/srv/irclogs.ubuntu.com/2014/05/21/#juju-gui.txt

hatchhuwshimi thanks for the QA, are you doing a review too?00:04
huwshimihatch: I'm not sure I know enough to give you a good review00:06
hatchalright no problem00:06
hatchthanks00:06
huwshimihatch: I mean, there's some red bits and some green bits and the green bits could be a little more green. Does that help?00:07
hatchlol00:07
rick_h_huwshimi: hey, side note and heads up. We'll be having the weekly AU call this week. And we'll be doing our 1-1 to go over annual review stuff00:11
huwshimirick_h_: Sure!00:11
rick_h_huwshimi: I know we've fallen a bit off the wagon on those and I want to get back on track after the demo/vegas and such00:11
bachey huwshimi00:11
huwshimirick_h_: It's all good :)00:12
bacrick_h_: i'm still working to get the charmworld jenkins working again.  lots of fiddly bits.  i've also requested production be completely restarted to see if that'll solve the unknown issue.00:12
rick_h_bac: ok, thanks for the heads up. If there's issues with the charmworld jenkins and we can't recover we can look at just taking it over and moving it into our CI. Of course that would probably mean a move to git as well, but wtf. 00:13
rick_h_kind of bummed to have lost a day on it now. 00:14
rick_h_bac: do we know what the issue is on the jenkins side?00:14
bacrick_h_: it should be recoverable, i've just never launched it from scratch before00:14
rick_h_bac: ah, ok. So you did have to rebuild the env from scratch?00:14
bacrick_h_: yes.  state server went awol so i had to tear down the env00:14
rick_h_:(00:14
rick_h_bac: when we do get it back up let me know. I recover from issues on the gui side by backing up all of /var/lib/jenkins and wonder if there's any backusp of that in the charmworld CI that would help shorten things up. 00:15
huwshimirick_h_: Should cancel in the deployer bar really clear all your uncommitted changes or is it just cancelling the deploy?00:17
huwshimirick_h_: It seems to me that there should be two actions. 'Cancel deploy' and 'clear uncommitted changes'00:18
bacrick_h_: jenkins seems to be mostly functional now.  looks like the migration may have introduced a spurious test failure.00:19
rick_h_huwshimi: hmm, not that cancel. There's supposed to be a cancel on the actual bar while you're working I thought00:25
rick_h_huwshimi: right, from the summary page, when you hit deploy, I would expect cancel to dump me back to my environment with uncommitted work still there00:26
rick_h_huwshimi: /me goes to look at designs00:26
huwshimirick_h_: https://drive.google.com/a/canonical.com/?usp=folder#folders/0B7XG_QBXNwY1WGlyVzN2LWtrcmM00:26
huwshimirick_h_: I can't see anything there.00:26
rick_h_bac: :/ yay, it's running but now tests don't pass? 00:26
rick_h_huwshimi: sorry, it was in some mockup I was using to make the cards I guess but isn't there now00:27
rick_h_huwshimi: maybe shoot luca an email just to make sure and put this card under 'needs specification' in the board00:27
huwshimirick_h_: Yeah, I thought there was a clear button as well.00:27
bacrick_h_: yes.  i'm pushing a branch to disable the new test.00:27
huwshimirick_h_: OK, will do.00:27
rick_h_huwshimi: cool thanks00:28
rick_h_bac: ok, and then create a card/bug to look into that new test?00:28
bacyes00:28
rick_h_bac: sounds like a good plan, thanks for dealing with the pita00:28
bacumm, pita00:29
rick_h_pain in the booty00:30
rick_h_:)00:30
rick_h_ci issues always strike me as a pita. it's there to just run and run and take a beating and not spend a bunch of time on. I guess everything breaks though. 00:30
huwshimirick_h_: I can't reproduce this card: "with il flag: deploy a service, click on a unit from the inspector for the unit details, the charm results shows up." 00:39
rick_h_huwshimi: ok, I think it's part of the work hatch is doing anyway00:39
rick_h_huwshimi: just remove that card then please00:40
huwshimirick_h_: The 'Upgrade service' links do open the charm details, but I'm pretty sure they're supposed to.00:40
rick_h_huwshimi: I think it came up during out mad dash around demo material and we must have cleaned it up some00:40
rick_h_huwshimi: right, it was in the sidebar that the charms shows up00:40
huwshimiAh I see00:40
rick_h_huwshimi: so you'd be in the inspector, click on a unit, and then the sidebar would turn into charm results vs inspector00:40
huwshimiYeah, that's not happening anymore00:40
rick_h_huwshimi: ok, cool then please remove the card00:41
huwshimiGreat!00:41
rick_h_huwshimi: have you done vegas expenses?00:54
huwshimirick_h_: Oh, no I haven't I'll do that now.00:54
rick_h_huwshimi: thanks00:54
rick_h_phew, almost done for tonight. One more task. 00:55
hatchto review my branch???? :P00:56
rick_h_hatch: ugh, sorry no. I'm going to pass tonight. Brain is about done 00:56
