[01:21:08] !log reedy synchronized closed.dblist 'Closing wikimania2011wiki' [01:21:13] Logged the message, Master [01:21:46] !log reedy synchronized wmf-config/InitialiseSettings.php 'Bug 36571 - Please lock wikimania2011 wiki' [01:21:49] Logged the message, Master [01:23:56] !log reedy synchronized wmf-config/InitialiseSettings.php 'Bug 36595 - Please enable Extention:NewUserMessage on ml.wikipedia' [01:23:59] Logged the message, Master [01:25:07] !log reedy synchronized wmf-config/InitialiseSettings.php 'Bug 36595 - Please enable Extention:NewUserMessage on ml.wikipedia' [01:25:10] Logged the message, Master [01:27:28] !log reedy synchronized wmf-config/InitialiseSettings.php 'Bug 36533 - Set sitename to Telugu Wiktionary' [01:27:31] Logged the message, Master [01:30:41] !log reedy synchronized wmf-config/InitialiseSettings.php 'wgShowUpdatedMarker enabled on anything that isn't enwiki or dewiki' [01:30:44] Logged the message, Master [01:36:56] !log reedy synchronized wmf-config/InitialiseSettings.php 'Bug 36663 - Please allow bureaucrats to add and remove autoreviewer status on pt.wiki' [01:36:59] Logged the message, Master [01:39:40] !log reedy synchronized wmf-config/InitialiseSettings.php 'Bug 36522 - Upload link should lead to UploadWizard instead of commons:Special:Upload' [01:39:43] Logged the message, Master [01:42:05] !log reedy synchronized wmf-config/InitialiseSettings.php 'Bug 36506 - Site logo for Tsonga Wikipedia -- ts.wikipedia.org' [01:42:09] Logged the message, Master [01:53:08] got a question regarding mediawiki and redirects leading to a subsection: Is it "wanted behaviour", that when I have a Redirect called "Site_A", leading to "Site_B#Something", that the URL in my browser reads "Site_A#Something" after I clicked a Link to A? [01:54:01] Example from de.Wikpedia: http://de.wikipedia.org/wiki/Bizarre_Love_Triangle Leads to http://de.wikipedia.org/wiki/New_Order#Singles, but the URL in the browser reads http://de.wikipedia.org/wiki/Bizarre_Love_Triangle#Singles [01:56:45] (figured, I would ask this here, since it's a somewhat technical problem ... I'd guess it has to do with the Redir being done via an AJax call or something ... [01:58:34] yes that is desired [01:59:19] ok [02:00:10] took me some 2 times back and forth till I figured, why a link to page a leads me to A#some, till I saw I wasn't on page a at all ... ^^ [02:13:10] !log LocalisationUpdate completed (1.20wmf2) at Wed May 9 02:13:10 UTC 2012 [02:13:14] Logged the message, Master [02:19:33] !log Running cleanupUploadStash.php over all wikis [02:19:36] Logged the message, Master [02:37:02] !log LocalisationUpdate completed (1.20wmf1) at Wed May 9 02:37:02 UTC 2012 [02:37:05] Logged the message, Master [02:44:05] Can I please get some help at https://secure.wikimedia.org/wikipedia/meta/wiki/Meta:Requests_for_help_from_a_sysop_or_bureaucrat#Lost_Password_for_DieWeisseRose ? [02:44:27] Please don't use secure.wm.o [02:44:38] use https://meta.wikimedia.org [02:44:56] I don't understand. [02:45:09] secure.wikimedia.org is deprecated [02:45:09] Okay, now I understand [02:45:32] I will repost the request to meta [02:47:38] Just checked meta and my request is arleady there. [02:48:32] yeah [02:48:35] It's the same site underneath [02:48:43] It's just the way you linked is the old secure method [02:49:36] Shall I link using meta.wikimedia.org or is it unnecessary? [02:50:51] Shall I repost the text of my help request here? [02:51:44] Per the replies, you've got to prove somehow it's your account [02:55:03] Okay, well I have the text and exact date/time of the watchlist token I set last year and I have access to the email address used by Wikimedia to contact me about