[01:22:07] (PS1) Milimetric: aggregates and dangerous queue fix [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/76871 [01:22:18] (CR) Milimetric: [C: 2 V: 2] aggregates and dangerous queue fix [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/76871 (owner: Milimetric) [08:29:09] hey qchrs [08:29:11] hey average [14:39:45] (PS1) Milimetric: 8 threads is the minimum for 6 rapid-fire reports [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/76919 [14:39:59] (CR) Milimetric: [C: 2 V: 2] 8 threads is the minimum for 6 rapid-fire reports [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/76919 (owner: Milimetric) [14:55:23] (PS1) Milimetric: fixed aggregation, added sum as default output [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/76922 [14:55:37] (CR) Milimetric: [C: 2 V: 2] fixed aggregation, added sum as default output [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/76922 (owner: Milimetric) [15:12:49] doods, I'm going to take the train into the city to meet up with yuri and sumana and some other wmf folks [15:12:50] see yas [15:43:15] (PS1) Milimetric: fixed csv output and errant user_id column [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/76930 [15:43:23] (CR) Milimetric: [C: 2 V: 2] fixed csv output and errant user_id column [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/76930 (owner: Milimetric) [15:56:39] (PS1) Milimetric: fixed decimal adding issue on aggregates [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/76932 [15:56:40] (CR) Milimetric: [C: 2 V: 2] fixed decimal adding issue on aggregates [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/76932 (owner: Milimetric) [16:02:30] poooiiiinnngggg [16:03:02] just the man i wanted to see drdee! [16:03:10] i'm working on a draft email to you all [16:03:12] with the updates [16:03:24] I'll send it just to you and you can forward it on after we do some testing [16:04:17] k [16:04:44] ok, sent email, let's brain bounce in the batcave drdee? [16:05:14] i am untangling my headset [16:05:15] 1 sec [16:08:35] :) [16:08:43] that must be an Extremely tangled headset [16:08:47] it is :D [16:34:41] qchris around? [16:34:46] yes. [16:34:49] awesome [16:34:57] how is the x-cs stuff moving along? [16:34:58] What do you want him to d? [16:35:04] Good. [16:35:11] i am still waiting for feedback from amit [16:35:26] Oh. :-) [16:35:34] average whats' up? [16:35:40] Seems it was not as pressing as he made us believ? [16:35:53] * drdee mumbles not saying anythong [16:36:47] thanks for providing feedback on https://gerrit.wikimedia.org/r/#/c/75349/ [16:37:02] not much, boiling some pasta, in the meantime trying to write some tests for dclass.. [16:37:12] apparently it can't find the lib on my system although it's installed.. [16:37:15] not sure why [16:37:44] average: ... you mean Java cannot load it? [16:38:05] average: or it's just not there? [16:38:33] qchris: the lib is on my system [16:38:41] qchris: however, it won't load [16:38:55] using System.load or System.loadLibrary? [16:38:56] did you look at the old loader logic? [16:39:02] it contains some debugging statements [16:42:31] I get UnsatisfiedLinkerError [16:42:40] UnsatisfiedLinkError [16:45:14] most likely a path issue [16:45:50] it's looking somewhere else than you think it is looking [16:46:09] average: before you add unit-tests please incorporate qchris his feedback [16:46:23] and get the patchset merged [16:46:33] once that has happened we might want to add some more unit-tests [16:46:39] but i am not sure if that is super urgent right now [16:46:55] ok [16:48:15] average: The variant of the code that I checked out yesterday had a try catch block for UnsatisfiedError ... [16:48:37] average: Is the exception coming from that part, or from a different part? [16:49:09] qchris: from that part yes [16:49:38] average: /usr/lib/libdclassjni.so exists, and can be read? ldd shows no 'not found' for it? [16:50:53] average: Does the exception stem from the System.load or System.loadLibrary line? [16:57:27] System.load("/usr/lib/x86_64-linux-gnu/jni/libdclassjni.so"); [16:57:49] qchris: that's what I tried and I still get the UnsatisfiedLinkError [16:58:16] might just be because the upstream jni has a class name that's completely different than the one I'm loading the .so in [16:58:54] qchris: System.load [16:59:35] average: Sholud we pair program after Scrum? [17:28:35] oh tnegrin, drdee, ottomata: I'll be offline in 32 minutes [17:28:43] but I'm all yours until