[05:18:40] wm-bot3 just died. Is that bad? [05:19:39] Oh hi wm-bot! [05:19:39] just? [05:19:39] !ping [05:19:40] !pong [05:19:43] huh_: i did reboot the host... [05:20:34] !ping [05:20:35] !pong [05:21:05] !ping [05:21:05] !pong [05:21:06] !log bots booted bots-labs and started boncers/wm-bot back up again [05:21:06] Logged the message, Master [05:21:07] !ping | mutante [05:21:07] mutante: !pong [05:21:21] !ping [05:21:21] !pong [05:21:44] !pong [05:21:44] !pang [05:21:49] !pang [05:21:49] !pung [05:21:53] Ping reply from wm-bot in 1 minute, 12.88 seconds. -> Ping reply from wm-bot in 1.24 seconds. [05:21:55] !pung [05:21:55] !derp [05:22:05] !derp [05:22:05] Ok enough. I don't want to do this all day. [05:22:15] !derpie [05:22:26] !derpie is !herpie [05:22:27] Key was added [05:23:00] !herpie [05:23:09] !herpie is !derpie [05:23:10] Key was added [05:23:42] seems to work fine now, thanks jeremyb [05:23:52] !herpie [05:23:53] !derp [05:24:00] huh_: ^^ :P [05:24:13] jeremyb: thanks!:) [05:24:18] how? [05:24:20] !herpie [05:24:20] !derp [05:24:27] !derpie [05:24:27] !herpie [05:24:35] I have no idea. [05:24:51] huh_: I was messing with you :P [05:24:53] mutante: bitte [05:25:12] huh_: fixed [05:25:17] !derpie [05:25:18] !herpie [05:25:22] !herpie [05:25:22] !derpie [05:25:48] !herpie [05:25:53] ... [05:25:58] I though _ was illegal in DNS. e.g. mysql -h be_x_oldwiki.labsdb be_x_oldwiki [05:26:00] !derpie [05:26:35] Dispenser: you mean like _dmarc? [05:26:39] !ping [05:27:13] huh_: you have inserted char \u000f [05:27:22] in your command [05:27:36] zhuyifei1999: how did you know? [05:27:51] my client showed it [05:27:54] oh [05:28:50] in most clients, that wouldn't be so obvious... [05:29:25] $ host -t any _dmarc.wikimedia.org [05:29:25] _dmarc.wikimedia.org descriptive text "v=DMARC1\; p=none\; sp=none\; rua=mailto:postmaster@wikimedia.org" [05:29:29] Dispenser [05:32:36] @trusted [05:32:36] I trust: .*@wikimedia/.* (2trusted), .*@mediawiki/.* (2trusted), .*@wikimedia/Ryan-lane (2admin), .*@wikipedia/.* (2trusted), .*@nightshade.toolserver.org (2trusted), .*@wikimedia/Krinkle (2admin), .*@[Ww]ikimedia/.* (2trusted), .*@wikipedia/Cyberpower678 (2admin), .*@wirenat2\.strw\.leidenuniv\.nl (2trusted), .*@unaffiliated/valhallasw (2trusted), .*@mediawiki/yuvipanda (2admin), .*@wikipedia/Coren (2admin), [05:32:52] quite a list [05:33:04] i think i'm not on there :P [05:33:20] nice way to ping a bunch of people at once [05:33:41] huh, wikipedia/.* is trusted? didn't notice that [05:33:55] but not wikispecies/.* [05:33:56] jeremyb: :P [05:33:57] :P [05:34:04] @whoami [05:34:04] You are root identified by name .*@wikimedia/jeremyb [05:34:09] @join #wikimedia-gs [05:34:11] huh_: mysql --help [...] Usage: mysql [OPTIONS] [database] [05:34:17] @help [05:34:17] I am running http://meta.wikimedia.org/wiki/WM-Bot version wikimedia bot v. 1.20.2.0 my source code is licensed under GPL and located at https://github.com/benapetr/wikimedia-bot I will be very happy if you fix my bugs or implement new features [05:34:51] jeremyb: I thought only petan and T13 are root [05:35:11] Dispenser: it doesn't work without the USE. Did you try running it? [05:35:58] zhuyifei1999: i did just reboot the box... [05:36:32] actually i granted myself root IIRC [05:37:51] So its redundant to supply `USE database`, unless you looking to cut interrupter start up time. Use `mysql --defaults-file=~/replica.my.cnf -h $dbname.labsdb -BcN ${dbname}_p` [05:38:44] hm [05:38:47] * huh_ tries [05:39:13] Is there a difference between $dbname and ${dbname} ? [05:39:18] no [05:39:31] one is nicer when it appears in a string [05:39:35] *not really [05:39:43] * jeremyb runs away [05:40:38] Why doesn't the skeloton `ln -s .my.cnf replica.my.cnf ` [05:40:52] zhuyifei1999: