[08:26:52] 3Wikimedia Labs / 3tools: Add some of the missing tables in commonswiki_f_p - 10https://bugzilla.wikimedia.org/59683#c17 (10Silke Meyer (WMDE)) Thanks, Coren! merl, can you please check and confirm this? [08:59:08] 3Wikimedia Labs / 3tools: Add some of the missing tables in commonswiki_f_p - 10https://bugzilla.wikimedia.org/59683#c18 (10merl) 5RESO/FIX>3REOP No, globalimagelinks is still missing (mentioned in comment #1 and #5): $mysql -hs5.labsdb commonswiki_f_p -e "show tables like 'globalimagelinks'" Empty Set [09:05:32] Hi Coren, seen this one? https://bugzilla.wikimedia.org/show_bug.cgi?id=59683#c18 Please have a look! [09:13:37] 3Wikimedia Labs / 3tools: struct::stack/TclOO version mismatch - 10https://bugzilla.wikimedia.org/66984#c1 (10Silke Meyer (WMDE)) p:5Unprio>3High Increasing importance. Coren, what do do say about Giftpflanze's suggestion? [09:22:52] 3Wikimedia Labs / 3tools: watchlist table not available on labs - 10https://bugzilla.wikimedia.org/57617 (10Silke Meyer (WMDE)) [09:22:52] 3Wikimedia Labs / 3tools: Missing Toolserver features in Tools (tracking) - 10https://bugzilla.wikimedia.org/58791 (10Silke Meyer (WMDE)) [12:15:40] 3Wikimedia Labs / 3tools: Cannot start java processes using the grid engine - 10https://bugzilla.wikimedia.org/67588 (10nosy) 3NEW p:3Unprio s:3normal a:3Marc A. Pelletier Running a tool via crontab with jsub results in the following error: Picked up _JAVA_OPTIONS: -Xmx150M Error: Could not create t... [12:27:22] 3Wikimedia Labs / 3tools: Cannot start java processes using the grid engine - 10https://bugzilla.wikimedia.org/67588#c1 (10merl) Have you tried to add more mem to sge runtime environment for the job? Also some of my simple java tools need at least 4G mem (-l h_vmem=4000M). [12:37:07] 3Wikimedia Labs / 3tools: Add some of the missing tables in commonswiki_f_p - 10https://bugzilla.wikimedia.org/59683#c19 (10merl) Also templatelinks problem on s5 select * from commonswiki_f_p.templatelinks limit 1; ERROR 1296 (HY000): Got error 10000 'Error on remote system: 1054: Unknown column 'tl_from_n... [13:03:08] 3Wikimedia Labs / 3tools: Cannot start java processes using the grid engine - 10https://bugzilla.wikimedia.org/67588#c2 (10nosy) Unfortunatelly the original parameters of SGE as well as you suggested one do not work. [14:20:01] 3Wikimedia Labs / 3tools: Add some of the missing tables in commonswiki_f_p - 10https://bugzilla.wikimedia.org/59683#c20 (10Marc A. Pelletier) globalimagelinks added, tl_from_namespace has issues (being worked on) [14:22:17] 3Wikimedia Labs / 3tools: Cannot start java processes using the grid engine - 10https://bugzilla.wikimedia.org/67588#c3 (10Marc A. Pelletier) Could you specify the exact command line you are using to run the tool? [14:29:31] 3Wikimedia Labs / 3tools: Add some of the missing tables in commonswiki_f_p - 10https://bugzilla.wikimedia.org/59683#c21 (10Marc A. Pelletier) BTW, now would be a good time to mention if there are other irregularities and not just after I fixed both of those. :-) [14:31:16] 3Wikimedia Labs / 3Infrastructure: Labs: Mails from cron should not go to ops-wikimedia - 10https://bugzilla.wikimedia.org/45827#c1 (10Andre Klapper) Anybody knows if this is still a problem, or obsolete? [14:35:16] 3Wikimedia Labs / 3Infrastructure: Labs: Mails from cron should not go to ops-wikimedia - 10https://bugzilla.wikimedia.org/45827#c2 (10Marc A. Pelletier) At the very least, that's no longer completely descriptive as the tools project at least doesn't send global root mail to the list. [14:43:26] Installing handbuilt packages are okay in Lab instance? [14:43:37] (with little changes) [14:50:28] kart_: Locally (to the tool) sure; but if it's reasonable to expect many other users would want to use it it