* rick_h_ checks morning schedule00:56
rick_h_hatch: I should be able to review it before you start tomorrow00:57
hatch:) np have a good night00:57
rick_h_hatch: assign it my way or something so I don't forget00:57
rick_h_man, I had my email under control this morning I swear00:58
rick_h_Makyo: any chance you're still around? Did you check out Menno's notes in the email?01:01
rick_h_huwshimi: approived01:24
rick_h_approved that is01:24
huwshimirick_h_: Thanks :)01:29
rick_h_huwshimi: thanks, amke sure to move the kanban card for that please01:30
huwshimiah yes01:36
hatchanyone have any book recommendations from oreilly? I have a free book and not sure what to get03:22
hatchhey kadams54  any book recommendations from oreilly? 03:30
kadams54Seems like I should have since I stopped by their booth and talked awhile at Pycon, but right now I'm drawing a blank… why?03:31
hatchI have a free book from gophercon03:32
hatchI'm trying to decide between Python, Clojure, or something-else-totally-not-related03:33
hatchlike R03:33
hatchcould maybe learn some webgl too03:33
hatchso many options03:33
hatchso little time03:33
hatchlearning python....1600 pages.....lol03:41
rick_h_morning frankban 11:19
frankbanhi rick_h_ 11:20
rick_h_frankban: talked with Ian yesterday and he's offered us any help we need in our refactor/pull apart modules work. 11:20
frankbanrick_h_: great11:20
rick_h_frankban: ian/wallyworld 11:20
rick_h_so I feel a nice bit better about that side of things. 11:20
rick_h_and sounds like work is already started as dave pulled out the 'thirdparty' stuff I think last night11:21
frankbanrick_h_: cool11:26
frankbanuhm freenode keeps disconnecting me11:36
frankbanrepeating my last message, not sure if I was disconnected11:37
frankbanrick_h_, redir: I wrote a simple bash script which can help migrating subfolders while juju-core is still bzr: http://pastebin.ubuntu.com/7496993/11:37
rick_h_frankban: very cool11:46
redirfrankban_: awesome12:06
redirfrankban_: I think that 73 can just be git filter-branch --subdirectory-filter store -- --all12:06
redirwhere store is the folder you want to promote12:07
redirso git filter-branch --subdirectory-filter $folder -- --all12:07
frankban_redir: cool thanks12:28
frankban_rick_h_: so, fo we want to preserve the history even for packages that we need to add to existing git project? e.g. juju-core/testing/filetesting -> github/juju/testing/filetetsting12:30
frankban_s/fo/do12:30
rick_h_frankban_: I think so if we can12:33
rick_h_frankban_: we can check with the core team as they've been pulling some other deps out how/what they've been doing for it12:33
rick_h_there's the conversation around cmd lately for instance12:33
frankban_rick_h_: I think we can: http://stackoverflow.com/questions/1425892/how-do-you-merge-two-git-repositories12:34
rick_h_frankban_: cool yea I think it's valuable to be able to go back and git blame/etc a file down the road12:35
frankban_rick_h_: +112:35
redirI haven't done it but I know you can. Which is how a repo can have to parent-less commits12:36
redirs/to/two 12:36
redirsorry more than one parent12:37
* redir drinks more coffee12:37
redirno wait yes two parentless (initial) commits is a repo that is a merge of projects12:38
rick_h_coffee sounds good, /me goes and gets some12:43
redirbrb12:49
bacredir, rick_h_: jenkins lander successfully tested and landed a branch.  woo.  qa.mjc is now on r51413:00
bacafter doing some qa on it i'll do a deploy to staging in a little while13:01
bacneed to step afk for a bit.  bbiab.13:02
rediruh, I'll ask abou the failing ES thing when you're back13:06
rick_h_bac: woot13:07
redirrick_h_: frankban_ fwiw: https://github.com/reedobrien/juju-core a core imort with store removed...13:12
frankban_redir: cool! what command did you use?13:15
redirfrankban_: should we share a doc or a repo or something?13:17
redirhttp://paste.ubuntu.com/7497303/13:17
frankban_redir: maybe so. I guess when we remove code from juju-core we don't want to rewrite its history13:20
redirfrankban_: maybe. I don't know. 13:21
redirI mean  probably not repeatedly13:21
frankban_redir: yeah, while I see a value in not having all the unrelated history when we create a new project from a package in core, I guess in juju-core we still want the rm change to be there13:26
rediryou could just `git rm store`13:26
rick_h_yea, I think you want all the history and such because you want to be able to git co 1.0 13:26
rick_h_and have it all there/happy13:26
rediroh tags13:26
redirI didn't push the tags13:27
rick_h_I don't see us as rewriting any history in core13:27
redirfine by me13:27
rick_h_what's the bonus or advantage? /me is still getting through coffee and not seeing the forest13:27
redirI don't know that there is. 13:28
redirless cruft13:28
rick_h_ok, then yea let's not worry about that13:28
