[00:37:10] (PS1) Jdlrobson: Update editing schema to include latest schema revision [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/119911 [00:43:32] (CR) JGonera: [C: 2 V: 2] Update editing schema to include latest schema revision [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/119911 (owner: Jdlrobson) [09:25:20] So wikmetrics1.eqiad.wmflabs is the host. [09:25:50] Let's see if there is some puppet on it. [09:25:51] I think to make wikimetrics not run as debug we need to override it here: ./manifests/roles/wikimetrics.pp [09:26:07] which is the wikimetrics definition that lives outside teh wikimetrics puppet module [09:26:50] The instance config has: Puppet Class role::wikimetrics [09:27:14] Yes, that is in "manifests/role/wikimetrics.pp" [09:27:33] So I'd say if you want to turn off debug mode, I'd start in that file and go down further. [09:28:18] modules/wikimetrics/manifests/init.pp has a debug parameter that defaults to true. [09:31:02] right. something needs to override it in labs deployment of wikimetrics. will ask otto as it is not clear to me where does the override goes [09:31:06] *go [09:35:16] Oh well. We can find that ourselves ... :-) [09:35:24] manifests/role/wikimetrics.pp:108 looks like a good place. [09:35:39] There the ::wikimetrics class gets instantiated. [09:35:50] We just need to set debug to false there. [09:35:53] (I guess) [09:36:41] You can do so unconditionally and thereby have all wikimetrics instantiations that base on role::wikimetrics have debug mode turned off [09:36:59] Or you could add a separate parameter for it to pass through to ::wikimetrics. [10:04:30] And how wouldwe turn it "on" for the vagrant setup only? Cause changing the default you are right it is easy [10:05:19] but what i do not see is how to do "dev env" overrides (or the other way arround, prod-only overrides) [10:12:18] If you want to turn it on/off depending on whether or not it is in vagrant, you can either add the debug parameter to the wikimetrics role [10:12:54] Or you can introduce yet another wikimetrics global (there are quite a number of them already) [10:13:06] And depending on this parameter/global, you would set debug in the instantiation of ::wikimetrics. [10:13:45] ::wikimetrics_web_mode is an example for a global [10:14:42] (CR) Nuria: [C: 1 V: 1] "I think what we are missing here is a test that recurrent reports are indeed running. I think it should be fine if a subsequent test adds " (2 comments) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/111623 (owner: Milimetric) [10:14:49] For the role with parameters, look for example at role::db::labsdb [10:19:59] (CR) QChris: [C: -1] Add scheduler mode to wikimetrics (1 comment) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/111623 (owner: Milimetric) [10:24:17] (CR) Nuria: "First, my apologies cause I never published my review comments on review #4. Sorry, I have corrected that now." (3 comments) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/111914 (owner: Milimetric) [11:05:01] (CR) Nuria: Add concatenated recurrent reports (2 comments) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/119693 (owner: Csalvia) [11:08:21] ok, qchris will look at role::db::labsdb , might not get to do this until standup [11:23:03] (CR) Nuria: "Sorry, forgot to include this in my prior comment: this change also needs to incorporate ability to "delete" coalesced(?) reports when the" [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/119693 (owner: Csalvia) [12:04:55] hello [12:05:11] nuria / qchris, how goes things? I'll jump on your code reviews right now [12:05:27] early start eh? [12:06:39] you know ..i had been putting comments on reviews for now about a month and a half w/o anyone seen them ... sad... fixed that now. [12:07:04] i am working on your comments on patch #21, [12:12:04] yeah nuria, I apologize, I feel like it's always my fault when new folks get screwed by our tools. But I can't help shaking my fist at the tools a little bit [12:12:30] * milimetric shakes his fist at gerrit [12:12:56] * gerrit shakes fist back [12:16:11] lol [12:16:34] see? what an unfriendly tool [12:17:27] qchris, I'm a little scared about what I'm about to do [12:17:49] I have that train of 3 Changes and I want to make them into four changes [12:17:51] my plan: [12:17:57] git rebase -i master [12:18:09] split the commit I want to [12:18:15] make it into two commits [12:18:19] Sounds