] rogpeppe1, will do, thanks [13:56] natefinch: explain that to me? I think I'm missing something about go comments [13:56] (or point me at doc page on it) [13:57] mgz: doc comments on the "package foo" declaration in multiple files will all get concatenated together, so you really only need it in one file (and if you put the same comment in multiple files, it'll be duplicated in the godoc) [13:58] brb [13:59] natefinch: everything at the top of each file? or is there some extra checking of bits before 'package' like there are for functions? [14:03] Which developer worked on the KVM provider? [14:05] mbruzek, a lot of people, myself included [14:05] mgz: just the doc "attached" to the package foo line (i.e. without a space before the package foo declaration) [14:05] dimitern: First of all thank you! It works wonderfuly. [14:05] mbruzek, thanks! (that wasn't what I expected frankly) :D [14:06] mgz: My godoc documentation (written in godoc) https://godoc.org/github.com/natefinch/godocgo [14:06] dimitern: I had a few questions about it. Can I bother you with them? [14:06] mbruzek, sure - shoot, and I'll try to help [14:07] natefinch: you coming? [14:07] dimitern: I ran out of disk on one machine. I don't see how I can add more disk in the set-constraints document. Is it possible to add more disk to the images? [14:07] natefinch: thanks! [14:08] dimitern: I have done speed testing between kvm, while it is slower than lxc it is not much slower. I am very happy because I can do more isolation with KVM. [14:08] mbruzek, let me check - IIRC constraints are supported, but most likely ultimately ignored by the provisioner [14:09] dimitern: Wait I found the devel document on kvm. https://jujucharms.com/docs/devel/config-KVM I actually see the answer in the docs. RTFM -> mbruzek! [14:10] The developer docs are not default so I had to do some searching for them. [14:10] mbruzek, I'm wrong apparently - root-disk=20G as a constraint should work [14:10] dimitern: it looks like root-disk is what I am looking for. Do you know what the default root-disk value is? [14:11] mbruzek, there are no defaults in the code, so I guess whatever uvt-kvm does by default (the tool we use to start kvm instances) [14:11] dimitern: I tried get-constraints but I get empty if I bootstrap a kvm environment. I would have hoped to see what the default constraints would be. [14:11] dimitern: OK fair enough. Thank you [14:13] mbruzek, well *bootstrapping* should in general have a 8G root-disk constraint unless given explicitly [14:14] mbruzek, but as the local provider is a bit special, YMMV [14:14] dimitern: My mileage has been great! I really needed this KVM provider for doing more container work. So thanks again! [14:15] mbruzek, glad to help and thanks! === natefinch is now known as natefinch-afk [14:29] dimitern, I don't suppose you have a recent overview of the details of agent upgrade? [14:31] fwereade, no, not really [14:38] Bug #1480942 opened: Need a `juju add-constraints' command [14:54] dimitern: I am about to go into a meeting, but http://reviews.vapour.ws/r/2274/ should be worth a look. [14:54] dimitern: I hope! === psivaa-lunch is now known as psivaa [15:00] dooferlad, sure, will have a look a bit later [15:11] dooferlad, reviewed! [15:14] jam, do you know if alexisb is around today? [15:16] I guess not [15:27] I'd expect her to be swapped from travelling still [16:03] dimitern: ping! hangout! [16:25] bbl, going to get my new glasses === natefinch-afk is now known as natefinch [18:16] https://bugs.launchpad.net/juju-core/+bug/1477281 does anyone have any idea of a workaround for this issue? it’s completely hosing every machine we have running juju [18:16] Bug #1477281: machine#0 jujud using ~100% cpu, slow to update units state

[18:17] moqq, what version of juju are you using? [18:18] 1.24.4-trusty-amd64 [18:18] and you are still seeing the issue? [18:18] moqq, ^^ [18:19] yes [18:19] moqq, we believed that issue had been fixed and needed someone to verify [18:19] moqq, can you please add a note to the bug [18:20] w/ logs if you have them [18:20] ok [18:20] we will get eyes on it asap [18:20] thanks moqq [18:22] HEY! i lied [18:22] alexisb: false alarm [18:23] i upgraded this morning and was still seeing the issue but manually restarted all the services and then i guess it just took awhile [18:23] but cpu usage eventually dropped back down. thank you thank you juju team!! [18:23] moqq, heh ok good, you had me scared [18:24] * perrito666 finishes reading backlog and breathes agai [18:24] n [18:25] moqq, can you still please add a note to the bug, so that we can work on getting it closed [18:25] i would add note [18:25] except when i try to log in [18:25] it says, “Bad bot, go away! Request aborted." [18:27] :| [19:49] bbl bike [20:25] >:( after a brief stint of relative quietness, juju is back to pegging the cpu [20:30] alexisb: ^ [21:32] moqq, do you have either logs or ability to give the dev team access to the env? [21:34] alexisb: doubtful on direct access but i can get some logs together shortly [21:34] moqq, that would be most helpful [22:05] hows it going everyone?? [22:05] core, devs, charmers: Is there a method by which juju can be forced to not overwrite changes to config files on node reboot?