[01:43] morning huwshimi [01:43] rick_h_: Hey [01:44] woot! crossed 2k views! https://www.youtube.com/user/celebrateubuntu [01:45] :) [01:46] rick_h_: The release seems to have gone well :) [01:46] huwshimi: yea, so far. Kind of more quiet than I expected, but we'll see. [01:46] 2k views but no comments and 5 thumbs up last time I looked heh [01:47] and no comments on any of the blog posts or reddit share [01:50] heh, ugh AVERAGE VIEW DURATION 1:42 [01:51] Makyo: has it beat AVERAGE VIEW DURATION 1:55 === uru_ is now known as urulama === fabrice_ is now known as fabrice|lunch [11:35] morning === fabrice|lunch is now known as fabrice [12:34] morning rick_h_ [12:41] morning lazyPower [13:08] https://bugs.launchpad.net/juju-gui/+bug/1375251 [13:08] Bug #1375251: Juju gui malforms display with missing x/y coordinates on existing env w/ charm upgrade [13:08] if there's any additional info you need feel free to ping me. This env is long-standing. [13:15] good morning jujugui. can i get reviews on https://github.com/juju/juju-gui/pull/589 and https://github.com/juju/juju-gui/pull/588, please? [13:17] jcsackett: taking a look [13:18] kadams54: thanks. [14:05] hatch: you're kill-mv branch bounced off the lander. :/ [14:05] s/you're/your/ [14:07] hatch: sauce is showing several failures that look valid, which is disturbing since it passed CI. [14:09] frankban: I've got the node entered and it's 'commissioning' [14:10] not sure how long this takes, will start to enter the second node and see how it goes [14:10] rick_h_: yeah I see, great [14:10] frankban: it mentions 'a properly configured dhcp server' can you figure out how to check if the maas install set up dhcp correctly or if there's something else to do? [14:12] rick_h_: there is a maas-dhcp-server upstart, but it says it's stopped [14:12] rick_h_: service maas-dhcp-server status [14:12] will take a look [14:12] jcsackett: yeah they are all 'real' failures [14:12] hatch: whoo. [14:12] hatch: this flag is a real bear. [14:13] I'm going to have to go through the entire test suite and fix every instance of env creation [14:13] hatch: you, me, and rick_h_ should talk, b/c the env/go.js updates have a similar issue. [14:14] well I would like to get this landed before any more go [14:14] it touches so many parts of the app [14:14] rick_h_: also note that "apt-get upgrade" shows lots of maas related packages ready to be upgraded, including maas maas-cli maas-cluster-controller maas-common [14:14] maas-dhcp maas-dns maas-region-controller maas-region-controller-min [14:15] not sure we want to do that [14:15] frankban: oh yea probably should. This install is a bit old [14:16] rick_h_, http://paste.ubuntu.com/8458814/ line 108, have you or anyone else run into this issue before? [14:16] hatch: i'm happy to coodinate landings and block on you getting yours done, but i would still like to chat if you've got a moment? [14:16] stokachu: lookihng [14:16] stokachu: yes! we got that in one of our upgrades. We had to manually remove the directory I think. [14:17] hatch: b/c i'm not entirely the best update for the env/go.js issue. :/ [14:17] stokachu: are you in a position to destroy/reploy the services? Is this on an upgrade? [14:17] s/i'm not entirely/i'm not entirely sure of/ [14:17] rick_h_, yea we could do that [14:18] stokachu: was this a charm upgrade? [14:18] just to help us narrow down and make sure it's something like what we've seen before? Or an initial deploy? [14:19] jcsackett: yeah sure just 2 mins, can't get the camera to work [14:19] hatch: dig; i'm chilling in the standup chat. [14:20] rick_h_, this is a fresh deploy [14:20] new bootstrap and everything [14:20] stokachu: oh, then no I've not seen it. [14:20] stokachu: /me digs into that log more [14:21] hmm, same error though...what kind of race can there be there ugh. [14:21] stokachu: will file a bug and start looking into it closer. Thanks for the traceback/log [14:22] rick_h_, thanks man really appreciate it [14:23] rick_h_, one thing that may help reproduce it is to