bac | rick_h_: you around? | 02:06 |
---|---|---|
=== gary_poster|away is now known as gary_poster | ||
gary_poster | hey huwshimi, I'm running away, but I'd like your review of https://codereview.appspot.com/22200045 sometime today when you get a chance. Would like your opinion of the effect in particular, but code thoughts are more than welcome as well. | 02:28 |
huwshimi | gary_poster: No problems, I'll take a look | 02:28 |
gary_poster | thanks! | 02:28 |
gary_poster | bye :-) | 02:28 |
=== gary_poster is now known as gary_poster|away | ||
huwshimi | gary_poster|away: Bye | 02:29 |
hatch | hey Makyo great video :D | 02:33 |
Makyo | hatch, thanks; working on finding a transcoding that doesn't artifact so much before I submit it proper. | 02:34 |
hatch | it sounds so 'official' :D | 02:34 |
Makyo | That maaaaay have been the goal? :P | 02:35 |
hatch | haha - well I'm glad to be back at work tomorrow - I'll try and put together my notes for everyone else | 02:36 |
Makyo | Rock on~ | 02:41 |
rick_h_ | huwshimi: sorry, been out | 03:10 |
rick_h_ | huwshimi: I'd say that it can just have a very generic initial template. | 03:10 |
huwshimi | rick_h_: It's all good, I've made progress :) | 03:10 |
rick_h_ | huwshimi: just a div container to stuff content into and other can deal with it/run with it | 03:10 |
rick_h_ | huwshimi: ah, cool stuff then. Going ok then? | 03:10 |
huwshimi | rick_h_: Yeah, just writing tests at the moment. | 03:11 |
huwshimi | rick_h_: It'll need a good review as there is stuff that is probably wrong, but I wasn't sure how else to do it :) | 03:11 |
rick_h_ | huwshimi: sure thing | 03:11 |
huwshimi | rick_h_: One question however, to add a new test file do I just add the name to test/index.html? Do I need to do anything else to get the test runner to see them? | 03:13 |
rick_h_ | huwshimi: yea, just add to the test/index.html | 03:13 |
huwshimi | hmmm... | 03:13 |
rick_h_ | huwshimi: make sure there are no syntax errors or otherwise in the console or it'll look like the file just doesn't load | 03:14 |
huwshimi | Ah right, that could be it | 03:14 |
huwshimi | oh, lint is not happy :) | 03:15 |
rick_h_ | lol | 03:15 |
huwshimi | Yep, that was it | 03:19 |
rick_h_ | party | 03:19 |
=== rogpeppe1 is now known as rogpeppe | ||
BradCrittenden | hi frankban | 10:04 |
=== BradCrittenden is now known as ba | ||
=== ba is now known as bac | ||
frankban | hi bac | 10:07 |
bac | hi frankban. i was going to ask you for a charmworld review, but it was simple and urgent so i self-reviewed. bad me. | 10:07 |
frankban | bac: np | 10:10 |
rick_h_ | bac: thanks for the link fix. I assumed urls would be generated with request.route_url and not need updating. Should have looked :/ | 11:22 |
rick_h_ | morning frankban, welcome back to the party :) if any of the cards don't make sense let me know. I wrote them up as we were hitting things yesterday so might not be all that clear headed. | 11:24 |
frankban | rick_h_: cool, I have to go for a while now, I'll ask when back, thank you | 11:43 |
rick_h_ | frankban: have fun | 11:43 |
rick_h_ | benji: started making cards out of the todo list and marked our 'debug' card as done. I've got to run the boy to day care this morning so will be in/out for a sec this morning. | 11:50 |
benji | rick_h_: k | 11:50 |
bac | rick_h_: i'm just glad that the fix was easy. after applying my bundle proofing stuff i was paniced when i saw everything produce 404s | 12:05 |
bac | rick_h_: it occurs to me we should give jcastro and others a chance to fix their bundles before we roll the proofing onto production. on staging only two bundles make it through proofing, none of jorge's | 12:06 |
rick_h_ | bac: well, we need it in production so when he runs proof locally it finds all the issues to fix | 12:35 |
rick_h_ | bac: imo, the only way to fix it all is to run charm proof on them and get the list of issues | 12:35 |
bac | rick_h_: i agree. my thought was he'd want a chance before they all disappear on production | 12:41 |
rick_h_ | bac: meh, I think with release pushed we'll just bribe him to push changes today/tomorrow :) | 12:41 |
rick_h_ | doh, he's not around atm for me to bug him about bribery | 12:41 |
bac | rick_h_: i just don't want to get grief like "i had this big demo and there was nothing there". | 12:42 |
bac | like the recent demo mark did without warning us | 12:42 |
rick_h_ | bac: ah, yea that I can get behind. | 12:42 |
rick_h_ | bac: good point | 12:42 |
rick_h_ | luca___: I now feel like I have permission to reply to all questions with "Use the force luca" | 12:51 |
rick_h_ | luca___: and why did that email make me very very afraid? lol | 12:52 |
luca___ | rick_h_: lol | 12:52 |
luca___ | rick_h_: you'll find out haha | 12:52 |
* rick_h_ goes to get his antacids | 12:52 | |
luca___ | rick_h_: a 4 hour meeting 1-2-1 meeting with Mark only leads to one thing | 12:52 |
rick_h_ | CHANGE! | 12:53 |
luca___ | rick_h_: bingo | 12:53 |
* rick_h_ is one smart cookie | 12:53 | |
rick_h_ | ok, so as long as I'm right to be afraid | 12:53 |
rick_h_ | benji: lol, looks like I should read the comments more. See #3 in http://bazaar.launchpad.net/~hazmat/python-jujuclient/trunk/view/head:/jujuclient.py#L41 | 12:57 |
benji | hmm | 12:59 |
=== gary_poster|away is now known as gary_poster | ||
frankban | rick_h_: ping | 13:11 |
rick_h_ | frankban: pong | 13:11 |
gary_poster | frankban, welcome back :-) | 13:11 |
frankban | rick_h_: I see you are working on the python jujuclient. I was thinking that, while we are there, we could also add an easy fix to deploy() to support ToMachineSpec (required by quickstart), but that's not a blocker now | 13:12 |
frankban | rick_h_: (and currently quickstart just overrides deploy) | 13:13 |
rick_h_ | frankban: ok, if it's not for the bundle release I say add a bug/card and happy to go back to it but right now it's release blockers or bust | 13:13 |
frankban | rick_h_: cool, ok | 13:14 |
rick_h_ | there's all these lovely disclaimers that jujuclient is "pre-alpha software...use at your own risk" but it's kind of in our productin story now so I assume we'll have to be spending some time getting it out of pre-alpha stage :/ | 13:14 |
bac | hi gary_poster. staging looks good and i think it is ready to roll out to production. but none of jorge's bundles pass the proof test. so, should we give him a chance to fix them before rolling out to production and having them disappear? (only two bundles are on staging, hatch's and my fake one.) | 13:16 |
frankban | rick_h_: yeah. do you have a minute to talk about those charm cards? | 13:16 |
rick_h_ | frankban: sure thing | 13:16 |
bac | frankban: how's the new laptop? | 13:16 |