then [17:28:53] you're in SF tomorrow? [18:00:53] ok guys, I'm outa here in a bit [18:01:11] k [18:01:13] call me on my cell 248 877 7079 if any buildings start collapsing :) [18:01:16] i'm about to start poking at ssl on wikimetrics [18:01:37] btw ottomata, https redirection from inside flask looks way harder than I thought [18:01:53] because it usually doesn't know anything about the protocol it's served on [18:01:56] hmm, you can't just send an http redirect header? [18:01:59] is there a router? [18:02:16] i could set a header for you and you could check it [18:02:21] hm [18:02:30] yeah, that would work right? [18:02:34] yup [18:02:35] well [18:02:35] no [18:02:43] because how would flask tell apache to serve https? [18:02:58] nono, it won't [18:03:03] like, apache goes: give me URL [18:03:04] you just redirect to https [18:03:09] yeah [18:03:17] apache says http:blalbla/page [18:03:20] in your router (or whatever) [18:03:21] you say [18:03:37] but i don't know blalbla [18:03:48] if ($page == 'page' && !$HTTP_HEADER['ssl']) send redirect; [18:03:50] or maybe flask knows that [18:03:55] you don't need blabla [18:03:56] you just need page [18:03:57] right? [18:04:19] don't i have to say "redirect to https:blalbla/page [18:04:21] "? [18:04:41] hm [18:04:47] i guess, but you can use the HTTP_HOST header [18:04:49] that hsould be set [18:04:59] right, true [18:05:00] or [18:05:00] HOSTNAME [18:05:02] whatever it is [18:05:07] yeah, should be there [18:05:22] ok, that might work, we'll have to test in production :) [18:05:39] until then, i think having https enabled is great [18:05:51] for now, i'm just going to make it available, and not force redirect [18:05:56] and you can code whatever you can [18:05:57] i'll get this working by monday [18:05:59] oh! [18:06:01] can apache go [18:06:17] if http://metrics/safe -> no redirect [18:06:28] else if http://metric/unsafe -> redirect to https://metric/unsafe [18:06:29] ? [18:06:30] yes [18:06:37] i can do in apache [18:06:38] with wild cards? [18:06:38] :) [18:06:42] ja [18:06:45] mod_rewrite no problem [18:06:45] really? [18:06:46] with full PCRE regexps [18:06:49] 'cause that'd be easier no? [18:06:52] ja probalby [18:06:55] if you know the urls [18:07:02] or, url patterns [18:07:02] ja [18:07:40] /reports/result/ [18:07:58] /cohorts/detail/ [18:08:06] those are the only two that are dangerous as far as I know [18:09:29] sorry ottomata, I meant: [18:09:39] /reports/result/* [18:09:39] /cohorts/detail/* [18:09:41] k [18:09:46] will see what I can do [18:09:54] cool, but just having it enabled works for now [18:10:04] milimetric: do you mind if I experiment with this on wikimetrics live? [18:10:07] and when i'm in the office tomorrow we can think about this redirect [18:10:15] k [18:10:20] just keep in the loop with jaime and liana [18:10:29] because they'll be testing and if you restart they might be confused [18:10:35] hmm ok i'll experiment locally first [18:10:37] just a bit more setup [18:10:39] s'ok [18:11:10] thanks very much sir [18:11:13] i'm taking off now [18:11:16] tty all later [18:12:24] k laatas [18:55:04] woot, check it: [18:55:05] https://metrics.wmflabs.org/ [18:55:09] http and https both work [18:55:19] will redirect if it hits one of hte URLs we are worried about. [18:55:22] to https [19:00:16] ottomata: I get a cert error :/ [19:00:33] yup [19:00:35] ottomata: metrics.wmflabs.org uses an invalid security certificate. [19:00:36] self signed [19:00:40] you ahve to say "ok ok fine" [19:00:43] The certificate is not trusted because it is self-signed. The certificate is only valid for wikimetrics.pmtpa.wmflabs [19:00:45] (Error code: sec_error_untrusted_issuer) [19:00:56] oh ok, I'll grant a temp exception :) [19:01:13] " error! [object Object]" [19:02:54] sumanah: try http://metrics.wmflabs.org/reports/results [19:03:15] The requested URL was not found on the server. If you entered the URL manually please check your spelling and try again. [19:03:27] https://metrics.wmflabs.org/reports/results [19:03:35] perfect danke [22:11:45] (PS1) Rfaulk: add. form request_obj from raw request. [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/77062 [22:12:46] (CR) Rfaulk: [C: 2 V: 2] add. form request_obj from raw request. [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/77062 (owner: Rfaulk) [22:23:24] tnegrin: I'm the one sort-of-hosting Yuri & Andrew at Kaltura today :) [22:23:49] that's cool -- Andrew was looking forward to it [22:24:08] aww yay :)