run the deploy and let it sit for a few hours [14:24] and then issue another service deploy to trigger that config-changed hook [14:24] see if fails then [14:25] stokachu: so you've got juju deploy juju-gui and then another command later? [14:26] rick_h_, yea like a relation join that triggers this config-changed hook [14:26] stokachu: gotcha, ok thanks! [14:26] stokachu: have added notes to #1373875 and will update there when we can replicate/thing we have a fix. [14:26] rick_h_, awesome thank you [14:26] stokachu: did you pass any specific options to the GUI? [14:27] frankban, just a plain juju deploy [14:27] no constraints etc [14:28] rick_h_, that a private bug? [14:28] stokachu: what triggered the config-changed hook at line 108? [14:28] stokachu: yes, it was started by IS when we hit the one upgrade snag [14:29] rick_h_, gotcha [14:29] frankban, we think it was mysql doing a db-relation-changed against nova-cloud-controller [14:29] and seemed that juju-gui randomly errored then [14:32] frankban: sometimes config-changed is run by juju and I think there's a missing safety check there or something [14:32] yea i dont know the rules in which config-changed gets triggered but i seem to recall it gets trigger by a lot of events [14:32] related or not [14:33] rick_h_: I see, I think we should just manually remove the symlink before re-creating it [14:33] frankban: rgr [14:34] frankban: probably easiest/safest 'always will work' hack [14:34] rick_h_: so, do you mind if maintenance is temporarily super saturated? we have *many* branches in play for kill-mv-flag and i'd rather not group them into the one card as we're closing in on done. [14:37] jujugui: can i get one more review on https://github.com/juju/juju-gui/pull/589 and https://github.com/juju/juju-gui/pull/588 ? both are relatively small and QA is already done (thanks kadams54!) [14:38] jcsackett: rgr, ok. As long as we can get it cleaned out during the day today [14:41] rick_h_: dig. [14:48] hatch: is notifier on IE a legit, or is that one that fails randomly? https://saucelabs.com/jobs/7fb5041fe9c44ec68b8636e981f659cc [14:48] looking [14:49] that's spurious [14:49] notifier tests are async in a synchronous world :) [15:01] jujugui call nowish. [15:01] bqh [15:01] bah [15:01] whoops, be there in a jiff [15:15] jujugui anyone know if you can buy sim cards in the airport in brussels? [15:15] hatch: so I saw a thing that there's a phone rental store in the airport for it [15:15] hatch: dunno. i'm curious too [15:16] http://www.brusselsairport.be/passngr/shops/airport_terminal/shopping/5253/ [15:16] jujugui ^ [15:16] hatch: spain and france are horrible for getting SIM cards. :( [15:17] cool thanks [15:21] stepping out to grab a coffee bbiab === fabrice is now known as fabrice|family [15:50] guihelp: I need reviews for https://codereview.appspot.com/147360043 . anyone available? thanks! [15:51] frankban: do you know much about debugging upstart scripts? [15:51] ugh this branch is so frustrating [15:52] lol [15:52] rick_h_: I am not an upstart expert, but logs should go to /var/log/upstart/.log [15:52] frankban: yea, they're not which is the fun part [15:52] :-/ [15:52] rick_h_: which upstart job are you trying to debug? [15:52] frankban: the maas-dhcp-server is not starting [15:53] frankban: I found some cases of old ip addresses and worked past a couple of issues in the setup [15:53] frankban: and now stuck on that one, I've gotten it configured, and it should work from what I can tell, but no log output about why it's failing to start [15:53] frankban: and the upstart script will need to be picked apart a bit to manually run/process and try to figure out what it's unhappy about [16:01] * rick_h_ goes to make up some lunch [16:02] kadams54: still need someone to look into this? I'm sending off another test run [16:03] hatch: I'm spinning up an EC2 instance, I'll ping you once that's ready to go. [16:03] kadams54: wasn't the issue with sandbox? [16:04] Yes, but I'm comparing sandbox with a real env to see where things