good. [12:18:32] then push for review [12:18:33] ok [12:25:20] after you can write this as a UGS [12:28:52] hey nuria, it just occurred to me that I might screw you up if I do this [12:29:03] in trying to do it i realized a bunch of stuff has to be rebased, etc. [12:29:14] let's keep it simple [12:29:33] there's no simple here :( I think I'm about to lose my mind again with gerrit [12:29:42] ok let's keep in mind "smaller changes" going forward [12:29:45] i'm not sure where to rebase the changes [12:29:46] Hey ... rebasing is not gerrit, but git :-) [12:29:52] but let's proceed what what we have [12:30:18] w/o rebasing [12:30:21] but qchris's review was that I should rebase [12:30:47] ? rebase? I voted to split the commit. [12:30:55] yea, which means I need to rebase [12:31:07] sorry, one of the rebases is "git" and the other is "gerrit" [12:31:18] Naw. You can split for the same parent if you want to. [12:31:21] one of them is me rebasing my commits to get new patches that people added on top of them [12:31:30] i'm SO confused right now [12:31:46] ok, no energy to argue about what we should or shouldn't do [12:31:53] let's split because it's the *right* thing to do [12:31:57] but I need help [12:32:08] Sure. What to help with? [12:32:12] two problems [12:32:52] 1. changes happened to the string of 3 commits since I last pushed [12:33:03] 2. changes happened to master [12:33:24] so I have a local branch that I would like to pull these changes all into before I split the first change [12:33:46] ? Ok. If you want to. [12:34:02] i thought i would need to do that, no? [12:34:17] i don't ever want to do anything that i don't need to, in general [12:34:32] Depends a bit how you want to start and how clean your commits are. [12:34:34] So help me. [12:34:40] Which commits are we talking about? [12:34:56] https://gerrit.wikimedia.org/r/#/c/111623/ [12:35:06] What are the others? [12:35:20] 111623, 111914, 112165 [12:35:38] For which of them do you have local modifications that are not yet in gerrit? [12:36:01] that doesn't matter, I'll stash it, it's a simple fix [12:36:36] ok. Problem 1 solved. [12:36:41] nono [12:36:53] sorry, problem 1 is: changes that OTHERS did to the 3 commits since I pushed them [12:37:26] Harrrr. Who didn't read the etherpad where I asked you to leave other's commits alone? [12:37:33] :-D [12:37:49] me [12:37:51] :( [12:37:58] well, everyone, but I was the instigator [12:38:05] Hahaha :-) [12:38:06] god I juuust want to get work done [12:38:07] :( [12:38:20] Well ... that's a different issue. [12:38:24] yes [12:38:26] sorry [12:38:30] Ok so 111623 still fits on master. [12:38:53] hm, weird, so does its child [12:39:00] so why is 112165 saying it's outdated... [12:39:01] 111914 still fits on 111623 [12:39:03] yes. [12:40:15] Because the parent of change 112165 patch set 24 is 8c8d8554c0135253e6956b8505e96771a31e4842 [12:40:55] That is patch set 8 of change 111914, but patch set 9 is current there. [12:41:21] oh I see [12:41:37] that's because of that extra layer of unrelated empty patches on top of all 3 changes [12:41:44] right... so we have to fix this :) [12:41:55] (PS25) QChris: [WIP] Run recurring reports using the scheduler [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 (owner: Milimetric) [12:42:14] you fixed it!! [12:42:18] what?! [12:42:23] So now we have patch set 25. [12:42:24] yes. [12:42:28] All three align nicely now. [12:42:29] I want to learn :( [12:42:31] I feel so stupid [12:42:53] git fetch ssh://qchris@gerrit.wikimedia.org:29418/analytics/wikimetrics refs/changes/65/112165/24 && git checkout -b c112165_24 FETCH_HEAD [12:43:04] git checkout -b c112165-tmp [12:43:13] git rebase 998710335f105b0f9d3dacd094cfccd392071763 [12:43:20] git push origin HEAD:refs/for/master [12:43:22] Done. [12:43:46] That fetches patch set 24. [12:44:02] so you fetched 24, put it on a temporary branch, rebased on the proper commit from 111914, then pushed [12:44:05] Creates a temporary branch for it (actually not strictly required. But happens out of habit) [12:44:11] Yes. [12:44:13] ok [12:44:34] Problem 1 solved? [12:44:51] actually I think I thought this was problem 2 [12:44:56] Ok. [12:45:02] What was problem 1? [12:45:08] "1. changes happened to the string of 3 commits since I last pushed" [12:45:11] problem 1 is getting these changes pulled into my local branch [12:45:24] Ok. [12:45:30] You have them in your stash. Right? [12:45:49] no, these