[12:42] jujugui: just noticed staging.jujucharms.com is down. looking into it. [12:42] thanks bac [12:42] frankban, may be a hair late for call, but will try not to be. [12:43] gary_poster: ack no problem [12:43] thx [12:54] jujugui: staging was in a hook config-changed error. re-running the hook cleared it up. can't explain. [12:54] odd [12:59] bac: do we have hook logs? [12:59] frankban: i'll look [13:00] bac: thanks [13:08] http://paste.ubuntu.com/6453180/ [13:37] marcoceppi: I'm sure you're busy with UDS again, but if you get a few minutes I'd like to have that chat [13:56] benji: we can do it during lunchtime at UDS [13:57] marcoceppi: cool, ping me when you're ready [14:01] jcsackett: did your charmworld deployment happen yesterday? [14:02] benji: I'll have to try to get out out today to get the download counts into the api for the gui branch I've got. [14:12] luca__: ping around? or antdillon [14:12] rick_h_, Hey, hows things? [14:13] antdillon: party party. I'm adding the download counts to the bundle token/details and wanted to make sure I'm just adding a 3rd row of metadata and copying the wording from charms "Deployed X times" [14:14] rick_h_, I'd love to say yes but havnt seen the design for a while. I chase down Luca and get back to ya [14:14] antdillon: thanks, appreciate it. [14:15] rick_h_, Are charm names always all lowercase? [14:16] antdillon: hmm, no, nothing enforcing it I can think of [14:19] we should [14:19] :) [14:20] rick_h_: bundle token is is meant to show "X services | X units" [14:20] rick_h_: as seen here: https://docs.google.com/a/canonical.com/file/d/0B7XG_QBXNwY1NFRjWW1MQkNkVWs/edit [14:20] rick_h_: and so are bundle details [14:20] rick_h_: as seen here: https://docs.google.com/a/canonical.com/file/d/0B7XG_QBXNwY1R0Z4YzVXYVlORVU/edit [14:21] luca__: right, but we've added the new ability to track downloads of bundles, which we could not do before [14:21] luca__: so I'm wondering where to stick it [14:21] rick_h_: interesting [14:22] bac: ping, I'm also having issues getting the counts to work in search. I get 0's even though when I hit the bundle directly I get a download count. [14:22] rick_h_: so for now put "X services | X units | Deployed X times on the details only [14:22] bac: nvm, looking a tthe wrong numbers [14:22] rick_h_: argh, I'm not sure where to put it [14:23] luca__: see, why ping you guys :) [14:23] rick_h_: ok [14:23] rick_h_: I might get Spencer to mock it up quickly [14:23] rick_h_: is that ok? [14:23] luca__: yep, not a problem. Thanks [14:30] benji: yeah, we deployed through r464. [14:30] cool, thanks [14:35] -19C feels like -28C [14:36] it's days like this that I'm glad I don't have to drive to work [14:37] rick_h_: have you worked on the charm store any? [14:38] benji: no, I found the code once when I was trying to figure out where it went but never messed with it. [14:38] benji: my understanding is that it's hazmat's baby and he talked about doing a bunch with it at last sprint [14:38] yeah [14:42] gary_poster: when you have a moment I think I have finally found a promise lib that checks all of the boxes https://github.com/petkaantonov/bluebird [14:42] hatch: http://imgur.com/HpgCyIQ [14:42] lol wow that would suck - the worst I've ever had is up to the windows [14:45] benji: looks nice [14:46] rick_h_: I know I do... wait, what? [14:46] benji: :) your link [14:47] rick_h_: which link is that? [14:47] oh sorry, that was hatch that linked it [14:47] doh [14:47] reading irc fail today [14:47] lol [14:47] rick_h_: you read the promise stuff? [14:47] well I'm not going to give hatch credit [14:47] haha [14:47] hatch: went through the readme [14:48] I didnt' expect much beyond "it's a promise, it sucks" from you :P [14:48] hatch: I've been burned by fire too many times, but it says it won't hurt me [14:48] just don't trust it :P [14:48] don't worry, it has candy [14:48] it won't hurt you [14:48] I don't believe [14:49] the cake is a lie [14:50] hatch, cool, will look soon [15:00] hey luca__ you want me in the meeting, or just watching & on irc? [15:01] gary_poster: come on the meeting [15:01] gary_poster: https://docs.google.com/a/canonical.com/presentation/d/1TDkZ5Sm76QEbpNKR5wPOcqyY-D6xu6wqfM-6qwdzULc/edit#slide=id.g7adc252f_0_33 [15:01] great luca__ will do [15:02] gary_poster: what do you think [15:02] gary_poster: ? [15:03] luca__, amazing! a lot of effort and looks great [15:03] gary_poster: join us in the hangout and speak out the dev effort [15:03] about^ [15:04] gary_poster: https://plus.google.com/hangouts/_/7ecpjj966m9ee4cgq98undnp0c?authuser=0&hl=en-GB [15:11] rick_h_: do you need a review? [15:12] hatch: not yet thanks though [15:12] OH KAY! [15:17] rick_h_: that's one heck of a diff lol [15:17] ~2400 haah [15:17] hatch: it's all one test/data/json file [15:17] :P [15:17] it's a tiny branch [15:17] oh lol I didn't open it [15:19] are we going to be pushing the standup because of vuds ? [15:19] yesterday he didn't say so, but might be a strange/light standup [15:22] yeah I'd be ok with pushing it an hour [15:29] brb [15:29] I gota step out and help boost a neighbours car, will be back in 10 [15:29] jujugui: review please https://codereview.appspot.com/29960045 [15:29] bac: I'll take a look. [15:29] bac: I can when I get back [15:29] ok cool [15:29] :) [15:29] heh [15:29] how does one boost a car? [15:30] bac: jumper cables? [15:30] no, those are for jumping a car. does he have booster cables [15:30] * benji was thinking "boost" as in "steal" [15:30] "can you give me a boost" [15:30] hell, he could be using a jack for all i know [15:31] I guess I've heard it said that way [15:31] boost it up so i can get the tyre off [15:31] hah [15:31] yea, I mean we've all seen gone in 60 seconds and boost if the #1 word from the movie I think [15:31] but guess I've heard it both ways [15:32] well, i haven't quite seen that yet. perhaps after my smokey and the bandit marathon. [15:32] what?! well now you've got a task this weekend. [15:32] http://www.cnn.com/2013/10/31/us/new-york-los-angeles-cannonball-speed-record/ [15:32] the remake isn't bad, little cheesy in places but the car scenes and mic'd well [15:35] rick_h_, Hey got a sec? [15:35] antdillon: sure thing [15:35] rick_h_, Is https://manage.jujucharms.com/api/3/search/interesting down? How long for? [15:36] Just started using it for testing lol [15:36] antdillon: oh hmm, interesting [15:37] antdillon: I did request a upgrade but it hasn't happened yet that I know. Looking into it [15:37] rick_h_, Awesome thanks [15:38] antdillon: looks like the upgrade is in progress and things are going boom atm. Will update you when I know more [15:38] rick_h_, Awesome thanks [15:41] gary_poster, bundle workflows in ~20 [15:42] jcastro, ack thx. [15:42] back [15:43] gary_poster: are we moving our standup for the uds session? [15:43] bac: rick_h_ benji to 'boost' is to 'jump start' a car :) [15:43] jujugui heads up the mjc deploy went boom and so the api and such are not returning atm. Working with webops [15:44] uh oh. let me know if i can help rick_h_ [15:44] rick_h_, urg and thank you [15:44] bac, I was going to suggest that we have standup without me [15:44] I'd like to watch the vuds [15:44] ok [15:44] then [15:45] standup as soon as it's done? [15:45] jujugui want to move standup to one hour later? [15:45] bac: did you have a migration or anything that hit ES in your branch? [15:45] any objections? [15:45] rick_h_: no siree [15:45] sure - the chrome dev summit also starts in 1:45 [15:46] bac: k, didn't think so but ES seems up and dead and can't figure out why. The deploy shouldn't have touched it that I can tell [15:46] hopefully our standup won't take more than 30 min ;-) [15:46] haha i'd hope :) [15:46] gary_poster: let's just go around here. nothing from me. :) [15:46] learnt some interesting stuff with it in the background yesterday [15:46] heh [15:47] would like to have call actually [15:47] ok, i'll send my people out for chicken [15:51] bac, I owe you one for yesterday. If you want to bow out for today's call np. [15:53] the link in the calendar wasn't working for me [15:53] http://summit.ubuntu.com/uds-1311/meeting/22004/t-cloud-juju-bundle-workflows/ this one works [15:56] gary_poster: thanks, but i'm eating now. not a problem. i didn't need a big plate of fried chicken and plantains after all. [15:56] ok, still owe you then ;-) [15:57] adeuring: can you join #webops please. ES on mjc went boom and I think it's a matter of marking an unassinged shard back to assignment. I'm looking for people that have messed with ES