go wrong [16:04] Someplace fakebackend.js is failing to properly mimic the real env. [16:04] ohh I thought you had said that was already solved [16:09] rick_h_: basic check: is there a line in the /etc/init/maas-dhcp-server file that says "log" ? [16:10] rick_h_: also, you could add a line which says "output" and it should output to console [16:12] jrwren: I don't see anything that says log [16:12] rick_h_: add it to make upstart log output to /var/log/upstart/.log [16:18] rick_h_: mind if I hop on that search box bug? [16:18] hatch: not at all ty [16:18] hatch: seems like a state fall down in there [16:19] yeah this issue used to exist [16:19] hatch: did you and kadams54 sync up ok then? [16:19] we fixed it, then apparently it's back [16:19] heh wheee [16:19] rick_h_: yeah I guess he is going to test some ec2 stuff first [16:20] rick_h_, hatch: trying to get an EC2 env setup to compare real env vs. fakebackend.js and see where fakebackend.js is going off the rails. [16:20] ok cool [16:20] rick_h_, hatch: unfortunately attempt #2 to set things up using quickstart just failed with a utf8 decoding stacktrace [16:21] Any ideas? [16:21] don't use quickstart? [16:21] file a bug? [16:21] eat a sandwitch? [16:22] sandwich even [16:22] I'm wondering if EC2 is wonky right now due to the Xen stuff that's hit the fan. [16:23] rick_h_: it is possible that dhcpd is failing because no subnet is present in /etc/maas/dhcpd.conf ? [16:31] Makyo: the card you're working on is already done [16:31] ...well, bonus :D [16:32] yeah it's part of the commitStatus stuff [16:32] Is it coupled in with other work? [16:32] Okay. [16:32] it was required to do the uncommitted/waiting on ack/committed [16:32] stuff [16:32] but there may be some extra stuff that could be cleaned up I supopse [16:32] I'll take a look to be sure. [16:32] Otherwise, I'll just move it along. [16:33] excellent [16:33] hope I saved you a ton of work haha [16:33] Yeah, I'll certainly take it. === alexpilotti_ is now known as alexpilotti [16:39] rick_h_: dhcpd is now running [16:40] rick_h_: I restarted maas and now the conf file includes the 10.0.0.0 subnet [16:40] frankban: rgr, rebooting the node here to see if it picks it up now [16:40] frankban: woot! [16:40] frankban: we have network booting [16:41] rick_h_: really? yay! [16:41] frankban: yes, fetching archivese/etc [16:41] rick_h_: it was not obvious at all [16:42] frankban: yea :/ thanks for hacking on that [16:42] rick_h_: I'd assume maas would regenerate the conf file when saving the network configuration [16:42] rick_h_: I think a celery job failed to update the dhcp conf [16:43] rick_h_: nuc2 is ready [16:43] frankban: woot! [16:43] * rick_h_ power cycles nuc1 [16:46] rick_h_: they are both ready [16:46] frankban: woot! [16:46] * rick_h_ does a happy dance! [16:47] rick_h_: should we start them, or is it juju work? [16:47] ok, so now the question, can we stop then and start them and stop then [16:47] frankban: going to try to stop them [16:47] ok, that didn't work, guess ready doesn't mean running [16:47] so trying to start them [16:48] yeah [16:48] hmmm, no luck there either [16:49] rick_h_: what's the error? [16:49] frankban: The action "Start selected nodes" could not be performed on 2 nodes because their state does not allow that action. [16:49] frankban: had the same with stopping the node [16:50] rick_h_: have you seen the popup that appears when you mouse over the start button in a node detail page? [16:50] rick_h_: something about registering a ssh key [16:50] frankban: oh, nope didn't see that [16:51] rick_h_: but shoulds't be juju in charge of provisioning (i.e. starting) machines in maas? [16:51] frankban: maybe? outside of my docs reading to date on this now [16:52] rick_h_: I very quickly try to do that (juju bootstrap -e maas) [16:52] I can [16:52] frankban: rgr [16:52] rick_h_: ok launched [16:55] https://code.launchpad.net/~jtv/maas/bug-986185-start-node-requires-ssh-key/+merge/103536 at least found the message :) [16:55] rick_h_: did you allocated