[00:50:21] <Pathoschild>	 With the recent replication changes, do we still need to connect to multiple slices for crosswiki queries or will all wikis be accessible from any slice?
[02:15:10] <Betacommand>	 Pathoschild: its about the same as the toolserver used to be
[02:15:28] <Betacommand>	 there are no more _f_p tables
[02:16:10] <Pathoschild>	 So we still need one connection per slice to query every wiki, then.
[02:17:12] <Betacommand>	 Pathoschild: its one connection per cluster
[02:17:28] <Betacommand>	 S1 S2....
[02:19:53] <Pathoschild>	 Yep.
[02:20:41] <Pathoschild>	 Thanks. :)
[02:58:37] <wikibugs>	 3Wikimedia Labs / 3Infrastructure: Database upgrade MariaDB 10: Engine / Option mismatch on table `user_properties` - 10https://bugzilla.wikimedia.org/68942#c2 (10Sean Pringle) https://mariadb.atlassian.net/browse/MDEV-5867
[03:05:04] <comets>	 O_O
[03:10:51] <wikibugs>	 3Wikimedia Labs / 3Infrastructure: Database upgrade MariaDB 10: Engine / Option mismatch on table `user_properties` - 10https://bugzilla.wikimedia.org/68942#c3 (10Sean Pringle) Have applied on labsdb replicas:  SET GLOBAL SQL_MODE=IGNORE_BAD_TABLE_OPTIONS;  This will only help when loading dumps that include...
[03:26:08] <wikibugs>	 3Wikimedia Labs / 3deployment-prep (beta): beta labs not responding; API shows 503 from varnish - 10https://bugzilla.wikimedia.org/68574#c7 (10Brett Simmers) 5NEW>3ASSI a:3Brett Simmers I'm working on this. I'm pretty sure it's a bug in hhvm's fastcgi server.
[04:33:03] <spagewmf>	 beta labs still not responding
[05:14:24] <swtaarrs>	 spagewmf: yeah it's hhvm having a bad time. I've figured out the issue but don't know when I'll be able to have a fix
[05:14:30] <swtaarrs>	 in the meantime, I'll restart hhvm on the beta servers
[05:17:31] <swtaarrs>	 !log deployment-prep restarted hhvm on deployment-mediawiki0{1,2} to unwedge them
[05:17:34] <labs-morebots>	 Logged the message, Master
[08:53:04] <swtaarrs>	 !log deployment-prep rebuilt and restarted hhvm on deployment-mediawiki02 with potential fix
[08:53:06] <labs-morebots>	 Logged the message, Master
[09:07:29] <qchris>	 Connecting to labsdb (e.g. host 192.168.99.3) does not work from i-00000325.eqiad.wmflabs, while it works from i-00000125.eqiad.wmflabs.
[09:07:53] <qchris>	 Some software on i-00000325.eqiad.wmflabs went wild before and mass connected to the database,
[09:08:30] <qchris>	 is there some automatic per-IP rate-limiting in place that
[09:08:47] <qchris>	 would exclude the previously misbehaving instance?
[09:53:06] <qchris>	 qchris: Sorry for the noise. The i-00000325.eqiad.wmflabs was missing DNATs. Can connect now again.
[10:02:07] <wikibugs>	 3Wikimedia Labs / 3Infrastructure: Database upgrade MariaDB 10: Engine / Option mismatch on table `user_properties` - 10https://bugzilla.wikimedia.org/68942#c4 (10metatron) Thanks. In addition I've set up a regex to filter this mismatch. This was mainly for https://tools.wmflabs.org/tools-info/optimizer.php...
[11:53:07] <Cyberpower678>	 springle, ping
[11:53:31] <Cyberpower678>	 @seen spagewmf 
[11:53:31] <wm-bot>	 Cyberpower678: I have never seen spagewmf 
[11:53:33] <Cyberpower678>	 @seen spagewmf
[11:53:33] <wm-bot>	 Cyberpower678: spagewmf is in here, right now
[11:53:42] <Cyberpower678>	 spagewmf, have a moment?
