[13:31:06] @info aawiki [13:31:15] @replag [13:33:00] @replag [13:33:01] Krinkle: No replag currently. See also "replag all". [13:33:04] @replag all [13:33:05] Krinkle: [s1] db32: 0s, db36: 0s, db12: 0s, db26: 0s, db38: 0s; [s2] db30: 0s, db13: 0s, db24: 0s; [s3] db34: 0s, db39: 0s, db25: 0s, db11: 0s [13:33:05] @info aawiki [13:33:06] Krinkle: [s4] db22: 0s, db31: 0s, db33: 0s; [s5] db45: 0s, db35: 0s, db44: 0s; [s6] db47: 0s, db46: 0s, db50: 1s; [s7] db16: 0s, db37: 0s, db18: 0s [13:33:07] Krinkle: [aawiki: DEFAULT (s3)] db39: 10.0.6.49, db34: 10.0.6.44, db25: 10.0.6.35 [19:42:26] Reedy: You there ? [19:42:49] Yus [19:43:06] I need a quick hand in something on integration.mw.org but I don't have cluster access yet [19:43:14] The swarm is down, the JSON api is returning invalid stuff [19:43:18] can you do a select query for me on it [19:43:56] SELECT * FROM jobs WHERE id=3012 [19:44:04] https://github.com/jquery/testswarm/issues/99 [19:44:17] right now all swarm clients are getting an invalid response, I suspect that row is malformed [19:44:34] r??ponse has two null values http://i.imgur.com/J85Dr.png [19:44:34] breaks [19:44:42] indeed [19:44:44] errrm, let me fine the mysql password [19:45:01] (typo fixed) [19:45:40] 3011 as well. [19:45:41] Hm.. [19:46:00] perhaps it's not the row but something in between. Maybe check php error log in a minute if that row doesn't' tell us anything [19:46:23] :o [19:47:13] eh, sorry, select from table "runs" not "jobs" [19:47:14] "SELECT * FROM jobs WHERE id=3012 [19:47:18] "SELECT * FROM runs WHERE id=3012 [19:47:50] it's alright, I haven't found an instance of the mysql password I can use yet :p [19:48:17] it should be in the config file [19:48:41] "the config file" [19:48:43] testswarm/config.ini [19:49:02] Doesn't help when I don't know where testswarm is [19:49:07] www root [19:49:13] It's not [19:49:36] depending on how that server is set up /var/www/testswarm otherwise there may be a Apache Alias for /testswarm to elsewhere [19:49:43] /srv/org/mediawiki/integration [19:49:52] right [19:49:59] or not [19:50:35] /usr/share/testswarm/site [19:51:56] I can't read /etc/share/testswam/config.ini [19:52:08] oh, wait [19:56:49] I can see the username, password and database name in cleartext [19:58:01] yup [19:58:17] mysql no like [19:58:23] :O [19:58:40] host ? [19:58:44] all local [19:59:24] the password is so simple too [20:00:10] hm.. looking at the source code of the testswarm getrun json api, a db error could make sense [20:00:28] undefined variables yield NULL in PHP right ? (aside from the E_NOTICE) [20:00:33] Has it ever worked? [20:00:38] yes for sure [20:00:39] 100% [20:00:43] an hour ago [20:00:47] lol [20:00:51] Just checking :p [20:00:51] used it all day, had a bunch of browsers open [20:01:06] anything in the php error log screaming db errors ? [20:01:30] Can't view the apache logs [20:02:13] can you hand-edit the files in testswarm/site ? [20:02:34] like enable error reporting or something ? [20:02:58] Nope [20:03:11] Let's see if the puppet manifests suggests I have sudo to some account [20:03:27] account testswarm perhaps [20:03:32] hasher made an account for it [20:03:43] hashar* [20:03:48] all files are root/root [20:03:51] ok [20:04:59] sudo_user { [ "demon", "hashar", "reedy" ]: privileges => [ [20:05:00] 'ALL = (jenkins) NOPASSWD: ALL' [20:05:00] ,'ALL = NOPASSWD: /etc/init.d/jenkins' [20:05:00] ,'ALL = (testswarm) NOPASSWD: ALL' [20:05:59] a new revision was committed, this one works again [20:06:05] grmpf. [20:06:22] after this one is fully tested it'll try the old one again [20:06:29] lets see what it does, brb [20:07:27] I'm not even sure what those accounts are for [20:07:56] for jenkins at least, presumably what the java processes run as [20:08:19] testswarm does periodic svn checkouts, sqlite database creations and mediawiki installs [20:08:27] yeah [20:12:24] Guess this probably means without pinging a root (which for obvious reasons I'd rather not do unless it's an emergency), you'd have to speak to Antoine [20:13:07] looks like it only affects clients that have nothing else to test [20:13:26] but the problem is that because it causes a TypeError in the client, they stop refreshing [20:13:31] nothing we can do, upstream bug [20:13:46] I thought it was something in our database [20:16:25] aha, it is a bug in our database. This run id is fairly old, at least 200 commits. Looks like hashar probably inserted a dummy row during testing last week without all columns and since that's now the oldest untested "job row" it'll keep trying that one as soon as a client completes all new tests. [20:16:59] for now we can ask swarm people to manually refresh their browser as soon as they've completed all tests [20:18:04] bailar??n [20:18:17] ops, wrong window [21:30:39] Here's a Merry-Christmas-Question I can't figure out: how do I tell from the database whether a page is deleted or not? [21:30:51] There's no indication of that information in the 'page' table. [21:36:15] are it's revisions deleteD? [21:40:24] omg, I should have thought of the 'revision' table. The obvious answer! Thanks Reedy.