gary_poster | bac, no. They are not a loss until they work. :-) | 13:16 |
gary_poster | bac, IOW, fire away | 13:17 |
rick_h_ | frankban: https://plus.google.com/hangouts/_/7ecpihg9purkm4gn74bkdgttpg?hl=en | 13:17 |
bac | gary_poster: ok, i've already composed a message. i'll just change the wording to be more "you're currently screwed, here's how to fix it." | 13:17 |
frankban | rick_h_: Charm tests do not install requirements.pip to be able to run make test. how to reproduce? it seems to me that test calls unitttest and ftest, and both depend on setup | 13:17 |
gary_poster | bac, lol ok | 13:17 |
gary_poster | bac, do you disagree with me? | 13:17 |
frankban | bac: the new laptop is enjoying its long trip from shangai, in venice now, hpefully it will arrive tomorrow | 13:18 |
bac | gary_poster: given the way high-level presentations are being done with no warning, i worry that someone will discover no bundles are available at an inopportune moment. | 13:18 |
gary_poster | :-) | 13:18 |
bac | gary_poster: i guess maarten is mostly at risk | 13:18 |
bac | sabdfl has already done his. | 13:19 |
gary_poster | bac, but they are already using comingsoon, which is tied to staging, which already has the characteristic you are worried about | 13:19 |
bac | ohhhh | 13:19 |
gary_poster | (and they were using the file dnd functionality) | 13:19 |
bac | ok, non-issue then | 13:19 |
bac | off to RT | 13:19 |
gary_poster | cool | 13:19 |
gary_poster | :-) thanks | 13:19 |
bac | will release r446 | 13:20 |
gary_poster | yay! | 13:21 |
bac | rt filed, jjo ack'ed | 13:26 |
gary_poster | cool. thank you very much for pushing this through (at all hours of the night), bac | 13:27 |
bac | was no problem. | 13:28 |
bac | gary_poster: there is another 'make jujucharms release' card. if it is need could you add a description? | 13:38 |
gary_poster | bac, yeah. the intent is "make [non-manage] jujucharms release" | 13:39 |
gary_poster | suggestion on spelling that? :-) | 13:39 |
gary_poster | rick_h_, do you have some time to review https://codereview.appspot.com/22200045/ ? You would be obvious choice, but can ask others if you are putting out fires or otherwise busy | 13:40 |
rick_h_ | gary_poster: sure thing. Looking now. I saw someone looking at it so moved on. | 13:41 |
gary_poster | cool thanks | 13:41 |
gary_poster | rick_h_, would you like a liberal dosing of comments? If so, can do. Was just wondering about that myself. | 13:54 |
gary_poster | rick_h_, and thank you for review | 13:54 |
rick_h_ | gary_poster: yea maybe. I mean it's pick-apartable just read dense. | 13:54 |
rick_h_ | gary_poster: I thought about trying to move the function out of linkify, but you're using the array in that scope so I guess you can't | 13:55 |
gary_poster | rick_h_, right. ok cool. I'll give a whirl with some comments and then land it. | 13:55 |
rick_h_ | gary_poster: maybe just whitespace around the inner function would help? | 13:55 |
gary_poster | rick_h_, lol ok | 13:55 |
rick_h_ | gary_poster: cool, thanks. | 13:55 |
hatch | morning all | 14:07 |
rick_h_ | morning hatch | 14:12 |
bac | wow, with 99.92% reporting the VA attorney general race is separated by 53 votes out of 2.2M cast. | 14:12 |
rick_h_ | ouch | 14:13 |
hatch | wait for the recount....lol | 14:14 |
hatch | rick_h_: did you get the lastpass v3 update? It not longer looks like someone threw up it's UI | 14:25 |
rick_h_ | hatch: yea, I went to ping you on it but you were afk | 14:25 |
rick_h_ | hatch: figured you'd be happy | 14:25 |
hatch | yup, it's a little slow now but oh well | 14:28 |
rick_h_ | hatch: meh, new release. I'm sure i'll pick up as they get it going | 14:28 |
hatch | I got the popup during my talk yesterday lol | 14:29 |
hatch | that was pretty comical | 14:29 |
rick_h_ | nice | 14:29 |
rick_h_ | incognito mode ftw | 14:29 |
rick_h_ | I've tried to start using that for talks since the great 'google voice text has arrived' debacle :) | 14:29 |
hatch | haha | 14:29 |
gary_poster | rick_h_, beddah? thoughts? http://pastebin.ubuntu.com/6370735/ http://pastebin.ubuntu.com/6370754/ | 14:31 |
rick_h_ | gary_poster: lol, nice | 14:32 |
gary_poster | :-) | 14:32 |
rick_h_ | gary_poster: hatch appreciate it when he gets sent in to fix something or add another bit of parsing :) | 14:32 |
gary_poster | :-) cool | 14:33 |
* hatch goes to look | 14:33 | |
hatch | haha | 14:33 |
hatch | thems a lot of comments | 14:33 |
gary_poster | heh | 14:33 |
hatch | ugh so many emails | 14:50 |
gary_poster | jujugui if you hard reload you will see I changed comingsoon to point back to manage | 14:59 |
rick_h_ | gary_poster: thanks! | 14:59 |
gary_poster | welcome | 14:59 |
hatch | I also learnt yesterday that you removed the charmworld flag during the demo....lol | 14:59 |
rick_h_ | lol | 14:59 |
rick_h_ | hatch: what? You didn't get a local running copy to demo with?! | 14:59 |
rick_h_ | hatch: boy you leaned a lot of lessons in one talk | 15:00 |
hatch | "and as you can see here when I do a search for 'hatch' I don't get....oh I do....well looks like they removed that flag" | 15:00 |
hatch | haha | 15:00 |
gary_poster | heh | 15:00 |
gary_poster | sorry | 15:00 |
hatch | rick_h_: ahh it was totally fine | 15:00 |
gary_poster | hey rick_h_ I have to head out, but "home" link is missing from charm browser in comingsoon. investigate | 15:00 |
hatch | I deployed to ec2 the trunk | 15:00 |
gary_poster | please? | 15:00 |
hatch | to demo how it worked | 15:01 |
rick_h_ | gary_poster: ugh, looking. | 15:01 |
gary_poster | rick_h_, wait I may be stupid | 15:01 |
rick_h_ | gary_poster: hmm, I see it /me hard reloads | 15:01 |
jcastro | hey fellas | 15:01 |
bac | hi jcastro | 15:01 |
jcastro | so the errors on "blah blah is not a type of string" and stuff | 15:01 |
gary_poster | rick_h_, nevermind me. sorry for the heart attack. as you were. ;-) | 15:02 |
jcastro | are the only errors in my bundles | 15:02 |
rick_h_ | gary_poster: ugh, but I'm hitting an issue where I can't get from browse to env and back :/ | 15:02 |
rick_h_ | ugh, and I'm out of it as well. Environment isn't a button oops | 15:02 |
gary_poster | rick_h_, that's a UX thing, yeah | 15:02 |
rick_h_ | jcastro: also remove constraints for now please. | 15:02 |
gary_poster | rick_h_, however, if you want to see a bug, try this. ;-) | 15:02 |
jcastro | I only have constraints in one | 15:03 |
rick_h_ | ruh roh... | 15:03 |
jcastro | but the other ones, those aren't my fault right? | 15:03 |
rick_h_ | jcastro: yea, we chased down a lot of bug fun with it. Appreciate it pointing out the issue | 15:03 |
rick_h_ | jcastro: 'other ones'? | 15:03 |