[12:53:37] <wikibugs>	 3Wikimedia Labs / 3tools: Install Linux kernel AIO access library: libaio (libaio1) - 10https://bugzilla.wikimedia.org/68615 (10metatron) 5PATC>3RESO/FIX
[14:58:40] <wikibugs>	 3Wikimedia Labs / 3tools: Install package: python-pygments (syntax highlighting) - 10https://bugzilla.wikimedia.org/69050 (10metatron) 3UNCO p:3Unprio s:3normal a:3Marc A. Pelletier Pls. install ubuntu package python-pygments (syntax highlighting) http://packages.ubuntu.com/precise/python-pygments
[16:03:28] <scfc_de>	 I'm seeing some random, apparently transient DNS strangities in Tools logs, e. g. 'server host resolves rdata host "tools-master" as "tools-master.eqiad.wmflabs"' (14:18Z), "unable to resolve host tools-webproxy.eqiad.wmflabs" (9:41Z), etc.  Is the DNS server well?
[16:36:14] <scfc_de>	 !log tools Manually rerouted mails for tools.mybot@tools-submit.eqiad.wmflabs
[16:36:16] <labs-morebots>	 Logged the message, Master
[16:39:32] <scfc_de>	 !log tools tools.mybot's crontab uses qsub without -M, added that as a temporary measure and will inform user later
[16:39:34] <labs-morebots>	 Logged the message, Master
[16:44:43] <ialex>	 scfc_de: No need to inform the user, it’s me :P
[16:51:09] <scfc_de>	 ialex: Ah, okay :-).  Well, you see in the crontab what I added.  You could also use jsub which does this automatically.
[16:52:51] <ialex>	 scfc_de: yup, thanks :)
[16:53:00] <Nemo_bis>	 hello ialex 
[16:53:07] <ialex>	 Nemo_bis: hey
[16:53:28] <Nemo_bis>	 so you stalk "mybot"? :D
[16:55:09] <ialex>	 Nemo_bis: well, it’s *my* bot :D
[16:59:30] <Nemo_bis>	 ;)
[18:00:47] <CP678>	 legoktm, I responded to the talk.  You can answer either there or here. :-)
[18:00:54] <legoktm>	 oh hey
[18:01:20] <legoktm>	 > I edit a flow board on dewiki. Will that edit show up on enwiki?
[18:01:26] <legoktm>	 the answer is "it depends"
[18:01:44] <Cyberpower678>	 On what?
[18:01:46] <legoktm>	 some flow discussions will be only on one wiki, and some will be cross-wiki
[18:02:27] <Cyberpower678>	 And say I wanted to count the flow edits made on enwiki, whether the board is cross wiki or not?  How can that be done?
[18:02:31] <legoktm>	 (I think, I'm not really on the Flow team anymore)
[18:02:39] <legoktm>	 I'm not sure
[18:02:49] <legoktm>	 most of the cross-wiki functionality hasn't been developed yet
[18:03:03] <Cyberpower678>	 I guess I have to push back development of the flow counter a bit.
[18:03:48] <Cyberpower678>	 legoktm, Are you good with JS.  Different topic.
[18:04:12] <legoktm>	 not really. I can hack my way around it if necessary though
[18:04:27] <Cyberpower678>	 :p
[18:04:50] <Cyberpower678>	 Thanks
[18:05:34] <YuviPanda>	 Cyberpower678: around?
[18:05:42] <Cyberpower678>	 Yes
[18:05:58] <YuviPanda>	 Cyberpower678: pm?
[18:14:08] <YuviPanda>	 !ping
[18:14:08] <wm-bot>	 !pong
[18:21:32] <rohit-dua>	 can i use scp to transfer files from tools lab to my computer?
[18:24:16] <YuviPanda>	 !Ping
[18:24:18] <YuviPanda>	 !ping
[18:24:18] <wm-bot>	 !pong
[18:24:24] <YuviPanda>	 rohit-dua: yeah, you should be able to
[18:25:06] <rohit-dua>	 is something like this correct:  scp -r rohit-dua@tools-login:/data/project/bub/public_html/BUB /home/rohit/Desktop
[18:25:56] <rohit-dua>	 I get a Permission denied (publickey,hostbased).
[18:26:18] <YuviPanda>	 rohit-dua: roht-dua@tools-login.wmflabs.org?
[18:26:26] <YuviPanda>	 rohit-dua: also hmm, I am unsure :| can you ssh in normally?