rick_h_redir: bac there are two cards in landing from yesterday. Are those up to date? Did they land or still waiting on something?13:29
rick_h_now that the charmworld lander is back up and running13:29
redirI am not fighting for it... just if I migrate something out, I would migrate it out13:29
rick_h_redir: yea I don't know it's migrating but moving13:30
rick_h_and moving doens't write history 13:30
redirwhen bac is bac I'll ask about that 027 card13:30
rick_h_ok cool13:30
redirrick_h_: true but histlry in store and history in core store may not line up anymore13:30
rick_h_what about the landing card about the 'migration for ngrams index change'13:30
redirwhich history is right?13:30
redirrick_h_: I think that is held up by 02713:31
rick_h_redir: ah, gotcha13:31
redirit may have landed but with that test skipped13:31
* rick_h_ follows links in card to investigate13:32
rick_h_yep, merged will move card13:32
frankban_rick_h_, redir: I just completed an investigation about migrating juju-core/testing/filetesting: http://pastebin.ubuntu.com/7497346/  What do you think?13:38
rick_h_frankban_: that looks good/sound to me. I think we should add that as a doc in the juju-core code on the first one of these we do13:40
rick_h_frankban_: and then we need to make sure we work with either sinzui or ourselves to get the new package into CI so that we run tests on branches/pull requests of that package13:40
bacredir: bac is back13:41
frankban_rick_h_: +1 on adding a doc to juju-core when we have a complete migration document (not just an investigation)13:42
frankban_rick_h_: could you please review the sanity of the vcs commands? especially git ones13:42
rick_h_frankban_: so I think the push would need to be git push frankban add-filetesting:add-filetesting13:45
rick_h_frankban_: well I'd so it that way at least. I'd git co -b add-filetesting13:46
rick_h_do the merge13:46
rick_h_and then push it to your fork13:46
rick_h_and then pull request that back to the origin juju/testing13:46
frankban_rick_h_: git push worked for me and pushed the changes to frankban/testing. yeah, in the real migration we want to then make a pull request and wait for the branch to land13:48
rick_h_frankban_: right, I think because you created the branch as a remote branch by adding the frankban/master to the checkout -b command13:49
rick_h_frankban_: I'm just stating in the more typical workflow you'd have it local, push it to your fork, and then pull request it back to the origin13:49
frankban_rick_h_: ack13:50
frankban_rick_h_: I'd ping Ian too for a review of that document if you agree13:50
rick_h_frankban_: sounds good to me. 13:51
frankban_rick_h_: cool thanks13:51
rick_h_frankban_: just make sure we ack that we'll have to adjust that to git-based soon13:51
rick_h_as he wanted to make that point clear, I think he'd prefer we didn't move forward until that work was done, but I don't think we can handle the delay 13:51
rick_h_frankban_: also make sure to include sinzui13:52
rick_h_frankban_: I think there was an issue where a moved dep broke older versions of juju from running because they referred to a dep in the wrong place or something13:53
rick_h_I don't think we'll hit this as we're moving from within core to outside, and not two different places outside core13:53
rick_h_but I'd like to make sure13:53
frankban_rick_h_: yes, we must pay particular attention when we propose the juju-core change13:54
redirbac: how can I reproduce that fail?13:56
redirlost scroll I thought it was just really quiet in here.13:57
redir:)13:57
bacredir: well, jenkins spins up a fresh system.  perhaps delete everything out of ES and then run 'make test'13:57
redirbac also a fresh mongo?13:59
bacredir: well it was specifically complaining about ES.  but, sure.14:00
redirbac passes14:07
bacredir: did you see in the output where a lot of the ES values were _na_?14:09
rediryeah14:09
bacodd, no?14:09
redirbac WAG at not being in a green state14:10
redirmaybe we can try waiting for that14:10
bacwe've been down that road...14:11
bacredir: try running via: bin/test -v -s -x --with-id  charmworld14:11
bacthat's what jenkins does.  maybe it runs in a different order14:11
bacredir: i'm going to proceed with a rollout to staging.  having r514 on qa.mjc does not show constant server restarts but staging and prod do.14:16
frankban_rick_h_, redir: email sent14:16
rick_h_frankban_: ty14:16
bacredir: if you want to follow along, the guide is at https://docs.google.com/a/canonical.com/document/d/1G6IoLyDz3VSw7lMRL7QUbaKKXropxyVm3uOFYsqKSeU/edit14:17
redirfrankban_: late now, but the paste LGTM...14:17
hatchI decided last night to go with Introducing Python http://shop.oreilly.com/product/0636920028659.do14:18
frankban_redir: thank you14:18
rick_h_hah, we'll make a python dev out of hatch yet14:18
