[00:00:28] I don't think those people did null edits, I think that edit conflicts or something gets saved [00:00:47] but still there are lots of cases where people "conflict with themselves" [00:01:17] they're technically null edits [00:01:45] yeah.. right [00:02:18] > [00:02:18] This query had a minor typo in it ("trcmin_rc_cur_id" --> "trcmin.rc_cur_id"). [00:02:20] Here are the results from enwiki_p (warning: 8004 results): [00:02:23] . [00:02:25] > [00:04:49] why are Brooke's tables prettier [00:05:23] people edit warring on Aeschylus, really? :) [00:05:26] mysql -BN [00:05:29] Because he used that. [00:05:31] Batch mode. [00:05:46] Which is vastly more helpful for moving the data around. But sucks for presentation. [00:06:53] ok /me ignorant [00:07:56] You don't have a Toolserver account? [00:08:15] Oh, you do. [00:08:21] If you play around with MySQL, it'll make sense. [00:08:54] well, there are also other things I can do and actually do with it [00:09:24] * ToAruShiroiNeko plays with mysql [00:09:28] drop table here and there [00:09:29] <# [00:09:30] <# [00:09:36] <3 Dammit1 [00:09:39] and yes sometimes I run others' MySQL queries sometimes [00:09:42] ... [02:24:46] !log LocalisationUpdate completed (1.20wmf5) at Mon Jun 25 02:24:46 UTC 2012 [02:24:57] Logged the message, Master [02:49:09] !log configured mediawiki-commits to discard mails from gerrit pending resolution of the "implicit destination" issue [02:49:15] Logged the message, Master [02:51:55] TimStarling: I had already redirected gerrit@ from root to :blackhole: [02:52:05] so at least root@ would not get those [02:52:16] but I guess mediawiki-commits moderator would [02:54:12] apparently the committer is in the envelope sender [02:54:36] so it was spamming our committers every time they committed anything [02:54:45] with rejection notices [03:11:48] that's... just wrong [03:13:29] just checked the headers, it looks like the committer is not in the envelope sender [03:13:46] but our committers are definitely complaining about backscatter [05:40:15] First time I can recall seeing this: "Due to high database server lag, changes newer than 55 seconds may not appear in this list." And on a Sunday night? That seems strange. [05:46:04] Pine: not unheard of [05:46:07] @replag [05:46:08] jeremyb: No replag currently. See also "replag all". [05:46:11] @replag all [05:46:12] jeremyb: [s1] db38: 0s, db36: 1s, db32: 0s, db59: 0s, db60: 0s, db12: 0s; [s2] db52: 0s, db53: 0s, db54: 0s, db57: 0s; [s3] db39: 0s, db34: 0s, db25: 0s, db11: 0s [05:46:13] jeremyb: [s4] db31: 0s, db22: 0s, db33: 0s, db51: 0s; [s5] db35: 0s, db45: 0s, db44: 0s, db55: 0s; [s6] db47: 0s, db43: 0s, db46: 0s, db50: 0s; [s7] db37: 0s, db56: 0s, db58: 0s, db26: 0s [07:38:07] hello [07:53:36] hashar: ! [07:53:41] !! [07:53:41] A red exclamation mark (!) in the recent changes or on your watchlist means that edit is not patrolled yet. Read more about patrolling at https://www.mediawiki.org/wiki/Help:Patrolled_edits and configuration at https://www.mediawiki.org/wiki/Manual:$wgUseRCPatrol [07:53:52] heh:) morning [07:54:13] still in Europe ? :D [07:54:19] hashar: we need to push out a changed redirects.conf and we need your help:) [07:54:25] good morning, hashar and mutante [07:54:28] oh really [07:54:29] to clarify changes svn/git [07:54:41] which change? is that on production? [07:54:47] I am not really willing to play with prod apaches [07:54:56] that can lead to some nasty squid cache corruption [07:54:58] ;D [07:55:03] sigh, that doesnt sound good [07:55:08] yes, production [07:55:11] and we need it today :p [07:55:17] ohh [07:55:17] and yes, there is a change in SVN [07:55:24] I though we were using git [07:55:30] operations/apache-config repo [07:55:40] that is exactly the question [07:55:43] the equivalent one in svn should be read only now [07:55:48] arr [07:55:56] but there is no deploy script? [07:56:01] Sam did the migration while we were in SF beginning of may [07:56:07] let me check on fenari [07:56:07] and still svn. and the apache-sync would now sync .git [07:56:12] as there is .git by now [07:56:16] there is no deploy for sure [07:56:33] but even Sam said to check with you :p [07:56:41] so conf are /home/wikipedia/conf/httpd [07:56:49] and if we have no deploy yet, how can we disabled the old [07:57:21] so hmm it is a bit of a mess actually [07:57:22] hashar: yes, i know. and try "svn status" and "git status" [07:57:25] it now has both [07:57:30] the idea is that the public files are in the git repo operations/apache-config [07:57:34] for people to tweak them [07:57:43] and run sync-apache-simulated [07:57:46] additionally we have kept the svn repo there to track the private files [07:57:52] you can see it rsying .git stuff to /usr/local [07:58:09] but nobody did that before [07:58:12] hmmm [07:58:32] there is a few changes in redirects.conf , untracked to git [07:58:37] can we use sync-file? [07:58:56] there were also changed to redirects.conf that are live, but were not commited to svn [07:59:00] ? [07:59:05] s/changed/changes [07:59:09] well [07:59:21] so if you look at git with git diff redirects.conf [07:59:26] you see several changes [07:59:36] but they have been committed to svn according to svn log | less [07:59:43] i wonder how that last addition has been pushed to cluster [07:59:47] r73 : robh added wiki-pedia.org redirects to wikipedia.org [07:59:52] the one that adds "educacao" -> pt.wp redirect [08:00:00] r74 : dzahn existing educacao redirect which was live ... [08:00:22] need to add them to git :) [08:00:29] ack, it was sitting around in "svn diff" , but already live on cluster [08:01:32] I will git add them all in the same commit [08:01:35] and publish that to gerrit [08:01:47] thats cool and thanks, but how does it help if we cant deploy? [08:02:29] i mean, do we need to change sync-apache to add an ingore for .git to the rsync command? [08:03:43] probably [08:04:02] hmm, and svn is not read-only yet [08:10:23] here are the live hacks : https://gerrit.wikimedia.org/r/#/c/12841/ [08:10:35] tracked in svn already [08:11:17] ack, those are good. but you know what i dont see yet? how did rob push his change on the cluster? [08:11:36] so that was the first step [08:11:37] because it doesnt look like he used sync-apache [08:11:41] sync the svn / git repo [08:11:53] as for syncing apache, I have really no idea how it is done nowadays :( [08:11:58] I would say sync-apache [08:12:45] sync-apache-simulated does the copy so I am not sure it is really simulating anything [08:13:20] i made that one [08:13:31] it is just a copy of the other script, or at least was at some point [08:13:39] and added an "-n" to the rsync command [08:14:04] -n, --dry-run perform a trial run with no changes made [08:14:04] ohhh [08:14:12] well hmm bla [08:14:35] so that is where i saw it _would_ rsync .git [08:14:46] well hmm symbolic link and testing if $0 == sync-apache-simulated would have been great :] [08:14:49] anyway [08:14:52] i guess we want to ingore .git and use the old way to sync [08:15:07] the git repo is only to make the apache conf public and available in gerrit [08:15:22] there might be a plan to have apaches uses git to fetch the files [08:15:29] but for now, just use the plain old rsync from fenari [08:15:53] I know Ryan has a plan to have MediaWiki deployed on apaches using git [08:16:01] i like the part of having it in