the Referendum last August. [02:55:26] hmm [02:55:32] If you got contacted, you must have an email somewhere.. [02:55:43] *an email set somewhere [02:57:47] I linked all of the accounts last year, I'm not sure of the terminology but they was some procdure for setting one password for all the various wiki accounts. [03:00:46] Yeah [03:00:56] So is it just your meta account? [03:04:38] I can't access any of my accounts and when I use my username to request a password, I get a message that there is no assocaited email address but I got the referendum message months after I had the password loss. [03:06:11] Which was your original wiki? [03:06:29] You've got an authed email, but no address set [03:07:23] Okay how can I set that email when I can't access the accounts? [03:08:03] You've got an email set on commons [03:08:23] And on enwikiquote [03:08:43] Try using one of those wikis to reset your password [03:10:13] I can set that email to your global account if needed [03:11:16] I just tried both of them but I'll try again. Can you tell me the username for en.wikiquote. When I tried to reset my password I got: "There is no e-mail address recorded for user 'DieWeisseRose' " [03:12:01] https://en.wikipedia.org/w/index.php?title=Special%3ACentralAuth&target=DieWeisseRose [03:12:04] It's the same username [03:12:51] One set on eswiki too [03:13:41] I've just set that email to your global account [03:15:47] Okay, thank you. I just tried a passwd reset via the URL you gave but got "Action throttled". I guess I'll have to wait. Do you what the reset time is ? [03:15:53] a few minutes [03:16:00] which wiki? I can probably override that [03:16:04] <10? [03:16:25] en wikipedia [03:19:15] I just invalidated your user cache [03:19:28] though I think the throttle is seperate in memcached [03:20:23] shall i try another reset then? [03:24:49] well, the reset time is apparently more than 7 minutes because I just another reset attempt and it's still too soon [03:31:36] 5 in 5 minutes apparently [03:31:39] or default is [03:41:12] I just tried again--no go. [03:41:53] would it make anydifference if I tried a different wiki? [04:16:51] Reedy, I just tried password reset but still get "There is no e-mail address recorded for user 'DieWeisseRose'." [07:07:08] Reedy, thanks for update markers; people on en.wiki are complaining about weird behaviou of watchlist without it, but hopefully it won't last long :) [07:27:47] jeremyb, you edit conflict my emails! [07:28:57] {{fact}} [08:03:42] DieWeisseRose2 Hi! see http://viaf.org/viaf/131008281/#Weisse_Rose_%28Resistance_group%29 [08:07:01] gangleri: DieWeisseRose2 ain't here [08:07:32] thanks [08:23:56] hi gangleri [08:24:10] was that JS on test.wiki fixed? [08:25:51] Hi! Nemo_bis hi! please note that an improved version of http://svn.wikimedia.org/svnroot/mediawiki/trunk/tools/viaf/index.html is available - this is a [[authority control]] tool [10:10:15] Hello, Houston, we have a problem ! [10:10:25] Try to edit the last section of the article Montreal. I can't. [10:10:34] http://en.wikipedia.org/w/index.php?title=Montreal&action=edit§ion=21 for instance [10:10:49] I get an error : [10:10:50] Request: POST http://en.wikipedia.org/w/index.php?title=Montreal&action=submit, from 91.198.174.41 via sq78.wikimedia.org (squid/2.7.STABLE9) to 10.2.1.1 (10.2.1.1) Error: ERR_ZERO_SIZE_OBJECT, errno [No Error] at Wed, 09 May 2012 10:06:12 GMT [10:12:13] other pages? [10:12:22] or (as far as you know) just that one? [10:15:21] Not all articles are broken. [10:16:40] very slow saving here [10:16:53] Request: POST http://en.wikipedia.org/w/index.php?title=Montreal&action=submit, from 208.80.152.75 via sq78.wikimedia.org (squid/2.7.STABLE9) to 