[14:00:08] <wm-bot>	 Technical Advice IRC meeting starting in 60 minutes in channel #wikimedia-tech, hosts:  - all questions welcome, more infos: https://www.mediawiki.org/wiki/Technical_Advice_IRC_Meeting
[14:50:06] <wm-bot>	 Technical Advice IRC meeting starting in 10 minutes in channel #wikimedia-tech, hosts: @tgr & @Tonina_WMDE - all questions welcome, more infos: https://www.mediawiki.org/wiki/Technical_Advice_IRC_Meeting
[15:00:43] <Tonina_WMDE>	 Hey everyone \o Me and @tgr are here to answer your questions in this week's Technical Advice IRC Meeting
[15:01:00] <tgr>	 hi all!
[15:01:49] <ato_01>	 Hi there!
[15:03:21] <ato_01>	 Is it possible to ask a question with toolforge too?
[15:03:54] <ato_01>	 I mean running bots and processes.
[15:04:15] <Tonina_WMDE>	 ato_01:  all questions are welcome
[15:04:16] <tgr>	 ato_01: anything is possible! the worst that can happen is that we can't answer
[15:04:30] <ato_01>	 Super.
[15:06:07] <ato_01>	 In the near past I could run python scripts through shell scripts through php post, 
[15:06:26] <andre__>	 o/
[15:06:47] <ato_01>	 And now I can only let run shell scripts to delete my logs
[15:07:16] <ato_01>	 I do not receive any error
[15:08:13] <ato_01>	 but my post request does not work. Do you know, why?
[15:08:44] <d3r1ck>	 o/
[15:08:57] <tgr>	 ato_01: so you are saying you have a PHP web application, you send a POST request to it, and the application executes a Python script via a shell command? And that has stopped working recently?
[15:09:18] <ato_01>	 yes, exactly
[15:09:58] <tgr>	 do you have logging in the PHP code? such as return status from shell_exec or whatever you use?
[15:10:35] <Tonina_WMDE>	 ato_01: can you paste the post request you send
[15:10:39] <ato_01>	 No, I can not log in php. 
[15:10:58] <ato_01>	 All I can do, only send a post request
[15:11:17] <ato_01>	 maybe you can help in it
[15:11:39] <ato_01>	 just a sec
[15:11:50] <tgr>	 cannot meaning that the application has logging code but it does not actually write anything to the logs? or that you are not sure how to set it up?
[15:14:59] <tgr>	 in the meantime, if anyone else has questions, feel free to ask, IRC works well for multiple parallel discussions
[15:15:19] <ato_01>	 I just sent a post request, like this
[15:15:24] <ato_01>	 <form action="./test-1.php"> 								    <input type="submit" value="Force"> 								</form>
[15:15:47] <ato_01>	 It has run a shell script
[15:17:13] <ato_01>	 which started a jsub command with a python script, which has a -o (log output) parameter
[15:18:11] <ato_01>	 and it was working befor this year for sure
[15:18:44] <Tonina_WMDE>	 ato_01: ok so what is the expected result in the end? logs being deleted?
[15:18:50] <ato_01>	 now all I can do with a shell script is to clear all my log files
[15:19:45] <ato_01>	 I still would like to let run or stop in case of emergency
[15:20:13] <ato_01>	 through web browser, without putty
[15:20:48] <Tonina_WMDE>	 ato_01: can you somehow show us the contents of test-1.php
[15:21:10] <ato_01>	 <?PHP  shell_exec("./scripts/test-1.sh"); header('Location: https://tools.wmflabs.org/ato/index-secret.html?success=true');  ?>
[15:22:04] <ato_01>	 And it is almost the same where I clear my log files.
[15:28:45] <Tonina_WMDE>	 ato_01: but clearing the log files works, right? The problem is in the python script or more precisely that it's not executed?
[15:29:02] <ato_01>	 clearing still works
[15:30:12] <ato_01>	 only python scripts (or jsub commands) what I can not any more
[15:30:37] <Tonina_WMDE>	 hmmm
[15:31:05] <ato_01>	 clearing is not more. Only the following:
[15:31:16] <ato_01>	 #!/bin/bash echo "" > /data/project/ato/public_html/log/test-1.txt 
[15:32:30] <tgr>	 ato_01: probably the first step is to check the job grid error logs
[15:33:18] <ato_01>	 yes, but I can not find anithing like this
[15:33:32] <tgr>	 probably with qstat, I'm not really familiar with the grid system though
[15:34:17] <tgr>	 so maybe its better to hop over to #wikimedia-cloud and ask there
[15:34:52] <ato_01>	 ok
[15:34:58] <ato_01>	 I will do that
[15:36:05] <ato_01>	 I am just seeing my error.log
[15:36:11] <ato_01>	 ./scripts/test-1.sh: line 2: jsub: command not found
[15:36:47] <ato_01>	 But the same shell script through putty works
[15:38:14] <tgr>	 the web app might be using a different user or different shell
[15:38:37] <tgr>	 you can try using the full path to jsub
[15:39:07] <ato_01>	 do you know the full path of it?
[15:40:41] <tgr>	 /usr/bin/jsub
[15:41:08] <ato_01>	 Thank you! I will try it
[15:45:20] <Tonina_WMDE>	 anyone else? :)
[15:45:56] <tgr>	 ato_01: triggering pywikibot jobs from a web interface seems like a common enough use case so it might be worth asking about that on the Cloud IRC or mailing list, they might suggest best practices or sample code
[15:46:03] <ato_01>	 ./scripts/test-1.sh: line 2: /usr/bin/jsub: No such file or directory
[15:46:35] <tgr>	 for example, sending the output of jsub to devnull might also set you up for trouble in the long run
[15:47:56] <tgr>	 maybe the web user has limited access?