[18:30:22] <rohit-dua>	 YuviPanda: scp -r rohit-dua@tools-login.wmflabs.org:/data/project/bub/public_html/BUB /home/rohit/Desktop  gives Connection timed out. yes i'm able to ssh normally.
[18:30:58] <YuviPanda>	 rohit-dua: uh, I'm not sure :|
[18:34:10] <rohit-dua>	 YuviPanda: got it. my connection issue.:p
[18:34:22] <YuviPanda>	 rohit-dua: ah :)
[19:03:37] <wikibugs>	 3Wikimedia Labs / 3tools: install GD perl lib - 10https://bugzilla.wikimedia.org/67199#c1 (10Andre Klapper) s:5blocke>3normal Hi Aka, could you please elaborate what this is needed for, and what's the usecase?  > currently it does'nt find the GD package.  Pasting the output is welcome.
[19:47:06] <YuviPanda>	 !ping
[19:47:06] <wm-bot>	 !pong
[19:59:15] <hedonil>	 YuviPanda: Hi, seen the new  *fancy* Labs SQL Optimizer? :-) 
[19:59:27] <YuviPanda>	 hedonil: saw the email, IS AWESOME
[19:59:34] <YuviPanda>	 hedonil: where are you running these?
[19:59:44] <hedonil>	 YuviPanda: now waitung for your quarry to complete the SQL suite! 
[19:59:50] <YuviPanda>	 hedonil: :D
[20:00:05] <YuviPanda>	 hedonil: can I see the source? :)
[20:00:07] <hedonil>	 YuviPanda: Where? right here on labs
[20:00:19] <YuviPanda>	 hedonil: oh? I thought you couldn't run 'EXPLAIN' queries?
[20:00:33] <hedonil>	 YuviPanda: Haha
[20:00:41] <hedonil>	 YuviPanda: we are hackerz
[20:00:46] <YuviPanda>	 iseez :)
[20:00:48] <hedonil>	 ;-)
[20:00:53] <YuviPanda>	 hence asking for source :)
[20:00:59] <hedonil>	 sec
[20:01:10] <YuviPanda>	 did you setup a db with some imported data and run explain on that?
[20:01:30] <valhallasw`cloud>	 or are you running on the non-censored databases? >:)
[20:02:02] <hedonil>	 YuviPanda: main parts are: /data/project/tools-info/create-optimizer-db.sh
[20:02:12] <YuviPanda>	 valhallasw`cloud: heh :)
[20:02:21] <GerardM->	 any news if the copying of the Wikidata dumps has been done ... 
[20:02:41] <hedonil>	 YuviPanda: and  /data/project/tools-info/public_html/optimizer.php
[20:03:28] <valhallasw`cloud>	 hedonil: the php is not readable
[20:03:38] <hedonil>	 valhallasw`cloud: oh
[20:03:52] <valhallasw`cloud>	 hedonil: no +x on public_html for 'everyone'
[20:04:04] <hedonil>	 ? it's 664
[20:04:10] <YuviPanda>	 GerardM-: no news, and I don't think we'll get any until Wikimania
[20:04:20] <valhallasw`cloud>	 hedonil: 774
[20:04:26] <valhallasw`cloud>	 should be 775
[20:04:38] <GerardM->	 I really need the statistics so that I can prepare my presentation
[20:04:47] <hedonil>	 valhallasw`cloud:  better?
[20:05:06] <valhallasw`cloud>	 drwxrwsr-- 5 tools.tools-info tools.tools-info 4096 Aug  1 04:33 public_html/
[20:05:18] <hedonil>	 ah
[20:06:55] <hedonil>	 now it should be ok
[20:07:09] * YuviPanda guesses it's running the query with some sort of sleep and then using show analayze
[20:07:58] <hedonil>	 YuviPanda: no, it's a set of custom databases based on running schemas
[20:08:04] <YuviPanda>	 ah :D
[20:08:36] <YuviPanda>	 hedonil: springle was suggesting we setup a labsdb machine that contains 'snapshots' of the views, dumped once a week or something
[20:08:40] <YuviPanda>	 should provide accurate EXPLAINs
[20:09:41] <hedonil>	 YuviPanda: anything that works is welcome
[20:09:57] <YuviPanda>	 hedonil: indeed, and you've created the first one that works. YAY!