are the changes others did [12:46:03] like, I last pushed, then people added patchets on top of all my commits [12:46:10] and now I'd like to push again, but not erase their work [12:46:54] Now I am confused. Where are the modifications that we need to incorporate? [12:46:59] Are they already in gerrit? [12:47:39] ok, assume nothing past this point, I'll try to clarify [12:47:49] ok. [12:48:05] 1. I made three commits, pushed them to gerrit [12:48:16] 2. I made some changes, so those commits each got several patchsets [12:48:40] 3. others made changes to all of the commits [12:49:00] 4. I now want to make a change to one of the commits [12:49:10] but I first have to get their changes [12:49:54] Sounds ok up to now. [12:50:14] right, so the problem I have is, how do I do that? [12:50:33] Which commits do you want to fetch? [12:51:39] oh ok, I see where we're misunderstanding each other [12:51:45] I want to fetch these into a local branch I have [12:51:48] that has all three commits [12:52:13] when I work on this string of changes, I check out that branch, make changes, rebase, whatever, then just push for review all of it [12:52:24] and it does the right thing to all the changes at the same time [12:52:32] Ok. [12:52:42] but you're saying that I can just delete that branch and get one commit at a time, right? [12:52:47] *one Change at a time [12:52:56] Not sure I understand. [12:53:18] it's ok, I won't waste any more of your time, I think I just have to delete my nice and cozy local branch [12:53:19] it's fine [12:53:33] For which changes in gerrit do you have local modifications that do not fit to the respective most recent patch sets in gerrit? [12:53:58] I don't have any local modifications [12:54:08] I never mentioned that, I'm not sure why you think so [12:54:29] Sorry. It seems I am really not understanding. [12:54:43] sorry too for explaining poorly [12:54:43] * qchris rereads what you wrote. [12:54:46] nonono [12:54:54] what is above is very confusing, don't read [12:55:16] it doesn't matter, I think I understand qchris, no need to bother you anymore [12:55:32] ? [12:55:33] Ok. [12:56:00] Before you go into some mode of working that does not feel productive, please, let's discuss again. [12:56:04] I can explain sometime if you're curious, or if we're having beer and I can't keep my dislike of gerrit bottled up inside me any longer :) [12:56:11] :-P [12:56:42] well, yeah, I think the mistake here was having two people work on a change [12:57:02] I think that's even worse than I thought last fall when we made the same mistake. I just thought we were better at gerrit since then [13:01:09] mmmm.. why wouldn't two people be able to work on a change? isn't gerrit a collaboration tool? [13:01:34] oh, I see what the problem is, charles was probably using git review [13:01:45] which is how that blanket of patches ended up on top of each of the changes in the chain [13:01:53] hahahahahaha nuria, nonono, you're totally misunderstanding gerrit [13:02:09] right, it's atool to make your life harder [13:02:12] now I GET IT [13:02:14] :) [13:03:04] the first step in understanding gerrit is understanding that it is extremely opinionated about how one should structure their commit history [13:03:15] like, opinionated about every little tiny detail [13:03:35] and all this opinion, if not understood and internalized, causes something like a culture clash [13:03:55] for me, this clash will never end, mostly because I don't like opinion in my tools [13:03:59] I like to be the one with the opinions [13:04:53] (PS26) Milimetric: [WIP] Run recurring reports using the scheduler [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 [13:08:27] nuria: fyi - I just pushed my fixes for the tests, so all tests should pass after that last patch ^ [13:08:41] you don't need to rebase if you don't want to, easiest thing is: [13:08:43] git stash [13:08:56] git fetch ... (the fetch command) [13:09:02] git stash pop [13:23:46] (PS11) Milimetric: Add scheduler mode to wikimetrics 7 [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/111623 [13:23:48] (PS1) Milimetric: Add utility functions for scheduling [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/119986 [13:32:00] (PS10) Milimetric: Allow reports to be rerun [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/111914 [13:32:44] nuria: I'm about to rebase this other change again [13:32:56] just letting