[00:09:54] <gwicke>	 YuviPanda: is your npm issue resolved?
[00:10:08] <gwicke>	 (sorry for the latency)
[00:10:10] <YuviPanda>	 gwicke: well, they resolved it by copying node_modules and checking it in :)
[00:10:22] <gwicke>	 hmm
[00:10:22] <YuviPanda>	 gwicke: but I found the issue - trusty ships with a really old version of npm...
[00:10:57] <gwicke>	 yeah; I'll be happy once we run Debian stable on physical hosts & SID in containers ;)
[00:11:00] <YuviPanda>	 gwicke: guess I'll have to find ways to package it :)
[00:11:02] <YuviPanda>	 heh
[00:11:27] <gwicke>	 there is a ppa for 0.11 as well
[00:12:09] <gwicke>	 we are generally okay with trusty's npm though; precise was much worse
[00:12:29] <YuviPanda>	 gwicke: heh, yeah. problem is toollabs doesn't support a local repo that works to gives us different precise / trusty packages.
[00:12:41] <YuviPanda>	 I need to investigate, and then just... get... a... new... nodejs... package.
[00:12:43] <YuviPanda>	 because this is stupid.
[01:33:58] <yug>	 ／
[06:48:45] <shinken-wm>	 PROBLEM - Puppet failure on tools-exec-09 is CRITICAL: CRITICAL: 44.44% of data above the critical threshold [0.0]  
[07:18:44] <shinken-wm>	 RECOVERY - Puppet failure on tools-exec-09 is OK: OK: Less than 1.00% above the threshold [0.0]  
[09:31:59] <petan>	 !petan-
[09:31:59] <wm-bot>	 There are multiple keys, refine your input: petan-build, petan-forgot, 
[09:32:03] <petan>	 !petan-bui
[09:32:03] <wm-bot>	 make -j `getconf _NPROCESSORS_ONLN` deb-pkg LOCALVERSION=-custom
[10:32:33] <liangent>	 !petan-forgot
[10:32:33] <wm-bot>	 !petan-build
[11:05:28] <wikibugs>	 3Continuous-Integration, Wikimedia-Labs-Infrastructure: integration-slave1001.eqiad.wmflabs can't start, mount.nfs yields failure in name resolution - https://phabricator.wikimedia.org/T76250#821022 (10hashar)
[14:53:12] <wikibugs>	 3Wikimedia-Labs-wikitech-interface: [Regression] Editing "Documentation" page for labs project not working - https://phabricator.wikimedia.org/T47519#821370 (10Aklapper)
[17:25:41] <shinken-wm>	 PROBLEM - Puppet failure on tools-webproxy is CRITICAL: CRITICAL: 100.00% of data above the critical threshold [0.0]  
[18:00:56] <bd808>	 !log wikimania-support Added Niharika29 as project member and admin
[18:01:02] <labs-morebots>	 Logged the message, Master
[18:54:47] <bd808>	 !log wikimania-support Updated labs-vagrant clone to head of precise-compat branch and installed all available apt-get updates.
[18:54:49] <labs-morebots>	 Logged the message, Master
[19:17:11] <multichill>	 YuviPanda: How it the labs monitoring coming along?
[19:23:29] <YuviPanda>	 multichill: not bad. shinken.wmflabs.org (guest/guest)
[19:26:39] <multichill>	 YuviPanda: Nagios clone?
[19:27:57] <guillom>	 Hi. I'm getting a "Lost connection to MySQL server during query" error on a test script that I'm pretty sure was working a couple weeks ago. Has anything changed in Labs that would require a change in the script itself?
[19:32:05] <multichill>	 guillom: Long query?
[19:35:37] <guillom>	 multichill: not even. I've tested several band some even work in Quarry, which has a lower threshold.
[19:36:05] <guillom>	 I get the error only a few seconds after starting it.
[19:36:45] * guillom continues to investigate.
[19:49:50] <YuviPanda>	 multichill: yeah, pretty much.
[19:58:42] <YuviPanda|zzzz>	 guillom: is this python?