jcastro | "blah blah is not a type of string" and stuff | 15:03 |
rick_h_ | jcastro: yea, the config values must be of the right type. That should be good to go now | 15:04 |
jcastro | ok, so I need to update something then | 15:04 |
rick_h_ | jcastro: so if it says it's an integer it has to be a interger not "9" or anything | 15:04 |
jcastro | also, if that's good to go then how come brad's tests came back with all of them flunking? | 15:04 |
rick_h_ | jcastro: there was a bug in the gui exporting them all as strings, but should be fixed | 15:04 |
gary_poster | rick_h_, (1) in quicksearch, choose "applications" (2) in quicksearch again, type a string to search for and then press return to make it happen. In query string you'll see you are now constrained by category. not cool because you can't escape category anymore | 15:04 |
jcastro | rick_h_, ugh, so basically | 15:05 |
rick_h_ | gary_poster: yea, we hit that at the sprints. | 15:05 |
jcastro | I have to redo all my bundles | 15:05 |
rick_h_ | jcastro: no, small tweaks. Remove quotes if the config is an int | 15:05 |
gary_poster | jujugui, need to head to do fam stuff for a bit (middle school thing I mentioned). back in 1.5 hours hopefully | 15:05 |
rick_h_ | jcastro: should be small/easy. Which one are you looking at? | 15:05 |
bac | jcastro: can you proof lp:~jorge/charms/bundles/mediawiki-scalable/bundle ? | 15:05 |
jcastro | wordpress | 15:05 |
jcastro | bac, ah, that one is crashing something | 15:06 |
bac | yeah, i'm looking at it | 15:06 |
bac | may just be indentation | 15:06 |
jcastro | rick_h_, constraints removed from discourse | 15:07 |
rick_h_ | jcastro: and wordpress please | 15:07 |
jcastro | huh weird, that one had blank ones | 15:08 |
jcastro | done | 15:08 |
rick_h_ | jcastro: yea, and they were of the wrong names to work | 15:08 |
jcastro | oh they all do | 15:09 |
jcastro | ok so how do I fix these bundles | 15:09 |
jcastro | if it's a string it looks like this: | 15:09 |
jcastro | "binlog-format": MIXED | 15:09 |
rick_h_ | jcastro: which fields are failing? | 15:10 |
jcastro | "block-size": 5 <-- and this one looks like an int | 15:10 |
rick_h_ | jcastro: it shold list all the ones it doesn't like | 15:10 |
jcastro | a bunch | 15:10 |
rick_h_ | dude, helpful, pastebin or something | 15:10 |
jcastro | working it! | 15:10 |
rick_h_ | :) | 15:10 |
jcastro | http://paste.ubuntu.com/6370941/ | 15:10 |
jcastro | http://paste.ubuntu.com/6370942/ for wordpress | 15:11 |
rick_h_ | jcastro: kthx, will look into it. Looks like something is still borked | 15:11 |
jcastro | yeah | 15:11 |
jcastro | that's why I'm like wait a minute, it's not my fault I'm not passing tests! :p | 15:12 |
jcastro | bac, oppressor! | 15:12 |
bac | ha | 15:12 |
bac | rick_h_: is this an issue that should stop our release? | 15:13 |
rick_h_ | bac: yes, the type checking in proof is bad still if that's the latest code on mjc | 15:14 |
rick_h_ | bac: did the release with the type check fixes go out? | 15:14 |
bac | rick_h_: manage has not been updated. i don't know what rev you're talking about, on staging i assume. | 15:14 |
rick_h_ | bac: ah so maybe it's not an issue then | 15:14 |
bac | manage is very old | 15:15 |
rick_h_ | bac: ah ok then. That would explain it. Ok, I thought I had fixed it. Will try to do a local test then | 15:15 |
bac | rick_h_: actually i say old but i mean it has not been updated as part of this effort. not sure what version of proof is on it. | 15:15 |
rick_h_ | bac: looking for the branch where I fixed the type checks | 15:16 |
* bac wishes heartbeat showed rev | 15:16 | |
bac | rick_h_: i don't know how to tell what is running on manage | 15:16 |
rick_h_ | bac: yea, well if heartbeat isn't showing rev then it's old enough I think. abel had that added I thought | 15:16 |
jcastro | hmm, should I update my charm tools now? | 15:16 |
rick_h_ | jcastro: yea, need to do that. bac's email had notes on the rev, but it sounds like the server isn't updated with the latest atm either | 15:18 |
* jcastro nods | 15:18 | |
bac | rick_h_: no, abel's follow-on branch for revno in heartbeat has not landed even in trunk yet | 15:18 |
jcastro | rick_h_, I was already in the car ready to drive over and kill you if you were going to make me remake every bundle | 15:18 |
rick_h_ | jcastro: heh, trying not to man. Trying | 15:19 |
rick_h_ | bac: I could have sworn I saw that go by with a LGTM, man I'm not keeping up today | 15:19 |
* rick_h_ isn't seeing the changes I expect to see | 15:20 | |
rick_h_ | bac: ok, I'm looking for changes in r437 | 15:22 |
rick_h_ | bac: I thought there was a deploy in progress yesterday after the rollback so had expected that to be on mjc? | 15:22 |
bac | no | 15:22 |
bac | mjc has not been touched. an RT is waiting | 15:22 |
rick_h_ | ok, so jcastro then ignore proof until the server gets updated sorry. | 15:22 |
rick_h_ | bac: rgr, thanks. Sorry for the confusion | 15:23 |
bac | rick_h_: i had a disconnect and thought charm-proof was hitting staging | 15:23 |
bac | actually i just didn't think too hard about it | 15:23 |
rick_h_ | bac: no, I thought proof hit mjc once he released charm-tools 1.0 | 15:23 |
rick_h_ | err 1.1 that is | 15:23 |
bac | i'm sure you're right | 15:23 |
bac | i just tried changing my /etc/hosts to point mjc to sjc but get SSL errors | 15:24 |
rick_h_ | yea, ugh | 15:24 |
rick_h_ | best thing is to hit proof with curl or httpie | 15:24 |
rick_h_ | err, the staging proof endpoint | 15:24 |
rick_h_ | bac: http://paste.mitechie.com/show/1058/ | 15:25 |
bac | cool | 15:25 |
rick_h_ | that's it, going back to bed. I give up. "bzr shelve --all ... bzr: ERROR: bzrlib.errors.MalformedTransform: Tree transform is malformed [('missing parent', 'new-6')]" | 15:26 |
frankban | guihelp: I need two reviews and one QA for https://codereview.appspot.com/22300044 . Is anyone available? Thanks! | 15:49 |
Makyo | frankban, on it. | 15:49 |
frankban | (quickstart + python) | 15:49 |
frankban | thanks Makyo | 15:49 |
Makyo | jujugui call in 10 | 15:50 |
bac | jujugui: manage.jujucharms.com updated. heartbeat shows failure on bundles but ingest just started at :45. http://manage.jujucharms.com/heartbeat | 15:50 |
rick_h_ | bac: cool, thanks! | 15:50 |
rick_h_ | jcastro: re-run your proof please. When I charm proof your wordpress bundle I get no output | 15:53 |
bac | \o/ | 15:54 |
bac | rick_h_: this one still has issues ~jorge/charms/bundles/mediawiki-scalable/bundle/ | 15:55 |
jcastro | rick_h_, CONFIRMED! \o/ | 15:55 |
bac | jcastro: ^^ | 15:56 |
jcastro | ok | 15:56 |
rick_h_ | bac: k, looking | 15:56 |
rick_h_ | bac: oh, it's got like blow up and die kind of issues :/ | 15:56 |
