[00:01:22] Thehelpfulone: Oh, awesome it worked! [00:01:30] I am debugging a dbbot-wm issue [00:01:39] heh :) [00:01:40] I didn't see it coming trough, looking in the wrong channel [00:22:55] getting close.. [00:32:11] Hey [00:32:15] Tech question [00:32:28] Is selective blocking something that's technically possible [00:33:59] "selective blocking" is pretty vague [00:34:07] and my crystal ball isn't working atm [00:34:20] Blocking a user from a namespace or set of titles [00:34:27] no it's not [00:38:00] It's not possible? [00:38:06] Can it be possible? [00:38:17] At the moment, it's not [00:38:18] if some one was to spend the time coding it [00:38:27] But a blocked user can still edit their talk page. [00:38:46] yes, because someone decided that was a feature worth coding [00:39:16] I've been doing a community survey on dispute resolution. [00:39:22] Common response [01:24:14] @externals update [01:24:28] Krinkle: Successfully updated externals! [01:24:30] @externals [01:24:30] Krinkle: [all.dblist] last modified: 2012-04-09 01:24:16 (UTC); [db.php] last modified: 2012-04-09 01:24:16 (UTC) [01:43:32] @info s1 [01:43:32] jeremyb: [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 [01:43:40] @info db1047 [01:43:40] jeremyb: [db1047: s7] 10.64.16.36 [01:45:43] Krinkle: investigating which issue? do you know why it ignored me at 15:28:04 UTC? [01:46:02] how many hours ago was that? [01:46:10] 09 01:46:02 < Krinkle> how many hours ago was that? [01:46:20] that's also UTC [01:46:34] Up until 2 hours ago dbbot-wm was a blank Kribo instance with no plugins [01:46:35] so, slightly over 10 hours [01:46:47] huh [01:46:52] somehow the wmfDbBot plugin was commented out in LocalConfig.php [01:47:00] I fixed that and restarted it [01:47:16] (and implemented the @externals command so that we can see how old it's local cache of noc.wm.o is) [01:47:22] i figured it was just in some weird state and needed a boot [01:47:22] @help [01:47:23] Krinkle: [all.dblist] last update: 2012-04-09 01:24:16 (UTC); [db.php] last update: 2012-04-09 01:24:16 (UTC) [01:47:23] Type @commands for list of commands. This bot is running http://meta.wikimedia.org/wiki/WM-Bot version wikimedia bot v. 1.1.4 source code licensed under GPL and located in wikimedia svn [01:48:19] that msg really should state a GPL version # [01:48:29] @quit [01:48:36] * Krinkle renames help to info [01:49:13] erm? info's already taken [01:49:16] @info [01:49:21] oh, shit [01:49:25] of course [01:50:20] @info [01:50:20] jeremyb: Invalid arguments [01:50:26] @help [01:50:26] Type @commands for list of commands. This bot is running http://meta.wikimedia.org/wiki/WM-Bot version wikimedia bot v. 1.1.4 source code licensed under GPL and located in wikimedia svn [01:50:33] @externals [01:50:33] jeremyb: [all.dblist] last update: 2012-04-09 01:24:16 (UTC); [db.php] last update: 2012-04-09 01:24:16 (UTC) [01:50:41] @docs [01:50:42] Krinkle: https://wiki.toolserver.org/view/dbbot-wm [01:50:48] @docs [01:50:48] Krinkle: https://wiki.toolserver.org/view/dbbot-wm [01:50:49] @info dewikisource [01:50:49] jeremyb: [dewikisource: DEFAULT (s3)] db39: 10.0.6.49, db34: 10.0.6.44, db25: 10.0.6.35, db11: 10.0.6.21 [01:50:57] I renamed help to docs [01:51:19] @info DEFAULT [01:51:19] jeremyb: [s3 (DEFAULT)] db39: 10.0.6.49, db34: 10.0.6.44, db25: 10.0.6.35, db11: 10.0.6.21 [01:51:25] oooh, that works ;) [01:51:34] Krinkle: i figured [01:52:42] Krinkle: how does it handle -1 replag? [01:53:34] maybe we could get coloring for -1 and high replag. also teach it somehow about in progress migrations with known high lag [01:53:42] @info db42 [01:53:42] jeremyb: [db42: s7] 10.0.6.52 [01:53:54] @replag all [01:53:55] Krinkle: [s1] db38: 0s, db36: 0s, db32: 0s, db52: 0s, db53: 0s; [s2] db13: 0s, db30: 0s, db54: 0s; [s3] db39: 0s, db34: 0s, db25: 0s, db11: 0s [01:53:56] Krinkle: [s4] db31: 0s, db22: 0s, db33: 0s, db51: 0s; [s5] db35: 0s, db45: 0s, db44: 0s, db55: 0s; [s6] db43: 0s, db47: 0s, db46: 0s, db50: 0s; [s7] db37: 0s, db16: 0s, db56: 0s, db26: 0s [01:54:07] jeremyb: I don't know of any -1 [01:54:10] something new? [01:54:27] Krinkle: -1 i think means replication stopped? not certain [01:54:30] Anyway, @replag shows all clusters with replace > threshold (which is 2 seconds) [01:54:34] replag* [01:54:49] I updated documentation [01:54:52] https://wiki.toolserver.org/view/dbbot-wm [01:54:55] https://wiki.toolserver.org/view/WmfDbBot [01:55:00] https://wiki.toolserver.org/view/Kribo [01:55:03] and committed to svn [01:56:16] db42 should be lagged. i don't see it listed there... ? http://ganglia.wikimedia.org/latest/graph_all_periods.php?title=mysql_slave_lag&mreg[]=^mysql_slave_lag%24&hreg[]=db42&aggregate=1&hl=db42.pmtpa.wmnet|MySQL%20pmtpa [01:57:04] i guess it just fetches from API though and API's been taught to ignore somehow [01:57:27] It uses Api instead [01:57:29] indeed* [01:58:43] hrmmmm, someone linked me to this relationship diagram of the DB hosts and now I don't know where it is... [01:58:57] http://mediawiki.org/wiki/DB [01:59:03] I created that one [01:59:15] not the schema. which replicate from which [01:59:19] oh [01:59:20] that [01:59:20] and which have which wikis [01:59:28] I don't know of a diagram [01:59:31] but the data is on noc [01:59:33] i'll find it [01:59:38] which dbbot-wm uses [01:59:46] http://noc.wikimedia.org/conf/highlight.php?file=db.php [02:00:05] @info db42 [02:00:05] Krinkle: [db42: s7] 10.0.6.52 [02:00:16] eswiki, huwiki, .. [02:01:25] https://fisheye.toolserver.org/browse/~br=trunk/krinkle/trunk/wmfDbBot/inc/Functions.php?hb=true#to252 [02:02:46] Krinkle: found it: http://noc.wikimedia.org/dbtree/ [02:03:24] @info db1039 [02:03:24] Krinkle: [db1039: s7] 10.64.16.28 [02:03:33] ah, it does know about the new 10xx ones [02:04:11] All those orgCharts are nice (there's a bunch of them for different purposes) [02:04:34] Except that they are all way too wide. The same data could be presented much clearer in less space and taller instead of wider [02:05:06] I'd rotate it -90 degrees [02:05:14] well the clusters aren't really related to each other [02:05:23] they don't have to be connected [02:05:46] Krinkle: but there's discrepency betwen your bot and the link for that last thing you asked it above [02:06:03] the chart says s5, bot says s7 [02:06:11] I trust mine [02:06:20] likewise it's wrong about db42 [02:06:29] which is s1 not s7 [02:07:03] The bot is using a semi-live copy of db.php, updated within the hour [02:07:15] unless noc.wm.o has been forgotten to sync, that is the right one [02:07:38] The chart may be saying that it's been updated within the last minute, but that refers to the replag values [02:07:42] not the association [02:07:47] right, but there might be cluster slaves that aren't in db.php [02:07:55] sure [02:08:05] and the chart has some better source of truth [02:09:02] well, if it cares to publish the data in a machine readable form, I can update the bot to use that instead, but in general for most cases when things like nagios complain, being able to search by key right away (e.g. dbbot-wm) seem useful. [02:09:18] and dbbot-wm is simply as good as the data it is fed, which comes from db.php [02:11:11] sure. it is quite useful. but it should at least be made to say it doesn't know when it doesn't instead of (apparently) just always saying s7! [02:12:06] well, db.php doesn't include relationships to the default cluster [02:12:49] so I don't see how it could distinguish between a "dbhost that is in db.php and is part of the default cluster" and a "dbhost that shouldn't be in db.php but is but is not in the default cluster" [02:12:51] yes it does [02:13:16] For instance, it doesn't say which cluster db1039 belongs to [02:13:23] right [02:13:43] sectionLoads (dbhost>cluster id) only contains the non-standard ones [02:13:49] but db.php does say some members of s7 and some of s1 and some of s5 [02:14:07] how do you mean non-standard? [02:14:59] I don't know [02:15:21] I just know it's missing information and wmf's load balaner makes the same assumption that I do: if not in that map, assume default [02:15:46] the reason the bot sometimes fails and mediawiki does not is because the bot is sometimes queries for dbhosts that mw will not use [02:15:56] idk about that. i'm thinking if it's not in the list then it's not in rotation at all [02:17:22] !log LocalisationUpdate completed (1.19) at Mon Apr 9 02:17:22 UTC 2012 [02:17:26] Logged the message, Master [02:19:14] probably [02:19:25] ah, Idoesn't assume default, s7 is not default [02:19:44] it uses a foreach loop with a break to see which section it belongs to and then passes that section to the irc bot [02:19:55] but if t's not in any section, the loop will continue to the end [02:20:06] and the last value set in section will be s7 [02:20:41] when the bot was written it didn't consider the case of dbhosts out of rotation in db.php [02:20:53] since those would never show up in @info or @info
, it rarely happends [02:20:57] unless you know the ID [02:21:14] in which case there is no need to know the info [02:21:29] ohwell, it could use fine-tuning, but not at 4am after easter :P [02:22:49] oh, wow i assumed you were in SF ;P [02:23:11] you should know better than that :P [02:23:41] I'm at least as crazy as you are :P [02:24:11] assuming your irc local time is correct [02:25:36] you are at least as crazy as i am. but i'm now America/New_York for the last 5 days in fact [02:25:53] aha [06:18:24] would someone with access to https://labsconsole.wikimedia.org/wiki/Help:Access#Giving_users_Labs_access.2C_if_they_don.27t_already_have_SVN_access please make the link at __leaving a request on this page (MediaWiki.org) (this is recommended) __ to protocol relative [06:18:53] protocol neutral [06:22:27] sDrewth: That is done by the interwiki map that labs uses, post a bug in BZ under Labs -> General [06:24:38] k [06:24:52] ppl must get sick of my bzs [06:26:09] p858snake|l: there is no general, so I will poke under Unspecified [17:36:39] !log nikerabbit synchronized wmf-config/InitialiseSettings.php 'Bug 35426 - WebFonts on mr.wikisource.org' [17:36:41] Logged the message, Master [20:03:07] can someone help me with a template? [21:29:54] bonne nuit =_= [22:32:24] !log asher synchronized wmf-config/db.php 'returning db12 as enwiki recentchange/watchlist db' [22:32:26] Logged the message, Master