might be better to install it globally. [14:50:42] kart_: (As was done with mwparserfromhell for instance) [14:51:21] kart_: Also, that complicates puppetizing your setup; but that only impacts you. :-) [14:56:02] 3Wikimedia Labs / 3tools: Add some of the missing tables in commonswiki_f_p - 10https://bugzilla.wikimedia.org/59683#c22 (10Marc A. Pelletier) 5REOP>3RESO/FIX templatelinks updated to track the recent schema changes. [14:56:36] Coren: just for one instance. Crazy bug. [14:57:22] kart_: Yeah, there's nothing wrong with the idea just the usual disclaimer applies: "If you break [your instance] you get to keep both pieces." :-) [14:58:48] hi, I want to use a regxp instead of a plain text in the crontab but I think it doesn't work [15:00:16] I have somthing like: [15:00:16] 0 10,11,18,19,22,23 * * * [[ "$(TZ=:Europe/Andorra date +\%H)" =~ 12|20|00 ]] /usr/bin/jsub -N t07-lst -once -quiet program option [15:00:49] but it generate an error [15:02:02] 0 10,11,18,19,22,23 * * * [[ "$(TZ=:Europe/Andorra date +\%H)" =~ 12|20|00 ]] && /usr/bin/jsub -N t07-lst -once -quiet program option [15:02:42] What error it generate coet|cawiki? [15:06:35] Miguel2706: http://pastebin.com/fScDwpaL [15:06:45] Ummm [15:09:23] it is based on https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help#Specifying_time_zones [15:09:46] * andrewbogott shakes fist at merlbot [15:10:04] The line works in bash normally, coet|cawiki? Using only [[ "$(TZ... [15:10:05] but using a regexp, my other expression in plain text work fine [15:10:17] yes [15:10:39] it does [15:10:52] coet|cawiki: I think cron will fail to properly parse bashisms. [15:11:37] It is "sh", no bash [15:11:53] Right. [15:12:44] Hence: no bashness; iirc [[ is a bash-specific extension [15:13:37] Hey guys, a friend wants use Tools shell, would you please add him on https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools ? His username is User:Calak and he had a toolserver account formerly [15:14:05] maybe i must write it as [ "$(TZ=:Europe/Andorra date +\%H)" =~ "12|20|00" ] [15:14:08] ? [15:14:13] (Also, syntactically, you should probably quote the regex, I'm pretty sure the | will be parsed as a quote operator) [15:15:27] ebraminio: What's his username? [15:15:29] He has shell permission also https://wikitech.wikimedia.org/w/index.php?title=Special:Log&page=User%3ACalak but unable to login on tools with it [15:15:32] andrewbogott: i only started all my merlbot scripts like it is done every night on toolserver [15:15:35] Coren: Calak [15:16:16] Merlissimo: I'm getting literally hundreds of 'job failed' messages all of which are from merlbot. Seems like it's crashing and restarting every second or two. [15:16:29] Merlissimo: pm me your email and I'll forward you some of the messages if you like. [15:16:50] they are restarting once and the second time they go into error state, so that they are not restarted [15:17:02] ebraminio: Yeah, I see; he didn't make a request to be added -- but I just added him now. [15:17:33] andrewbogott: error are ERROR 145 (HY000) at line 5: Table './mysql/proc' is marked as crashed and should be repaired or ERROR 2013 (HY000) at line 8: Lost connection to MySQL server during query [15:17:37] Coren: Oh, that would be great. Thank you [15:18:20] my other doubt is: how i can remove some user from the mail_list of the jsub? [15:18:36] Merlissimo: Hm, it seems like I only ever get email notices from your bot, I wonder what's different [15:19:12] Merlissimo: anyway, I'm in the middle of a scheduled window -- I'll return to you in a bit :) [15:19:14] Merlissimo: What database is that? [15:19:48] Merlissimo: btw, you're saying that by design your tool involves running hundreds of concurrent jobs in the queue? [15:19:49] andrewbogott: you got