[20:10:41] <hedonil>	 YuviPanda: if springle provides snapshots, we have already a GUI for that ... ;-)
[20:10:45] <YuviPanda>	 :D
[20:10:47] <YuviPanda>	 indeed
[20:11:06] <YuviPanda>	 hedonil: I might integrate this into quarry at some point into the future. Until then, I'll just link to the tool from the UI
[20:12:33] <hedonil>	 YuviPanda: my thoughts! if quarry is operational + you have a repo + springle provides the snapshots  we can work on that together!
[20:12:46] <YuviPanda>	 hedonil: yay :)
[20:12:55] <YuviPanda>	 hedonil: quarry is kinda operational now, you can run queries, etc.
[20:12:58] <YuviPanda>	 hedonil: am tuning it over the weekend
[20:13:40] <YuviPanda>	 hedonil: quary repo is on gerrit, btw. analytics/quarry/web
[20:19:13] <hedonil>	 YuviPanda: ah analytics! my friends :P
[20:19:17] <hedonil>	 hehe
[20:19:20] <YuviPanda>	 hedonil: :D
[20:20:50] <hedonil>	 YuviPanda: I just submitted a patch https://gerrit.wikimedia.org/r/#/c/151295/
[20:21:51] <hedonil>	 YuviPanda: so thaht this syntax highlighting will be available and everything can be written in python
[20:21:51] <scfc_de>	 Don't we have now EXPLAIN functioning for running queries?
[20:22:21] <hedonil>	 scfc_de: yep It's SHOW EXPLAIN
[20:22:37] <hedonil>	 but you have to be fast, sometimes
[20:23:09] <scfc_de>	 Well, if it's too fast to catch it, no need to do optimization?!
[20:23:59] <hedonil>	 scfc_de: and one of the crucial parts of this tool is to provide the running definitions / indices / sanitized views at a glance
[20:25:08] <hedonil>	 fresh, no outdated references ! ;-)
[20:25:15] <YuviPanda>	 scfc_de: indeed, sometimes you want to plan, tweak and then execute :D
[20:25:45] <hedonil>	 YuviPanda: or play around a bit...
[20:26:46] <YuviPanda>	 yeah
[20:46:55] <YuviPanda>	 !ping
[20:46:55] <wm-bot>	 !pong
[20:49:28] <YuviPanda>	 hedonil: re: your patch, I updated the commit messages. we try to keep them not more than 80chars per line :)
[20:50:01] <hedonil>	 YuviPanda: Just saw it ;
[20:50:28] <YuviPanda>	 hedonil: also the 'prefix' would be the module this was updating, in this case 'tools'
[20:51:06] <hedonil>	 YuviPanda: btw. I recall Krinkle (and his message documentation guide)  "bytes are cheap - use them" :P
[20:51:22] <hedonil>	 hehe
[20:51:29] <YuviPanda>	 halfak: :D indeed. just not on the same line :D  
[20:51:32] <YuviPanda>	 err
[20:51:33] <YuviPanda>	 hedonil: ^
[20:51:48] <YuviPanda>	 hedonil: "Your commit message is not too large until it causes someone's computer to swap"
[20:51:49] <halfak>	 :P
[20:52:08] <valhallasw`cloud>	 YuviPanda: just make your commit message call git gc --aggressive :-p
[20:52:12] <Krinkle>	 Yep. Subject line < 80. Other lines wrap between 80-100 and write clearly and all you like. There's no need to worry about the length. A reviewer might trim it down if it seems appropiate.
[20:52:42] <YuviPanda>	 valhallasw`cloud: hehe
[20:53:01] <YuviPanda>	 Krinkle: you still have the record for commit message length vs patch length, don't you? :)
[20:59:47] <hedonil>	 Krinkle: btw. thanks for your git /gerrit guides, for me this is like a candle in the dark :-D
[20:59:59] <YuviPanda>	 \o/
[21:00:09] <YuviPanda>	 hedonil: puppet is fun when you get used to it :)
[21:00:38] <hedonil>	 YuviPanda: ...and to be *mingy* the modul is called toollabs :P
[21:00:53] <YuviPanda>	 hedonil: heh, true, but we've used tools as prefix all the time so far
[21:01:11] <hedonil>	 YuviPanda: tsstsstss
[21:02:03] <YuviPanda>	 :)
[21:02:10] <YuviPanda>	 some people use toollabs too
[21:02:11] <hedonil>	 Obey the rules! or make your own ones :P
[21:02:12] <YuviPanda>	 I think either is ok
[21:02:19] <hedonil>	 Ha!