there is root.. and then there is ..actual root [05:41:19] and then there is actually actual root [05:41:21] mutante: root in irc and root in shell [05:41:25] err... ln -s replica.my.cnf .my.cnf [05:42:18] Dispenser: works, thanks [05:42:19] zhuyifei1999: yea,one is admin of a project, another is labs admin who can add himself to all projects, yet another is root on the server that runs labs [05:43:55] Or `echo "!include replica.my.cnf" > .my.cng` (Untested) [05:44:18] Dispenser: .cnf [05:45:46] Last thing before I sleep. If they don't want to provide a Toolserver.wiki/Toolserver.namespacename it could be setup as a project :-) [05:51:39] Dispenser: {{support}} [06:00:19] zhuyifei1999 / Dispenser https://bugzilla.wikimedia.org/show_bug.cgi?id=48625 http://pastebin.com/zddwGPpa [06:02:36] huh_: I've already cced [06:06:15] @seen whowas [06:06:15] zhuyifei1999: Last time I saw whowas they were talking in the channel, but they are not in the channel now and I don't know why, in #cvn-sw at 12/24/2013 1:04:38 AM (5h1m36s ago) [06:22:09] For the record, whowas=huh_=Wikimedia User:PiRSquared17=me [06:27:32] /whois whowas , maybe one should be called "seen" [09:01:06] Hi! http://tools.wmflabs.org/url-converter/test.html is not reachable, what could be the reason for this? (I guess it's my fault, because tools of others still work) [09:06:40] strange. the access.log says [24/Dec/2013:09:02:29 +0000] "GET /url-converter/test.html HTTP/1.1" 200 25 [09:08:05] but if I try to reach that page, I have to wait several minutes, before getting "Service Temporarily Unavailable" [09:12:22] lustiger_seth: try "webservice start" on shell [09:12:54] tools-webserver-01 is overloaded [09:12:59] Starting webservice... started. [09:13:27] still not working [09:15:21] so, do I just have to wait (because admins will get some kind of automatical notification, if a server is overloaded?)? [09:17:00] lustiger_seth: slow but seem to work for me [09:17:20] now it seems to work for me, too [09:17:28] and it's overloaded for a long time: http://ganglia.wmflabs.org/latest/?r=week&cs=&ce=&c=tools&h=tools-webserver-01&tab=m&vn=&mc=2&z=medium&metric_group=ALLGROUPS [09:17:43] sort of the usual delay of everything, i guess ;) [09:18:24] hmm, but my cgi-scripts still don't work [09:18:26] http://tools.wmflabs.org/url-converter/index.cgi [09:19:20] but in november everything worked [09:22:07] lustiger_seth: you sure you ran it in your tool account? [09:23:10] what do you mean? it's a web-script, so everybody can use it. [09:24:26] I mean did you ran the script as local-url-converter? [09:24:38] yes, that works fine [09:24:49] maybe the server configuration has changed, and perl-cgi-scripts are allowed in cgi-bin only? [09:26:23] I never used perl, sorry [09:27:24] lustiger_seth: but I don't fund your httpd running: http://tools.wmflabs.org/?status [09:27:30] *fing [09:27:34] *find [09:29:06] hmm, now http://tools.wmflabs.org/url-converter/test.cgi and http://tools.wmflabs.org/url-converter/index.cgi seem to work, but one has to wait server minutes [09:29:17] s/server/several/ [09:31:11] most of the others seem much faster [09:31:27] right, that's the strange thing. [09:32:18] I don't understand that status-page. of course bots and other permanent jobs are listed there. but web pages are not running constantly. or has everybody got his own webserver? [09:32:36] lustiger_seth: kind of [09:32:54] that's what "webservice start" does [09:33:31] ok, I see. I've never used that command before on the lab server, I guess. [09:33:31] https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help/NewWeb [09:33:39] lustiger_seth: ^^ [09:35:09] "webservice status" says: 'queue instance "task@tools-exec-05.pmtpa.wmflabs" dropped because it is