[00:18:35] 10Quarry: Add database selector - https://phabricator.wikimedia.org/T76466#1959492 (10Krenair) [00:18:38] 10Quarry: Build a schema explorer - https://phabricator.wikimedia.org/T71192#1959491 (10Krenair) [00:20:34] 10Quarry: Build a schema explorer - https://phabricator.wikimedia.org/T71192#739158 (10Krenair) Spoke to Yuvi about this, thinking about a box to the right hand side of the query editor displaying a list of tables. For each table, a collapsible table showing fields and their types. Obviously this would depend on... [00:23:07] error reason 1: 01/22/2016 15:20:10 [51118:11172]: can't stat() "/data/project/liangent-py/updatedyk.out" as stdout_ [00:23:15] NFS problem again? [00:23:30] job id 2587488 [00:24:20] how can I know which host it is/was running on? [00:33:31] 10Quarry: Cannot download data from a query with Unicode characters in its title - https://phabricator.wikimedia.org/T123031#1959503 (10Krenair) From `quarry-main-01:/var/log/uwsgi/app/quarry-web.log`: ```Traceback (most recent call last): File "/usr/lib/python2.7/dist-packages/flask/app.py", line 1836, in __c... [00:42:54] 10Tool-Labs-tools-Other: Please fix labs/tools/wikidipendenza - https://phabricator.wikimedia.org/T124560#1959510 (10Ricordisamoa) [00:50:44] 10Quarry: Cannot download data from a query with Unicode characters in its title - https://phabricator.wikimedia.org/T123031#1959513 (10Krenair) a:3Krenair [00:57:27] 10Quarry: Excel does not recognize Quarry CSV output as UTF-8 - https://phabricator.wikimedia.org/T76126#1959520 (10Krenair) So a patch for UTF-16 was merged, is Excel happy now? [01:07:43] iirc i used to be able to boot jessie instances in labs, but i'm not getting that option right now. Took a quick search through wikitech but didn't find anything, is this intentional? [01:08:29] trying because i need something that uses systemd [01:19:06] 6Labs: No debian image available for booting new instances - https://phabricator.wikimedia.org/T124566#1959522 (10EBernhardson) 3NEW [01:25:57] ebernhardson: CentOS ? [01:26:17] darkblue_b: well sure, but there are only two options in the labs list to boot up. both ubuntu :) [01:26:35] 6Labs: No debian image available for booting new instances - https://phabricator.wikimedia.org/T124566#1959530 (10scfc) This is a temporary configuration that will be fixed in the next few days. [01:27:09] I admit I have not gotten back to solr yet this year [01:27:20] .. tangentially related [02:09:04] 10Tool-Labs-tools-Other, 10Gerrit: Please fix labs/tools/wikidipendenza - https://phabricator.wikimedia.org/T124560#1959545 (10TTO) [02:25:42] 10Tool-Labs-tools-Other, 10Gerrit: Please fix labs/tools/wikidipendenza - https://phabricator.wikimedia.org/T124560#1959567 (10Krenair) 5Open>3Resolved a:3Krenair Done [02:35:56] 10Tool-Labs-tools-Other, 10Gerrit: Please fix labs/tools/wikidipendenza - https://phabricator.wikimedia.org/T124560#1959577 (10Ricordisamoa) >>! In T124560#1959567, @Krenair wrote: > Done Oh, thank you so much! :heart: [03:10:44] 10Quarry, 5Patch-For-Review: Query counter increases but draft query is not accessible when window is closed and query doesn't have a title - https://phabricator.wikimedia.org/T101394#1959599 (10Krenair) a:3Krenair [03:16:31] ebernhardson: no jessie images in the list looks like a bug of some kind. I'll open a phab task [03:18:25] bd808: i opened a phab task alreadsy :) someone else responded its a known temporary issue [03:19:31] *nod* I figured that would be the case. probably some issue getting an image with the right kernel built or something [03:45:47] 6Labs, 10Tool-Labs: Linkwatcher spawns many processes without parent - https://phabricator.wikimedia.org/T123121#1959618 (10Beetstra) @valhallasw - the bot crashed (no clue why, it seems to have troubles with MySQL). I restarted it this morning, and it is now on 1215 [04:29:17] Do I have to keep my token on my profile once I verify my account at https://tools.wmflabs.org/tusc ? [04:53:43] Do I have to keep my token on my profile once I verify my account at https://tools.wmflabs.org/tusc ? [04:53:43] Do I have to keep my token on my profile once I verify my account at https://tools.wmflabs.org/tusc ? [04:53:43] AmandaNP: Do I have to keep my TUSC token on my profile? [04:54:16] it copied my message... [06:00:23] is there an issue with tools webservice? [06:00:37] I’m getting the following: [06:00:38] error: commlib error: got select error (Connection refused) [06:00:38] error: unable to send message to qmaster using port 6444 on host "tools-grid-master.tools.eqiad.wmflabs": got send error [07:32:41] 6Labs, 10MediaWiki-Vagrant, 10VisualEditor: VisualEditor not working in labs-vagrant - https://phabricator.wikimedia.org/T124575#1959713 (10Yurik) 3NEW [07:35:17] 6Labs, 10MediaWiki-Vagrant, 