mail because if my script fails i returd exit code 100 which sets the job to error state, so that they cannot be restarted by e.g. cron. i think no other is using this [15:20:02] Coren: Thank you very much :) [15:20:59] andrewbogott: minor patches? https://gerrit.wikimedia.org/r/#/c/144330/ and https://gerrit.wikimedia.org/r/#/c/144331/ [15:21:01] andrewbogott: yes, mostly one array job with one task for every wiki [15:21:14] YuviPanda: later please, I'm deploying atm [15:21:19] andrewbogott: ah, ok [15:21:52] Coren: can you merge https://gerrit.wikimedia.org/r/#/c/144331/? is preventing https://gerrit.wikimedia.org/r/#/c/144360/ from being merged (despite +2) [15:22:40] andrewbogott: on toolserver i limited the maximum numbers of job running in paralell by using user_slot resource, but this is not available on labs, see https://bugzilla.wikimedia.org/show_bug.cgi?id=52976 [15:23:15] Merlissimo: That's on my todo for today. [15:24:48] 3Wikimedia Labs / 3tools: struct::stack/TclOO version mismatch - 10https://bugzilla.wikimedia.org/66984#c2 (10Marc A. Pelletier) Unable to reproduce: marc@tools-login:~$ tclsh % package require struct::stack 1.5.1 [15:25:03] o.o [15:25:09] Coren: I just deployed the new servicegroup code on wikitech. Looks ok to me, but let me know if anything seems amiss. [15:26:07] gifti: No shit; that's copy-paste from my ssh. Perhaps you have a locally installed version that's confusing your tcl? [15:26:13] Coren: can you repeat with tool account? [15:26:40] gifti works tools.giftbot doesn't [15:27:09] Coren: i certainly don't have tcllib installed locally … [15:28:02] tools.csbot@tools-login:~$ tclsh [15:28:02] % package require struct::stack [15:28:02] 1.5.1 [15:28:15] huh [15:28:35] gifti: Some environment variable, perhaps, which tclsh uses? [15:28:59] i guess 8.5/8.6 [15:29:02] something [15:29:07] Also, what does 'which tclsh' give you? [15:36:17] YuviPanda: since this deletes labs volumes, do you mind reading through this and confirming that it won't run amok? https://gerrit.wikimedia.org/r/#/c/144063/ [15:36:19] Coren: 8.5.11 doesn't give error because it doesn't have tcloo, 8.6.0 has and chooses different implementation [15:38:26] gifti: Aha. Hmm. I honestly don't know how to fix that for you then, I know epsilon about tcl. [15:39:12] andrewbogott: looking [15:39:26] Coren: i suggested two things, how complex would they be? [15:39:29] gifti: Short of rebuilding tcl_8.6.0 if you tell me what needs to be changed. [15:41:17] (remember tcl 8.6 isn't actually supported in Precise, I had to built it blindly) [15:41:39] it's not about tcl, it's about tcllib [15:42:37] gifti: ... which is not (supposed to be) version-dependent; at least I don't have a 8.6 specific version. [15:42:45] this old version suboptimally has a fixed version requirement [15:42:56] and tcl8.6 has a newer version of tcloo [15:43:09] I see. [15:43:13] can you install a newer version of tcllib? [15:45:16] gifti: The currently installed version (1.14-dfsg-1) is the latest Precise version. I'd have to build a new one and hope nothing breaks in the system tcl. Not impossible, but an "interresting" challenge given my limited knowledge of tcl. If you build one that you test to work right with 8.6 and the system tcl, I have no issue with installing it though. [15:45:55] ok, will test it [15:46:32] gifti: It _might_ be possible to just backport the Trusty one if it's recent enough. If I were you, I'd start with that. [15:46:54] ok, what would i have to do for that? [15:55:00] Coren: ↑ [16:10:43] gifti: Well, the first step is to just download the deb and see if it installs at all or if it introduces dependency hell. [16:20:04] 3Wikimedia Labs / 3tools: Performance problem on database server s5 using commonswiki - 