to help get it back up please? [15:58] rick_h_: ack [16:35] jujugui snow is throwing a wrench in plans, can someone adopt lp:~makyo/juju-quickstart/ensure-juju-lxc to address frankban's concerns? [16:35] rick_h_: who requested the mjc deploy? [16:36] Makyo: just submit it with adopt? [16:36] hatch, with concerns addressed - that's the part I won't have time for. [16:36] Makyo: if there are changes to be made i can handle it this afternoon [16:36] bac, thank you [16:37] Makyo: this is for the card "Install repo and juju" ? [16:37] bac, yes [16:37] cool. [16:38] gary_poster, for the summary for UDS, what do you consider the top 3 things your team is working on for this cycle? [16:39] benji: are you looking at my review? [16:39] bac: I'm trying to QA and having issues (with Juju itself I think) [16:39] oh, ok. [16:39] that QA is tiresome [16:39] jcastro, (1) not clear yet, unfortunately; and (2) it may not be public once it is. [16:39] bac: how long does the gui take to come up after the juju set juju-gui "juju-gui-source=lp:~bac/juju-gui/fix-dnd-bundles-wrt-id" command? [16:40] oh, a bit [16:40] a few minutes. not crazy slow [16:40] gary_poster, ok, so I can just say "continuing Juju GUI improvements" [16:40] jcastro, +1 [16:42] antdillon: it's up, don't know for how long and don't know why, but it's working atm [16:42] jujugui mjc is up and running, it's a confusing mess, but it's alive [16:42] thanks rick_h_ [16:42] rick_h_: anything we can do to fix this? It seems to happen more often than we'd like :) [16:42] rick_h_, good news at least. thank you. [16:43] will bring up on standup [16:43] for the record I know absolutely nothing about how it's setup :D [16:43] Pretty sure it is hatch's fault :-P [16:43] hatch: the best approach is to have a pre-deploy environment similar to production in prodstack [16:44] gary_poster: probably! They should have written it in javascript [16:44] :-) [16:44] for some reason it keeps complaining about the {} I put in there [16:44] :P [16:44] :-) [16:46] ignoring the content of the article.....this plane is bonkers huge http://www.wired.com/autopia/2013/11/dreamlifter-wrong-airport/?cid=co14766074 [16:48] NPR claims they are going to try to take off at 13:00 ET [16:48] I love that 'try' I am going to guess they only get a single chance at this [16:48] it's not like they can abort 3/4 of the way down [16:49] hopefully they can find a nice headwind [16:49] to help lift that thing [16:49] hatch: i saw the antonov a few years ago: http://www.flickr.com/search/?w=77711790@N00&q=antonov it is a bigger plane but the dreamlifter has large cargo volume [16:49] s/large/larger/ [16:50] jeesh [16:51] I saw one of those double decker planes nosed in at the airport once....standing at the window in the terminal and we still didn't even come up to the middle of the nose [16:51] I can't even imagine the amount of power those engines need to generate to move that thing [16:52] this one http://upload.wikimedia.org/wikipedia/commons/d/d4/Airbus_A380_Paris_Air_Show.jpg [16:57] jujugui call in 3 [16:57] ah, thank you [16:57] hah I forgot too [16:57] bac: QA didn't go well (the count didn't increment) [16:58] benji: did you give it some delay? [16:59] bac: (as in between the deploy and the 'juju set juju-gui "charmworld-url=http://staging.jujucharms.com"') yep [17:00] benji: no, as it, check the counts again now [17:00] s/it/in/ [17:01] bac: it incremented! [17:01] yay [17:01] benji: i really don't understand the lag [17:01] is there a cache in front of the app? [17:01] that is odd [17:01] probs [17:03] yea, there's apache and squid in front [17:15] when is your turkey day? [17:15] thurs [17:15] of next week [17:15] ahh ok [17:16] hey frankban one of our italian friends gave us one of these (early) for christmas. do you eat them? any good? http://en.wikipedia.org/wiki/Panettone [17:17] looks like a Babka bread [17:17] (Ukranian) [17:18] bac: it depends, some of those are very good, others are just like eating paper... yeah it's a very typical christmas sweet in italy [17:18] probably the same :) if so, then it should be sort of a sweet bread [17:19] oh frankban you want to talk? sorry [17:19] I was about to head out [17:19] hatch: yes, usually with raisin and candy [17:19] gary_poster: reading the diff, quick