10VisualEditor: VisualEditor not working in labs-vagrant - https://phabricator.wikimedia.org/T124575#1959726 (10Yurik) [08:15:41] ugh, i get that as well [08:15:50] the grid seems down [08:16:32] 6Labs, 10MediaWiki-Vagrant, 10VisualEditor: VisualEditor not working in labs-vagrant - https://phabricator.wikimedia.org/T124575#1959811 (10Yurik) [08:18:44] 6Labs, 10Tool-Labs: Grid Engine down - https://phabricator.wikimedia.org/T124586#1959812 (10Giftpflanze) 3NEW [08:20:14] YuviPanda: ^ [09:28:38] looks like there’s an issue with grid engine again… [09:28:50] qstat returns : [09:28:52] error: commlib error: got select error (Connection refused) [09:28:52] error: unable to send message to qmaster using port 6444 on host "tools-grid-master.tools.eqiad.wmflabs": got send error [09:29:51] and my jobs have been down since 4 AM (UTC) [09:29:58] it's sunday morning … [09:30:21] maybe this will be resolved today, maybe tomorrow … [09:59:52] 6Labs, 10Tool-Labs: Grid Engine down - https://phabricator.wikimedia.org/T124586#1959893 (10valhallasw) 5Open>3Resolved a:3valhallasw Last information in `messages`: ``` 01/24/2016 04:00:36| timer|tools-grid-master|E|Corrupted database detected. Freeing all resources to prepare for a reconnect with reco... [10:00:04] 6Labs, 10Tool-Labs: Grid Engine down - https://phabricator.wikimedia.org/T124586#1959899 (10valhallasw) [10:00:06] 6Labs, 10Tool-Labs, 5Patch-For-Review: GridEngine down due to bdb issues - https://phabricator.wikimedia.org/T122638#1909517 (10valhallasw) [10:03:28] yay [10:27:23] valhallasw`cloud: Mogge, grid engine weer kaduuk? [10:27:33] multichill: en weer heel als het goed is [10:28:02] Ah, het viel me op omdat ik Cron qsub $HOME/bin/tvpupdater > $HOME/log/cron/tvpupdater krijg [10:38:22] bd808: Hi, I'm trying to enable role for this new labs instance (https://phabricator.wikimedia.org/T120516) and I get this error when I try to list the available roles https://dpaste.de/Sjs6 .. could you check it please ? [11:09:51] How many precise continuous hosts are there on tools? [11:10:39] Nemo_bis: there are 15 or so precise exec hosts [11:11:34] Ok. I wonder why BUB always gets tools-exec-1404 and -1410 [11:12:19] those are trusty hosts, but iirc there's also around 15 of them [11:14:55] Yeah sorry [11:15:55] how far apart in time did you try? [11:16:39] SGE tries to schedule the jobs on the least loaded hosts, so if you try a few times in a few minutes, chances are the same hosts will be the least loaded [11:17:02] 6Labs, 10Tool-Labs: Linkwatcher spawns many processes without parent - https://phabricator.wikimedia.org/T123121#1960010 (10valhallasw) Thanks, I have rescheduled the other jobs. [13:46:12] Cyberpower678: http://tools.wmflabs.org/xtools-ec throws 404 [14:15:00] 6Labs, 10MediaWiki-Vagrant, 10VisualEditor: VisualEditor not working in labs-vagrant - https://phabricator.wikimedia.org/T124575#1960143 (10scfc) AFAIUI, this is working as coded: You change the role configuration and afterwards you have to run `vagrant provision` to have the configuration take effect. But... [14:22:25] 10Tool-Labs-tools-Other, 7I18n: [[Intuition:Raun-help p4/en]] i18n issue - https://phabricator.wikimedia.org/T124605#1960157 (10Purodha) 3NEW [14:27:32] 6Labs, 10MediaWiki-Vagrant, 10VisualEditor: VisualEditor not working in labs-vagrant - https://phabricator.wikimedia.org/T124575#1960174 (10scfc) My MediaWiki-Vagrant box gives: ``` [tim@passepartout /var/tmp/vagrant]$ vagrant roles enable visualeditor Ok. Run `vagrant provision` to apply your changes. [tim... [14:32:45] 10Tool-Labs-tools-Other, 7I18n: [[Intuition:Raun-help p4/en]] misses a "the"? - https://phabricator.wikimedia.org/T124605#1960186 (10Aklapper) [14:34:02] 10Tool-Labs-tools-Other, 7I18n: [[Intuition:Raun-help p4/en]] misses a "the"? - https://phabricator.wikimedia.org/T124605#1960157 (10Aklapper) @Purodha: Please make sure to [[ http://tools.wmflabs.org/ | subscribe maintainers ]] so they can be are aware of this task. Thanks! [14:36:21] bd808: Currently here? [14:39:57] 10Tool-Labs-tools-Other, 7I18n: [[Intuition:Raun-help p4/en]] misses a "the"? - https://phabricator.wikimedia.org/T124605#1960199 (10Kenrick95) Thanks @Purodha and @Aklapper :smile: [14:44:41] 6Labs, 10Tool-Labs, 10DBA, 6Stewards-and-global-tools: Throttling linkwatcher tool user as it is consuming 100% CPU - https://phabricator.wikimedia.org/T121094#1960222 (10MarcoAurelio) Hi. Is it possible to have at least COIBot reporting again at Meta-Wiki? Our anti-spam efforts ain't much productive if we... [14:46:46] 10Tool-Labs-tools-Other, 7I18n: [[Intuition:Raun-help p4/en]] i18n issue - https://phabricator.wikimedia.org/T123214#1960226 (10Kenrick95) @MtDu and @Macofe This is Raun, maintained [[ https://github.com/kenrick95/Raun | here ]]. I'm the Tool maintained and