[15:49:15] <tgr>	 if the task is not time-sensitive you could rearrange it so that the PHP script just pushes into some simple queue (a DB or text file) and a periodically scheduled grid job takes the tasks from there
[15:49:48] <tgr>	 but I'm just guessing, you should ask people more familiar with toolforge
[15:52:24] <ato_01>	 It should be some security update in the near past. I will ask somebody on the toolforge channel. Thank you anyway!
[15:59:53] <tgr>	 sorry we couldn't be of more help
[16:00:04] <Tonina_WMDE>	 :/
[16:00:34] <ato_01>	 no problem. you tried it. :)
[16:00:44] <ato_01>	 have a nice evening!
[16:04:42] <tgr>	 this concludes the tech advice hour, there will be one the same time next week
[16:05:57] <tgr>	 also since next Wednesday is first of the month, there will be an Americas-friendly second meeting at 11 PM UTC
[16:06:11] <tgr>	 o/
[16:56:18] <[1997kB]>	 Hello, why there's no join date on these user pages https://en.m.wikipedia.org/w/index.php?title=User:Shoy https://meta.m.wikimedia.org/wiki/User:FloNight in mobile view like this one https://en.m.wikipedia.org/wiki/User:1997kB below username?
[16:59:03] <Krenair>	 MariaDB [enwiki_p]> select user_name, user_registration from user where user_name in ('Shoy', 'FloNight', '1997kB');
[16:59:03] <Krenair>	 +-----------+-------------------+
[16:59:04] <Krenair>	 | user_name | user_registration |
[16:59:04] <Krenair>	 +-----------+-------------------+
[16:59:04] <Krenair>	 | 1997kB    | 20170802154554    |
[16:59:05] <Krenair>	 | FloNight  | 20051011002941    |
[16:59:07] <Krenair>	 | Shoy      | NULL              |
[16:59:09] <Krenair>	 +-----------+-------------------+
[16:59:11] <Krenair>	 3 rows in set (0.00 sec)
[17:00:17] <Krenair>	 oh, FloNight's link is metawiki, not enwiki
[17:00:41] <Krenair>	  MariaDB [metawiki_p]> select user_registration from user where user_name = 'FloNight';
[17:00:42] <Krenair>	 +-------------------+
[17:00:42] <Krenair>	 | user_registration |
[17:00:42] <Krenair>	 +-------------------+
[17:00:46] <Krenair>	 | NULL              |
[17:00:48] <Krenair>	 +-------------------+
[17:00:50] <Krenair>	 1 row in set (0.00 sec)
[17:00:57] <Krenair>	 [1997kB], yeah basically if there is no user_registration timestamp it won't show that line
[17:01:07] <[1997kB]>	 and why is that?
[17:01:17] <Krenair>	 why might user_registration be NULL?
[17:01:28] <Krenair>	 Well the field hasn't always existed
[17:02:49] <[1997kB]>	 oh so when these accounts were created there was no user registration field?
[17:02:51] <Krenair>	 it was added in MW 1.6
[17:02:54] <Krenair>	 yeah
[17:03:00] <Krenair>	 the docs for that table literally say
[17:03:05] <Krenair>	 "For old users, they may have a value of NULL for this field."
[17:03:24] <[1997kB]>	 but it doesn't happen with old users?
[17:03:29] <[1997kB]>	 all old*
[17:03:31] <Krenair>	 https://www.mediawiki.org/wiki/Manual:User_table#user_registration
[17:03:42] <Krenair>	 [1997kB], some are older than others
[17:07:48] <[1997kB]>	 aha. got it.. Thanks Krenair :)
[17:10:51] <Krenair>	 http://mediawiki.org/wiki/Special:Code/MediaWiki/12207
[17:13:31] <Krenair>	 [1997kB], actually enwiki does appear to have had some back-filled
[17:13:38] <Krenair>	 MariaDB [enwiki_p]> select min(user_registration) from user;
[17:13:39] <Krenair>	 +------------------------+
[17:13:39] <Krenair>	 | min(user_registration) |
[17:13:39] <Krenair>	 +------------------------+
[17:13:39] <Krenair>	 | 20010121021221         |
[17:13:40] <Krenair>	 +------------------------+
[17:13:41] <Krenair>	 1 row in set (51.06 sec)
[17:13:48] <Krenair>	 just not all
[17:13:51] * Krenair shrugs
[17:14:44] <Krenair>	 Bugzilla was a thing by this point so you may be able to find a ticket for it
[18:02:11] <legoktm>	 [1997kB]: enwp had user_registration backfilled, but I don't think any other wiki had
[18:08:11] <[1997kB]>	 legoktm: but it still doesn't show in some enwiki accounts, one is above.
[18:09:23] <James_F>	 The back-fill was just a lie, right? We put in the date of first edit.
[18:09:36] * James_F has vague memories.
[18:10:51] <legoktm>	 correct
[19:17:17] <Krenair>	 James_F, legoktm: so maybe it missed those who registered before registration started getting recorded but had not edited before the backfill?
[19:18:10] <James_F>	 Yeah, and accounts registered in ~2003 which were never used to edit are still `null`.
[19:18:38] <James_F>	 Also the script whoever wrote might have died due to a hiccough. 
[19:24:36] <Krenair>	 James_F, hm?
[19:24:40] <Krenair>	 the script whoever wrote?
[19:24:53] <Krenair>	 I think Tim Starling wrote the backfill script
[19:25:17] <James_F>	 OK.
[19:57:46] <legoktm>	 there's a task somewhere about running it for the rest of the wikis
[19:58:08] <legoktm>	 I remember looking into this during SULF