bac | rick_h_: did you say having 'constraints' with defaults causes problems? | 15:56 |
bac | yeahs | 15:56 |
rick_h_ | bac: not with proof, but in the real world yes | 15:56 |
bac | oh, so worst case scenario | 15:57 |
bac | yay for false positives | 15:57 |
rick_h_ | bac: well we don't proof constraints right now. and if they're not 100% correct for Go the deployer will die in a fire | 15:57 |
rick_h_ | bac: this was the big result of yesterday's barn burning joy | 15:58 |
Makyo | jujugui call in 2 | 15:58 |
rick_h_ | bac: oh hmm, that json error is out of the remote proof call. So something isn't happy in our end after all perhaps | 15:59 |
benji | rick_h_: I'll look at your branch and then we can touch base as to what to do next. | 16:11 |
rick_h_ | benji: did you want to catch up then and coordinate. Did you still have the diff we had going on in the charm stuff, the branch for the deployer and such? | 16:11 |
benji | rick_h_: yep, I have all those diffs around | 16:11 |
rick_h_ | benji: cool, sorry about that. You had most of the other stuff so i took the 'easy' one right off. :/ | 16:11 |
frankban | macbook users, any suggestion for installing saucy on a mac? I've found http://blog.kylebarlow.com/2013/05/installing-ubuntu-1304-raring-ringtail.html but each guide is different | 16:12 |
hatch | frankban: are you going to do it on metal or vm? | 16:12 |
bac | hi evilnickveitch did you see the comments jcastro made in the Google doc on bundles we delivered to you in SF? i was wondering if that was something i needed to pursue or if you were adding the extra info. | 16:13 |
frankban | hatch: metal | 16:13 |
hatch | hmm sorry I don't have any input then :) | 16:13 |
bac | rick_h_: how shall we pursue the proof issue jcastro's branch raises? | 16:14 |
frankban | hatch: any reason to prefer fusion? | 16:14 |
hatch | frankban: I just find it removes any driver issues | 16:15 |
bac | frankban: i prefer fusion b/c managing vms is much easier and less stress when upgrading | 16:15 |
hatch | so things like the touchpad, camera, etc work as expected | 16:15 |
bac | and it just works | 16:15 |
hatch | bac: you're using fusion? I've been using parallels but it doesn't seem to support new Ubuntu well at all | 16:16 |
bac | snapshot the vm, upgrade to the next release, and if it doesn't work just rollback or restore from backup | 16:16 |
bac | yep, i'm running mavericks, fusion 6, and saucy | 16:16 |
bac | only issue is i have to manually set my display size when rebooting | 16:16 |
hatch | ahh | 16:16 |
bac | i like 1920x1200 and it doesn't stick and picks another | 16:17 |
hatch | my issue is that parallels doesn't release the parallels tools for Ubuntu very fast so the display drivers get corrupted | 16:17 |
frankban | bac: interesting, yeah, snapshots are cool | 16:17 |
hatch | it's kind of a mess | 16:17 |
hatch | so now I'm running it w/o X | 16:17 |
rick_h_ | bac: sorry, drink refill. Need to trace what's going on. I'm debugging a local charm-proof instance to see what mjc is returning right now | 16:17 |
bac | frankban: when running metal i'd have to mentally clear the next four hours before upgrading in case something went crazy. | 16:17 |
bac | rick_h_: ok, let me know if i can help. | 16:18 |
rick_h_ | bac: ok, mjc is throwing a 500 internal server error at that bundle | 16:18 |
rick_h_ | bac: so the thing to do from here would be to toss it at a local charmworld instance like that pastebin I showed you and see what is failing | 16:19 |
bac | frankban: the touchpad is also a big issue. everything works mac-like even when in fusion. on metal i found it unusable as brushing it would jump the cursor. may have improved by now. | 16:19 |
bac | rick_h_: i wish proof had a hidden option for specifying the server | 16:20 |
rick_h_ | bac: +1 | 16:21 |
rick_h_ | bac: so I pdb;d int the proof and this is the string it's sending up http://paste.mitechie.com/show/1063/ | 16:21 |
rick_h_ | bac: so I'll run charmworld locally and use that + httpie to test it and see what's up | 16:22 |
rick_h_ | bac: strange, I get a "could not parse the yaml file" error returned when I test locally :/ | 16:23 |
rick_h_ | bac: ok, it passes on mjc as well | 16:24 |
bac | rick_h_: he has a 'services' option on one of his services that looks dubious to me. removing it didnt' help though | 16:24 |
rick_h_ | http://paste.mitechie.com/show/1064/ <- bac | 16:24 |
rick_h_ | bac: so it must be something in how the charm proof is sending the values that it doesn't like ? | 16:24 |
bac | yep that's the unhelpful error i've seen | 16:24 |
rick_h_ | bac: because if I pdb stop and print the bundle_file contents then I get the correct error message | 16:25 |
rick_h_ | bac: well, it's what we've got to show though. It's unparseable yaml. It doesn't know why though. | 16:26 |
rick_h_ | bac: but the proof code is getting a 500 server error. I wonder if sinzui is still setup to get emails of those | 16:26 |
bac | rick_h_: yes | 16:26 |
rick_h_ | bac: yea, then we should try to get a copy of that email then. I can't dupe it locally so would be helpful to see the traceback in the email | 16:29 |
rick_h_ | bac: ooh, changed the url in the proof library to my localone and got the error | 16:29 |
rick_h_ | bac: http://paste.mitechie.com/show/1065/ | 16:30 |
rick_h_ | bac: can you add type support in proof for float then? | 16:30 |
rick_h_ | bac: looks like I just missed that as a valid type for a config value. I didn't see that in the docs. :/ | 16:31 |
bac | oh | 16:32 |
bac | brb | 16:33 |
Makyo | Ffffffff | 16:34 |
Makyo | Forgot to destroy environment. | 16:34 |
Makyo | I'm so bad at this. | 16:34 |
hatch | lol | 16:35 |
hatch | Makyo: maybe you should create a script to destroy the environment after a 24h timeout lol | 16:35 |
Makyo | I'm seriously pondering that, yeah, just a cron job that runs nightly with something like `juju destroy-environment -e aws-goju -y` | 16:37 |
hatch | haha | 16:39 |
bac | Makyo: i've found my little notification-based watcher very helpful. | 16:44 |
Makyo | Oh? | 16:44 |
bac | i put something on our blog ages ago | 16:45 |
bac | not super smart but gives me enough of a nudge to know when things are running | 16:45 |
bac | Makyo: http://jujugui.wordpress.com/2013/04/09/alerts-for-running-ec2-instances/ | 16:46 |
Makyo | Oh! Right! | 16:47 |
Makyo | Will look into that. Thanks bac | 16:47 |
rick_h_ | thanks benji | 16:54 |
hatch | rick_h_: http://bgrins.github.io/devtools-snippets/#console-save | 16:54 |
rick_h_ | bac: so added a card for the float type not supported in the bundles section. If you get a chance to grab it before me yay if not I'll try to get to it. It's not killer atm, because it'll fail to ingest, but needs fixing | 16:54 |