[19:58:53] <guillom>	 YuviPanda|zzzz: yeah
[19:59:22] <YuviPanda|zzzz>	 bearND: I usually call the .ping method at start of a loop or somrthing. Connections die easily 
[19:59:31] <guillom>	 I'm going to try the same query from the console just in case.
[19:59:36] <YuviPanda|zzzz>	 Brrr
[19:59:50] <YuviPanda|zzzz>	 guillom: not bearnd
[20:00:25] <guillom>	 But… a few seconds?
[20:03:26] <guillom>	 hmmm, command line works
[20:04:11] <guillom>	 I'll figure it out
[20:12:09] <guillom>	 Bleh, I get "MySQL server has gone away" as soon as I start that query in the mysql console.
[20:12:19] * guillom needs more sql skillz.
[20:30:34] <Coren>	 guillom: I know Sean put in some query memory limits to avoid the bigger problem of database desync - you may be hitting against it if your query is particularily heavy.
[20:31:38] <guillom>	 Coren: Thanks. It turns out the query eventually runs., albeit slowly (I need to go through Commons' 24 millions files).
[20:32:17] <guillom>	 I'm going to try to decouple the sql part from the python script
[21:26:28] <multichill>	 guillom: That sucks, looks like some of my queries also stopped. For example https://commons.wikimedia.org/w/index.php?title=Commons:Monuments_database/Categorization/Statistics&action=history hasn't been updated for a while
[21:26:34] <multichill>	 Probably something different....
[21:33:04] <ebernhardson>	 anyone used the eventlogging role in vagrant?  It appears to setup eventlogging to connect to //localhost:8100/event.gif,  but i dont see anything else in the vagrant puppet that ever sets up port 8100 
[21:34:21] <bd808>	 ebernhardson: tgr has used it I think. I know it does not setup the python server but I thought that at least at one time it added apache config to answer the port.
[21:35:21] <ebernhardson>	 bd808: makes sense,  i'll check with him and see if we can add whatever is necessary to get it forwarding and just writing to the log file(without the full stack of event logging backend)
[21:35:59] <ebernhardson>	 tgr: would love to ask about eventlogging in vagrant if you have a chance some time
[21:36:30] <bd808>	 There may be a bug for that actually. I think I remember Ori saying you could just look at the js console to see what it was trying to do too.
[21:36:53] <ebernhardson>	 well, sorta.  if you can decode http://localhost:8100/event.gif?%7B%22event%22%3A%7B%22entrypoint%22%3A%22reply-post%22%2C%22action%22%3A%22save-success%22%2C%22schema%22%3A%22FlowReplies%22%2C%22isAnon%22%3Afalse%2C%22sessionId%22%3A%22C1xRQJxHVTEEPr576QbPgOx8pJf0EHGU%22%2C%22funnelId%22%3A%22rl7NrVQX79x6IKAwolp38UJXK3f40KDa%22%2C%22pageNs%22%3A3%2C%22pageTitle%22%3A%22Zomg%22%7D%2C%22clientValidated%22%3Afalse%2C%22revision%22%3A10561
[21:37:10] <bd808>	 urldecode ftw :)
[21:37:15] <bd808>	 Also https://phabricator.wikimedia.org/T56042
[21:37:35] <bd808>	 looks like I just talked about setting up the apache proxy
[21:37:45] <bd808>	 It shouldn't be too hard to do
[21:38:54] <ebernhardson>	 i would be tempted to put a php script in /var/www/event.gif, make it runnable by apache, and write out the json to a file
[21:38:57] <ebernhardson>	 should be like, 5 lines
[21:39:04] <bd808>	 ebernhardson: Also https://old-bugzilla.wikimedia.org/show_bug.cgi?id=43601
[21:39:32] <bd808>	 Ori actually wanted a non-varnish solution for the extension itself
[21:39:45] <guillom>	 Hmm, so, newbie question: Can I run a query from the console and not from the sql interactive shell?