gerrit and public, way better [08:16:11] but yeah, we should have new sync scripts [08:16:20] one major issue we ahve [08:16:22] that pull from git to fenari, and then do the old stuff [08:16:33] is that we never know the change we are pushing -;( [08:16:36] and also remove the svn client [08:16:49] well we still need svn to track the private files! [08:17:04] hrmm, true, yea [08:17:06] and some tools are only in svn for now :( [08:17:12] but yeah long term [08:17:18] we need an apache-config-private or something [08:17:46] * mutante nods [08:18:00] so hmm [08:18:32] I got lot of pending changes on fenari : [08:18:32] diff -u -w -b /home/wikipedia/conf/httpd /usr/local/apache/conf [08:18:56] maybe fenari no more receive conf updates :5 [08:19:22] how did the educacao redirect get live on cluster?;) [08:19:38] well someone edited the redirects.conf on fenari [08:19:41] did a svn commit [08:19:44] then sync-apache [08:19:45] (then i could do the same for the ones we need today), and then take our time to do it all right [08:20:00] it looks like sync-apache does not deploy the conf on fenari :-(((( [08:20:06] hashar: but thats the thing, it was NOT committed in svn, yet live [08:20:16] hashar: that is exactly what made ME commit it , see above [08:20:23] OHH sorry :( [08:20:28] I am misunderstanding you [08:20:33] yea, and it says something about not working on puppet hosts [08:21:20] so svn revision r74 seems to add a educacacao.wikimedia.org server alias to redirects.conf [08:21:30] cd /home/wikipedia/conf/httpd && svn diff -c r74 [08:21:40] r74 : dzahn existing educacao redirect which was live [08:21:41] ack [08:21:53] when i did that the redirect already worked [08:21:59] ohhh [08:22:00] hence "existing" [08:22:04] did you just commit? [08:22:08] yes [08:22:14] so someone did the change and did not commit [08:22:19] and just because i wanted to commit MY change as well [08:22:24] and this was sitting there in "svn diff" [08:22:40] yeah need to blame whoever did the change and forgot svn commit [08:22:43] yes, and someone also did not rsync from there to /usr/local/ [08:22:48] OHH [08:22:51] but someone managed to get it on the cluster [08:23:07] definitely not me [08:23:30] yeah, i know. just wondering how it was pushed out [08:23:38] * hashar wishes we had an audit of who modify which files: -] [08:23:47] so once you modify a file in /home/wikipedia/conf/httpd [08:23:57] to deploy it you just have to ask apaches to rsync [08:24:02] git / svn is irrelevant [08:24:09] they are just for logging purposes [08:24:13] and not used to deploy the configs [08:24:22] so that person did edit then sync-apache [08:24:27] he just forgot to svn commit [08:24:31] i know. but deploying is running sync-apache [08:24:42] and sync-apache rsyncs from /h/w/ to /usr/local [08:24:50] and that would rsync the .git files [08:24:56] and they are not in /usr/local [08:25:18] unless it was done when there was no .git yet [08:25:27] or somebody ran a manual rsync command to ignore them [08:25:30] hashar@srv278:/usr/local/apache/conf$ grep caca * [08:25:30] redirects.conf: educacao.wikimedia.org \ [08:25:47] the file belong to robh dated 2012-06-15 19:05 [08:26:02] are you sure rsync attempt to sync the dotfiles ? [08:26:12] maybe it ignores them just like ls does by default? [08:26:21] or maybe rsync is even clever enough to ignore .svn .git .. ? [08:26:27] it said in the -n dry-run [08:27:28] well I have no idea [08:27:55] hmm [08:28:12] on srv278 the .git directory is there, though it is dated from earlier (aka May 09) [08:28:24] well, ok, i will just make sure the changes i want are in the /usr/local/ path [08:28:32] and we have merged in gerrit now... [08:28:47] i