will fix this together with T124605. Thanks :smile: [14:51:34] 10Tool-Labs-tools-Other, 7I18n: [[Intuition:Raun-help p4/en]] misses a "the"? - https://phabricator.wikimedia.org/T124605#1960237 (10Kenrick95) 5Open>3Resolved a:3Kenrick95 Updated & deployed new message string. [14:51:49] 10Tool-Labs-tools-Other, 7I18n: [[Intuition:Raun-help p4/en]] i18n issue - https://phabricator.wikimedia.org/T123214#1960240 (10Kenrick95) 5Open>3Resolved Updated & deployed new message string. [14:55:01] 6Labs, 10MediaWiki-Vagrant, 10VisualEditor: VisualEditor not working in labs-vagrant - https://phabricator.wikimedia.org/T124575#1960244 (10Yurik) I did run the provision after enabling visualeditor role, and it did install parsoid, but I suspect it didn't start the parsoid service. Regardless, I suspect tha... [15:19:38] 6Labs, 10Tool-Labs, 10DBA, 6Stewards-and-global-tools: Throttling linkwatcher tool user as it is consuming 100% CPU - https://phabricator.wikimedia.org/T121094#1960266 (10Beetstra) I'm working on that @MarcoAurelio. [15:25:06] hey guys ! in terms of performance on the labs db, how bad is this sql query ? http://alphos.fr/wikidata/wikilinks-redirects-conflicts-report [15:26:24] Alphos: make sure to always provide page_title /and/ page_namespace [15:26:26] i'd rather ask first than apologize later for tying mysql for a long time ^^' [15:26:40] because the index is on (page_namespace, page_title) [15:26:54] valhallasw`cloud right [15:28:00] i'm having a brainfart though - where would i need to provide it ? [15:28:17] ON source.ips_site_page = elwiki_p.page.page_title [15:28:51] not sure how i'd provide it :/ [15:29:16] doesn't wb_items_per_site provide namespace? [15:29:41] bah, no, it doesn't, it just prefixes the page name if there's a namespace [15:29:45] yup [15:29:57] Alphos: just add a where elwiki_p.page.page_namespace = 0 [15:30:05] because you won't be able to match other namespaces anyway [15:30:13] won't i ? [15:31:10] let me guess : page_title doesn't have the namespace in it ? [15:31:52] yup... ah, well... [15:34:32] Alphos: right, mediawiki uses page_namespace / page_title, and e.g. a template has page_namespace = 14, and no prefix in page_title [15:34:44] yup [15:36:12] now with http://alphos.fr/wikidata/wikilinks-redirects-conflicts-report : how bad is it perf-wise ? [15:37:35] i know the basics of cross-db joining, but i really have no idea of how bad it is ^^' [15:38:01] (nor as to how fine labs will handle) [15:38:29] Alphos: this is the only thing that really stuck out for me, but I'm also not super experienced with it [15:38:39] try with a small wiki first, I'd say [15:39:06] yup, that was my intention ; used elwiki because i *know* it has what i'm looking for [15:39:37] and it's the smallest wiki i have stumbled upon that has what i'm looking for, so far [15:44:15] Alphos: also there's https://tools.wmflabs.org/tools-info/optimizer.py , but it seems.. broken [15:44:28] won't hurt trying [15:44:35] oh, i see :D [15:46:01] 10Tool-Labs-tools-Other: tools-info optimizer.py: "Access denied for user 's51772'@'10.68.18.46' (using password: YES)" - https://phabricator.wikimedia.org/T107571#1960302 (10valhallasw) [15:46:03] 10Tool-Labs-tools-Other: tools-info connecting to wrong database server - https://phabricator.wikimedia.org/T105911#1960304 (10valhallasw) [15:47:44] does 3500 pages seem small enough a wiki to test with ? [15:48:24] (trying one not too small with an alphabet i can reasonably type on my keyboard) [15:49:07] Alphos: you can always kill the query if it runs for very long. In a second terminal, connect to the sql server again, SHOW PROCESSLIST; then KILL ; [15:49:22] thanks :) [15:51:46] 10Tool-Labs-tools-Other: tools-info connecting to wrong database server - https://phabricator.wikimedia.org/T105911#1960314 (10valhallasw) Note that the IP is the IP of the webgrid host, not of the database server. I've moved the replica.my.cnf to replica.my.cnf.old, which should re-generate the file with the c... [15:55:41] it's actually surprisingly fast compared to what i expected. not lightning fast (0.68 seconds for rmwiki), but still faster than i expected [15:56:09] still, not sure how well it's going to scale for a multi-million wiki :/ [15:56:50] if it's linear i can definitely live with it ; if it's square, not so much ^^' [15:56:54] that's always difficult to say, but it won't be lightning fast in any case [15:57:10] obviously [15:57:36] heuristically I would say it will scale maybe as n^4? as you join four tables that all have a relevant number of entries that scales roughly as n [15:57:37] i'll try with a 10k+ wiki [15:57:50] valhallasw`cloud except indexes [15:57:55] but then lookups are only log n, so maybe it's more (log n)^4 [15:58:35] except lookups on wikidatawiki_p.wb_items_per_site