[21:02:40] <YuviPanda>	 valhallasw`cloud: btw, I switched to halfak's MWOauth library
[21:03:01] <YuviPanda>	 valhallasw`cloud: mostly because I initially thought there was a unicode bug in yours, and then it turned out to be in mine :|
[21:03:10] <YuviPanda>	 valhallasw`cloud: also it has py3 compatibility, and I intend to switch at some point soon
[21:03:52] * halfak is meaning to submit a pull request to add mwoauth to flask-mwoauth
[21:04:40] <halfak>	 That is the *only* library where I tried to support python 2 and 3.  
[21:04:43] * halfak shivers
[21:04:47] <halfak>	 not worth it. 
[21:04:48] <YuviPanda>	 heh
[21:05:16] <YuviPanda>	 halfak: I'll probably switch quarry to py3 once everything else stabilizes
[21:05:39] <halfak>	 Hokay.  Always easier to start on 3 though.  
[21:06:01] <halfak>	 I just took a flying leap about 6 months ago.  I took some flack for not supporting 2 for a while. 
[21:06:31] <YuviPanda>	 halfak: yeah, I did, and then found that weird bug I filed on your mwoauth library, and had to go back to 2, and now I can go back, except I've to find out how to package the libraries I'm using for py3
[21:06:34] <YuviPanda>	 they all support py3
[21:07:01] <halfak>	 YuviPanda, I still haven't had the actual error message for that :P
[21:07:06] <halfak>	 Also, i can't replicate it. 
[21:07:13] <YuviPanda>	 halfak: yeah, because when I put mwoauth back it just worked
[21:07:28] <halfak>	 hmm.  I'm gonna close that bug. 
[21:07:32] <YuviPanda>	 halfak: ya
[21:08:43] <YuviPanda>	 halfak: also, YGM
[21:08:50] <halfak>	 YGM?
[21:09:03] <YuviPanda>	 halfak: ah, You've got mail
[21:09:12] <YuviPanda>	 halfak: I picked it up from Oliver, and presumed it was well known
[21:09:57] <legoktm>	 {{YGM}}
[21:10:25] <YuviPanda>	 heh
[21:10:32] <halfak>	 Ahh!  I would have got it if you put the brackets in.  
[21:10:37] <halfak>	 I like to {{done}} things.  
[21:10:50] <halfak>	 You may see {{done}} appear in my Markdown.
[21:10:54] * halfak shrugs
[21:10:57] <YuviPanda>	 indeed, I like {{done}} too
[21:22:35] <Krinkle>	 YuviPanda: Haha, yeah, I think I do.
[21:22:39] <Krinkle>	 (hold that record)
[21:22:42] <YuviPanda>	 :)
[21:22:47] <Krinkle>	 Though i wasn't aware of it when I reached it.
[21:22:52] <Krinkle>	 Any particular one you have in mind?
[21:23:06] <Krinkle>	 hedonil: Thanks, that's nice to hear.
[21:27:00] <valhallasw`cloud>	 yeah, would be good to merge mwoauth and flask-mwoauth
[21:27:31] <valhallasw`cloud>	 might not be trivial, though, as flask-mwoauth depends on flask-oauth
[21:28:23] <YuviPanda>	 Krinkle: not sure, Nemo_bis dug it up
[21:28:35] <YuviPanda>	 valhallasw`cloud: no reason for flask-mwoauth to depend on flask-oauth tho
[21:28:50] <YuviPanda>	 valhallasw`cloud: I got it on quarry pretty easily, should be not too hard to just turn that into a blueprint
[21:28:57] <valhallasw`cloud>	 YuviPanda: not necessarily, but flask-oauth does the grunt work :-p
[21:29:05] <valhallasw`cloud>	 I basically only changed url's
[21:29:37] * Nemo_bis doesn't remember anything
[21:32:32] <halfak>	 valhallasw`cloud, hard to comment without digging in, but mwoauth should do all of the lifting.  You'll need to manage tokens inbetween page loads.  That's the hardest bit. 