10https://bugzilla.wikimedia.org/67602 (10merl) 3NEW p:3Unprio s:3normal a:3Marc A. Pelletier Some queries are currently not runnable on s5.labsdb because of performance problems. Probably this is caused by federa... [16:20:17] 3Wikimedia Labs / 3tools: Missing Toolserver features in Tools (tracking) - 10https://bugzilla.wikimedia.org/58791 (10merl) [16:20:18] 3Wikimedia Labs / 3(other): (Tracking) Database replication services - 10https://bugzilla.wikimedia.org/48930 (10merl) [16:22:31] 3Tool Labs tools / 3[other]: merl tools (tracking) - 10https://bugzilla.wikimedia.org/67556 (10merl) [16:22:31] 3Wikimedia Labs / 3tools: Performance problem on database server s5 using commonswiki - 10https://bugzilla.wikimedia.org/67602 (10merl) [16:27:31] Coren: i'm a bit lost; which command would i use that doesn't require being root? [16:28:45] gifti: That will require you being root; so it'll need a project and an instance for this. I can attempt to make time for this in the coming weeks, but it's not going to be easy or fast for me to do it. [16:29:20] gifti: Sorry, I though I was clearer but clearly I left too much implicit. :-) [16:29:37] hm, i guess the best way forward is to just build it locally and close the bug [16:40:01] Coren: btw, diamond logs have been made *much* less noisy, should only log errors now :) [16:46:47] 3Wikimedia Labs / 3tools: struct::stack/TclOO version mismatch - 10https://bugzilla.wikimedia.org/66984#c3 (10Giftpflanze) 5NEW>3RESO/WOR I just installed tcllib 1.16 locally [16:51:20] Merlissimo: I'm trying to answer to your bug in a way that doesn't take twenty paragraphs of tl;dr and failing. :-) The short of it is that what you're doing isn't supported and you should do this with trivial application logic instead, but explaining how in detail hardly fits in a bug report. :-) [16:52:44] Merlissimo: If you want, I can write example queries for you; a code snippet is worth a tousand words? :-) [16:53:49] Coren: we had the application logic think last year and i told WMDE that i have not time to rewrite hundrets of existing scripts developed over years. [16:54:34] the response by Silke_afk was that these scripts will run on labs [16:55:25] my bot is quite essential for dewiki community [16:56:37] Merlissimo: We have been clear that federation has specific limits, and help for adaptation was offered for the past 18 months. I'd be glad to help you adapt your bots, still, though obviously making "hundreds" of ports will not be instantaneous. [16:57:28] That said, they _will_ run, as-is, but clearly not at any sort of reasonable speed (which you have already noticed) [16:57:59] It may be possible to adapt your query so that it can still be done in SQL directly, but adaptation is nonetheless required. [16:58:39] (And if there is going to be changes, might as well do the app logic thing which is going to be much more efficient to begin with than the combined query ever could) [16:59:30] For instance, "select page_id from categorylinks, page where cl_to LIKE 'Deletion_requests_%' and page_id=cl_from;" works fine through federation (0.04s) because it uses a strict index. [17:00:13] Coren: If you joined that with, say, a query from dewiki it's be just as fast as running it locally. This is the kind of scenario that works fine with federation. [17:00:34] it is one query. solving only this one is less than 1% of the solution. rewriting it in application logik ist not skill problem, it is a time problem because i can do this only in my freetime. [17:00:39] Where it breaks down is if you attempt to do a join between two federated tables on the external database. [17:02:08] Merlissimo: We have had people available to help you with adaptation for over a year; WMDE made part of their development team available exactly for