[00:10:02] AaronSchulz: I'm off now; are you on top of that? should I open a bug tomorrow? [00:10:21] or no need? [00:10:51] meh, no need [00:11:04] okay! [00:11:12] see you [00:12:21] bye [00:13:12] !log copying MW log archives from NFS to fluorine via rsync on fenari [00:13:21] Logged the message, Master [01:03:41] http://guardianlv.com/2012/08/why-america-needs-gary-johnson/ http://www.garyjohnson2012.com Hi! I'm sorry I'm spamming, but we the people have to fight against the mockery our politicians are making of the US presidency [01:10:07] !log aaron synchronized php-1.20wmf9/includes/filerepo/file/File.php 'deployed 08ea8adc95a4edba0b1f2f13c34f2a57725f1a58' [01:10:28] !log aaron synchronized php-1.20wmf9/includes/filerepo/backend/TempFSFile.php 'deployed 08ea8adc95a4edba0b1f2f13c34f2a57725f1a58' [01:10:48] aaron cleared profiling data [01:12:25] Logged the message, Master [01:12:35] Logged the message, Master [01:19:09] TimStarling: ben and I could tell from graphite/ganglia that there more getLocalReference()/sec than should [01:20:49] swift can handle it for now, but those should really be tracked down [01:21:04] I added a bit of profiling too, but I'm about to head out for today [01:21:42] maybe add the caller to the profiling section name? [01:22:02] yeah, I was thinking I might end up doing that [01:22:13] if you want to mess with this tonight please do so :) [01:22:29] * AaronSchulz hopes the doubling of apache network "in" doesn't spam the network too hard [01:22:52] * AaronSchulz also wishes the media handler code was less convoluted [01:24:23] TimStarling: if you see any trouble come up you can switch $readFromSwiftShared/$readFromSwiftLocal [01:24:39] ok [01:25:02] 'shared' is the worst offender [01:25:21] of course I'll be looking at this first thing tomorrow [01:25:40] ok, see you [01:25:44] bye [01:28:29] asher@fluorine:/a/mw-log$ zgrep -c " 22:" archive/memcached.log-20120816.gz [01:28:29] 555 [01:28:32] asher@fluorine:/a/mw-log$ grep -c " 22:" memcached.log [01:28:32] 18218 [01:29:04] asher@fluorine:/a/mw-log$ zgrep -c "Error connect" archive/dberror.log-20120816.gz [01:29:04] 15031 [01:29:12] asher@fluorine:/a/mw-log$ grep -c "Error connect" dberror.log [01:29:12] 441874 [01:30:10] i think crappy foundry switch related packet loss is up a bunch since traffic to the apaches increase :( [02:37:38] !log LocalisationUpdate completed (1.20wmf9) at Fri Aug 17 02:37:38 UTC 2012 [02:37:49] Logged the message, Master [02:42:12] !log aaron synchronized php-1.20wmf9/includes/filerepo/backend/SwiftFileBackend.php 'Profiling' [02:42:22] Logged the message, Master [02:43:11] aaron cleared profiling data [03:08:47] !log aaron synchronized php-1.20wmf9/includes/filerepo/backend/FileBackendStore.php 'Profiling' [03:08:57] Logged the message, Master [03:09:58] aaron cleared profiling data [03:10:46] !log aaron synchronized php-1.20wmf9/includes/filerepo/backend/FileBackendStore.php 'Profiling' [03:10:55] Logged the message, Master [03:13:59] !log aaron synchronized php-1.20wmf9/includes/filerepo/backend/FileBackendStore.php [03:14:09] Logged the message, Master [03:14:18] aaron cleared profiling data [03:16:12] * Aaron|home goes up the stack [03:16:16] !log aaron synchronized php-1.20wmf9/includes/filerepo/backend/FileBackendStore.php [03:16:25] Logged the message, Master [03:21:01] aaron cleared profiling data [03:34:09] !log aaron synchronized php-1.20wmf9/includes/media/SVG.php 'deployed 3ae632c3c9111b16f019574bcafec58fa3ee408d ' [03:34:19] Logged the message, Master [03:34:34] !log aaron synchronized php-1.20wmf9/includes/media/DjVu.php 'deployed 3ae632c3c9111b16f019574bcafec58fa3ee408d' [03:34:43] Logged the message, Master [03:35:21] aaron cleared profiling data [03:36:36] TimStarling: http://ganglia.wikimedia.org/latest/?r=hour&cs=&ce=&tab=v&vn=swift+frontend+proxies [03:38:30] I guess you fixed it [03:39:39] at least we know we won't have to worry about handling load for a long time to come ;) [03:40:21] yeah :) [03:42:33] Aaron|home: that's just a cherry pick of 20265, right? [03:43:17] would be soooo nice to have those 2 gerrit changes linked to eachother somehow ;/ [03:44:03] the fix was just one change, 20265 [03:45:29] Aaron|home: and 20266 too? [03:45:32] i assumed [03:46:22] * 3ae632c (origin/wmf/1.20wmf9, refs/changes/66/20266/1) Don't get local file copies for svg/djvu files unless needed. [03:47:31] well, 20266 is the cherry-pick into wmf9, I guess you can add a comment or something [03:48:45] jeremyb: hmm, cherry pick has an -x flag [03:48:55] I guess I can try to remember to use that [03:49:19] i see! [03:49:34] or we could add it to the cherry-pick docs [03:49:37] on wikitech [03:49:40] "When recording the commit, append a line that says "(cherry picked from commit …)" to the original commit message" [03:50:35] well, i'm going to add more context to this diff when i review it. (/me thinks Tim said to do that... ;P) [03:53:04] i've never seen those files before but it looks reasonable [03:54:02] Aaron|home: anyway, i was thinking you could even just keep the change id as is and it might work. idk for sure [03:54:11] (see the "included in" section) [03:55:23] Aaron|home: you have a new comment on bug 39221 [03:56:27] oh, huh, it *is* the same change