too... [16:00:09] wow [16:00:15] 1.11 seconds on alswiki ! [16:00:31] now that's even faster than i expected based on rmwiki ^^' [16:02:47] skipping from 3500 pages in NS_MAIN to 20500 pages in NS_MAIN, both with empty resultsets : less than doubling the time O_O [16:03:46] valhallasw`cloud whaddaya think, elwiki with 114k pages in NS_MAIN ? [16:04:03] Alphos: just try it :-) [16:05:00] i'm feeling, well, maybe not proud, but i'm feeling SOMETHING ! :D [16:05:17] 3.68 seconds, 10 items in the resultset [16:05:43] pattern of progression looks like log(n) so far [16:06:48] and for what i see, it works as intended ! \o/ [16:10:57] well, almost as intended... [16:18:46] that's odd, i find other legitimate results than the one i stumbled upon, but not the one i stumbled upon... [16:27:04] oh, i think it's going to be a very stupid thing [16:27:21] underscore versus space [16:32:58] Luke081515, hi [16:33:55] Cyberpower678: hi [16:34:06] Looking [16:36:12] Luke081515, try now [16:37:00] works :) [16:38:43] damn, index breakage with REPLACE() really makes things longer :-( [16:38:57] 2min30 for elwiki [16:39:10] uh, 2min 7.30 [16:53:50] 10Tool-Labs-tools-Other, 7I18n: Find or create a CLDR data parser for Intuition or PyIntuition - https://phabricator.wikimedia.org/T102231#1960385 (10Nemo_bis) CLDRPluralRuleParser was split in https://gerrit.wikimedia.org/r/#/c/237924/3 BTW [16:54:25] (03CR) 10Nemo bis: "This is a labs tool, not MediaWiki ; is PLURAL really supported? Can use the PHP library maybe https://gerrit.wikimedia.org/r/#/c/237924/3" (031 comment) [labs/tools/crosswatch] - 10https://gerrit.wikimedia.org/r/265804 (owner: 10MtDu) [17:14:47] 10MediaWiki-extensions-OpenStackManager, 10MediaWiki-Authentication-and-authorization: ApiDocumentationTest failure: Undefined property: AuthPlugin::$boundAs - https://phabricator.wikimedia.org/T124613#1960399 (10Nemo_bis) 3NEW [17:54:01] (03PS1) 10BryanDavis: Add proper HTML structure to error pages [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266075 (https://phabricator.wikimedia.org/T119830) [17:54:03] (03PS1) 10BryanDavis: Add viewport meta tag to generated content [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266076 (https://phabricator.wikimedia.org/T119830) [17:54:05] (03PS1) 10BryanDavis: Cleanup error pages [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266077 [17:54:07] (03PS1) 10BryanDavis: Cleanup code style of index.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266078 [17:54:09] (03PS1) 10BryanDavis: Cleanup code style of list.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266079 [17:54:11] (03PS1) 10BryanDavis: Cleanup code style of status.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266080 [17:54:13] (03PS1) 10BryanDavis: Cleanup code style of tool.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266081 [17:54:15] (03PS1) 10BryanDavis: Reduce font-size for web interface link on tool.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266082 [17:57:19] (03CR) 10jenkins-bot: [V: 04-1] Add proper HTML structure to error pages [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266075 (https://phabricator.wikimedia.org/T119830) (owner: 10BryanDavis) [17:58:46] (03CR) 10jenkins-bot: [V: 04-1] Add viewport meta tag to generated content [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266076 (https://phabricator.wikimedia.org/T119830) (owner: 10BryanDavis) [18:00:13] (03CR) 10jenkins-bot: [V: 04-1] Cleanup error pages [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266077 (owner: 10BryanDavis) [18:01:43] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of index.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266078 (owner: 10BryanDavis) [18:03:14] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of list.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266079 (owner: 10BryanDavis) [18:03:34] 6Labs, 10MediaWiki-Vagrant, 10VisualEditor: VisualEditor not working in labs-vagrant - https://phabricator.wikimedia.org/T124575#1960432 (10bd808) This sounds a lot like the problem described in {T119731} [18:04:50] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of status.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266080 (owner: 10BryanDavis) [18:06:20] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of tool.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266081 (owner: 10BryanDavis) [18:07:52] (03CR) 10jenkins-bot: [V: 04-1] Reduce font-size for web interface link on tool.