hatchrick_h_ haha, well I'm interested in machine learning stuff and it seems like people use Python a lot for it14:19
hatch(for some reason)14:19
redirbac jenkins flags pass too14:19
rick_h_hatch: +114:19
redirhatch: NLTK FTW14:19
rick_h_redir: heh, using that for our suggested tags stuff in bookie14:19
hatchI'll have to check that out14:19
redir:) 14:20
hatchR is also very popular for it but I figured Python would do better for my career :)14:20
jcsacketthatch: yeah, R is not really a growth language. :p14:20
jcsackettoutside of actual labs, anyway.14:21
hatchclosure was also on the list....but maybe after python & go14:21
jcsackettclojure is neat14:22
hatchThe book hasn't arrived so I'm finally getting to reading Functional Javascript 14:22
hatchI'll let y'all know how it is in a couple days14:23
rick_h_cool14:23
redirbac reading14:24
bacredir, i created the RT then went to #webops on the canonical IRC and called it to their attention14:25
bacredir: for your bug, the next thing i can think to do would be to spin  up a fresh container and see if it fails there14:26
redirbac hal14:26
redirbac on 12.04 right?14:26
bacyeah14:26
redirwill do14:27
bachal?14:27
redirbot in #webops14:27
bacoh, yes, i hadn't paid attention to its name. poor hal, toiling unappreciated.14:36
rick_h_guihelp tiny review please https://github.com/juju/jenkins-github-lander/pull/1114:38
kadams54Looking…14:38
hatchlgtm14:39
hatchnice the lander has tests14:39
hatch:)14:39
redirbac should be completed in 1-2 weeks14:39
bacredir: the lxc?14:39
rick_h_hatch: :)14:39
redirbac the ticket14:47
redirbac where can I see what the lander "does"14:48
rick_h_jujugui fyi, updated the lander with that fix. The next person to try to submit a pr and merge/landing please keep an eye out for strange things14:50
redirjujugui call in 1014:50
rick_h_monsters and errors and the like14:50
bacredir: bzr+ssh://chinstrap.canonical.com/~abentley/jenkins-lander14:52
bacredir: i've pulled the good parts here: https://pastebin.canonical.com/110543/14:54
redirgreat b/c I get permission denied..14:54
rick_h_redir: to what? I'll work out the missing team/etc memberships14:54
rick_h_oh, chinstrap, not sure about that one14:55
redirnp just trying to reproduce the 027 fail in a container...14:57
redirlike the lander does14:57
rick_h_jujugui call in 114:59
bacredir: you should get access to chinstrap14:59
rick_h_redir: ^15:00
redirlooks like I have it not on the container15:01
redirbut on my actual machine. bac rick_h_ 15:01
rick_h_redir: call now?15:01
redirdoho15:02
bacredir, rick_h_: we should talk about how much effort to put into this test fix.  the migration has worked on prodstack so spending time on that test may not make great sense.15:08
rick_h_bac: rgr15:08
redirbac #192 Putting new mapping without creating the filter and analyzer ... ok15:14
redirin precise container15:15
redirI love the standups. They are so efficient15:16
redirand fast15:16
rick_h_redir: bac so is that ok? let's timebox the test issue to today and if we can't dupe or get anywhere we take another look at the value of the test and move on?15:18
rick_h_redir: heh, we try to suck everyone's time away 15:18
bacrick_h_: so do we wait on this test to be fixed before deploying to production?15:19
redirrick_h_: bac if I had to guess:) Because we like WAGs15:19
rick_h_bac: no, I think if we've got this working in practice and on the prodstack staging we can call this a follow up card?15:19
bacrick_h_: b/c after deploying the migration will have been done and the test has no value.15:19
redirrick_h_: bac: I would guess there is an environmentdifference where that failure occurs15:19
rick_h_redir: yes, which we should be aware of for sure. However if that env difference doesn't effect prodstack deployments it's an env issue we can solve out of band 15:20
rick_h_imo and all that 15:20
bacrick_h_: my biggest issue for now is the constant server restarting on production and staging, but not QA.  staging and qa have the same code base.15:21
rick_h_bac: right, so we'll see if a deploy catching something? 15:22
* redir has an idea15:22
rick_h_bac: shall we jump back in the standup hangout?15:22
bacrick_h_: i didn't understand what you wrote15:22
bacrick_h_: ok15:22
rick_h_redir: ^15:22
frankban_rick_h_: I forgot to add a quickstart card to the backlog. It might be interesting to have some integration tests for quickstart. This way, having the suite included in the juju-core CI, we will be able to actually test quickstart on live environments using both current and development versions of juju. It could be a nice idea to include this work in the future development (midterm?)15:26
hatchhas anyone used the Kobo aura hd? http://www.kobo.com/koboaurahd15:30
rick_h_I have a friend that had a kobo for a few years and they angered him and he has left them15:32