call in 5 minutes? [17:20] frankban, going to lunch w/ wife. will see if you are back when I return [17:20] jujugui: the branch i'm landing now for juju-gui has some retry logic to handle the lbox-404 problem. let me know if you see it again. [17:20] if not email? [17:20] gary_poster: sounds good [17:20] ok ty [17:20] btw, lbox-404 has a new album coming out soon [17:20] haha [18:14] jujugui can I get a review please? https://codereview.appspot.com/30210044/ [18:14] sure [18:14] claiming small review since 99% of the changes are the test data [18:18] * hatch waits for the 3 or so more patch sets rick_h_ will commit before he is done his review :P [18:18] hatch: :P [18:18] haha [18:19] it's why you don't review until i ask. I like to push up and self-review and usually find something to fix [18:19] ahh [18:19] I usually try to `bzr diff --old ../../trunk | vim -` before I propose [18:19] saves me from doing the huge lbox delay multiple times [18:20] I like the context change and verification that lbox is happy [18:20] frankban: did you QA matthew's branch? [18:20] well, lbox is slow, but 2:32 for the last run [18:20] bac: no, I was waiting for the fixes [18:20] so not enough to put off getting the nice diff/review space [18:22] hatch: http://nuclearsquid.com/writings/reporttime-in-zsh/ :) [18:22] oh boy I wish [18:22] I'm probably closer to 4m [18:23] on 4m <---- this side of the 4m :D [18:31] rick_h_: have you ever used ZNC for irc? [18:31] hatch: no [18:31] alright - last night I was looking at various irc bouncers [18:32] hatch: no, I just keep my irc window open on a side monitor. I don't get ping's or anything [18:32] ahh - yeah I was hoping to get something that keeps me always logged in so I can move from computer to computer and keep the history [18:33] yea, I get that by running irssi on my ec2 server, but it's disconnected [18:41] frankban: i've made the changes to matt's branch if you have time to QA. lp:~bac/juju-quickstart/makyo-ensure-juju-lxc [18:42] frankban, I'm back but you should be gone. :-) [18:43] gary_poster: heh :-) [18:44] frankban, do you want to talk, or call it a night? [18:48] rick_h_: so this 'downloads_in_past_30_days' when is that passed in? [18:48] I dont' see it in the current diff [18:49] hatch: from the api, it's dumped into the model as the whole object [18:49] hatch: that was updated in the back end, so cfg just *has* it [18:50] oh ok [18:50] so the issue there is that if there are any updates to that attribute from the api, cfg won't be updated [18:50] is that not a concern? [18:51] huh? I don't see the concern [18:51] say the model is created [18:51] then we update the model [18:51] downloads_in_past_30_days will be the new value, but recent_download_count will be the old one [18:51] it's read only, not sure how that would be updated [18:52] so no, not an issue since it's based off of 15min ingest data and isn't written [18:52] ok so we will never update the model once it's been loaded once [18:52] if we start writing bundle/charm models then we've got a LOT more work to figure out [18:52] hatch: right [18:52] ahh ok - so this will eventually be a problem when we implement a proper client side database [18:52] the absolute worst case, is that we cached the model and ingest happened in the last 10min and you don't see the updated count atm because we hit our cache [18:52] but until then I suppose it's ok [18:52] but welcome to caching [18:53] bac: well over my EOD, can do it tomorrow, or please feel free to do the QA and merge if you want. [18:53] frankban: ok. if it isn't merged please take a look tomorrow. [18:53] hatch: yea, this is the same way the current charm model works. I'd like to keep it in sync and how to deal with a client-side-db is interesting, but yagni [18:53] bac: sure! thanks [18:55] done [18:56] gary_poster: so should I be working on my ci chops or my 'how does elasticsearch' work chops? [18:56] rick_h_, ...ci. [18:56] gary_poster: k [18:56] thank you [19:02] hatch: thanks for the review. No good reason to not use value vs valueFn. Updating now. [19:03] cool [19:03] just skipped the brain [19:03] :) sok === gary_poster is now known as gary_poster|away [19:04] using juju lxc if the install script modifies /etc stuff will that be restricted to the lxc container