nuc1, or is it juju going? [16:56] frankban: didn't touch anything [16:56] rick_h_: cool, this can be a good sign [16:56] wahoo! [17:03] rick_h_: http://pastebin.ubuntu.com/8460003/ [17:03] frankban: ok, so ssh keys next [17:03] rick_h_: it seems so [17:04] https://maas.ubuntu.com/docs/nodes.html [17:04] frankban: so there's some ssh key notes there, but it seemed like part of the virutal machine docs [17:04] frankban: worth a quick try? [17:05] rick_h_: the popup says to add ssh keys from http://maas.jujugui.org/MAAS/account/prefs/ [17:06] frankban: oh missed that sorry. [17:06] too many convos at once [17:08] frankban: so reading https://maas.ubuntu.com/docs/juju-quick-start.html [17:09] frankban: it says "There is no need to explicitly add an SSH key to MAAS when using Juju; it will automatically put your public key on any hosts that it starts up." [17:09] frankban: that reads to me that the ssh key should match up with the juju one? can you add your public key on there and retry your test? [17:10] rick_h_: heh, reading here it seems that my keys are "useful" but not strictly required: https://juju.ubuntu.com/docs/config-maas.html#edit-or-create-the-configuration [17:10] rick_h_: but I can try [17:10] frankban: hmm ok, /me goes back to that paste to read that error more carefully [17:12] jujugui it appears that the BrowserSearchView class is not used anywhere in the app....can someone verify this? Just want to make sure my grepping didn't fail [17:13] rick_h_: retrying with a public key in the envs.yaml [17:13] frankban: k [17:14] hatch: at first look search.js and charmbrowser.js are not used [17:14] hatch: we did update that and such, but surprised we never got around to removing the old code [17:14] yeah that's what I was thinking - I just didn't want to cut it out without a second eye on the grep :) [17:14] hatch: yea, git grep says the two classes there aren't used [17:16] rick_h_: in that same page, one of the steps for juju is Adding an SSH key... [17:17] frankban: on which page, the juju one or the maas on? [17:17] rick_h_: perhaps the meaning of the sentence you quoted is something like "there is no need to add your (as a juju user) ssh keys in order to access the nodes via ssh" [17:17] rick_h_: https://maas.ubuntu.com/docs/juju-quick-start.html#adding-an-ssh-key [17:18] jujugui: still looking for reviews on https://github.com/juju/juju-gui/pull/590 and https://github.com/juju/juju-gui/pull/591. the test failure on the first of those is spurious--a successful build is linked to from the comments. [17:19] frankban: k yea. I think we can try this out with your ssh public key side and see if it works. If it does, we should probably add a shared one for the team or something. Maybe it'll make sense to add a user account for each person on the team with their own public key [17:19] jujugui for those going to brussels https://www.youtube.com/watch?v=2fw7ls-53RM :) [17:19] jcsackett: I'll look at those as well. [17:19] kadams54: awesome, thanks. :) [17:20] hatch: you mean I've been learning French for nothing?!? [17:20] hatch: dutch? [17:21] apparently it's bilingual [17:21] it's like 80% french in brussels. [17:21] Je suis une chaussure rouge [17:21] oh really? [17:21] See?!? [17:21] so "parlez-vous anglais?" is a bit more useful. :p [17:21] well damn you interent [17:21] I'm all set. [17:22] kadams54: you're a red shoe? [17:22] lol [17:22] well if it's mostly french then yay [17:22] I know infinitely more French than Dutch lol [17:22] Damn Canadians. [17:22] eh, there's still 20% dutch, hatch. :p [17:23] well if they don't speak english then I'm sol no matter what language they choose [17:23] so...just making sure I can switch the language program to english :P [17:23] Hablas español? [17:24] un poco [17:24] i wish i spoke spanish; would be very handy here. [17:25] kadams54: a tumar classes en Duolingo [17:25] rick_h_: added a key to maas, and trying to bootstrap juju with another one [17:25] frankban: rgr [17:25] Yeah, I'm hitting up Duolingo for the French