[21:40:04] <bd808>	 guillom: sure use `mysql`
[21:40:17] <bd808>	 I think the .my.cnf should be setup to connect
[21:40:20] <ebernhardson>	 guillom: echo "select 1 from revision" | mysql
[21:40:21] <guillom>	 I must be doing something wrong
[21:40:37] <guillom>	 .my.cnf works fine
[21:40:54] <guillom>	 I want to run a query written in a text file
[21:41:18] <ebernhardson>	 guillom: cat filename | mysql
[21:41:26] <ebernhardson>	 or mysql < filename
[21:41:45] <guillom>	 when I try mysql -h commonswiki.labsdb < commons_query.txt I get an Access denied error
[21:41:57] <guillom>	 even though I can do the same query from the interactive shell
[21:42:03] <guillom>	 Sorry for the newbiness.
[21:42:06] <ebernhardson>	 hmm
[21:43:26] <ebernhardson>	 guillom: sorry i'm not sure, if `mysql -h commonswiki.labsdb` connects and queries, but `mysql -h commonswiki.labsdb < filename` doesnt, that is some oddity i've never encountered :S
[21:44:28] <guillom>	 ebernhardson: Thanks; I'm assuming it's something Labs-specific because the Internet doesn't seem to know either.
[21:45:21] <Coren>	 guillom: When you say "from the interactive shell" you are using 'mysql' to compare, not the 'sql' wrapper, yes?
[21:45:28] <guillom>	 Coren: yes
[21:45:56] <guillom>	 mysql -h commonswiki.labsdb works (I can connect)
[21:45:58] <Coren>	 Because sql uses replica.my.cnf directly, whereas mysql defaults to .my.sql which may not contain the same credentials.
[21:46:27] <Coren>	 Hm.  And you're sure you have read permission on commons_query.txt yes?
[21:46:38] <guillom>	 AHAH! Good question
[21:46:42] <guillom>	 checking
[21:47:05] <guillom>	 yes, it's all the tool's
[21:48:16] <Coren>	 Because that literally doesn't make sense.  mysql simply accepts commands from stdin; it should make no difference whether it's from the tty or a pipe.
[21:48:29] <guillom>	 hmmm
[21:49:56] <darkblue_b>	 hi Coren.. still lurking here.. but let me recommend Postgres for certain things :-)
[21:49:59] <guillom>	 tools.mrmetadata@tools-login:~$ mysql -h commonswiki.labsdb < commons_query.txt 
[21:49:59] <guillom>	 ERROR 1045 (28000) at line 2: Access denied for user 's52317'@'%' (using password: YES)
[21:50:18] <darkblue_b>	 sorry about the perms issue at hand, no suggestions there
[21:50:43] * Coren ponders.
[21:50:54] <Coren>	 Where is that txt file?
[21:51:13] <guillom>	 in the tool's root folder. Do I need an absolute path?
[21:51:26] <Coren>	 I need a tool's name.
[21:51:29] <Coren>	 :-)
[21:51:37] <guillom>	 oh, sorry :) mrmetadata
[21:52:11] <guillom>	 Coren: As I said, sql newbie here. Ugly things in that folder.
[21:54:45] <guillom>	 Could the issue be the outfile?
[21:54:52] <Coren>	 guillom: That query doesn't actually work by hand either, and it's clear why:  you realize that INTO OUTFILE tries to create a file on the *server* right, not the client?  :-)
[21:55:30] <guillom>	 Coren: Well, no, I didn't, or I would be even more of an idiot :P
[21:55:44] <Coren>	 You want to add -e to mysql, and redirect the output instead.  Like:  'mysql -e -h commonswiki.labsdb < filename >outputfile.txt'
[21:56:02] <guillom>	 Coren: Thank you so much!
[21:57:11] <guillom>	 I think I need to get someone to teach me mysql properly, instead of just making do with stackoverflow
[21:57:51] <guillom>	 I had forgotten I had added that last line after testing it by hand.
[23:15:19] <guillom>	 Does anyone have a working example of a job running a mysql command?
[23:15:49] <guillom>	 I've got to the point where the command works, but I can't seem to manage to jsub it
[23:26:31] <guillom>	 (Fixed with a shell script)