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266082 (owner: 10BryanDavis) [18:10:12] out of curiosity, would it be possible to request a FILE mysql grant for my project ? looking to SELECT INTO outfile ^^' [18:10:28] if not, i'll manage, but why use php when you can do it all in mysql ? :p [18:30:13] 10MediaWiki-extensions-OpenStackManager, 10MediaWiki-Authentication-and-authorization: ApiDocumentationTest failure: Undefined property: AuthPlugin::$boundAs - https://phabricator.wikimedia.org/T124613#1960451 (10Legoktm) ``` public function getAllowedParams() { return array( 'project' => array( ApiB... [18:38:50] 10Tool-Labs-tools-Other: 504 error for Autolist 2 on https://tools.wmflabs.org/autolist/ - https://phabricator.wikimedia.org/T124280#1960458 (10Aklapper) 5Open>3Resolved a:3Aklapper Seems to be back up [18:55:16] Hey everyone, i have two gerrit accounts pointing to thesame email and i want to delete one of the gerrit account [18:55:25] please is it possible to delete one the of the gerrit accounts? [19:29:50] could i interest anyone in a tiny bit of code review ? :-D less than 2KB php, 1.5KB SQL https://github.com/alaefin/wikidata-redirects-conflicts-reports [19:32:15] bd808: ooooh, better admin page! [19:32:49] Alphos no; SELECT INTO writes to the database server, so it's not actually useful [19:33:25] isn't it to the mysql-client server ? [19:33:34] nope [19:33:50] ok, so not everything i once knew has come back :D [19:34:07] still, good to know i can't do much about that [19:34:29] Alphos: https://dev.mysql.com/doc/refman/5.7/en/select-into.html " The file is created on the server host, so you must have the FILE privilege to use this syntax." [19:35:04] i guess i *could* mysql 'query' > localfile.sqlout, but that's less parsable than csv [19:35:42] Alphos: iirc if you pipe the output, it changes to tab separated [19:36:03] worth a try with a tiny wiki [19:37:08] 10Tool-Labs-tools-Other, 10DBA: tools-info connecting to wrong database server - https://phabricator.wikimedia.org/T105911#1960516 (10valhallasw) That doesn't seem to work, though. Maybe it's a time issue, but I don't know the details of the replica password system well enough to be sure. @jcrespo, could you f... [19:37:25] 10Tool-Labs-tools-Other, 10DBA: tools.tools-info credentials are not functioning - https://phabricator.wikimedia.org/T105911#1960518 (10valhallasw) [19:38:47] nicely spotted [19:39:07] won't be any of the sprintf magic though, but i can live with that [19:39:31] d3r1ck: no, but you can change the email address of one of them (https://gerrit.wikimedia.org/r/#/settings/contact or maybe on wikitech, not sure) [19:39:55] then again, $(printf)... [19:45:55] 6Labs, 10Tool-Labs: toollabs dh_test (debian-glue) fails - https://phabricator.wikimedia.org/T124618#1960520 (10valhallasw) 3NEW [19:46:43] (03CR) 10Merlijn van Deen: [C: 032 V: 032] Add proper HTML structure to error pages [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266075 (https://phabricator.wikimedia.org/T119830) (owner: 10BryanDavis) [19:47:19] (03CR) 10Merlijn van Deen: "I don't have submit privileges, though, so I don't think I can actually override jenkins :/" [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266075 (https://phabricator.wikimedia.org/T119830) (owner: 10BryanDavis) [19:48:29] (03CR) 10Merlijn van Deen: [C: 032 V: 032] Add viewport meta tag to generated content [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266076 (https://phabricator.wikimedia.org/T119830) (owner: 10BryanDavis) [19:49:32] (03CR) 10Merlijn van Deen: [C: 032] Cleanup error pages [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266077 (owner: 10BryanDavis) [19:49:50] (03CR) 10jenkins-bot: [V: 04-1] Cleanup error pages [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266077 (owner: 10BryanDavis) [19:49:53] (03CR) 10Merlijn van Deen: [C: 032] "Yuck, tabs ;-)" [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266078 (owner: 10BryanDavis) [19:50:03] (03CR) 10Merlijn van Deen: [C: 032] Cleanup code style of list.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266079 (owner: 10BryanDavis) [19:50:22] (03CR) 10Merlijn van Deen: [C: 032] Cleanup code style of status.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266080 (owner: 10BryanDavis) [19:50:33] (03CR) 10Merlijn van Deen: [C: 032] Cleanup code style of tool.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266081 (owner: 10BryanDavis) [19:50:44] (03CR) 10Merlijn van Deen: [C: 032] Reduce font-size for web interface link on tool.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266082 (owner: 10BryanDavis) [19:50:48] (03CR) 10jenkins-bot: [V: 04-1] Cleanup error pages [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266077 (owner: 10BryanDavis) [19:50:50] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of index.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266078 (owner: 10BryanDavis) [19:50:58] (03CR) 10jenkins-bot: [V: 04-1] Cleanup error pages [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266077 (owner: 10BryanDavis) [19:51:00] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of index.