can do a manual rsync and ignore .git [08:28:58] well just copy the .git ? [08:29:02] oh, .git is on servers? [08:29:03] it is already there anyway [08:29:12] and that doesnt bother us? well, ok [08:29:15] ok [08:29:22] + it is in /usr/local/apache/conf/ which is not accessible from outside [08:29:25] so I guess that is fine [08:29:32] the .git only contain public data anyway [08:29:37] kthx [08:29:40] so just go ahead and rsync as usual [08:29:49] looking at the man page, rsync has a -C option [08:29:58] which ignore a bunch of files such as .svn .gitignore .git ... [08:30:05] we might want to use that [08:31:58] "in the same way CVS does". oh, lemme try that, yep. sync-mutante-coffee though [08:32:39] good idea [08:32:46] I am rewriting your sync-apache-simulated :-] [08:33:04] thanks, heh, my diff was _one_ character:) [08:37:02] diff /h/w/conf/httpd/redirects.conf /usr/local/apache/conf/redirects.conf [08:37:16] way too many changes for a recent rsync [08:37:29] brb [08:43:38] mutante: I think /usr/local/apache/conf on fenari is not updated by rsync / sync-apache [08:43:49] fenari might no more be in the 4 groups we sync files too [08:44:01] https://gerrit.wikimedia.org/r/12842 [08:44:13] that change avoid code duplication between sync-apache and sync-apache-simulated [08:44:34] added you as a reviewer [08:45:36] $ grep -c fenari apaches image_scalers snapshot searchidx [08:45:36] apaches:0 [08:45:38] image_scalers:0 [08:45:39] snapshot:0 [08:45:40] searchidx:0 [08:45:47] bahhhhh [08:49:01] ok, hmm. but the good thing about it is i can still just run sync-apache and get it on the cluster? [08:49:38] the out-of-sync with fenari doesnt seem to be a very new thing then [08:50:27] so yeah sync-apache to deploy thing [08:50:31] then I have no idea how you test it [08:50:48] for the last redirects i used Jeffs test script [08:50:53] maybe reload config on the test.wikipedia.org server [08:50:56] ohhh [08:50:56] and manually deployed on srv193 [08:51:10] and tested it on that [08:51:11] were Are D4 tests SCrript ? [08:51:12] ;) [08:51:36] gotta ask Jeff to copy to public place [08:53:54] Help needed [08:54:19] Can anyone update me how to add my name as an entry to Facebook Directory on wikipedia : http://en.wikipedia.org/wiki/Wikipedia:Facebook [08:54:51] Aminuddinshroff: we are not Facebook ? :-D [08:54:58] Aminuddinshroff: ask in #wikipedia-en please [08:55:19] I have added a talk discussion for this ! [08:55:29] Thanks for the info [08:55:56] I am new to Wikimedia, can you update me with how to operate Wikipedia ? [08:57:59] Aminuddinshroff: well this channel is about system administrators [08:58:05] Aminuddinshroff: who maintains the servers :-] [08:58:27] Aminuddinshroff: fro the Wikipedia project, you really want to ask in #wikipedia-en I think [08:58:30] and welcome :-] [08:58:40] Thanks ! [08:58:44] I will do that ! [08:58:57] Aminuddinshroff: http://en.wikipedia.org/wiki/Wikipedia:GLAM/BeginnersGuide [08:59:31] (yeah, it is for the "glam" project, but most of it is pretty good for any beginner, links to tutorials etc) [09:35:02] I am out. Will be back a bit after 2pm (CET) [09:36:23] hashar: it works :) thanks for your time [09:36:54] \O/ [09:36:56] see you later! [09:36:58] redirects seem to work per ticket and LiAnna should be happy. just-in-time production [09:37:01] cya [10:44:14] all dead? DNS problems? "Firefox can't find the server at meta.wikimedia.org." [10:45:00] "ping: unknown host sv.wikipedia.org" [10:46:10] hm, works from the university, not from home [10:54:02] can't reproduce from Europe and http://www.downforeveryoneorjustme.com/meta.wikimedia.org [10:58:30] "ping 91.198.174.225" works fine, but my local