you know in case you're about to push too [13:33:04] I'm talking about https://gerrit.wikimedia.org/r/#/c/112165/ [13:33:24] because the split of the commit from the previous one is unfortunately bubbled up all the way [13:38:23] (Abandoned) Milimetric: Add utility functions for scheduling [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/119986 (owner: Milimetric) [13:41:17] (PS27) Milimetric: [WIP] Run recurring reports using the scheduler [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 [13:41:19] (PS1) Milimetric: Add utility functions for scheduling [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/119992 [13:42:43] ok nuria, I'm done making changes to all these ... Changes [13:43:17] 1 hour, 41 minutes so I could re-organize my commits in one simple way [13:48:35] (CR) QChris: "Got merged upstream." (1 comment) [analytics/camus] (wmf) - https://gerrit.wikimedia.org/r/119868 (owner: Ottomata) [13:49:02] (Abandoned) Ottomata: maven-shade-plugin generates this file and changed it for Hadoop 2 [analytics/camus] (wmf) - https://gerrit.wikimedia.org/r/119868 (owner: Ottomata) [13:49:20] ottomata: How can you be so fast? [13:49:39] Are you doing nothing but lurking in the IRC channel to see if anything happens? [13:49:45] :-P [13:51:43] are we using the free batcave again? [13:55:18] nuria: I hope so :-) [13:57:42] ok, the free batcave it is [13:59:36] qchris: nuria: ottomata: csalvia: tnegrin: I can't get into the new batcave :( [13:59:44] whaa [13:59:48] https://appear.in/wmf-analytics-batcave [13:59:55] it worked so well yesterday though [13:59:56] what's your problem? [14:00:21] "failed to get access to your camera" [14:00:26] I think it's because I'm docked at my desk now [14:00:48] I'd have to undock and work from my bed... [14:01:31] I'm in the new batcave but if you guys insist, I can try restarting [14:02:01] hmmm [14:02:09] milimetric: We're also in the new batcave, but I cannot see you. [14:02:15] Ah... you're gone :-( [14:03:44] anybody around? I wonder if my change to the X-Analytics header yesterday caused any problems with log parsing.... [14:21:53] i qchris^ [14:23:40] yurik: in a meeting right now. I'll ping you afterwards. [14:23:59] kk [14:29:11] yurik: Fortunately, neither the change nor the introduced bug actually made the machines stop. [14:29:28] qchris, introduced bug? [14:29:35] (At least not for the machinery I use. Others use different machinery and I haven't heard back from them) [14:29:50] I linked it at the change. Let me find it ... [14:30:22] https://bugzilla.wikimedia.org/show_bug.cgi?id=62922 [14:30:28] "Doubled zero tags in varnish logs" [14:30:49] It affects us in at least 7 places. [14:31:06] But luckily the machinery did not stop because of it. [14:31:07] bleh :( [14:31:12] cached issues probably [14:31:26] Do not know. [14:31:48] But regardless of whether caching issues or not... [14:32:03] It's unpleasant to see it happen :-) [14:32:10] sorry about that :( [14:32:21] agree, not pleasant [14:32:30] No worries. No machines stopped. And I caught it soon enough. [14:32:44] thx, making a temp workaround [14:32:57] Cool. Thanks. [15:01:33] qchris_away, https://gerrit.wikimedia.org/r/#/c/120010/ [15:12:27] (CR) Milimetric: [C: 2] Force provision when reloading vagrant [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/119742 (owner: QChris) [15:14:57] omg, I'm sorry I forgot to sign on since restarting [15:15:01] csalvia: I'm back if you tried to ping me [15:15:03] nuria: I'm back if you tried to ping me [16:51:55] (CR) Milimetric: Add scheduler mode to wikimetrics 7 (7 comments) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/111623 (owner: Milimetric) [16:56:31] (PS28) Milimetric: [WIP] Run recurring reports using the scheduler [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 [16:56:57] nuria: I had to change the migration, drop_constraint is not allowed, we need to use drop_index ^ [16:57:39] ok [17:17:18] (PS12) Milimetric: Add scheduler mode to wikimetrics [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/111623 [17:17:32] (PS13) Milimetric: Add scheduler mode to wikimetrics [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/111623 [17:17:49] (PS11) Milimetric: Allow reports to be rerun [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/111914 [17:19:43] (CR) Milimetric: "To those reviewing. Due to the refactor in change https://gerrit.wikimedia.org/r/#/c/112165/, any