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266078 (owner: 10BryanDavis) [19:51:02] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of list.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266079 (owner: 10BryanDavis) [19:51:09] (03CR) 10jenkins-bot: [V: 04-1] Cleanup error pages [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266077 (owner: 10BryanDavis) [19:51:11] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of index.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266078 (owner: 10BryanDavis) [19:51:13] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of list.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266079 (owner: 10BryanDavis) [19:51:15] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of status.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266080 (owner: 10BryanDavis) [19:51:19] (03CR) 10jenkins-bot: [V: 04-1] Cleanup error pages [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266077 (owner: 10BryanDavis) [19:51:21] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of index.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266078 (owner: 10BryanDavis) [19:51:23] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of list.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266079 (owner: 10BryanDavis) [19:51:25] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of status.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266080 (owner: 10BryanDavis) [19:51:27] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of tool.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266081 (owner: 10BryanDavis) [19:51:34] uh. help? [19:51:41] (03CR) 10jenkins-bot: [V: 04-1] Cleanup error pages [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266077 (owner: 10BryanDavis) [19:51:43] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of index.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266078 (owner: 10BryanDavis) [19:51:45] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of list.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266079 (owner: 10BryanDavis) [19:51:47] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of status.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266080 (owner: 10BryanDavis) [19:51:49] (03CR) 10jenkins-bot: [V: 04-1] Cleanup code style of tool.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266081 (owner: 10BryanDavis) [19:52:00] OM MY GOD SHUTUP [19:52:19] Cyberpower678: relax. [19:52:19] valhallasw`cloud: I think it's just overloaded with the dependancy tree [19:52:31] valhallasw`cloud: I'm not sure what jenkins is hating about those. Something to do with the deb parts I didn't touch [19:52:41] valhallasw`cloud, it makes a noise for every message. [19:52:44] bd808: yeah, and it ignores the debian test when submitting anyway it seems [19:52:47] Cyberpower678: then fix your irc client [19:53:14] bd808: but then it goes back between OK/not OK, so... not sure. [19:54:19] valhallasw`cloud, Yes I can ignore grrrit-wm in the future, but in the now, it just irritated the hell of me. [19:54:19] (03CR) 10Merlijn van Deen: [V: 032] "Let's try again." [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266077 (owner: 10BryanDavis) [19:54:38] I'm also confused about when I'm allowed to submit and when I'm not [19:55:01] oh, jenkins votes -2, and then I can't force-submit [19:55:14] there we go. [19:55:35] (03CR) 10Merlijn van Deen: [V: 032] Cleanup code style of index.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266078 (owner: 10BryanDavis) [19:55:53] (03CR) 10Merlijn van Deen: [V: 032] Cleanup code style of list.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266079 (owner: 10BryanDavis) [19:56:13] bd808: do you know how to deploy the website? [19:56:43] (03CR) 10Merlijn van Deen: [V: 032] Cleanup code style of status.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266080 (owner: 10BryanDavis) [19:56:46] nope. I'm not sure if it had to be built into a new deb or if there is another way [19:56:56] no, it's just a git pull for the 'admin' tool [19:57:06] it's in the .deb for historical reasons [19:57:14] ah. that should be easy enough then. [19:57:29] lemme write a few notes on Nova_Resource:Tools/Admin [19:57:46] (03CR) 10Merlijn van Deen: [V: 032] Cleanup code style of tool.