cases like this. I wish you had raised your lack of time earlier, we'd have been able to help you adapt everything in time. [17:02:52] Merlissimo: Now the best I can do is try to help you to fix things as fast as possible for the most critical ones; we can also try to see if WMDE still can spare the resources to help. [17:04:09] Merlissimo: So the first thing I need is the list of all of your queries. With much, many (most?) of them will require little to no adaptation. [17:04:55] Merlissimo: Obviously, any that already only operate on just dewiki or just commons won't need any change. [17:06:55] see all jobs currently in sge error state. they all beling to one script. [17:09:19] Coren: what abould the expresstrain solution that was used on toolserver? [17:09:19] BilderWunsch_Vorschlag.weekly2.qsub [17:09:22] ? [17:09:37] for example, yes [17:10:59] Merlissimo: That solution caused 95% of the replication problems in tool labs. We'll have a comparable system once we switch to MariaDB 10 (except that it won't break things) but that's not going to be available for several months unless I manage to get our DBA to do this mostly exclusively. [17:11:13] s/in tool labs/on the toolserver/ [17:11:27] * Coren sighs. [17:11:57] I wish we had started this months ago; we'd have been ready by now if we had. Lemme try to see what resources I can put together to do this as quickly as possible. [17:15:43] i was not able to test my script earlier because of missing features. the last year i have rewritten a free java libary only because it was not open source. and i am in contact with wmde for more than a year because of this. so nothing i have done wrong [17:21:46] 3Wikimedia Labs / 3deployment-prep (beta): db error on beta labs "centralauth.renameuser_status' doesn't exist" - 10https://bugzilla.wikimedia.org/67485#c4 (10Greg Grossmeier) So, obvious question: The schema change this represents has been made/will be made before the complaining code goes to production, ri... [17:28:49] 3Tool Labs tools / 3[other]: Fatal error: Allowed memory size exhausted - 10https://bugzilla.wikimedia.org/67606 (10Helder) 3NEW p:3Unprio s:3normal a:3None When I try to use http://tools.wmflabs.org/pathoschild-contrib/catanalysis/?title=!Artigos+destacados&cat=1&wiki=ptwiki I get this error: Fatal... [17:30:01] Merlissimo: I'm not assigning blame, I'm trying to scrounge up resources to help you. [17:31:32] 3Wikimedia Labs / 3tools: Performance problem on database server s5 using commonswiki - 10https://bugzilla.wikimedia.org/67602#c1 (10Marc A. Pelletier) That query is unlikely to be usable as-in with federation, at least not in its current form (using a federated table requires that the rows selected during a... [17:35:06] Coren: yes, thx. [17:45:14] hi! im new to labs, and i have troubles connecting to database. actually, i never connected, and im wondering if theres anyone who could help me? [17:50:01] 3Wikimedia Labs / 3tools: Performance problem on database server s5 using commonswiki - 10https://bugzilla.wikimedia.org/67602#c2 (10Sam Reed (reedy)) *************************** 1. row *************************** id: 1 select_type: SIMPLE table: templatelinks type: ref possible... [17:52:54] sanyi4: What happens if you try? Indeed, how are you trying? :-) [17:53:44] sanyi4: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help#Database_access probably has all you need. [18:03:17] i know that page, but its not helping much. im not an expert. i know php and sql, im able to write mysql queries in a php script, and thats it. so my genaral question is: what is the database name, user, and password i have to write in my php script? [18:05:41] sanyi4: The username in password are in your tool's home directory (in the file replica.my.cnf). The