10.2.1.1 (10.2.1.1) [10:16:56] Error: ERR_ZERO_SIZE_OBJECT, errno [No Error] at Wed, 09 May 2012 10:16:44 GMT [10:16:59] yep [10:17:32] but before that it takes ages [10:17:36] Can someone fix that please ? [10:17:45] so just the one article? [10:18:11] can you edit the article as a whole? or does that also fail? [10:18:19] That fails too [10:18:22] ok [10:18:57] trying https://en.wikipedia.org/w/index.php?title=John_F._Kennedy&action=edit§ion=47 ... saving slow as well [10:19:24] but saved... [10:19:43] I cannot reproduce the error [10:20:31] huh [10:20:57] Nicolas_: can you try with another browser? (I'm trying to eliminate possible causes here one at a time) [10:21:33] Rollback works too [10:21:40] but amazingly slow [10:26:02] I just tried with another browser. Same problem. [10:26:08] also, is it slow to load pages or only slow to edit? (I assume you are logged in) [12:02:06] I am still uploading that 1.4 gb file since yesterday [12:02:20] I wonder if it is still continuing [13:35:24] *ahem* [13:35:48] * apergos gives emijrp_ the ol' hariy eyball :-P [13:35:51] *eyeball [13:36:19] evil has been done haha [13:36:20] * Nemo_bis hides new hard drive [13:36:38] emijrp_, it's getting full quickly without an uploader for the wikis ;) [13:37:23] i heard your are buliding a Internet Archive mirror at home. [13:37:47] god damn 20.5GB :P [14:07:13] Hydriz: 20.5GB? [14:07:25] hehe [14:07:46] some image dumps by Wikimedia :) [14:17:51] we are getting so so so so so so much bot operation at all of the wikis, the lack of security is a concern [14:18:32] the authentication seems to be so piss weak [14:42:05] Does Wiki Mail really hides senders information, like IP? [14:49:37] I have a problem with a page, it won't assess properly [14:50:58] Anyone online? [14:54:06] Specifically, http://en.wikipedia.org/wiki/Category:B-Class_United_States_articles_of_Unknown-importance [14:54:19] http://en.wikipedia.org/wiki/Talk:Choctaw [14:54:32] I've assessed the quality of this page yet it will not get out of the B-class of unknown importance category [14:58:12] Anyone? [15:39:03] Hi everyone. There's a Meta and mediawiki.org translation tools bug triage at 16:00 UTC in #mediawiki-i18n. That's in about 20 minutes. [16:16:33] Anyone on? [16:16:37] I need help with a broken page [16:20:10] ChrisGualtieri: seems fine to me [16:20:22] jdlrobson: ping [16:20:27] hey jeremyb [16:21:17] jdlrobson: you might want to fix https://otrs-wiki.wikimedia.org/w/index.php?title=List_of_accounts&diff=33371&oldid=33360 ;-) [16:21:58] err, https://otrs-wiki.wikimedia.org/w/index.php?diff=33380&oldid=33374 rather [16:23:04] Gyg? [16:23:05] you mean don't call myself AGUILAR Allan ? :) [16:23:07] Huh? [16:23:17] is that a default name? I don't remember ever editing this page [16:23:24] not a default [16:23:32] Its still at http://en.wikipedia.org/wiki/Category:B-Class_United_States_articles_of_Unknown-importance though [16:24:06] Though Choctaw is at a different catagory and it won't show up in the sorted catagory [16:24:35] jdlrobson: you has /msg [16:25:53] jeremyb: I think its an error because it doesn't reflect the change made, and I've refreshed it [16:27:33] ChrisGualtieri: refresh harder? i don't see what you're seeing [16:28:25] I just wiped out my cache again [16:28:45] how many articles do you see at that url above? [16:28:53] None or one? [16:30:53] @replag [16:30:56] jeremyb: No replag currently. See also "replag all". [16:30:58] @info enwiki [16:30:58] jeremyb: [enwiki: s1] db38: 10.0.6.48, db36: 10.0.6.46, db32: 10.0.6.42, db52: 10.0.6.62, db53: 10.0.6.63 [16:31:12] Its been like this since last night, several hours [16:31:32] !log aaron synchronized