automated tests written here would just" [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/111914 (owner: Milimetric) [17:21:15] (PS2) Milimetric: Add utility functions for scheduling [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/119992 [17:21:33] (PS29) Milimetric: [WIP] Run recurring reports using the scheduler [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 [17:43:48] yurik: Thanks. The double zero tags seem gone \o/ [17:44:20] qchris, awesome! damn cache, it strikes again :) any way to clean it up retroactivelly? [17:45:01] Yes, we can. [17:45:22] But since tomorrow's files are affected as well, I guess it's best to hold our horses until tomorrow. [17:46:02] Then we need to start the fixing machinery only once instead of twice. [18:00:17] hmmmm hey i need a quick brain bounce [18:00:18] anybody? [18:00:31] i'm in hangout batcave [18:00:51] coming ottomata [18:00:54] danke [18:01:49] :) wait which batcave are you in? [18:01:52] google hangout [18:01:55] k [18:01:59] https://plus.google.com/hangouts/_/calendar/d2lraW1lZGlhLm9yZ19jYjM3bXU0OGNuaHRkN2hybmE4czI3b25hb0Bncm91cC5jYWxlbmRhci5nb29nbGUuY29t.c6j7qidqs491nhi7ovk9pi4h14?authuser=1 [18:02:14] i think we're doing musical batcaves... [18:02:31] google claims I'm in that one... [18:02:33] can you hear me? [18:02:34] i see you [18:02:40] https://plus.google.com/hangouts/_/calendar/d2lraW1lZGlhLm9yZ19jYjM3bXU0OGNuaHRkN2hybmE4czI3b25hb0Bncm91cC5jYWxlbmRhci5nb29nbGUuY29t.c6j7qidqs491nhi7ovk9pi4h14?authuser=1 [18:02:42] not your video, but that you've joined [18:02:43] you see me? [18:02:45] no, I see nothing [18:02:52] i see just myself [18:03:05] trying again [18:03:09] oh it says "it's taking too long to connect" [18:03:17] ok i'm in [18:03:23] now it says "you are not allowed to join this call" [18:03:26] hm [18:03:28] I think google's jealous... [18:03:33] new batcave? [18:03:50] k [18:03:59] https://plus.google.com/hangouts/_/76cpjkssa3mh1mgb5op16uo1ns?authuser=1 [18:03:59] ? [18:04:28] try now [18:07:12] (PS1) Nuria: Adding placeholder variables for public report directory for puppet managed environments. [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/120042 [18:43:57] (PS2) Nuria: Making sure public report directory for public reports does not get into git [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/120042 [18:48:46] (PS3) Nuria: Making sure public report directory does not get into git [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/120042 [18:50:54] (CR) Milimetric: [C: 2] "some people would complain about the blank lines, but maybe we can just do a "spacing and syntax cleanup change sometime later"" [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/120042 (owner: Nuria) [19:02:30] nuria: https://metrics-staging.wmflabs.org/static/public/hi.txt [19:02:32] cool! [19:02:44] milimetric: should I apply thta on production wikimetrics too? [19:04:09] I say yes, ottomata [19:04:40] let me make sure I get the change [19:04:44] (the puppet one) [19:06:25] ok ottomata, looks good, I think it's fine in production [19:07:49] k [19:11:46] milimetric, nuria: https://metrics.wmflabs.org/static/public/ho.txt [19:12:14] i have tested pretty throughly so i am confident about the change [19:13:14] woa ottomata - why you gotta call me a ho? [19:13:46] heh [19:13:48] hi oh! [19:13:50] hi ho! [19:14:01] oh [19:14:02] :) [19:49:07] (CR) Milimetric: WIP Removing cohorts from database (6 comments) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/119343 (owner: Terrrydactyl) [20:11:12] (CR) Nuria: [WIP] Run recurring reports using the scheduler (1 comment) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 (owner: Milimetric) [20:27:32] (PS2) Ottomata: Remove Java part of kraken-etl [analytics/kraken] - https://gerrit.wikimedia.org/r/119724 (owner: QChris) [20:27:41] (CR) Ottomata: [C: 2 V: 2] Remove Java part of kraken-etl [analytics/kraken] - https://gerrit.wikimedia.org/r/119724 (owner: QChris) [21:03:15] (PS30) Nuria: [WIP] Run recurring reports using the scheduler [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 (owner: Milimetric) [21:03:19] (CR) jenkins-bot: [V: -1] [WIP] Run recurring reports using the scheduler [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 (owner: Milimetric) [21:27:42] (CR) Milimetric: [WIP] Run recurring reports using the scheduler (4 comments) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/112165 (owner: Milimetric)