database name depends on the database (heh) - it's going to be in the form enwiki.labsdb (for the English Wikipedia, for instance) [18:08:29] is there a way to create a separate database for a certain tool? [18:09:49] sanyi4: you really should read through that link Coren gave :) [18:11:12] ok, more carefully this time :) [18:11:24] sanyi4: :) [18:28:33] 3Wikimedia Labs / 3deployment-prep (beta): db error on beta labs "centralauth.renameuser_status' doesn't exist" - 10https://bugzilla.wikimedia.org/67485 (10Greg Grossmeier) [18:32:02] 3Wikimedia Labs / 3deployment-prep (beta): db error on beta labs "centralauth.renameuser_status' doesn't exist" - 10https://bugzilla.wikimedia.org/67485#c5 (10Kunal Mehta (Legoktm)) So, about 5-10 minutes after I merged the patch into master, The DJ came on IRC and reported that the table was misssing, so I... [18:39:32] Echo! [19:14:13] Coren: i think something on tools-login is full [19:14:36] "something"? [19:14:40] memory [19:14:59] Yeah, it's trashing a little. Someone must be cheating. [19:16:09] Coren: want me to kill it? 'convert' from apper [19:16:23] I've already killed a bot. [19:16:37] 3228 apper 20 0 2148m 1.6g 2448 S 0 42.5 0:08.04 convert [19:16:51] sudo kill 3228 [19:17:24] Yeah, that was next on my hitlist (clearly) [19:17:42] :) cook [19:57:47] (03PS1) 10BryanDavis: Add dummy ssl key for stream.wikimedia.org.key [labs/private] - 10https://gerrit.wikimedia.org/r/144504 [19:58:21] (03CR) 10BryanDavis: [C: 032 V: 032] Add dummy ssl key for stream.wikimedia.org.key [labs/private] - 10https://gerrit.wikimedia.org/r/144504 (owner: 10BryanDavis) [20:07:58] !log deployment-prep Ran `apt-get dist-upgrade` on deployment-analytics01 to upgrade hadoop, hive, pig, etc which were failing to update via puppet. [20:08:02] Logged the message, Master [20:20:28] !log deployment-prep Fixed puppet on deployment-analytics01 with manual apt-get commands. [20:20:31] Logged the message, Master [20:55:54] !log deployment-prep Killed stuck `apt-get update` job on deployment-jobrunner01 started on Jun17 [20:55:56] Logged the message, Master [20:56:38] bd808: btw, I've been working on http://tools.wmflabs.org/giraffe/index.html#dashboard=ToolLabs+Basics&timeFrame=1h for tools, can make one for betalabs too if you want [20:57:00] bd808: we also have a 'time since last puppet run' metric in labs graphite now, so that can potentially be used to report on puppet failures [20:57:47] YuviPanda|zzz: Getting some dashboards started for beta would be nice. I made https://logstash-beta.wmflabs.org/#/dashboard/elasticsearch/puppet%20runs to look at puppet failures. [20:58:10] bd808: uid/pwd? doesn't seem to be ldap [20:58:11] There is no alerting in my logstash view however [20:58:26] YuviPanda|zzz: No, not ldap because labs :( [20:58:31] bd808: right. I've no alerting at all now either. All the graphite based alerting solutions I've looked at seem a bit, ugh. [20:58:44] YuviPanda|zzz: Password is at https://office.wikimedia.org/wiki/User:BDavis_(WMF)/logstash [20:59:54] bd808: I'm going to guess that alerts are more important than dashboards... [21:00:38] YuviPanda|zzz: Possibly, although at the moment I'm not sure who puppet failure alerts should go to. [21:00:48] hehe :) [21:01:52] I really would like them to be tied to the gerrit commits and bitch at the people who authored and merged :) [21:02:43] heh, but that's doable outside of graphite, I'd suppose [21:02:55] instead of a rebase, cherry-pick one by one, run puppet, see if things fail, bitch appropriately [21:03:00] that probably gets slow tho [21:13:40] !log graphite stopped carbon on diamond-collector to investigate io issues [21:13:42] Logged the message, Master [21:51:57] Coren: Erm [21:52:04] * a930913 ducks. [21:52:07] Why is Catscan