hatcheveryone I know has kobos and has sold their kindles for them, but none have the aura hd15:33
hatchwhat did they do to anger him?15:33
hatchbad hardware?15:33
rick_h_they had some software bugs and they at first tried sending different devices and after still having issues basically told him to $$$$ off15:34
hatchahh - that's no good15:34
hatchI dislike reading on a tablet but I have so many ebooks now I need some way to read them that doesn't suck15:34
hatchand so far I've not been impressed with the eink hardware15:35
hatchbut this aura looks fancy15:35
hatchit looks like amazon has dropped eink? https://kindle.amazon.com/ only a big image for the tablet15:36
rick_h_heh, paperwhite is the best kindle15:36
hatchspec wise the aura beats anythign else on the market it looks like15:38
hatchquite a bit heavier than the paperwhite though15:38
hatchbut bigger and higher resolution screen on the aura15:39
hatchheh15:39
rick_h_<3 my paperwhite though for tech books the size isn't the best15:39
rick_h_I mainly use the paperwhite for fiction/wordy reading vs tech code and the like15:39
hatchoh kindle doesn't do epub?15:41
hatchwell that's a deal breaker right there hah15:41
rick_h_no :(15:41
hatchall of my books are epubs15:41
Makyojujugui quick review/QA https://github.com/juju/juju-gui/pull/33715:45
redirall your books are belong to us15:45
hatchMakyo no qa instructions? *sadface*15:45
hatchredir no DRM here!15:45
Makyohatch, Oops, sorry, forgot that.  ONe sec, will add a comment.15:45
kadams54jujugui Anyone know how the ID on the ServiceUnit model gets set?15:46
Makyohatch, added, see https://bugs.launchpad.net/juju-gui/+bug/129245415:47
_mup_Bug #1292454: Subordinate relationship lines show green until moved on the canvas <juju-gui:Triaged> <https://launchpad.net/bugs/1292454>15:47
redirbac should that 207 fail be attached to a bug or anythign15:48
redir>15:48
* Makyo makes another coffee, since the first one didn't do the job15:48
redir?15:48
hatchkadams54 when it's added to the modellist15:48
* redir need another pot of coffee15:49
rediror tea15:49
frankban_kadams54: it gets set when the mega-watcher data arrives, in handlers.js15:49
bacredir: 207?15:49
redirbac sorry 027 migration fail15:49
bacredir: there is a card for it.  sure be sufficient15:50
bacs/sure/should/15:50
hatchyay no more renderTo15:50
bacredir: bugs are nice but i think we agreed they are only required if it is a longer-term issue or if it won't be started pretty much immediately15:50
kadams54hatch: I didn't see any code (specifically in _setDefaultsAndCalculatedValues) that would set the ID when adding to the ModelList. frankban: in the unitInfo method (line 183)?15:52
bacrick_h_: the charmworld charm start hook just calls charmworld/scripts/run which uses gunicorn_paster -D -- does that automatically restart workers?15:52
hatchkadams54 yes 183 is correct15:52
hatchsee the db.units.process_delta() method15:52
hatchfor how it processes the delta15:53
kadams54Got it, thanks…15:53
hatchbut you see in the unitInfo method that it assigns the value from change.Name to id15:53
rick_h_bac: looking, the -D it just daemonize. I don't think it does any auto restarting15:54
redirbac cool, just wanted to make sure before I commited15:55
redirpush is hanging...15:55
rick_h_bac: ok, so the scripts/run in charmworld has a max-requests of 25015:56
rick_h_bac: so ever 250 requests the worker is restarted15:57
rick_h_bac: and there's supposed to be 3 workers running at any given time15:57
bacrick_h_: yep15:57
rick_h_bac: but this restarting thing seemed bigger than just gunicorn doing worker management?15:57
rick_h_bac: I guess we can test it by watching top for hte worker processes and hitting the site with 1k requests and seeing if we can force the issue15:58
bacrick_h_: it is registering a new init every 2-3 minutes15:58
rick_h_bac: is it exact or just roughtly? what is a 'new init'?15:58
rick_h_bac: maybe we're getting hammered from outside traffic? Seems odd though. Ingest maybe? hitting for proof?15:59
bacrick_h_: charmworld init puts the server start time into mongo.  by looking at /heartbeat you can see the last start time.  it seems to be every couple of minutes.  any new worker would overwrite that time15:59
bacrick_h_: but i'm not just chasing this academically.  we have seen performance issues.16:00
rick_h_bac: ok, so I'd be curious to look at the apache request logs in front of the service?16:00
rediris bazaar.launchpad.com:22 not working? 16:03
redircan't push or pull16:03
rick_h_redir: testing16:03
rick_h_redir: bzr branch lp:charmworld just worked for me16:03
redir-v does nothing16:05
redirmmm16:05
bacredir: what cmd are you using?16:06
redirpretty complicated one16:06
redirbzr push16:06
baccan you try using the lp: addressing?16:07
rick_h_redir: can you just bzr branch lp:charmworld? 16:07