rick_h_ | hatch: cool | 16:55 |
bac | rick_h_: ok. i'm doing a low priority feel good branch right now but will be free after my late lunch | 16:55 |
rick_h_ | hatch: my console's been messed up a while actually. I need to look into it | 16:55 |
rick_h_ | bac: rgr, thanks. | 16:55 |
evilnickveitch | bac, its okay, i am taking care of it... | 16:56 |
bac | evilnickveitch: ok, that's wonderful | 16:57 |
benji | grr, I somehow destroyed my entire bzr config | 17:11 |
bac | rick_h_: when you have a moment i think you'll appreciate this: https://codereview.appspot.com/21930046 | 17:11 |
rick_h_ | bac: just LGTM'd :) | 17:11 |
rick_h_ | bac: ty! | 17:11 |
rick_h_ | benji: :( | 17:11 |
bac | :) | 17:12 |
gary_poster | ok, so after consuming the backlog, this is my understanding. Could people please confirm or correct? (1) frankban has Makyo reviewing https://codereview.appspot.com/22300044 and needs a second reviewer. (2) bac got m.j.c updated. It mostly works except we discovered that proof needs to bless floats, and this breaks one of jorge's bundles. bac and/or rick_h_ are working to fix. (3) rick_h_ and bac are also worki | 17:20 |
gary_poster | ng with jorge to make his bindles ingestable otherwise. (4) benji is working alone on the constraint whitelist charm deployer integration work, and I'd love to know status & expected timeline. | 17:20 |
gary_poster | s/bindles/bundles | 17:20 |
gary_poster | sound right? | 17:20 |
rick_h_ | gary_poster: somewhat. rick is trying to punt to bac to do the float fix and that should take care of jorge's bundles. Rick messed up and did the whitelist bit and it's up for review to hazmat and benjie has reviewed it. | 17:21 |
gary_poster | ah cool | 17:21 |
rick_h_ | gary_poster: and benji and rick are scheduled to sit down and go over what's next since rick grabbed the whitelist owrk on accident | 17:21 |
benji | gary_poster: I am working on getting all the branches squared away for proposing | 17:21 |
rick_h_ | but benji seems to be having bzr issues atm | 17:21 |
gary_poster | benji, ok great | 17:21 |
benji | I am, but have decided to ignore it as much as possible for the moment. | 17:22 |
gary_poster | bac, are you, uh, catching rick_h_'s punt of the float issue? I'm assuming that means a fix to proof, an upgrade to charmworld, and another deployment? | 17:22 |
rick_h_ | gary_poster: yea, if benji and I are in good shape I can grab it, but since it's only blocking one charm atm figured it was lower priority. | 17:23 |
gary_poster | rick_h_, agreed | 17:23 |
gary_poster | frankban, giveing you second review of branch. | 17:24 |
gary_poster | Makyo, how is video? | 17:25 |
gary_poster | giving | 17:25 |
Makyo | gary_poster, http://www.youtube.com/watch?v=vbRkfejThkk I'm investigating a way to get less artifacts in the video. | 17:25 |
hatch | Makyo: yeah I was wondering why the HD versions looked SD | 17:26 |
Makyo | The video capture is HD, but transcoding screwed it up. | 17:27 |
hatch | hmm interesting | 17:27 |
Makyo | RecordMyDesktop records at 15fps and kdenlive hates that :P | 17:28 |
hatch | can you upload the raw file to youtube? | 17:28 |
frankban | gary_poster: great thanks | 17:29 |
Makyo | hatch, the raw file has no sound. | 17:29 |
frankban | Makyo: thanks for the review, destroy the environment! | 17:30 |
Makyo | frankban, haha, thanks :D | 17:30 |
gary_poster | Makyo, cool! Did you send to mramm? Getting something sooner ratheer than later might be important | 17:31 |
hatch | Makyo: ahhh gotcha | 17:31 |
Makyo | gary_poster, will do, it's uploading to u1 now | 17:31 |
gary_poster | Makyo, awesome. please cc me | 17:31 |
gary_poster | frankban, you have two LGTMs and a QAOK :-) | 17:41 |
frankban | gary_poster: thanks cool | 17:41 |
hatch | Makyo: ie link branch is proposing so fire-up-that-ie!!!!!!!!! | 17:41 |
Makyo | IE is NEVER worth nine exclamation points >:/ | 17:42 |
hatch | but...but.... | 17:42 |
Makyo | Gear down there, big rig. | 17:43 |
hatch | ERMAHGERD IE!!!!!!!!! | 17:43 |
frankban | gary_poster: nice suggestion thanks, I can include an example in the command line, but the charm version is required, e.g. cs:~juju-gui/precise/juju-gui-116. juju-core does not accept an unversioned CharmURL | 17:44 |
gary_poster | frankban, ah ok, cool | 17:44 |
gary_poster | and lol hatch | 17:45 |
hatch | haha | 17:45 |
* Makyo boots ie | 17:46 | |
hatch | hmm I'm getting a number of errors in IE on comingsoon | 17:49 |
hatch | 'all' is undefined and 'modules' is undefined | 17:49 |
gary_poster | :-( | 17:49 |
hatch | investigating... | 17:49 |
hatch | oh they are the source map url errors | 17:50 |
hatch | Makyo: while you're in your IE can you also see if you can repro https://bugs.launchpad.net/juju-gui/+bug/1246971 (I cannot) | 17:51 |
_mup_ | Bug #1246971: Connecting service fails the first time in IE10 <juju-gui:Triaged> <https://launchpad.net/bugs/1246971> | 17:51 |
hatch | ohh | 17:56 |
hatch | our source maps use //@ which looks like cause errors in IE10 | 17:56 |
hatch | was this switched recently or has it always been like this and noone has noticed" | 17:57 |
rick_h_ | hatch: I thought that came up before. /me is trying to recall. | 17:57 |
rick_h_ | hatch: does it only effect sourcemaps though? | 17:57 |
rick_h_ | didn't we turn off sourcemaps at one point because of this? | 17:57 |
gary_poster | hatch no change afaik | 17:57 |
benji | hatch: I don't know of any recent source map work (I did the original work but I don't remmeber if the //@ was there from the start) | 17:58 |
rick_h_ | I remember the //@ | 17:58 |
hatch | rick_h_: yes teh application works as expected | 17:58 |
rick_h_ | hatch: ok, so maybe it was just ignored. I do recall the //@ causing IE issues though | 17:58 |
hatch | ok this must be a recent change in IE10 then | 17:58 |
hatch | we should create a high/urgent card to fix | 17:58 |
hatch | apparently it being in the source will break all debugging in IE11 | 17:59 |
hatch | according to some of the VS blogs | 17:59 |
rick_h_ | IE11? ugh | 17:59 |
hatch | I'm going to create an urgent card for this as it'll block release for ie10 | 18:01 |
Makyo | hatch, can't repro. | 18:01 |
hatch | Makyo: thanks, closing | 18:01 |
rick_h_ | hatch: I'm confused. App works as expected but blocks release? | 18:01 |
gary_poster | hatch, same q :-) | 18:01 |
hatch | rick_h_: gary_poster: IE10 pops up alerts() asking to debug the error if debugging is turned off | 18:02 |
hatch | but if you ignore those, it works fine :) | 18:02 |
gary_poster | hatch ugh :-( | 18:02 |
gary_poster | ok hatch, you tackling that? :-) | 18:02 |