right now. [17:25] which is probably wrong - I suck at the male/female parts of spanish [17:26] why can't languages be single sex [17:27] jujugui quick clean-up branch for reviews/qa: https://github.com/juju/juju-gui/pull/592 [17:28] Makyo: on it [17:28] rick_h_: what's the series running in the nucs? [17:28] frankban: I set it to trusty [17:28] frankban: in the maas ui [17:28] rick_h_: cool [17:28] frankban: http://maas.jujugui.org/MAAS/nodes/node-149f5df2-47e2-11e4-9745-eca86bffcfed/edit/ [17:28] I was confused as it seems like it should be juju determining that I guess, but :/ [17:28] Makyo: i'm looking too. [17:29] Makyo: already gave a +1 :P [17:29] Makyo: nm, i'll let hatch do #2 then. :p [17:32] Makyo: commented on the pr - lemme know your thoughts [17:33] Makyo: actually....a machine is an object not a model....bleh [17:33] I'll delete that comment :) [17:34] well it could go on the model list [17:34] I suppose [17:34] hatch: it's not ideal but it's grep-able and usable and better than 'new' [17:34] machineList.isNew(model) [17:34] hatch, yeah, though this method of always using .commitStatus gives us something to search for, as opposed to indexOf or regex on 'new' [17:34] Bleck, I don't like it. [17:34] oh no question it's better than regex on 'new' heh [17:35] alright, I'm ok with leaving it as is [17:37] rick_h_: last attempt, same key.pub on the client and on the server [17:37] frankban: ok [17:39] the latest installment in the saga that is Installing Fiber - they are now trenching my yard to run the conduit to the house [17:40] this fiber better be a religious experience after all this headache lol [17:42] nnvffhgrbjvniidukdfhedbvbcnlttejvb [17:43] Sorry :P [17:43] yeah I agree!! [17:43] :P [17:54] I'm having no luck in tracking down what handles the autocomplete token clicks [17:54] hatch: bwuahahahaha [17:54] :P [17:54] hatch: it's the same event that catching clicks on the tokens in the sidebar [17:54] lol hey it's all your fault - you and your events :P [17:54] hatch: autocomplete tokens are...token [17:55] tokens [17:55] if I recall correctly [17:55] that's what I thought too - but the handler for those aren't being called when I click on the category tokens [17:55] * rick_h_ dusts off the old memory [17:56] hatch: looking [17:57] in charmbrowser.js there is a .token click hanlder [17:57] that handler doesn't get called when you click on the category tokens [17:57] hatch: oh, but it's different because the AC widget has a selection made [17:58] ohhhhhhh [17:58] rick_h_: now I have to go. I'll try again later, I see that one of the steps in the docs is juju sync-tools, executed it, but still no luck. this is the failing (repeated) command: http://pastebin.ubuntu.com/8460289/ [17:58] good night! [17:58] frankban: ok, maybe tomorrow seeif you can catch any of the #maas guys to help out if possible [17:58] rick_h_: thanks you got me going again :) [17:58] frankban: if not, we'll look into it more tomorrow [17:58] cya frankban [17:59] hatch: this.ac._onItemClick = function(ev) { [17:59] rick_h_: I'll do [17:59] yup found it now thanks [17:59] totally forgot that categories were actually a list item [18:01] guihelp: Working sans quickstart… juju deploy --to 0 juju-gui && juju expose juju-gui [18:01] kadams54: correct [18:01] juju status reports: public-address: ec2-54-69-157-181.us-west-2.compute.amazonaws.com [18:01] kadams54: and the services is listed as 'started'? [18:01] But I'm not seeing anything when I pull that up in a browser [18:01] kadams54: or still 'pending' [18:02] kadams54: https://ec2-54-69-157-181.us-west-2.compute.amazonaws.com/ [18:02] kadams54: works for me [18:02] Yeah… I just realized juju status was reporting exposed as false [18:02] So I re-ran and that fixed. *sigh* [18:02] gotcha [18:04] :) [18:04] kadams54: also if you change the juju-gui-source it'll take a while but the status reports will always be green [18:04] so you will just have to keep trying every few mins [18:04] Yeah, I watch the debug log for that [18:06] jujugui off to