=== schwuk_away is now known as schwuk [09:20] luca__: http://10.20.69.166:2464/ [09:48] rick_h, done [09:48] hatch: woot! [09:51] .c [09:51] Makyo, qa'ing now [10:17] bcsaller: https://codereview.appspot.com/13309044 thanks! [12:33] rick_h: What's the nose package I need to install to be able to run single tests? [12:35] huwshimi: https://pypi.python.org/pypi/nose-selecttests/0.4 [12:35] rick_h: Cheers [12:35] huwshimi: so you can bin/pip install nose-selecttests and then you can run tests with -t XXXXX [12:42] Makyo: https://codereview.appspot.com/13441046/ [13:02] bcsaller, I landed your test fix as part of this branch, since my last one broke ci [13:02] ahh, ok, I merged trunk and didn't see it so I did finally propose [13:02] Still lboxing :) [13:34] orangesquad, bac or benji: Could you please review https://code.launchpad.net/~abentley/charmworld/use-latest-charm/+merge/183661 ? [13:34] * sinzui looks [13:35] wow, fast on the trigger this morning [13:35] benji: [13:35] benji: Was basically done on Friday. [13:37] abentley, r=me [13:56] curses, no IE love https://developer.mozilla.org/en-US/docs/WebAPI/Using_Web_Notifications [14:18] orangesquad: i have returned from having my stitches taken out. sinzui, i'll be free for our normal 1x1 time. [14:21] jcsackett, orangesquad: let's hangout in 9 minutes to do some orienteering [14:23] sinzui: +1 [14:32] jcsackett, adeuring https://plus.google.com/hangouts/_/538341c5065bcf63afe4dfc783cbb34db8a826b1?hl=en [14:32] sinzui: thanks; fighting G+ [15:03] sinzui, abel: Test runs used to take 4 minutes. Now, they take 3 hours: https://jenkins.qa.ubuntu.com/job/charmworld-merger-trunk/buildTimeTrend [15:04] abentley, adeuring Lp's test runs take 35 minutes. I think we need to save face and make runs fast again [15:05] sinzui: agreed... [15:33] hey guys [15:33] any word on when we'll have more readable URLs land? [15:33] jcastro: ummm...not at the moment. [15:34] ugh, this sucks [15:34] ok then, I'm going to use the github URLs! [15:34] hatch: https://ec2-54-226-194-198.compute-1.amazonaws.com/:flags:/serviceInspector/ [15:44] jujugui review please? https://codereview.appspot.com/13443044 [15:48] hatch: when you have a minute can you look to see if you think I missed any topics in https://codereview.appspot.com/13415045 [15:57] sinzui: I think we could avoid Tarmac by using this: https://wiki.jenkins-ci.org/display/JENKINS/URLTrigger+Plugin [15:59] Yummy [16:00] call? [16:01] benji: i'm here [16:02] bac: what you talking about, Willis? [16:02] jujugui: are sprinters going to participate in our daily call? [16:02] benji: i mean, i'm here and willing to have a call [16:02] oh [16:02] I thought you were saying you were on the call already [16:03] no sir [16:03] we are really busy trying to push back on having to do work ;) [16:04] bac: benji gary says yes we'll do a hangout shortly and will update [16:05] hangout url will be offered in the near future? [16:05] hatch: I did send you a link before [16:05] rick_h: eta on "shortly"? this hour? [16:06] bac: I *think* so [16:06] bac: I interrupted his current conversation to bring this up and he said yes, he wants to chat and will send a url [16:06] but no specific timeline given atm [16:10] rick_h: https://codereview.appspot.com/12323043/ has the instructions you wanted [16:11] benji: i'm getting F and E in trunk tests. :( [16:11] :( [16:12] adeuring: benji and i notice 'make test' on charmworld is insanely slow. is that a result of your es test fixes perhaps? [16:12] bac: I'm afraid so :( [16:12] bac: I made wait_for_green_status a noop and my subset of the tests went from 186 seconds to 6 seconds. [16:12] It's hard t [16:13] pproblem for me right now is that the tests a _cirrently_ fast again [16:13] without any apparent related change... [16:15] benji: i did the same and ran 'make test' in trunk resulting in [16:15] Ran 721 tests in 53.682s [16:15] FAILED (errors=19, failures=6) [16:16] adeuring: with trunk my tests ran for 10 minutes and didn't get past the first line of dots [16:17] crazy... abentley suggested to s/green/yellw/ in ElasticSearchLient.wait_for_green_status(). That might help butmight reintroduce the previous psurious failures... === schwuk is now known as schwuk_away [16:36] bac: have time for a review? https://code.launchpad.net/~benji/charmworld/bundle-feature-ui/+merge/183709 [16:36] benji: sure === schwuk_away is now known as schwuk [16:36] bac: I assume its just you and me on charmworld now, right? [16:37] (after the orange guys finish up any branches they are working on) [16:37] benji: yeah. glad we don't require two reviews [16:37] heh === TheRealMue is now known as TheMue [16:59] abentley, sinzui: in the wait_for_green method, we're passing a timeout to client.health as an integer. the docs for es looks like it wants a string that includes units, e.g. '30s'. does the python client convert those? it doesn't look like it to me. [17:01] bac: timeout – Number of seconds to wait for each request before raising Timeout [17:01] http://pyelasticsearch.readthedocs.org/en/latest/api/#module-pyelasticsearch [17:02] bac: abentley oh, but I see that some methods do want a string like '5m' [17:03] I think you are correct bac. The health check wants a string with units [17:03] sinzui: thanks for the link to docs! i hadn't found them before. === schwuk is now known as schwuk_away [17:05] sinzui: even if i wanted an integer, those docs say 'number of seconds'. it looks like our code thinks it is milliseconds. [17:05] i.e. the default is 30,000 [17:06] We have created a very patient test suite [17:06] sinzui: but we have impatient developers [17:20] oi, bin/es-update takes forever now too [17:29] benji: how do i see your new page? [17:32] bac: you have to go to a URL like /bundles/wiki-bundle/wiki/featured/edit [17:33] as far as I can tell there is no way to click through to a charm featured edit page so I didn't make a way to do it for bundles either [17:34] benji: so it has to be promulgated? [17:34] bac, benji, hey. I'm sorry I missed the call. Do you want to have a quick call now to catch up, or talk tomorrow? [17:34] gary_poster: i'm free to talk now, if you have time. [17:34] bac: yep; that is the same as for charms (only promulgated charms can be featured) [17:35] benji: so do you have test data? i've just been loading the two bundles from LP and neither are promulgated [17:36] bac: I did the same and it worked for me; maybe my understanding of how it works is flawed [17:37] gary_poster: I had just started lunch. I don't think I have anything significant to discuss. [17:52] benji: when i try to go to http://127.0.0.1:2464/charms/precise/apache2-passenger/featured/edit i get 403 - Forbidden [17:52] when i try to login i get 'resource not found.' [17:52] benji: when you're back from lunch we should talk [17:55] bac, benji, I don't know if you replied: the network kicked my off almost immediately after saying hi to you :-P [17:55] sorry [17:56] gary_poster: we said we'd be glad to talk to you tomorrow since it is so late over there and benji is eating lunch. [17:56] bac, cool, thanks. :-) talk you you tomorrow, then [17:56] gary_poster: he and i are talking amongst ourselves frequently and don't have much to say [17:57] ok cool bac. I wanted to let you know what is going on here [17:57] so will talk tomorrow [17:57] gary_poster: ok, have a good evening. [17:57] thanks, have a nice afternoon! [17:57] bye [17:58] sinzui: my critical bugfixes have landed. I would like to do a release. Is it safe to do that? [18:07] abentley, it is [18:26] bac: I'm back [18:31] jcastro m_3 marcoceppi: I believe the issues you reported Friday are all fixed. [18:31] Well, the ones I was working on. I don't know what other issues you might have reported :-) [18:31] abentley: which issue? [18:32] oh, the rack/rails one? [18:32] jcastro: Yes. [18:37] sinzui, benji, bac: I've deployed r377 to production. [18:37] k [18:37] thank you abentley [18:42] benji: sorry i had to run an errand [18:42] no worries [18:43] benji: guichat? [18:43] sure [18:55] sinzui: I am setting up a jenkins instance, and trying to run the test suite, but it is not installing pip. Any ideas? http://162.213.35.27:8080/job/Test%20charmworld/ws/bin/ [18:58] * sinzui thinks [18:59] benji: when you resolve your conflicts be sure to import pymongo into helpers [19:00] abentley, check the charm. I think there was a hack for pip...well maybe to install a specific pip lib [19:00] * sinzui visits a meeting [19:01] sinzui: pip is installed on the host, and the only reference to pip in the charm is a one-liner in config-changed that uses pip. [19:17] abentley, I have pip installed as /usr/bin/pip not /bin/pip [19:17] oh, but this in our tree [19:17] sinzui: I have both. [19:22] benji: try this: http://staging.jujucharms.com/login/openid [19:24] abentley, sinzui: we have a problem: http://manage.jujucharms.com/login/openid gives a 404. any idea what might've caused this problem? [19:25] bac: no, but I see staging is also affected [19:26] sinzui: I wonder if it's related to your routes work? [19:26] bac: hmm, that's not good [19:27] abentley, sinzui: i'm trying r372, before sinzui's branch [19:27] bac, abentley, benji, surely this is caused by my URL changes to support juju-gui promulgated charms [19:27] 372 works [19:28] 373 breaks [19:28] bac: confirmed. [19:28] bac Since login can happen against anything, we might need to specify series or "login" specifically [19:30] bac: we don't have a route for login, only logout. So routing's promulgated charm rule will always win [19:33] bac, we can bac my rev out, remove "# Unnamespaced promulgated charms" from routing, add routing for login add a factory/custom_predicate that knows charm series [19:33] sinzui: how does login work in the absence of the promulgated charm routing rules? [19:34] sinzui: is there a default that it hits? [19:34] I was just going to read how it works [19:35] sinzui: so there is an 'add_ubuntu_sso' after the routing is built. i guess that adds a default [19:37] bac, I see, and it is always