wmf-config/CommonSettings.php 'Removed backend config conditional now that everything was switched over.' [16:31:37] Logged the message, Master [16:34:34] jdlrobson: would you mind getting a cloak? [16:34:48] ChrisGualtieri: idk what's up there [16:34:55] i was looking at the wrong page [16:35:00] i see the same as you [16:35:08] (i was looking at the member page not the cat page) [16:35:09] When I edit the page I get an error [16:35:16] what error? [16:35:33] One sec I'll grab it [16:36:10] Our servers are currently experiencing a technical problem. This is probably temporary and should be fixed soon. Please try again in a few minutes. [16:36:23] jdlrobson: and /join #wikimedia-otrs #wikimedia-otrs-en [16:36:25] Request: POST http://en.wikipedia.org/w/index.php?title=Talk:Choctaw&action=submit, from 10.64.0.128 via cp1011.eqiad.wmnet (squid/2.7.STABLE9) to 10.2.1.1 (10.2.1.1) Error: ERR_ZERO_SIZE_OBJECT, errno [No Error] at Wed, 09 May 2012 07:06:49 GMT [16:36:47] ChrisGualtieri: apergos ping [16:36:49] gah [16:36:57] apergos: same as [[montreal]] earlier? [16:37:09] which seems like no one ever responded to you? [16:37:15] no one did [16:37:34] anyway, have to run [16:37:49] yeah, I'm mostly not here either [16:38:30] i hear you don't have a parliament yet. or whatever you call it [16:39:49] no government again today [16:39:53] its true [16:41:10] we should introduce legislations that allows apergos to travel [16:41:18] legislation* [16:41:31] I have a feeling that error is going to take awhile to fix [16:42:08] ChrisGualtieri: can you edit other pages? [16:42:17] Yes without problems [16:42:19] earlier when I looked through the various apache and syslogs, the only thing useful I saw were a pile of (newly generated) complaints about symlinks to skins-1.19 [16:42:29] nothing else remotely helpful [16:42:37] (for whoever else peeks at this) [16:43:24] i just got an unfancy (black and white) "Error reading from remote server" [16:43:44] didn't take long either [16:43:49] again [16:43:52] yuck [16:44:56] apergos: when i edit w/ no summary it takes a few secs and then it comes back with, you should use a summary (i have it enabled). when i save again then it takes a few more secs and comes back with that above [16:45:20] if you edit with a summary it works? [16:45:27] cause that sounds like a mw bug in that case [16:46:58] edit did go through actually: https://en.wikipedia.org/wiki/?title=Talk:Choctaw&diff=491627753&oldid=491543477 [16:47:45] w/ summary same error but edit did go through [16:47:48] (same page) [16:48:56] and the cat's still wrong [16:49:07] but that's a different issue i assume [16:51:10] then i edited choctaw again and it gave no error. and then talk:choctaw again and it did give the same error again [16:51:17] Well that's my issue actually [16:51:33] It breaks, won't update on the rest of Wiki, but the page catagory reflects the change [16:51:56] Rest of wiki doesn't put it where it belongs. [16:52:16] Not sure if it is a malformed WP:USA box or something triggering it with a template issue [16:52:44] But I tried updating it as it reflects to the other wikiprojects in hope of fixing it. Other pages sometimes have that problem, didn't try anything else. [16:54:08] !log aaron synchronized wmf-config/CommonSettings.php 'Make sure Swift backend will have journaling too.' [16:54:11] Logged the message, Master [17:23:06] robla: ping? [17:23:15] what's uP? [17:23:27] aren't we supposed to have a meeting right now? [17:23:37] or ~8 minutes ago :) [17:26:17] gah..sorry, yes [17:26:48] don't worry, free pizza's here [18:03:41] Reedy: where should we start deploying? [18:03:46] Dunno [18:03:53] Did you do it in stages before [18:04:20] 299 