DNS seems to be broken [11:01:01] you can try using Google DNS 8.8.8.8 or another one temporarily [11:06:25] all DNSes are responding [12:17:55] bla bla back [12:26:43] !log hashar synchronized wmf-config/InitialiseSettings.php '(bug 37699) Chage logo on uzwiki' [12:26:49] Logged the message, Master [13:16:15] It's deployment day again isn't it? [13:17:23] It would be if I set the calendar correctly [13:17:46] /I read it correctly [13:18:49] Reedy: You mean it's break everything day again? :D [13:19:06] Not everything [13:19:12] I can't directly upset enwiki today [13:19:22] Reedy: Branching 1.20wmf6 soon? [13:19:29] Exactly [13:19:34] Before that... [13:19:49] Ensure that git submodule update --init will work with anonymous checkouts [13:20:03] thanks :) [13:20:04] it does [13:20:23] Wait, why does that matter? [13:20:36] I am checking it out on Labs anonymously [13:20:43] i think we checkout extensions as anon, but mw is ssh maybe [13:20:57] hmm [13:21:08] the error I am getting so far is just the extensions [13:52:24] !log Killed php-1.20wmf4/cache/l10n from mediawiki-installation hosts [13:52:29] Logged the message, Master [14:40:17] i [14:40:20] hi [14:40:41] what is lastrevid and how i could to know what is the current revision id for a article_ [14:40:43] ? [14:41:09] I want to know the actual history id for a article [14:48:43] wilfredor: http://www.mediawiki.org/wiki/Manual:Database_layout should give you the answer [14:49:05] wilfredor: it explains all the tables and their fields. You want to look at `page` and `revision` tables [14:49:36] hashar: page_lasted? [14:51:49] wilfredor: what does it say? [14:52:01] hashar: nothing [14:52:08] This is a foreign key to rev_id for the current revision. It may be 0 during page creation. [14:52:08] [edit] [14:52:09] :) [14:52:25] you can also tell using the PHP classes provided in MediaWiki [14:53:23] mmm, interesting [14:56:40] hashar: thanks [15:43:53] !log Copying php-1.20wmf6 from /tmp to NFS /home on fenari [15:43:59] Logged the message, Master [15:47:39] goddamnit [16:03:29] !log reedy synchronized php-1.20wmf6 'Syncing php-1.20wmf6' [16:03:34] Logged the message, Master [16:08:47] !log reedy Started syncing Wikimedia installation... : Scapping to rebuild message cache for 1.20wmf6 [16:08:52] Logged the message, Master [16:28:12] !log reedy Finished syncing Wikimedia installation... : Scapping to rebuild message cache for 1.20wmf6 [16:28:17] Logged the message, Master [16:28:27] wmf6? [16:28:33] yup [16:28:37] every 2 weks [16:28:50] So, they have already passed? [16:29:16] * vvv can't find the deployment schedule [16:29:21] https://www.mediawiki.org/wiki/MediaWiki_1.20/Roadmap [16:30:12] !log reedy synchronized php-1.20wmf6/extensions/EducationProgram/ 'sync education programf iles' [16:30:17] Logged the message, Master [16:40:29] scapscapscapscap [16:42:41] !log reedy synchronized wmf-config/InitialiseSettings.php 'Disable EducationProgram on enwiki per request' [16:42:46] Logged the message, Master [16:43:36] !log reedy Started syncing Wikimedia installation... : Take 2 [16:43:42] Logged the message, Master [16:53:53] We need some games on fenari to play while scap runes [16:53:55] *runs [16:55:30] <^demon> Reedy: http://www.makeuseof.com/tag/play-games-inside-your-linux-terminal/ [16:56:30] 20 minutes for scap last time.... ~7 to go [16:56:43] <^demon> Look up cat videos on youtube? [16:57:57] ruins? [16:58:12] didn't you bookmark google's pacman doodle? [16:59:08] do yuo know of the html version of survivor? [17:01:00] apergos, does the timing of https://bugzilla.wikimedia.org/show_bug.cgi?id=37225#c48 light a bulb in your head? [17:01:00] is it perhaps related to some