last :( [19:38] sinzui: but it looks like that call can be made in routes.py before the promulgated [19:39] bac, I agree. I don't know why it wasn't done like that. Maybe charmworld was intended to be separate [19:40] http://paste.ubuntu.com/6060075/ [19:40] sinzui: i think that may work but i don't have any promulgated charms in my test data to see if that routing still works [19:41] sinzui: did you add test coverage for those routes? [19:41] not that our tests work or anything [19:41] bac, I certainly did! [19:41] \o/ [19:42] bac, login is from the first week we joined the project. [19:42] * sinzui find children [20:02] I guess I just assumed that the last thing you add would have the highest priority. [20:10] abentley, I think the order of registration is the order of matching. I had to move rules when adding the new urls to not collide with historic urls [20:17] bac, did you tell charmers that they cannot qa/feature any charms at the moment? [20:18] sinzui: i did not [20:18] sinzui: i'm been caught up trying to get a fix [20:19] I will tell them, Thy might no ever know since featuring and QA is a rare activity [20:20] sinzui: you will send an email to juju or to ~charmers directly? [20:22] bac charmers is a highlight word in #juju. [20:24] benji: could you review and QA/test https://code.launchpad.net/~bac/charmworld/fix-login/+merge/183757 [20:26] bac: sure [20:30] sinzui: if you could glance at ^^^ that would be nice too. [20:31] sure [20:37] bac: [20:37] The branch looks good and QA passed. [20:37] Reply [20:38] s/Reply// [20:38] benji: cool. did you try http://127.0.0.1:2464/precise/bonnie ? (it passes) [20:38] nope, I did apache [20:38] bonnie looks fine too [20:39] benji: and all of the tests passed? [20:39] bac: oh, I forgot the tests (I'm really not feeling well all of a sudden). Startin gth etests now. [20:41] benji: when that lands and is merged i can QA your branch. [20:41] bac: I didn't comment out the slow function, should I have? [20:42] benji: don't know. nothing works on my vm [20:42] bac, can't you change the tests to wait for yellow? [20:42] sinzui: didn't help [20:42] the tests are slowly passing, so I'll give them some time [20:43] sinzui: my VM seems to be pathological wrt these timeouts [20:43] egads! [20:43] benji: in the good old days of say r370, how long did it take 'make test' to run on your machine? [20:43] mine was always about 74 seconds [20:43] I don't know exactly, but not much more than a minute. [20:44] yeah, so clock time mine isn't that slow. i'm not sure why it causes the es stuff to go nuts [20:45] abentley: thanks!! [20:45] I'll test it this afternoon [20:48] benji: i also have this review stagnating: https://code.launchpad.net/~bac/charmworld/bundle-icon/+merge/183730 [20:50] bac: looking [20:52] benji: did the tests ever finish? [20:52] still running [20:53] oi goi oi [20:53] bac: sorry, I don't have the mental accuity to understand that branch right now; the only thing I can think right now is that I'm surprised that much code is needed bceause we are already storing all the bundle's file at ingest time [20:53] benji: they werent' being put into gridfs [20:54] benji: but it can wait until tomorrow. [20:54] thanks [20:55] there are about 100 dots thus far [20:55] (no errors or failures) [21:01] bac: tests still running; I bet it takes 20 more minutes; I'll tell you when it finishes [21:03] sinzui: do you think it wise to try to push this branch to production tonight? i'm unsure about webop coverage. [21:03] I can [21:03] sinzui: i'd like to shepherd it if you think it is feasible. [21:04] sinzui: gotta get it tested, landed, and on staging first [21:26] 200 dots or so [21:26] sinzui: after changing green to yellow i got [21:27] Ran 724 tests in 105.753s [21:27] FAILED (errors=7, failures=4) [21:27] benji: fwiw, i stopped elasticsearch, removed the storage in /var/lib/elasticsearch, and restarted it. my test sped up a lot. [21:27] yuck [21:28] bac, I will have time to run your branch in about 10 minutes [21:34] bac: do you want me to try that or let this run finish? [21:34] benji: how many dots to go? [21:35] benji: and there are no failures yet? [21:35] bac: no failures, 220 dots; I don't know how many are left [21:35] oi. well 724 tests. [21:35] yow [21:35] ok, I'm trying the speedup [21:41] bac: I don't think this is any faster. I'm giving up. I vote you merge your branch as-is and you and I work on the test issue as critical in the morning. [21:41] benji: ok. [21:41] benji: have a good night [21:42] thanks [22:16] sinzui: how often does tarmac run? my branch was approved 31 minutes ago but has not merged [22:19] bac. every 10/15 minutes, but as abentley noted in the channel, Jenkins takes 3h to run the test suite [22:19] sinzui: oh, my [22:20] sinzui: sorry i missed that comment [22:20] I favour rolling back adeuring's changes. It takes too long to respond to a real problem...Lp's test suite takes 35 minutes...thank you former yellow squad [22:28] sinzui: i'll do that tomorrow [22:28] it is insane