hatch | yeah sure | 18:03 |
gary_poster | ok thanks. hatch, would really like to release tomorrow. if expedient solution is possible, lemme know. For example, in theory, we don't serve sourcemaps for IE. Probably not easy to do, but if it were, I'd be fine with it for tomorrow. | 18:04 |
gary_poster | Fix would be nicer, of course | 18:04 |
gary_poster | and something we'd need soon after | 18:04 |
hatch | oh yippee Makefile | 18:05 |
hatch | ;) | 18:05 |
gary_poster | heh | 18:05 |
gary_poster | hatch can pair if you'd like | 18:06 |
hatch | I should be ok - I was just in a talk yesterday about Grunt and how nice the Gruntfile was to read :D | 18:08 |
gary_poster | :-P | 18:08 |
rick_h_ | hatch: http://uploads.mitechie.com/books/Managing_Projects_with_GNU_Make_Third_Edition.pdf enjoy | 18:09 |
gary_poster | lol | 18:09 |
hatch | haha yeah I have that already | 18:09 |
hatch | doesn't make Make any nicer to read :P | 18:09 |
rick_h_ | come on, make shell your new language to learn this year :) | 18:09 |
hatch | ontop of Go and Python eh? | 18:10 |
rick_h_ | sounds like a plan #moretolifethanjs | 18:10 |
hatch | pretty soon I'll have to forget how to speak English to fit any more languages in there | 18:10 |
rick_h_ | benji: heads up, uploading a dist for the updated jujuclient here: http://uploads.mitechie.com/lp/dist/jujuclient-0.14.tar.gz so we can pull it down and test things out in the charm since it's a local dep. | 18:11 |
benji | rick_h_: cool, I'll test it soon | 18:12 |
rick_h_ | benji: ok, I'm going to go fix the charmworld bug then. Let me know if you need a hand with anything on the other side. | 18:13 |
benji | k | 18:13 |
hatch | oh nice the source map code is in node not bash | 18:16 |
hatch | haha | 18:17 |
hatch | go benji | 18:17 |
benji | I don't always write JS, but when I do it is to avoid writing shell script. | 18:17 |
hatch | hahaha ^5 | 18:17 |
benji | ^5 | 18:17 |
benji | we need an IRC symbol for a fist bump... or not | 18:18 |
hatch | =B ᗺ= | 18:19 |
hatch | lol | 18:19 |
benji | heh | 18:22 |
hatch | Makyo: here is the link to the branch if you didn't already have it https://codereview.appspot.com/21430043/ | 18:34 |
bac | rick_h_, benji: we seem to have a bit of a problem: http://manage.jujucharms.com/heartbeat look at the charm/basket queue | 18:51 |
* benji looks | 18:51 | |
benji | hmm | 18:51 |
rick_h_ | bac: yea, was just going to ask on that. | 18:52 |
bac | i don't really know how to proceed. we have no visibility into prodstack | 18:52 |
rick_h_ | bac: yea, we've got to ask is for log files. | 18:52 |
rick_h_ | bac: and hope something jumps out from there. | 18:53 |
rick_h_ | bac: no vangaurd right now :/ | 18:53 |
bac | hi sinzui. we did a release to manage.jujucharms.com today. it seemed to well but then i noticed on http://manage.jujucharms.com/heartbeat the charm queue has grown crazy | 18:54 |
sinzui | well is ingest running? | 18:57 |
sinzui | hmm, I think Abels unlanded branch would tell us a little more | 18:57 |
rick_h_ | looks like no vanguard right now and it's not dead in the water to go emergency on it. | 18:58 |
rick_h_ | I've got a branch that needs a deploy up for review so we can watch it, get another deploy and get logs from it at that time? | 18:58 |
rick_h_ | bac: have time to review https://codereview.appspot.com/22370044 | 18:59 |
rick_h_ | bac: for the float/config validation fix in proof | 18:59 |
bac | sinzui: i'm not sure if ingest is running | 18:59 |
sinzui | bac, how long has it been since the deploy? | 19:01 |
sinzui | and if someone did the deploy, do we know the heartbeat was sane afterwords | 19:02 |
bac | a good while. let me look | 19:02 |
bac | sinzui: 15:27UTC. no, the heartbeat page showed the two failures b/c the bundles were removed by the migration and are to be repopulated by ingest | 19:03 |
sinzui | yep, I expected that. Ingest takes about 15 minutes, but we are hours later | 19:04 |
bac | sinzui: so queue runs will just keep adding on? | 19:06 |
bac | if ingest were started now it may not be able to empty the queue | 19:06 |
bac | rick_h_: jjo is not the vanguard but he is around. | 19:07 |
bac | he did the deploy | 19:08 |
bac | i'm not sure what to ask for | 19:08 |
bac | rick_h_: i did your review but don't understand why an unknown type is not an error | 19:08 |
rick_h_ | bac: we'd ask for the logs. There's app.log I think in the charm directory var/app.log (there's a few othersin there) | 19:08 |
rick_h_ | bac: yea, I wasn't sure what to do on that so went with ignore, but updating it to throw an error. | 19:08 |
bac | rick_h_: yeah, the log dir is a confusing jumble to me | 19:09 |
rick_h_ | I think that's the best way to keep that up to date | 19:09 |
rick_h_ | bac: yea, we'll figure it out though | 19:09 |
* bac looks at staging | 19:09 | |
sinzui | bac, The last item in the heartbeat indicates we are queue work, but not ingesting, webops can empty queues easily enough | 19:10 |
rick_h_ | bac: sinzui ping'd jjo but looks like he ran away. | 19:13 |
rick_h_ | bac: sinzui will keep an eye on vanguard | 19:13 |
sinzui | baskets are going down | 19:13 |
bac | rick_h_: thanks | 19:13 |
bac | sinzui: you mean the basket count? | 19:13 |
sinzui | yes | 19:13 |
rick_h_ | sinzui: oh, interesting | 19:13 |
bac | yes, berry berry | 19:13 |
rick_h_ | crazy queue | 19:13 |
sinzui | We can see that something is pulling | 19:13 |
rick_h_ | yea, we expect most bundles to fail so maybe it's got backed up and taking a few to get to one of the good ones :/ | 19:14 |
bac | the count is going down but it hasn't found any valid bundles yet | 19:14 |
bac | meanwhile on staging there are 11 bundles that are valid | 19:14 |
sinzui | I saw | 19:15 |
rick_h_ | ok, well keep an eye on the queue and see if we can get down to 0 | 19:15 |
bac | the entire set is about 14 baskets so something more than 14 bundles | 19:15 |
rick_h_ | bac: right, but I'm thinking the queue was already huge from something? The charms queue is a LOT larger than the number of charms in the store as well | 19:16 |
rick_h_ | 32K ? | 19:16 |
bac | rick_h_: i don't have the initial numbers. 32k is nuts. | 19:16 |
rick_h_ | bac: right, hmm the countvs just jumped up again | 19:17 |
bac | we should ask to have the queue process paused | 19:17 |
rick_h_ | yea | 19:17 |
bac | there are 2410 charms. queue has been run ~16 times since launch. 16 * 2410 = 38560 | 19:18 |
rick_h_ | bac: heh, well good to know we got through 500ish | 19:18 |
bac | well, ok, maybe it was only run 15 times | 19:19 |
sinzui | bac, rick_h_ I think this can help bring some sanity to the queue numbers http://pastebin.ubuntu.com/6372276/ | 19:21 |