rick_h_into a tmp dir?16:07
redirsays: Using saved push location: lp:~reedobrien/charmworld/027test-fails16:08
redirand hangs16:08
redirwhup16:09
redirhangon16:10
rick_h_bzr: ERROR: Not a branch: "bzr+ssh://bazaar.launchpad.net/~reedobrien/charmworld/027test-fails/"16:10
rick_h_when I try to pull it down16:10
redirrick_h_: was just pushing for the first time16:10
rick_h_redir: ah, gotcha16:10
redirssh controlsocket issue16:11
redirjust pushed again16:11
redircontrolmaster socket16:11
* rick_h_ lunches16:17
redirbac: http://bit.ly/1nf0yAo16:17
redir^merge request16:18
rediror should I just do that myself16:18
hatchrick_h_ are you ok with me leaving caching out of this conversion since it's to be re-written this cycle?16:21
redirthanks for testing that rick_h_16:21
redirbac: self approved...16:23
hatchjujugui has anyone run into an issue where the linter thinks something is undocumented when it is documented?16:29
kadams54nope16:32
rick_h_hatch: yes, I think so. There are cards to revisit it and I think pulling it is legit for this brfanch16:32
hatchrick_h_ ok thx - now I just need to figure out why the heck I can't get lint to pass16:33
rick_h_hatch: if it's not obvious throw up the pull request and reviewer comment the issue and can help look/get eyeballs on it16:34
hatchwhats the make command for generating the undocumented list?16:35
hatchahh undocumented16:36
redirbac wait for green status FTW16:38
hatchrick_h_ https://github.com/juju/juju-gui/pull/338/files no tests yet but 100% functional 16:40
hatchwell....80% (no cache)16:40
rick_h_hatch: cool, will look in a few min16:40
hatchsure, no rush, I'm going to review/qa Makyo's branch now16:40
rick_h_hatch: existing tests pass?16:41
hatchonly because I haven't removed the old code16:41
hatchthey will need to be entirely re-written16:41
hatchwell....ported I guess16:41
rick_h_ok, i like that wording better :)16:41
hatchhaha16:42
hatchMakyo the qa is doing the same issue as it does before....did I need to use any of the flags?16:47
hatchie) the lines don't turn grey until they are moved16:47
hatchand it's making requests for the relation icons all the time again....I thought this was fixed?16:48
hatchanywho's lemme know when you pop back16:48
rick_h_hatch: check the icons on comingsono and see if it's just a regression in the current branch16:48
hatchappears to be doing it on comingsoon as well16:50
rick_h_hatch: k, bug/card please and we'll address it. I thought it was fixed. I think I reviewed the branch. 16:51
rick_h_but we might have regressed on it and will see16:51
hatchok now it's not doing it on comingsoon16:52
rick_h_lol16:52
hatchthe lines are also grey on comingsoon16:52
hatchiunno what's going on16:52
MakyoMan, what the hell.17:03
MakyoIt was working this morning.17:03
MakyoWonder if the diff got messed up.17:03
Makyohatch, it's a race with some async stuff, so if comingsoon is running slower, it'll appear to work there.17:06
hatchMakyo ok cool17:06
hatchcan you reproduce that it doesn't work on your branch though?17:07
Makyohatch, rick_h_ ditto the relation icons.  It will appear to load them twice - as soon as the env receives data that the relation is actually a subordinate (that info isn't part of the bundles), it will re-request the icons for its relation lines.17:07
MakyoYeah.  My guess is that my computer was running slower this morning when it was working, and now things have speeded up.  I haaaate this >:/17:07
hatchMakyo with the simulator running it will continue to request the icons over and over17:07
MakyoNot seeing that.17:08
MakyoOne thing at a time, though, please.17:08
hatchwatch the terminal logs with the simulator running with the complex web example17:08
MakyoOne thing at a time, though, please :)17:08
hatchor the network tab for that matter17:08
hatchright - I'm just saying this branch (somehow) re-introduced that17:08
MakyoSure, but I can't focus on that right now.17:08
MakyoSomething goin on with charmworld?17:22
MakyoI'm getting random undefined charms.17:23
Makyojujugui ^17:23
rick_h_Makyo: it's got a restarting issue that is being investigated17:24
hatchMakyo I haven't seen that here yet, but yesterday charms were disappearing and appearing 17:24
MakyoOkay.17:24
hatchnow to....port tests17:24
hatchugh17:24
hatchI suppose the easiest way would be to just delete the file and see what fails :D17:25
redirbac the test passes. I don't know why it runs twice and fails itself the second time, apparently because there isn't really a merge.17:26
redirI am going to take myself for a walk before tackling the next thing. bbiab.17:26
MakyoOkay, I'm going to make food.  Disappearing charms making me crazy.17:27
MakyoAs soon as I say that, they come back.17:27
hatchhaha17:28
hatchrick_h_ replied to your comments17:31
rick_h_hatch: coolio17:33