[21:32:42] <halfak>	 I assume there's a session thingie in flash-mwoauth
[21:32:47] <valhallasw`cloud>	 yeah
[21:33:08] <halfak>	 Docs: https://pythonhosted.org/mwoauth/
[21:33:18] <valhallasw`cloud>	 is either of you at the hackathon next week?
[21:33:26] <halfak>	 I'll be there
[21:33:28] <halfak>	 :) 
[21:34:57] <YuviPanda>	 valhallasw`cloud: me too
[21:35:27] <valhallasw`cloud>	 halfak: I'm confused -- how would I POST something using your library? do I need to do request signing etc myself?
[21:37:04] <halfak>	 I'm not sure what you mean.  Are you talking about after you have the access token?
[21:37:27] <valhallasw`cloud>	 yes
[21:37:32] <halfak>	 Check this out: https://github.com/wikimedia/MediaWiki-OAuth/blob/master/examples/request-oauthlib.py
[21:37:58] <halfak>	 There, I use the access token to construct a Auth1() for requests to make use of. 
[21:38:11] <valhallasw`cloud>	 I see
[21:38:41] <YuviPanda>	 yeah, and you can expose the underlying 'requests' instance to the users as well if you want, so they can customize requests if they want
[21:39:06] <halfak>	 Boo.  I don't want people to know that I'm using requests on the inside. 
[21:39:23] <halfak>	 So that I can change the underlying lib without causing issues. 
[21:39:56] * halfak realized YuviPanda might me talking about flask-mwoauth
[21:40:11] <YuviPanda>	 halfak: yeah
[21:40:12] <YuviPanda>	 I am
[21:40:21] <valhallasw`cloud>	 I think flask-oauth uses urllib2, actually
[21:40:30] <YuviPanda>	 requests ftw instead :)
[21:40:33] <halfak>	 :) 
[21:40:46] <valhallasw`cloud>	 flask-/oauth/.
[21:40:57] <valhallasw`cloud>	 not exactly within my control :-p
[21:41:12] <YuviPanda>	 valhallasw`cloud: indeed, so the proposal is to switch flask-*mw*oauth to use halfak's mwoauth and  request-oauthlib
[21:41:19] <valhallasw`cloud>	 uhuh
[21:41:47] <YuviPanda>	 I don't think it should be too much work, and can do the legwork sometime after wikimania if you won't have the time - provided you're open to merging it, of course
[21:43:44] <valhallasw`cloud>	 YuviPanda: should be possible without changing the outward-facing API, so I'm all for it
[21:43:50] <YuviPanda>	 valhallasw`cloud: wheee
[21:45:19] <valhallasw`cloud>	 the only issue is that I have completely forgotten how it works :-p
[21:47:36] <halfak>	 Woo!  
[21:47:58] * halfak is stoked about consolidating on solving MW Oauth weirdness in one stop instead of two.
[21:48:13] <halfak>	 e.g. that identify() call.  
[21:48:31] <halfak>	 Sat down with milimetric and csteipp to make sure all the critical checks were in place. 
[21:49:07] <YuviPanda>	 halfak: what does wikimetrics use?
[21:49:35] <YuviPanda>	 valhallasw`cloud: heh :)
[21:49:47] <halfak>	 wikimetrics is using a custom fork of flask-mwoauth right now
[21:49:51] <YuviPanda>	 heh
[21:49:58] <valhallasw`cloud>	 oh god :P
[21:50:08] <halfak>	 It incorporates all of the nonsense we worked out in mwoauth into flash-mwoauth
[21:50:13] <YuviPanda>	 heh
[21:50:15] <YuviPanda>	 heheh
[21:50:19] <halfak>	 But it predated mwoauth
[21:50:23] <halfak>	 so it's not that bad. 
[21:50:41] <YuviPanda>	 I'm glad I don't have to do this :| I still don't fully understand the underlying protocol, but 'tis ok
[21:51:21] <halfak>	 YuviPanda, https://www.mediawiki.org/wiki/Extension:OAuth/Usage
[21:52:01] <YuviPanda>	 I know, just haven't had the time to read thrugh it fully :D
[21:52:15] <valhallasw`cloud>	 the basic concept it simple :-p the actual protocol in practice, though, with signing the right bits of information at the right time...
[23:35:06] <YuviPanda>	 !ping
[23:35:06] <wm-bot>	 !pong