having problems? [22:00:02] Hello, #wikimedia-labs. I am an developer from pre-Gerrit days now trying to start up in the post Gerrit world. [22:00:22] I seem to have a wikitech account, but it's probably tied to neilk@wikimedia.org, which I don't have access to. [22:00:33] Is the simplest thing just to regain access to neilk@wikimedia.org email? [22:00:42] * YuviPanda|zz waves at neilk_ [22:00:44] welcome back! [22:00:47] hi Yuvi [22:01:06] neilk_: :) Easiest is to regain that email, I'd say [22:01:16] YuviPanda: ok who's the person to see [22:01:16] neilk_: and then your account can be migrated from svn to ldap, if it already hasn't [22:01:29] neilk_: someone in OIT to regain access to that email, I'd suppose [22:02:15] YuviPanda: OIT? [22:02:21] neilk_: oh, Office IT, I mean. [22:02:27] hm [22:02:28] neilk_: they administer @wikimedia.org emails [22:02:32] ok [22:02:34] so [22:03:20] What does one have to do to qualify for an @wikimedia email? [22:03:44] a930913: some minor blood sacrifice [22:03:54] :o [22:04:17] YuviPanda: incorrect. MAJOR blood sacrifice [22:04:38] neilk_: shouldn't let that out until it is too late [22:05:01] The extent of the sacrifice depends on the moon cycle [22:05:12] not to mention the release cycle. [22:05:27] But since the moon is in the house of Orion, it's relatively minor [22:05:37] * marktraceur bullshitting (like all people talking about astrology) [22:07:15] Hmm, does it have to be my blood? [22:07:23] a930913: Not sure [22:07:34] I thought it had to be a virgin [22:07:55] Yes but the species is not specified [22:08:08] Indirectly killing server kitties is probably sufficient [22:08:19] Qcoder00: Like that makes a difference to a computer geek ;) [22:08:36] marktraceur: Oh, so I can just add some transclusions? [22:09:02] No no, add millions. And then change the template hundreds of times, touching each page in between. [22:09:08] * YuviPanda is killing server kitties atm [22:09:14] Wait this is bad advice. [22:10:03] * a930913 rescues YuviPanda's server kitties and puts them up for adoption. [22:11:21] TOO LATE [22:11:34] Noo! [22:11:41] * a930913 harvests the blood anyway. [22:12:27] huh, is Chip Deubner working there or is it all Julie Trias? [22:12:31] For adoption: Most of a kitty [22:12:45] neilk_: chip and julie are both working there, yes [22:12:47] neilk_: Send the mail to the tech support email...sec, PM [22:12:48] marktraceur: Sold for parts? [22:13:00] Yuvi: ok. The staff page is confusing, it looks like cdeubner is a bot or something [22:13:04] Hahaha [22:13:19] He basically is [22:13:22] neilk_: no no, *harvested* for parts [22:13:37] * YuviPanda wails at graphite [22:13:40] WHY GRAPHITE, WHY [22:13:48] neilk_: heh :D what marktraceur said [22:19:54] !log graphite rebooted diamond-collector [22:19:56] Logged the message, Master [22:53:14] !log graphite delete all the metrics in diamond-collector, investigating performance issues still [22:53:16] Logged the message, Master [22:58:30] !log graphite puppet disabled on diamond-collector for now, trying graphite with a more coarse storage-schema [22:58:31] Logged the message, Master [23:04:20] ok so another Gerrit question... from its perspective, the beginning of time is March 22, 2012? [23:05:17] <^d> Beginning of review time was then. Various repos had earlier commit history converted. [23:22:38] ^d: okay so it seems that it is not expected that usernames in the git repo carry over from the svn era. Pre-git users are @users.wikimedia.org, and current users are whatever they are (for instance chadh@wikimedia.org or [23:22:55] s/or/)/ [23:23:15] <^d> @users.wikimedia.org doesn't exist and never did. [23:23:25] <^d> It was just a fake catch-all we created when converting svn history. [23:23:30] ^d: I get it. Thanks