wikis to switch [18:04:27] yeah, with big wikis like this, it's probably good to do at least a couple individually, and then do the rest in one swoop [18:04:55] let's do dewiki first [18:05:20] de, fr, ja, ru are hte next big 4 [18:05:50] that works....let's do those in 5 min increments, and then the rest [18:06:03] de and fr should have users about [18:06:05] ru might... [18:06:09] I suspect ja won't [18:06:26] Let's start with DE then [18:07:49] !log reedy rebuilt wikiversions.cdb and synchronized wikiversions files: dewiki to 1.20wmf2 [18:07:52] Logged the message, Master [18:09:04] I notified in #wikimedia-de [18:09:30] * jeremyb was just thinking that de was already done. but i was confusing with the sharded containers [18:10:03] lol [18:10:19] Reedy: alrighty. fr now? [18:12:08] !log reedy rebuilt wikiversions.cdb and synchronized wikiversions files: frwiki to 1.20wmf2 [18:12:11] Logged the message, Master [18:14:48] Go for ru next [18:14:52] Reedy: yup [18:16:03] !log reedy rebuilt wikiversions.cdb and synchronized wikiversions files: ruwiki to 1.20wmf2 [18:16:06] Logged the message, Master [18:20:34] !log reedy rebuilt wikiversions.cdb and synchronized wikiversions files: jawiki to 1.20wmf2 [18:20:38] Logged the message, Master [18:23:47] * AaronSchulz watches Reedy blaze through the list [18:24:03] I'm going to migrate ALL of the wikis next [18:24:21] Reedy: yeah, I even made a test edit on ja....all looks good there [18:24:31] <^demon> Can we put MediaWiki.org back on 1.16 plz? [18:24:32] I say do everything else now [18:24:43] ^demon: can that wait 15 min or so? [18:24:51] <^demon> I guess :( [18:25:20] we should resolve the question of whether 1.15 or 1.16 would be better before rushing off [18:25:33] !log reedy rebuilt wikiversions.cdb and synchronized wikiversions files: 295 other wikipedias over to 1.20wmf2 [18:25:38] Logged the message, Master [18:25:49] * robla starts hitting all of the home pages [18:27:09] error log is still quiet [18:27:10] why limit to those 2 versions? [18:27:22] 1.6 would be better [18:27:24] maybe just have a different version per wiki? [18:27:30] pfft [18:27:32] <^demon> robla: After that, I've got a bug abotu running MediaWiki on Window 95 I'd like to discuss ;-) [18:27:39] 100k wikis look good. on t the 10k wikis: http://www.wikipedia.org/ [18:27:39] random wiki version per request [18:28:02] Reedy: ewww, that would be too evil [18:30:11] 10k wikis look fine [18:32:37] for a short while I wonder "what wikis is he talking about, we don't have so many" [18:32:59] * Nemo_bis archiving 7000 wikis [18:34:04] 1 PHP Warning: call_user_func_array() expects parameter 1 to be a valid callback, class 'languages' does not have a method 'getMessage' in /usr/local/apache/comm [18:34:04] on-local/php-1.20wmf2/includes/StubObject.php on line 59 [18:34:04] 1 PHP Warning: call_user_func_array() expects parameter 1 to be a valid callback, class 'languages' does not have a method 'getCode' in /usr/local/apache/common- [18:34:04] local/php-1.20wmf2/includes/StubObject.php on line 59 [18:34:04] 1 PHP Fatal error: Call to undefined method languages::getCode() in /usr/local/apache/common-local/php-1.20wmf2/includes/cache/MessageCache.php on line 616 [18:35:07] brion: can you cr https://gerrit.wikimedia.org/r/#/c/7023/ ? [18:35:30] * AaronSchulz debugs UW slowness [18:36:53] * Reedy waits to see if those errors are just transient [18:38:12] \o/ [18:42:28] robla: no change in what's appearing in the error logs [18:42:48] still lots of fatals? [18:44:57] "Cluster sql-s6: WARN - MNT-1228 Corrupted ruwiki." <-- Just got this error from a TS tool. Might be problematic? [18:47:14] The error message isn't useful [18:54:39] !log