bacredir: yeah, i saw that before.  unsure what jenkins is thinking.  thanks for getting the test to work.17:36
kadams54AFK for a lunch break.17:36
bacredir: you'll see jenkins did the same for 75/7617:39
hatchrick_h_ re bug #1321838 I have no idea what you're saying :) 17:39
_mup_Bug #1321838: gui exports a export.yaml vs a bundle.yaml <juju-gui:Triaged> <https://launchpad.net/bugs/1321838>17:39
rick_h_hatch: hit export and the file is called export.yaml17:39
rick_h_hatch: but to submit a bundle you need a file named bundle.yaml17:40
rick_h_hatch: it's one more step for no good reason 17:40
rick_h_to rename the file17:40
rick_h_hatch: wording advice welcome 17:40
hatchohhh ok now I get it17:40
hatchrick_h_ reworded for my feeble brain 17:43
rick_h_hatch: ty17:43
redirbac yeah I looked back through a couple and saw that17:56
redirchromium keeps crashing today:(18:09
kadams54guihelp: is a unit's ID changed in any way once it gets placed on a machine?18:16
bacrick_h_: it looks like the server start issue may just be a red herring.  over the last 11 hours apache has logged about 152 req/min, so with three workers and 250 max-request we'd expect to see a natural death every 4.9 minutes.  of course those requests are not evenly spaced, probably bunched up after US came on-line today.  so, i think what we are seeing is not an issue.18:17
Makyokadams54, no.  Units and services can't be renamed.18:17
bacredir: ^^18:17
kadams54Makyo: thanks18:18
bacrick_h_: it *did* seem to coincide with performance issues...18:18
redirbac OK18:18
bacredir: that make sense or am i mad?18:18
redirstill weird that the app thinks it restarted18:18
redirnothing in the logs?18:18
redirno errors?18:18
bacno18:19
baclogging on charmworld is a huge wart18:19
redirthough there wouldnt' be app errors i nteh log if not using mod_wsgi18:19
bacredir: quick hangout?18:19
redirbac There must be a witch under all those warts, since it "works".18:19
redirsure18:19
bacredir: normal bat channel18:19
hatchjcastro are you around?18:36
jcastroYO18:40
hatchjcastro: if a guy asks "how do i get started with Juju" where do you point them?18:41
jcastrohttps://juju.ubuntu.com/docs/getting-started.html18:41
jcastrocome on dude, the URL is literally getting started. :)18:41
jcastrohas inline videos and everything18:42
rick_h_lol18:42
hatchhaha, well it's a few steps to get there from juju.ubuntu.com18:42
jcastroyeah, but I don't have control over that. :-/18:44
hatchit's a known problem? :)18:45
hatchI've never really looked before heh18:45
jcastrohatch, if you guys can get me quickstart on osx and windows I can really trim that page18:46
bacjcastro: os x is underway18:46
rick_h_bac: is our hero18:47
rick_h_osx hero18:47
redirbac, rick_h_ : I can see the start time update each time I send a slew of requests at it locally18:55
redirI think rick_h_ was right, each 250 requests the procs get killed and restarted18:55
redirwith a new start time18:56
bacthanks for confirming redir18:56
baci'm making a slack task for cleaning up /heartbeat to be less confusing.  includes rewording that entry and removing the API2 featured entry since it is data-driven and not a measure of health18:57
rick_h_redir: bac yea, we should get the apache logs and make sure we're not getting spammed or something as well19:00
bacrick_h_: i have the logs and posted summary a little while ago19:00
rick_h_redir: bac especially as something is causing the site to misbehave. charms were leaving/arriving again and Makyo was hitting issues with getting responses19:00
bacrick_h_: 70% of the traffic is from gsa-crawler19:03
redirbac is that google-bot?19:04
rediror the gov't crawler?19:04
bacgoogle search appliance19:05
redirI can see the problem of 3 simultaneous requests coming in and all 3 procs getting killed and restarted impacting performance too bac, rick_h_ 19:05
redirOIC internal19:05
bacredir: gsa implies it is an internal appliance?19:07
redirbac my only experience with gsa is -- it is a blue box you buy^H^H^Hlease from google, stick in a rack and point at sites to crawl19:10
redirunliess this is a different appliance19:10
bacredir, rick_h_: confirmed it is an internal search appliance that is pounding charmworld19:12
redirbac IIRC you pay by documents indexed, so I haven't heard of anyone pointing it outside their own domain19:13
redirwe used them in gov't to index non public docs...19:13
redirso what to work on next19:14
* redir goes to look at kan ban19:14
rick_h_redir: can you setup a hangout? I've got something for you :)19:27
hatcha cookie? can I have a cookie?19:42
hatch:P19:42
* redir hands hatch a cookie19:47
hatchyussss!19:47
hatchhttp://i.imgur.com/7BbY4.gif19:48
bacrick_h_: the internal appliance was updated last friday.  IS is turning it down from '4' to '2'19:51
rick_h_bac: ok19:53