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266081 (owner: 10BryanDavis) [19:57:53] (03CR) 10Merlijn van Deen: [V: 032] Reduce font-size for web interface link on tool.php [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266082 (owner: 10BryanDavis) [20:00:57] bd808: https://wikitech.wikimedia.org/w/index.php?title=Nova_Resource%3ATools%2FAdmin&type=revision&diff=269242&oldid=264810 / https://tools.wmflabs.org/ :-) [20:03:16] valhallasw`cloud: sweet. https://tools.wmflabs.org/?status isn't working. Was that me or is the exec not working right now? [20:03:40] mmm, lemme check [20:04:43] bd808: crappy gridengine xml, I think [20:04:52] hrm [20:04:56] although I thought you just realigned the code [20:05:06] yeah I thought I did too [20:05:15] I converted backticks to exec() [20:05:39] I'll debug it some. should be an "easy" fix [20:05:42] does exec use the shell? there's a pipe in there [20:05:57] and exec only returns "The last line from the result of the command." [20:06:26] *nod* should have used shell_exec() I bet [20:06:33] yeah, or system() [20:06:55] * bd808 writes a patch [20:08:32] (03PS1) 10BryanDavis: Use shell_exec() instead of exec() to replace backticks [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266096 [20:09:59] valhallasw`cloud: ^ [20:10:01] (03CR) 10Merlijn van Deen: [C: 032 V: 032] Use shell_exec() instead of exec() to replace backticks [labs/toollabs] - 10https://gerrit.wikimedia.org/r/266096 (owner: 10BryanDavis) [20:10:14] * valhallasw`cloud dodges jenkins [20:10:46] https://tools.wmflabs.org/?status works now [20:11:05] hello [20:11:13] it's a YuviPanda! [20:11:34] Aren't you con vacation or something YuviPanda? [20:11:43] s/con/on/ [20:11:55] * valhallasw`cloud arrests YuviPanda to make it a con vacation [20:12:46] bd808: as for the error page design, I think we can re-use the WMF 'things are broken' page, i.e. https://tools.wmflabs.org/.error/errorpage.html [20:13:17] Did I mess up the free vmem calculations too? The numbers being shown look large [20:13:22] bd808: no that's next month [20:13:27] I did sleep 16h [20:13:34] bd808: no, that was already broken :-D [20:13:43] valhallasw`cloud: <3 for handling that outage too [20:13:47] although the numbers might be OK [20:13:51] I'm wondering if we should do the dump / reload now [20:13:59] YuviPanda: yeah, I think that would be the best option [20:14:27] ok, let me get out of bed and put some tea up, I'll brb in like 10mins to start doing it [20:14:28] maybe schedule downtime (i.e. mail labs-l) for tomorrow to do that? [20:14:32] hmm [20:14:42] well [20:14:50] let's just make a copy of it and do it elsewhere [20:14:53] and see what happens [20:15:02] shouldn't cause any downtime yet [20:15:08] YuviPanda: chasemp has tested things, so it would be good to have him involved [20:15:13] ah, yes, famous last words :P [20:15:20] hehe [20:15:21] well [20:15:23] valhallasw`cloud: I can add some pretty to those error pages for sure. It looked to me like the inital design was hoping to use something like https://tools.wmflabs.org/admin/?404 [20:15:25] a cp shouldn't cause downtime [20:15:36] yeah [20:15:37] YuviPanda: it's bdb. I don't trust bdb. [20:15:46] to do a cp? [20:15:48] hmm [20:15:52] that's a good attitude probably [20:16:22] bdb was the state of the art in 1988 when I worked on Novell networks [20:16:58] bd808: either works for me, but the current one feels cluttered to me. #notadesigner, though ;-) [20:21:07] it's a bit dumb that we load all of jquery just for the table sorter js too [20:21:16] lots of things that could be better [20:21:45] 6Labs, 10Tool-Labs: toollabs dh_test (debian-glue) fails - https://phabricator.wikimedia.org/T124618#1960559 (10scfc) a:3scfc That's due to `jsub` & Co. now trying to log things as part of the test suite and the logging executable not being there at all. [20:21:47] * YuviPanda would just rewrite the whole thing [20:22:28] meh. [20:22:50] Not until we find a much better system [20:23:06] like maybe just using hay's index or something similar [20:23:31] 6Labs, 10Tool-Labs: toollabs dh_test (debian-glue) fails - https://phabricator.wikimedia.org/T124618#1960562 (10yuvipanda) Whelp, that's probably all me since I force merged them when jenkins was failing since I couldn't figure out what was wrong and glue succeeded locally on a tools instance... [20:23:35] yeah [20:23:39] not rewrite from scratch [20:23:41] but.. [20:23:42] but his "send it all the browser and let them figure it out" design isn't my favorite implementation [20:24:33] * YuviPanda is kindof ok with that these days [20:25:21] * bd808 runs no-script and also doesn't trust tools.wmflabs.org since anyone can put js on there [20:25:59] * YuviPanda nods, but by dint of having an android phone has given up on most of those things anyway seemingly. [20:26:21] trusting javascript is just a tiny bit better than using `curl | sudo sh` as an install mechanism [20:26:44] I wouldn't say 'tiny', it has to escape a couple of sandboxes [20:26:45] * bd808 adjusts his tinfoil hat [20:27:03] I wonder if I can run my browser safely in a container [20:27:09] not as good as running it in another VM ofc [20:27:13] No need if you are just going to launch XSS attacks [20:27:24] sandboxes don't stop that [20:27:35] true [20:27:41] and the web is one giant pile of XSS vulnerabilites [20:27:45] yes [20:27:49] one of the things I've always wanted [20:27:51] is [20:27:54] toolname.tools.wmflabs.org [20:28:00] since I bet there are people leaking cookies [20:28:10] and being totally unaware of xss stuff [20:28:13] right now [20:28:15] we'd need a wildcard tls cert [20:28:17] that'll at least be a bit more difficult [20:28:24] that's doable, I checked [20:28:38] we've the budget [20:28:43] or letsencrypt ;-) [20:29:01] they aren't doing wildcards yet [20:29:06] although I can imagine nginx not liking hundreds of certificates [20:29:09] that too, yeah, although that'll complicate things since you'll have one certificate per tool [20:29:11] yeah [20:29:24] but that'll be fine, I think - we can get the wildcard [20:30:44] and tools.wmflabs.org/ will still work [20:30:54] we can have a mechanism in place for people to pick one or the other [20:31:00] and will setup redirects [20:33:59] we can also just use X.wmflabs.org? I'm not sure if the tools really adds something [20:34:16] and I'm wondering whether we can use it to show it's a 'stable' tool (multiple maintainers, etc) [20:34:47] the tools. adds that this is under toollabs and hence doesn't get raw IP access and stuff [20:36:15] mmm, right [20:41:02] valhallasw`cloud: re: downtime, I guess chasemp should be involved but I feel bad announcing downtime without actually asking him [20:41:55] well, do it on tuesday then ;-) [20:42:23] yeah [20:47:25] hey keep getting irc pings what's up? [20:47:39] hah [20:47:51] chasemp: gridengine died again this morning, and a restart brought it back up. [20:48:00] I Just want to try the 'stop, dump, reload' routine [20:48:04] hmm [20:48:05] yeah and I saw the email [20:48:24] valhallasw`cloud suggests we announce downtime before trying to do that but didn't want to announce and dump it on you without asking first [20:48:40] as in, I presumed you'd want to be involved as well, etc. [20:48:53] yeah sure [20:49:00] looks like doing it tuesday then? [20:49:27] yeah ok [20:49:31] I'll mail out [20:49:55] I figured from the email it was teh same old crappyness [20:50:05] not good but not sunday urgent [20:50:49] yeah [20:51:04] should I say 2h? [20:51:31] yeah, I'll say 2h should be enough... [20:51:34] are we talking strictly queue db only? [20:51:53] I think if we are going to do it we should hope that a dump and rebuild works [20:52:07] and if not we should plan on wiping the queue totally and sge will rebuiled it on restart fresh [20:52:12] that gets this all on even footing [20:52:16] yeah, I would just stop gridengine, dump, reload, start, but then without having time pressure [20:52:18] but the fallout from that I'm not entirely sure about [20:52:32] if that fails, send another email that we have to take the nuclear option [20:52:38] and plan that in for later in the week? [20:53:10] the logistics of announcing I'm open on [20:53:29] my first thought is to say the queue is corrupt and we ahve to fix it, this may involve wiping the queue and stay tuned [20:53:33] and then anything we can keep is a bonus [20:53:34] rather than [20:53:45] hey we need to fiddle and then another email that says oh sorry we lost your business [20:54:43] right. Yes, that might be a sane option. [20:55:31] hmm [20:55:32] on the other hand, we have to wait for a bit before we even know if the bdb dump cycle solved the issue [20:55:36] so the dump and reload might not work and cause the queue to be wiped [20:55:36] valhallasw`cloud: by chance is there a day you could be around on in teh coming week? :) [20:55:51] but that's definitely worst case [20:55:53] the dump and reload may not work, and if not we should wipe the queue [20:56:04] 10MediaWiki-extensions-OpenStackManager: ApiDocumentationTest failure: Undefined property: AuthPlugin::$boundAs - https://phabricator.wikimedia.org/T124613#1960639 (10Nemo_bis) [20:56:21] do we know how to wipe the queue 'safely'? [20:56:24] I am around-ish the next week, but monday is full [20:56:26] as in to not lose config and stuff? [20:56:28] YuviPanda: safe as in? [20:56:32] oh yes [20:56:35] ok! [20:56:36] I believe we do [20:56:39] that's cool then [20:56:55] tldr version is there are two indepedent db files [20:57:00] one for config one for queue [20:57:12] we can send out the email tomorrow then, since this probably has too many caveats for me to write out now [20:57:29] in my tsting when I stop the master proc and rm the queue db and rull restart sge just rebuilds a fresh queue db with no content [20:57:36] that did completely diconnect running jobs though [20:57:43] right [20:57:45] so if we rebuild queue [20:57:50] we've to restart all the exec nodes too [20:57:52] we need to restart all exec hosts [20:57:52] ya [20:58:04]