id. interesting [04:47:48] TimStarling: Hey, can I run the purge script on a few thousand files? [04:48:47] I guess so [04:49:20] if it is for a template update I can just run that job queue insertion thing [04:50:33] I found some files that failed the update from last time, maybe you want to investigate? [04:50:58] ok [04:53:20] http://p.defau.lt/?af3BHsSDNLtTyZC0P9cX9w [04:56:48] A related link was switched to a protocol-less, the primary method remains complicated as I'm discovering new rare templates on commons [04:57:47] some of the jobs must have failed [04:58:00] the pages seem to fall into clusters of page_id values [04:58:24] that's how jobs are divided up [04:59:30] Do you want to fix it or should I do it? [05:00:48] I'll queue them up again, individually this time [05:04:04] BTW, is there be any benefit testing more than the 3 million images needed by WMA? https://commons.wikimedia.org/wiki/Commons:Village_pump#Image_scaler_problems [05:07:09] we can make logs of that sort of thing if it interests you [05:08:25] Might be good. I was just asking since I already had a procedure down for it. [05:10:22] I've read the bug report now, it looks like the bug is probably the best place to discuss it [05:10:42] you can ask Aaron if he wants you to scan images [05:12:48] !log queued refreshlinks jobs for 1330 commonswiki image description pages with coord templates which weren't refreshed in the previous manual attempt, probably due to failed jobs [05:12:58] Logged the message, Master [13:48:51] !log robh ran sync-common-all [13:49:00] Logged the message, Master [14:27:23] !log robh synchronized wmf-config/CommonSettings.php [14:27:32] Logged the message, Master [14:27:54] !log robh synchronized wmf-config/InitialiseSettings.php [14:28:04] Logged the message, Master [15:47:50] hi zil [15:57:50] Zil: Hmm, I'm quite busy with Wiki Loves Monuments so I'm not sure if I have time for it [15:57:59] Sorry..... [15:58:30] Sure [15:58:39] That would be after the busy period :-) [15:58:47] (or in another one, you never know) [17:12:12] anyone present her who is experienced with mailinglists of wikimedia? [17:13:17] Romaine: what do you need? [17:13:29] still having problems with timeouts? [17:13:34] (was it you?) [17:13:35] no [17:13:37] ok [17:13:46] I experience a strange situation with one mailinglist I am admin of [17:14:35] I send an e-mail and then I got a message I need to approve my e-mail, while it should be send on the list without this approval :S [17:19:49] Romaine: it's a bit generic [17:20:13] are you sure the list has no been put on moderation, you've sent it with the correct address etc. etc. [17:20:58] I think so [17:22:24] Romaine: was the list the only recipient of the message and in the To: or cc: field? (no bcc) [17:22:36] I don't think we have spam checks [17:24:06] !log robh synchronized wmf-config/CommonSettings.php [17:24:15] Logged the message, Master [17:24:39] !log robh synchronized wmf-config/InitialiseSettings.php [17:24:49] Logged the message, Master [17:31:43] Romaine: what list? [17:34:59] wikimediabe-l [17:35:17] * maplebed looks [17:36:36] the case seems that I am moderator, I am not moderated myself, but still get a message that I have to approve my message [17:37:35] when you go to approve your message is it pending or did it get delivered? [17:38:42] As list administrator, your authorization is requested for the [17:38:42] following mailing list posting: [17:38:50] ...to approve or deny the request. [17:39:03] my own e-mail [17:39:40] yes, but is it actually pending or does the message lie (and so when you go look at pending requests your email is actually held waiting approval) [17:41:42] another admin had approved my email het told me just a moment ago [17:41:53] so yes, the e-mail needed to be approved [17:42:00] but let me test it again, moment [17:42:01] ok, so it acttually was held. [17:42:09] yeah, please test and don't approve the message. [17:49:16] it seems now the message got through... strange [17:50:40] well, if it happens again, we can look more closely. I don't see anything about the list that looked wrong. [17:51:25] ok [17:51:29] thanks for helping! [20:13:27] !log Deleting php-1.20wmf8/cache/l10n from mediawiki-installation [20:13:37] Logged the message, Master [20:15:13] !log Pointing php to php-1.20wmf9 [20:15:23] Logged the message, Master [20:18:18] we hardly knew ye [21:30:01] are the squid logs having problems? There is a page with 100 000 pageviews before creation o.O [21:30:13] ^ this on http://stats.grok.se/pt/latest90/Omega_Tauri# [21:41:09] !help [21:41:09] http://www.mediawiki.org/wiki/Help:$1 [21:41:09] !(stalk|ignore|unstalk|unignore|list|join|part|quit) [21:41:28] :-/ [21:41:44] !help why can't i dl enwiki-20120802-pages-articles-multistream.xml.bz2 or enwiki-latest-pages-articles.xml.bz2 ? ; the connection closes, and now i get a ERROR 403: Forbidden. ; i have tried in IE and using wget , and wget -c ; i tried an older wget -E --ignore-length [21:41:44] http://www.mediawiki.org/wiki/Help:why [21:41:45] !(stalk|ignore|unstalk|unignore|list|join|part|quit) [23:39:23] jeremyb: wikitech is fixed, rebuilt localization cache, would not fix it, then Roan helped and had a look, just deleted all the cache files, fixed it