and not spill over to the host? [19:05] hatch: rgr [19:05] hatch: it has it's own /etc dir in the lxc container [19:05] same goes for /var and /tmp etc? [19:05] hatch: yep [19:05] cool - I am jut going to test a guys charm while I'm working and didn't want to run the risk of mucking up my host machine [19:05] yea, lxc ftw [19:05] just watch out for your home dir if you bind that to your lxc containers [19:06] not unless juju does that [19:06] no [19:07] kewl [19:11] hmm debug-log is broken [19:11] that's awesome [19:12] debug-log doesn't work in lxc [19:13] ohh [19:13] well then! [19:16] wheels up [19:30] yeah? [19:30] I missed it I guess heh [19:32] boring. yay. [20:07] that's the best kind of flight :) === gary_poster|away is now known as gary_poster === gary_poster is now known as gary_poster|away [20:35] sometimes the right way....was the first way [20:36] jujugui when debugging model data use `console.table(data)` it works.....awwwweeeesuuummmmm [20:36] cool [20:36] hey blog that or something [20:37] I can do that [20:37] https://www.evernote.com/shard/s219/sh/5629e020-c68a-4d60-a839-72de7228b0d3/61cacd5857605826f28179505d7811d3 [20:37] is what it gives you [20:37] so it's not 'perfect' but very helpful [20:37] looks good [21:15] hey gary_poster|away back yet? === BradCrittenden is now known as bac [21:21] jujugui does anyone remember why we hide the unit action buttons when the unit is pending? [21:21] Don't we want people to be able to retry/remove them? [21:32] hatch: yea, I was playing with the chrome dev snippets and a lot of them use the console.table and very cool for thigns like header dumping and such [21:32] yeah really a fan now [21:32] :) [21:33] hatch: if it's pending can you retry/remove? Don't you have to wait for it to be done and in error before you can retry/resolve? [21:33] ooh, kitkat for my n10 just arrived, be back after install/reboot [21:34] rick_h_: well sometimes it can get stuck in pending [21:34] so that's why I'm curious [21:39] hatch: hmm, yea if it's stuck in pending then I'd think there's a bug there. Curious what the log is for that then. [21:39] well it's not really a bug [21:40] if I have an infinite loop in my install hook [21:40] it'll pend forever [21:40] right, that's a charm bug [21:40] yeah [21:40] but maybe not [21:40] it could be an issue with anything really [21:40] looped internets or something [21:41] I'm just saying we should probably give the user the ability to something about it [21:41] hatch: yea, but let's make it something useful [21:41] I don't know how we can tell if it's useful or not [21:41] if there's something retry can offer, then I'm all for it, but I'm not convinced it's just giving users false hope [21:42] oh....right, but there is no way we can tell that [21:42] hatch: find me a use case where retry would have been useful and maybe showing it can be done :) [21:42] anyway, I'm not sure on why it's not visible. It might just be a juju limitation [21:42] I'm just talking through why it might be one way or another [21:43] right - I can't think of one [21:43] hah [21:43] I've just seen charms stuck in pending before [21:43] then let it be I say. If it turns out to be something useful it can probably change. [21:43] hatch: right, I've seem people say in #juju about charms pending. I just can't say that retry/resolve helped them at all [21:43] so I'd want to verify the right way to fix a 'stuck' charm is exposed to the user vs just "here's some more buttons" [21:44] imo [21:46] right but there is no way we can offer that to the user [21:46] because pending CAN take hours [21:46] or days really hah [21:46] heh, yea. [21:46] once we get logging then that will be awesome [21:47] but for now I think I'll show the buttons [21:47] it's trivial to remove them [21:47] unless of course juju doesn't support retry/remove on pending units [21:47] hmm who would know [21:49] I asked in #juju-dev [21:49] we'll see [22:10] ugh I hate implementing sorting [22:17] hatch: :P [22:17] well the damn things never work as they are supposed to [22:18] I need to go and just write sorting algorithms for a month or so to get better at it [22:18] haha [22:19] annnd of course I missed an = sign [22:19] * hatch shakes fist at javascript [22:19] I had it right all along [22:19] lol