the doc, biab [18:10] I need some help with $(shell...) in make. The Makefile has APT_BASED := $(shell command -v apt-get ; echo $$?) and later compares to zero, but this is always false because of the output from command -v [18:10] so I change to APT_BASED := $(shell command -v apt-get >/dev/null ; echo $$?) and make just hangs [18:12] so I tried APT_BASED := $(shell { command -v apt-get ; echo $$?; }|tail -1) and that also just hangs. [18:13] yet APT_BASED := $(shell { command -v apt-get ; echo $$?; }|head -1) puts the value I expect into the APT_BASED variable (just not the value I want) [18:13] hrm. I should have asked before rick_h_ left. I know how much he loves make. [18:14] haha yup [18:29] hatch: Ready when you are to chat on this card. [18:31] guihelp: how are the various delta handler methods in handlers.js invoked? [18:31] kadams54: ok just eating lunch [18:31] so in a few [18:32] kadams54: from the delta [18:32] it's all synchronous so you should be able to traceback [18:32] look at the parse_delta methods [18:32] there are many [18:32] The deltas come back via WS? [18:33] find the issue. it was not where I thought. Make kind of sucks somtimes. [18:33] correct [18:33] jrwren: only sometimes? :) [18:39] hatch: :) more often than not, I am a make fan. [18:39] hatch: I'm even crazy enough to tell the grunt/gulp crowd that make does all that, better. === uru_ is now known as urulama [18:57] jrwren: well since those are typically just a wrapper over shell commands :) [18:57] but make is sure ugly.....like real ugly [18:58] kadams54: ok still need to chat? [18:58] jcsackett: holy crap that branch landed [18:59] jcsackett: https://github.com/juju/juju-gui/pull/587/files here is the final diff if you're curious [18:59] hatch: whoo! [18:59] hatch: just one more hell-branch to go. :p [19:00] hatch: that's a much smaller diff than i expected. nice. [19:02] yeah I just hope it doesn't block other peoples branches from landing with some wako new spurious falure [19:02] so I ran it a few times before landing [19:02] just to be safe [19:11] hmm this search bug is a little bugger to fix [19:12] technically it's working as it should.... [19:15] ahh stupid #$%^&(*^(*& javascript [19:20] kadams54: just ping if you still need to chat [19:58] jujugui lf a review and qa on https://github.com/juju/juju-gui/pull/593 [20:07] hatch: looking. [20:07] thanky [20:08] now which card to pick [20:08] hmmmmm [20:15] hatch: notifier [20:16] hatch: or the inspector cleanup of old code [20:16] oh that damn notifier [20:16] sure I'm up for a challenge [20:18] hatch: yea, those are the two big things for this cleanup week imo [20:18] hatch: especially as the new work on the sidebar will be all around the inspector code [20:18] jcsackett: if you don't mind https://github.com/juju/juju-gui/pull/594 here is one just removing old unused search code [20:19] rick_h_: ahh right [20:19] rick_h_: mind taking a quick glance at my branches in PR so I can land em? [20:19] they are both small [20:19] hatch: in a mtg atm but can later tonight [20:19] alrighty [20:20] hatch: i can look at that one too, sure. [20:21] thx [20:21] rick_h_: is there a notifier card? Or am I blind? [20:21] hatch: not yet no [20:22] alright I'll create an investigate card [20:23] hatch: peek at it, get a plan, and let's pow wow later [20:23] a o k [20:56] brb taking dogs out [21:14] back [21:25] jrwren: make huge bank from the advert on your blog? [21:26] and your about page is incorrect :) [21:31] jujugui: https://github.com/juju/juju-gui/pull/595 is ready for review and QA. [21:40] kadams54: made a request to review [21:40] in the * [21:54] rick_h_: you still pokin around? [21:54] hatch: kind of what's up? [21:54] just wondering if you wanted to chat about this notifications business [21:54] hatch: cool if you're ready, let's do it fast [21:55] standup room? [21:55] yup [22:54] Makyo: didn't you fix this? https://bugs.launchpad.net/juju-gui/+bug/1365260 [22:54] Bug #1365260: deployer bar commit button stays active after confirm