sinzui | rick_h_, bac: I timed ingest to take between 13 and 17 minutes last august. I think we want to watch the numbers to learn what the current time is | 19:22 |
rick_h_ | sinzui: rgr, thanks. Will see if we can get a vanguard to take it up | 19:22 |
rick_h_ | sinzui: bac jjo is chatting now | 19:24 |
sinzui | rick_h_, bac: Note that http://manage.jujucharms.com/tools/store-missing says it has not completed a single charm ingest for 4 hours | 19:28 |
bac | interesting | 19:28 |
rick_h_ | sinzui: yea, I think this might lead to getting logs and finding an issue | 19:28 |
sinzui | Did we do something mad like change ingest to get every history revision? | 19:29 |
bac | sinzui: queue finds the last 10 revisions | 19:29 |
rick_h_ | sinzui: well there's the backfill that does 10 revisions back | 19:29 |
rick_h_ | so not every, but 10 | 19:29 |
bac | rick_h_: can jjo determine if ingest is running? | 19:30 |
rick_h_ | bac: not asked. he's EOD'ing in 30min so get anything we need now | 19:30 |
bac | it may make sense to turn off the queue cron and run it once manually | 19:31 |
bac | rick_h_: he seems to be responding to your queries | 19:31 |
rick_h_ | asking for the logs now | 19:31 |
sinzui | rick_h_, We discusses this last July and August. back history has to be done in a separate proc because we need to guarantee responsive ingestion. I think 10 rev is less that 150 minutes but certainly way to long the the 15 minute balance we setup | 19:33 |
bac | rick_h_, benji: is this deploy to production the first with 10 revision history turned on? | 19:36 |
rick_h_ | bac: I didn't think so. :/ it's been a long while | 19:36 |
bac | rick_h_: i don't know the last rollout to prod | 19:36 |
rick_h_ | bac: ok, so charm queue is 0 | 19:36 |
bac | so ingest is running | 19:37 |
sinzui | bac, rick_h_ http://manage.jujucharms.com/tools/store-missing | 19:37 |
sinzui | ^ happy | 19:37 |
Makyo | gary_poster, ping | 19:38 |
bac | sinzui, rick_h_: is there anything about the proofing on production that would hang? egress firewall rules? | 19:38 |
sinzui | let watch heartbeat for 15 minutes | 19:38 |
bac | wait, we're talking to ourselves so it can't be an issue | 19:38 |
rick_h_ | bac: no, proofing doesn't hit the net. Just the db | 19:38 |
gary_poster | Makyo, sorry! was on phone with Jeff and lost track o' time. coming | 19:39 |
Makyo | gary_poster, ah, okay! | 19:39 |
bac | rick_h_: well, ingest hits a cw url which hits the db | 19:39 |
bac | rick_h_: i'm just trying to think about differences with prod that could be hanging ingest | 19:39 |
sinzui | bac. I am not current with proof changes. I think the last change made in cw was to include the version of proof in the cache to ensure the right version is always installed | 19:40 |
rick_h_ | bac: ah, understood | 19:40 |
* hatch sorry those of you who import your coke from Mexico http://gizmodo.com/mexican-coke-is-ditching-cane-sugar-for-high-fructose-c-1459437477 | 19:40 | |
sinzui | hatch, you have just ruined my day | 19:41 |
* sinzui goes back to steaming his brownies to make them soft again | 19:41 | |
hatch | lol | 19:41 |
bac | sinzui, rick_h_: baskets just dropped from 16 to 13. no bundles added though | 19:43 |
rick_h_ | bac: cool. | 19:43 |
sinzui | bac, why is basket ingestion so slow less than 1 a minute | 19:43 |
bac | sinzui: unclear | 19:44 |
bac | hatch: that makes no sense. 1 peso/liter isn't that much. i think this is just another 'New Coke' moment to force a change to cheaper corn-based coke. glad i don't drink the stuff anymore | 19:45 |
rick_h_ | sinzui: not sure, you'd think it'd be faster. One file to process really | 19:45 |
hatch | bac: You're probably right | 19:45 |
sinzui | rick_h_, bac, given the immediate fix we saw in the missing list. It is clear that the queue must be drained to get updates to the db and es. I think we might need to slow down the supervisord proc enqueue less often | 19:45 |
rick_h_ | bac: sinzui yea, just got the queue refilled, but didn't get through the bundles :/ | 19:46 |
sinzui | rick_h_, I see the bug. | 19:46 |
bac | rick_h_: but it doesn't make complete sense. the items are processed in order, right? so it will eventually get to all of the bundles. yet we see none being added | 19:46 |
bac | or is it not ordered? | 19:47 |
rick_h_ | benji: so my jujuclient fix got shot down. He's prefer we band-aid it in the deployer middleware itself and not in the deployer/client. | 19:47 |
benji | rick_h_: ooh | 19:47 |
bac | sinzui: and? | 19:47 |
sinzui | rick_h_, bac, we only queue charm tips, but the queue size indicates we are ingesting revisions. that is F*CKING crack | 19:47 |
rick_h_ | bac: right, but there were 16? so we're at 27, so we've got through 5 of them so far. | 19:47 |
benji | rick_h_: well, I guess we can resurect my branch, since that's what I was doing | 19:47 |
rick_h_ | benji: +1 | 19:47 |
sinzui | production will is on crack and may die of complications in a few hours | 19:47 |
rick_h_ | benji: he's not a fan of the limited 'approved' set of stuff and doesn't want the band-aid in the client | 19:48 |
bac | sinzui: but this scheme has been running on staging with no issues | 19:48 |
rick_h_ | benji: and he's all on board for fixing the communication/deployer as a library issues | 19:48 |
benji | rick_h_: re. library: cool! | 19:48 |
bac | rick_h_: my point is since deployment hours ago no valid bundles have been ingested. that makes no sense | 19:49 |
rick_h_ | bac: understood | 19:49 |
sinzui | bac, I see staging trying to do that. At least it gets little cycles. | 19:49 |
bac | rick_h_: sinzui: do you want to move to a hangout in 10 minutes? | 19:49 |
rick_h_ | bac: sure | 19:49 |
bac | feel free to fire one up now but i can't join for a few | 19:50 |
rick_h_ | benji: let me know what you need to help move this forward then. gary_poster fyi ^^ | 19:50 |
sinzui | bac. Nonetheless. There are only about 130 good charms and 600 bogus charms. The queue looks at those for new revisions. I think charmworld is asking if a revision ever changes, which is a violation of reality | 19:50 |
gary_poster | rick_h_, was following along as best I culd while on a call. could you give me a quick hangout so I can make sure I understand, before you have the bigger powow? | 19:51 |
rick_h_ | gary_poster: https://plus.google.com/hangouts/_/7acpjqjf6evpv7503n0uq4psjg?hl=en | 19:52 |
sinzui | rick_h_, bac, when the charm queue clear we got updates, since we cannot ever clear the bundle queue we don't get updates. I makes sense to me | 19:53 |
bac | rick_h_: ping when you're readu | 19:55 |
bac | s/readu/ready/ | 19:56 |
rick_h_ | bac: https://plus.google.com/hangouts/_/7acpjqjf6evpv7503n0uq4psjg?hl=en | 19:56 |