rick_h_bac: I still think something else is up. I'm not sure why this would cause charms to appear/disappear as Makyo and hatch have noticed the last two days19:54
redirbac where is your localcharm doc?19:55
redirrick_h_: will login access be needed?19:55
bacredir: i'm not sure what you want19:56
redirmmm19:56
redirme neither19:56
rick_h_redir: login to what?19:56
bacthe doc for deploying charmworld in an lxc?19:56
rediryou added the ability to imoprt localcharms.19:56
rick_h_bac redir the local charms laoded into charmworld19:56
bacah, that19:56
redirthat is doc'ed somewhere right, bac?19:56
redirrick_h_: will charmworld login access be needed?19:57
redirby the dude19:57
bachttps://docs.google.com/a/canonical.com/document/d/1KZoFKN7-Qo9AIyBrPnSak-1iAX4QepfbEO3qtfn2xxs/edit#heading=h.m1v3i5yotjq319:57
bacredir, rick_h_: ^^  -- it references a doc that was specific to the ATL demo, i.e. deploying to MAAS19:57
rick_h_thanks bac, redir needs to crib it for a partner that needs to be able to do something like this19:58
bacredir: local ingestion is also discussed briefly in charmworld/docs/index.rst19:58
redirmerci bac19:59
bacde nada19:59
rick_h_bah, I looked all over but not in index19:59
kadams54guihelp: https://github.com/juju/juju-gui/pull/339 is ready for QA and review20:39
hatchkadams54 I can, but I'm powering on my own tests right now, so if noone steps up I'll get it done before my EOD20:45
rick_h_kadams54: hitting my EOD, if needed I can look in the morning20:47
rick_h_to which I say, have a good night all, /me runs away20:47
bacredir: could you QA the ngram stuff against staging.manage.jujucharms.com?  i'm hoping to release r154 first thing in the a.m.20:51
bacs/r154/r514/20:51
redirbac that means setting up juju gui and pointing it at that?20:52
redirrightt?20:52
bacredir: or just curling20:52
redirduh20:53
* redir pulls foot from mouth20:53
* bac walk21:06
redirbac are there any bundles up there?21:16
redirbac so ngrams work21:32
redirhttp://staging.manage.jujucharms.com/api/3/search?text=ngrams:sql21:33
redirgets me a bunch of sql21:34
redirhttp://staging.manage.jujucharms.com/api/3/search?text=ouc does not however21:34
redirhttp://staging.manage.jujucharms.com/api/3/search?text=ngrams:sql%20AND%20series:trusty 21:35
redirgets me all kinds of sql things, but only from the trusty distro series21:36
redirand http://staging.manage.jujucharms.com/api/3/search?text=ngrams:sql%20AND%20series:trusty%20NOT%20categories:databases21:37
redirgets me things with sql in the name from the trusty series not listed as databases in categories21:37
hatch13 test downs, 5 more to go22:22
hatchI've found that if I run this mbp on the highest resolution setting chrome doesn't tear while scrolling22:31
hatchinteresting...22:31
=== jcsackett_ is now known as jcsackett
rick_h_redir: bac so we need to create a follow up card to enable ngram search in the gui it looks like?22:34
rick_h_redir: bac how does a search for charm:sugar get to work with ngrams? 22:34
=== _mup__ is now known as _mup_
huwshimiMorning23:03
rick_h_morning huwshimi 23:04
hatchmorning huwshimi 23:14
hatchI have finally finished the first leg of this conversion23:15
hatchphew*23:15
hatchjujugui I'm looking for a review/qa on the first step of converging the curated(editorial) and search results views https://github.com/juju/juju-gui/pull/33823:21
hatchrick_h_ I won't be able to get to kadams branch today, sorry you'll have to do it in the am :)23:22
hatchor someone :)23:22
hazmathatch, where was that conversation? #general?23:36
hatchhazmat are you in the Slack community?23:51
hazmathatch, no idea what that is23:51
hazmathatch, also btw +1 awesome23:52
hatch:-) thanks23:52
hatchhttps://slack.com/23:52
hatchthis is slack23:52
hatchone sec while I get you the info to get signed up23:52
hatchyou need to email them23:52
hazmatoh.. this like hipchat?23:52
hatchhttp://blog.gopheracademy.com/gophers-slack-community23:53
* hazmat is constantly amazed at the many reimplemnations of irc ;-)23:53
hatchyeah, or like the 100 other versions of the same thing23:53
hatchhaha23:53
hatchyeah - I'm not a huge proponent of it, but it has some cool features and the community is great, so if it allows them to keep the riffraff out I'm ok with it :)23:53
* hazmat hands +mod23:54
hazmatevery startup i've talked to in the last year is using hipchat.com .. basically as more universal irc23:54
hatchahh, when I was contracting a lot I worked with were using https://www.flowdock.com/23:55
hatchI'm not really sure what the difference is, but Slack has this cool bot which asks you questions in a chat to fill out your profile hah23:56
hatchI thought it was pretty cool23:56
hazmatnice23:56
hatchit's great when people who are totally unrelated to the project use it and are blown away23:57

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!