dumps processing? [17:01:23] my lightbulbs are mostly off for the day but lemme look [17:03:25] I can't imagine how dumps would have anything whatsoever to do with this [17:03:51] by causing lag maybe [17:04:44] some snapshot hosts were updated on May 28 [17:04:45] I don't think so [17:04:55] yes they werem no that's not relevant [17:04:59] maybe there was some runner update at the same time [17:05:01] new kernel. security updates [17:05:11] and php code? [17:05:21] nope [17:05:40] snapshot mw is updated when all the rest are, via scap. [17:06:17] minor fixes to the python scripts would have had no impact [17:06:29] they didn't touch mysql queries or anything even close [17:06:40] was it updated on May 28, with testwiki and mediawikiwiki ? [17:07:27] scap updates all apps servers and snapshot hosts [17:07:39] if there was a scap on may28 then that will have gone to the snapshot hosts too [17:08:03] en dumps didn't run til 6-2 though [17:08:12] hey run the first 8-9 days of the month and then are done [17:08:15] *they [17:09:56] the table dumps don't lock tablles [17:10:30] but which of the multiversion do they use? [17:10:39] the one of the wiki being dumped? [17:12:13] yes, dumps run the same code branch as a request for a page form the wiki [17:12:35] we clearly broke something on May 28-29 [17:12:47] but if the code wasn't even touched... :S [17:14:17] yeah, not me [17:14:38] I guess you are scouring the system admin log looking for possibilities? [17:16:22] may 30 was the beginning of the move to wmf4 [17:16:50] there was an instance of the failure reported on 29 [17:16:54] !ops do not fuck with krool Tbloemink arab haterfpmpfgpx zjimuupz ksctpx qgd [17:16:55] !ops do not fuck with krool Tbloemink arab hatery pgocehzeeb w qmpad rxfmkqqrl bs jrsijdn [17:16:56] !ops do not fuck with krool Tbloemink arab hatervd [17:16:57] !ops do not fuck with krool Tbloemink arab haters gau mtgfu dhhw ckswcus dmcmisjkh glxusb xn [17:16:58] !ops do not fuck with krool Tbloemink arab haterhkbn syhgq e khat ljvstxxh x rrunmhpwrc aljnuwjlte byaghnz [17:16:59] !ops do not fuck with krool Tbloemink arab haterljgoiomb slaed t [17:17:00] !ops do not fuck with krool Tbloemink arab haterpwsdrirraa hxcdt [17:17:01] !ops do not fuck with krool Tbloemink arab haterrclx ahjggd og x [17:17:03] !ops do not fuck with krool Tbloemink arab hatervujwyn czxsbdemap ub mrapsiz ewkmhzno zgup vmgzosj wvodcv vdhidweda mlhhgtlwjh [17:17:05] !ops do not fuck with krool Tbloemink arab haterscjch rgqa ppzfyvnyam evagbl jfyardul yuvj [17:17:05] !ops do not fuck with krool Tbloemink arab haterwiafhtizj hlwq [17:17:05] !ops do not fuck with krool Tbloemink arab hatercsrt ysfm nry w ssr yewdabbh [17:17:06] oh, this silly spammer [17:17:06] !ops do not fuck with krool Tbloemink arab haterndtifzb [17:17:07] !ops do not fuck with krool Tbloemink arab haterngyynkomtk znaca civ nyogybdr ppkin te dawihggb l umj tjelwvlc [17:17:07] Shhhh. [17:17:08] !ops do not fuck with krool Tbloemink arab haterwgsyzrgjke op [17:17:09] !ops do not fuck with krool Tbloemink arab haterfualyyholp xsugvmzwa y gj pepfy tnqx rkbyt ljazf nvhkrmfg [17:17:10] !ops do not fuck with krool Tbloemink arab hatererjpgbmdcd baor zs zogrjbfzp [17:17:11] !ops do not fuck with krool Tbloemink arab haterhnjs b i xttnr iaukfaholo atgetbltj [17:17:12] hi marienz [17:17:12] !ops do not fuck with krool Tbloemink arab hatercgow mlqjyvgddw vvtpbyklfc fgwfnm spxz n alyxt [17:17:13] !ops do not fuck with krool Tbloemink arab haterjqy jhvbjbrk slzas arxhiydvv azrmjujqf qlbeh [17:17:14] You're disrupting the channel. [17:17:23] Thanks marienz. [17:17:24] :DDD [17:17:24] Thanks [17:17:25] cross-channel abuse [17:17:34] np [17:17:35]