rick_h_ | bac: you're right, sorry didn't think it through | 19:56 |
hatch | Makyo: another ie branch to qa :) sorry https://codereview.appspot.com/22450044/ | 20:03 |
sinzui | bac, rick_h_ there are exactly 764 branches charms and bundles to process in Lp. Production might know about 20 more deleted charm branches. Any queuing greater than that is sign of insanity. I the db is corrupt with bogus charms. It might be fair to say that the queue/ingest is being feed revisions, not charms and bundles. It would be okay to queue revisions if the queue could just ask for changes since a last date, but I am | 20:03 |
sinzui | sure that code was not written | 20:03 |
rick_h_ | sinzui: we think bundles is hitting egress filtering and causing it to timeout/take long | 20:07 |
bac | rick_h_: https://rt.admin.canonical.com//Ticket/Display.html?id=65726&results=564596e05af8d34ac09865c9e1dfbc69 | 20:08 |
rick_h_ | bah | 20:08 |
bac | sinzui: i agree with your thoughts on queueing up revisions and we need to fix that. as rick_h_ said, the immediate problem causing the delay is the bundle proofing is being blocked by an egress firewall and timing out on each proof request. | 20:19 |
rick_h_ | bac: worse than that, it's not running the right version of charmtools! | 20:20 |
sinzui | rick_h_, charmtools in in the download cache. Are you saying the wrong cache was deployed to production? | 20:21 |
rick_h_ | sinzui: so we tried to change the proof url since it was hard coded to manage.jujucharms.com. The version installed is 1.1 | 20:21 |
rick_h_ | sorry, 1.1.0 and it should be 1.1.2 | 20:22 |
* rick_h_ is checking download-cache list right now | 20:22 | |
rick_h_ | bah, there's something bigger there. the wrong version is installed, why the runs aren't throwing errors/exceptions to the wind I don't know. | 20:24 |
sinzui | Abel added charm-tools to the download cache to ensure the code we deploy has a matching proof. LTS was running ancient charm-tools | 20:25 |
rick_h_ | sinzui: yes, and the download-cache and requirements.txt are updated to 1.1.2 | 20:26 |
rick_h_ | sinzui: somehow it's not updated on production. I asked for a change to bundles.py and there was no bundles.py on the machine because it was still at 1.1.0 | 20:27 |
rick_h_ | https://pastebin.canonical.com/100009/ was the result of looking for anything charmtools | 20:27 |
=== BradCrittenden is now known as bac | ||
rick_h_ | we lost our webops and waiting for a new person to torture. approx 2hrs according | 20:27 |
rick_h_ | according to jjo | 20:27 |
hatch | I'm going to run and grab a bite/run an errand, be back in < 1h | 20:28 |
sinzui | rick_h_, I think this step in deploy is supposed to guarantee we get the download cache ./build-charmworldgo ${REVNO?} | 20:34 |
rick_h_ | sinzui: ok, will look into that. thanks for the heads up | 20:35 |
rick_h_ | sinzui: is that something in the webops add-on? I can't find that file in the charm/source? | 20:39 |
sinzui | rick_h_, yep | 20:40 |
rick_h_ | sinzui: ah ok. Then definitely good to know since I can't see it atm | 20:40 |
sinzui | rick_h_, they make a fat charm: https://wiki.canonical.com/InformationInfrastructure/WebOps/CDO/Charmworld | 20:40 |
sinzui | They pull the revno we request and pull download-cache, then pack them into the charm for upgrade-charm to extract | 20:41 |
rick_h_ | sinzui: rgr, thaniks | 20:42 |
sinzui | rick_h_, I need to get children from school. I am searching for a branch that might contain the script that makes the fat charm | 20:43 |
rick_h_ | sinzui: ok, thanks. I'm sure we'll be able to move forward once the new vanguard arrives. | 20:44 |
bac | rick_h_: proposed change to marco | 20:55 |
bac | rick_h_: we're in a lull so i'm going to step out for a bit before it rains again. that ok with you? | 20:56 |
hatch | bak | 20:57 |
benji | rick_h_: I have to go to my appointment now, but here is a WIP branch: lp:~benji/+junk/sanitize-constraints-and-other-fun | 21:00 |
benji | rick_h_: as best I can tell we just need to fill in the tests and implementation for the sanitize_constraints function and we'll be good to go | 21:01 |
benji | I invision the remaining tests to be in their own class and directly excersize the function | 21:01 |
benji | after that we just have to do some test deploys and we should be done | 21:02 |
hatch | Makyo: not sure what happened it didn't get the tests before https://codereview.appspot.com/21430043/ here is a proposal with the tests | 21:02 |
hatch | but thanks for both qa's | 21:02 |
rick_h_ | benji: rgr, I'm working on broken charmworld deploy right now, but will see what I can do | 21:03 |
benji | rick_h_: thanks | 21:03 |
hatch | gary_poster: so looks like all of the urgents are done - shall I now switch to the spreadsheet? | 21:11 |
gary_poster | hatch, lemme take a look at the kanban...while you review this, please? :-D https://codereview.appspot.com/22500043 | 21:11 |
hatch | it's on! | 21:12 |
Makyo | hatch, ah, cool, thanks | 21:13 |
gary_poster | hatch, hook up the import control? That's on the spreadsheet *and* in the bundle lane *and* hopefully fast. :-) After that, yeah, spreadsheet. | 21:14 |
hatch | got it! | 21:15 |
gary_poster | cool thank you | 21:15 |
hatch | gary_poster: in this branch I deleted the importexport.js file that confuses everyone | 21:47 |
gary_poster | hatch sounds good. It had confused me too :-) | 21:48 |
gary_poster | benji, rick_h_, I need to go in 9. Can be back later. Where are we? https://plus.google.com/hangouts/_/calendar/Z2FyeS5wb3N0ZXJAY2Fub25pY2FsLmNvbQ.vq3atlo18hgrrbfqbii5cgqch4 ? | 21:52 |
* gary_poster back in 2 hours or so | 22:02 | |
huwshimi | Morning | 22:09 |
hatch | morning huwshimi | 22:14 |
huwshimi | hatch: Welcome back :) | 22:15 |
hatch | spanks - glad to be back | 22:15 |
hatch | albeit only for a couple days haha | 22:15 |
huwshimi | hatch: Are you off next week? | 22:18 |
hatch | huwshimi: only Friday and Monday, going snowboarding | 22:20 |
huwshimi | Oh nice! | 22:20 |
hatch | yeah I'm horribly addicted to it | 22:21 |
huwshimi | hehe | 22:25 |
hatch | huwshimi: this morning https://plus.google.com/118445028821328031751/posts/R5RRM1QRxNt | 22:25 |
huwshimi | hatch: Woah! | 22:27 |
huwshimi | hatch: Is that at your house? | 22:27 |
hatch | right beside | 22:28 |
huwshimi | wow | 22:28 |
hatch | yeah it's awesome here in the winter :) | 22:29 |
hatch | if you can handle the temperatures that is haha | 22:29 |
huwshimi | hatch: I think we had about two minutes of snowflakes amongst the rain this year | 22:30 |
hatch | haha - most would prefer that :) | 22:30 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!