aaron synchronized php-1.20wmf2/includes/filerepo/file/LocalFile.php 'deployed fa1a8d5119e1174f7458eb9516287f4867c46484' [18:54:42] Logged the message, Master [18:57:04] binasher: I wish the stats sampling could be a bit more than 1/50 sometimes [18:58:49] site_stats is so terrible [18:59:06] Krenair: that's the TS, not relevant here. (you should probably read MNT-1228 if you want to know more about it) [18:59:08] we are in agreement :) [18:59:14] so so so bad [18:59:19] binasher: all it takes is one bad caller [18:59:23] and all others get blocked up [18:59:40] everything about it is bad [19:00:07] the db schema is terrible [19:00:08] binasher: it's like a stadium of people standing in line to write their name on a list [19:00:29] if you want to do counters in mysql, it should just be one row per stat, primary key of stat name and a value column [19:01:03] and nothing with high volume potential should be updated synchronously in a request [19:01:12] binasher: even then it might help to insert into some journal that is polled and updates the single value columns [19:01:32] ideally, job queue, but thats a mysql insert in the same db, lol [19:01:46] synchronously update a memcached counter [19:01:56] using compare and swap? [19:02:05] * AaronSchulz hugs the pecl memcache client [19:02:07] have a scheduled job that reads and sets the mc counters to 0 [19:02:16] update once per minute or something like that [19:02:38] i'm tempted to rip this apart as a live hack [19:03:01] if you needed multiple job procs, cas would be nice perhaps [19:03:03] * AaronSchulz wonders [19:03:07] anyway, lots of options [19:03:13] I'm amused there's php-memcached and php-memcache [19:03:36] binasher: Hmm I guess you're right that having separate stats in separate rows would improve things ^^ [19:03:40] I think memcached is the better one [19:03:42] * AaronSchulz can't recall [19:03:51] Not sure why anyone ever thought it was a good idea to put them in named columns [19:03:59] RoanKattouw: not enough, but a decent and obvious start [19:04:01] binasher: why not just use udp for the journal? idk what's being counted though [19:04:14] Hah, OOM in SqlBagOStuff.php [19:04:21] jeremyb: memcached does counters very efficiently [19:04:39] binasher: ohh, duh, atomic counter requests [19:04:41] binasher: idk how it works... is atomic? [19:04:42] i kinda want to look up who originally wrote SiteStats [19:04:46] AaronSchulz: yup [19:04:52] binasher: and what happens when you lose a node? [19:04:55] don't care? [19:05:07] we use them for "failed password" attempts and such counters [19:05:24] if the db is updated ever X seconds, i think we can afford to lose that amount of stats updates [19:05:39] way way better than having edits fail on lock timeouts [19:05:40] ok [19:05:55] https://graphite.wikimedia.org/render?from=-1weeks&until=now&width=800&height=380&target=query-m.UPDATE.site_stats_SET_ss_total_edits_ss_total_edits_N.tp99&uniq=0.03382439159984463&title=query-m.UPDATE.site_stats_SET_ss_total_edits_ss_total_edits_N.tp99 [19:06:01] that makes me rage [19:06:03] binasher: we can always have a monthly cron that fixes stats or such [19:06:21] binasher: puh, file uploads are getting blocking for 70s :) [19:06:27] sometimes 50 seconds of an edit save attempt is waiting for the lock on site_stats [19:06:29] sometimes, not always, thank goodness [19:06:31] wow.. [19:06:49] this is all.. [19:07:16] that's only 99%? sometimes it's worse then, right? [19:08:11] Author: Tim Starling Date: Tue Nov 21 09:53:45 2006 +0000 [19:08:17] binasher: 99% of dberrors.log was lock waits on delete for site stats [19:08:28] * AaronSchulz will keep an eye on that [19:08:30] yeah [19:08:32] well [19:08:51] we should redo this all right now [19:09:02] *right now*? [19:09:03] or turn it off [19:09:09] and I mean 99% as in 99/100 times, not percentile [19:09:12] did ya'll see the issues saving on [[montreal]] and [[talk:choctaw]] mentioned here earlier? [19:09:14] 99/100 errors I mean [19:09:49] its around 20% of the current log [19:10:18] yeah it fluctuates of course [19:10:28] plus I just committed some stuff ;) [19:10:39] Wed May 9 19:04:57 UTC 2012 srv258 commonswiki LocalFile::recordUpload2 10.0.6.41 1205 Lock wait timeout exceeded; try restarting transaction (10.0.6.41) UPDATE `site_stats` SET ss_images = ss_images+1 [19:10:55] hopefully those went down at least, hard to tell yet [19:11:22] binasher: complain to robla ;) [19:11:37] binasher: can we put everything in memcached? [19:11:39] it'll be quicker [19:11:56] * robla bumps around under the bus [19:11:57] Reedy: even files? [19:12:07] I said everything [19:12:14] robla: careful! [19:12:19] even the job queue :p [19:12:22] We can store our PHP code files in memcached too [19:12:36] also git [19:12:40] gerrit's db [19:12:45] woo [19:12:47] Have you guys looked at beanstalkd for a memcached-like queue? [19:12:50] Is memcached webscale? [19:13:13] me dropping tables is webscale [19:14:01] Domas might get jealous [19:14:52] i wish he was on right now [19:15:33] binasher: are you turning into domas? [19:16:20] I think I'm turning Lithuanian [19:16:20] I really think so [19:16:20] I'm turning Lithuanian [19:16:22] binasher: you want to get domas started? [19:16:24] i do wish i was a little bit taller [19:16:38] y'all [19:16:38] and heavier [19:18:59] AaronSchulz: that's easier ;) [19:20:45] Reedy: did you reforming CommonSettings? [19:20:49] *reformat [19:21:24] AaronSchulz: no... [19:21:34] weird, $wgFileBackends suddenly looks like ass [19:23:21] looks ok vim, so it must be ok [19:23:30] just the public noc one is ugly [19:26:25] hello domas – we missed you [19:27:00] binasher, robla: so we can work on a "quicker" fix of breaking site stats into multiple rows or work on something more long term [19:27:16] the current way looks pretty unsustainable [19:27:29] i just started on a bugzilla ticket fwiw [19:27:51] seemed more reasonable than 'drop table' :/ [19:28:13] heh....I guess you're not turning into domas :) [19:28:28] joking about dropping tables doesn't count [19:29:33] i thought i was getting domas level awesome but i can only dream [19:29:54] also you can be fired...which might be a factor [19:30:06] ;) [19:35:53] =) [19:37:50] you're not domas cool till you break the site a few times (and whilst on a plane iirc) [19:38:13] I fix sites while I'm on a plane [19:38:29] so it evens out [19:45:04] one way not to have contention on site stats is not to have insane amount of transactions hanging [20:37:22] CREATE TABLE /*_*/site_stats ( [20:37:24] -- The single row should contain 1 here. [20:37:25] ss_row_id int unsigned NOT NULL, [20:37:27] binasher: :) [20:38:29] AaronSchulz: are you rewriting it? :) [20:38:30] * AaronSchulz wonders what a good name for a new table would be [20:38:41] it can still be site_stats [20:39:09] wikistats seems fitting [20:39:11] I tend to prefer new names if totally redoing a table [20:40:30] binasher: first thing is to change direct write queries to go through a class [20:40:40] binasher: Do you see any issues with bumping the memory limit up a bit for MW processes? [20:40:52] Just wondering if it'll help with some of the OOMs... Or they'll just fail later [20:41:24] even moar?? [20:41:37] All of the memorys. [20:41:44] default is currently 120MB [20:41:54] zhwikis get 160 [20:41:58] unless it uses LC [20:42:01] err, 180 [20:42:05] yeah, 180 [20:42:16]