temporarily not available' [09:35:34] and 'Your webservice is not running.' [09:36:17] (I did a webservice restart) [09:37:09] mh, there's something wrong, and we encountered it before (what was the solution? no idea) [09:39:09] shall I file a ticket at bugzilla? [09:39:26] we could grep chat logs first ;) [09:39:40] but of course you can [09:39:47] helps tracking things [09:42:43] zhuyifei1999: try this hacked version: [09:42:45] qsub -e ~/error.log -o ~/error.log -i /dev/null -q webgrid -l h_vmem=4g -b y -N "httpd-url-converter" /usr/local/bin/tool-lighttpd >/dev/null 2>&1 [09:43:19] this is so beautiful :o [09:44:25] I'm not sure about its beauty (and I'm a perl programmer...) [09:44:50] well, i program in terse tcl … [09:45:38] condolence [09:45:43] but tool labs crontabs/jobs invocation tops it all [09:46:32] I'm a python programmer [09:46:47] I used that magic qsub-line that I don't understand. What should I do next? [09:47:08] check if it works [09:47:33] 'Your webservice is not running.' [09:47:38] :( [09:47:54] then I have no idea why [09:48:32] @wakeup Coren [09:49:18] it's night in canada, anyway [09:49:40] giftpflanze: Coren is in Canada? [09:49:56] i very much think so, yes [09:51:04] afaics the problem has nothing to do with the fact that the code is written in perl, am I right? The problem is that the webserver won't start. [09:51:26] probably [09:51:43] it is the problem that the qsub doesn't send the job to tools-webgrld-01 [09:51:48] * lustiger_seth want's to find out a good summary for the bugzilla ticket [09:51:49] *webgril [09:51:55] *webgrid [09:55:41] lustiger_seth: could you give us an output of "qstat"? [09:55:54] empty [09:56:31] empty = no jobs running [09:56:43] = webservice not started [09:56:52] if i repeat the "beutiful" command, the output of qstat is: [09:57:04] 1961692 0.00000 httpd-url- local-url-co qw 12/24/2013 09:56:05 1 [09:57:23] 1961692 0.25000 httpd-url- local-url-co r 12/24/2013 09:56:16 webgrid@tools-webgrid-01.pmtpa 1 [09:57:31] and then again empty [09:57:52] aha! [09:58:23] could you read the contents opf error.log? [09:58:29] *of [10:00:05] '(log.c.166) server started' followed by '(log.c.118) opening errorlog '/data/project/url-converter/access.log' failed: Permission denied' [10:00:19] and '(server.c.938) Configuration of plugins failed. Going down.' [10:00:53] the rights for access.log are -rw-r----- 1 root local-url-converter 10985 Dec 24 09:53 access.log [10:02:06] qsub -e ~/httpd.err -o ~/httpd.err -i /dev/null -q webgrid -l h_vmem=4g -b y -N "httpd-url-converter" /usr/local/bin/tool-lighttpd >/dev/null 2>&1 [10:02:31] lustiger_seth: try ^^ before a root remove that log file [10:03:24] I have no idea why a root decided to create that file [10:04:11] is anybody here a root? [10:05:29] Coren [10:05:47] and maybe others [10:06:07] I meant somebody, who is awake. ;-) [10:07:35] if they are not afk, they should answered already :) [10:07:42] *have [10:12:02] ok, I'll finish creating that ticket [10:13:49] qsub -e ~/httpd.err -o ~/httpd.err -i /dev/null -q webgrid -l h_vmem=4g -b y -N "httpd-url-converter" /usr/local/bin/tool-lighttpd >/dev/null 2>&1 [10:14:05] lustiger_seth: ^^ also doesn't work? [10:15:23] (note that I've changed the log file name) [10:16:07] that command seems to do nothing [10:17:39] :( [10:18:09] (httpd.err is empty, qstat says nothing) [10:21:15] https://bugzilla.wikimedia.org/show_bug.cgi?id=58931 [10:22:08] However, thanks a lot for your time and help! :-) [10:24:55] :) [11:35:50] who do i have to kick to make beta work again? [12:22:04] Hi. I have problems in bastion. I can connect but looks Like I have lost my rights there. [12:22:45] E,g. I cannot create new files or creating ssh Id:rsa files [12:23:42] I used to be able to ssh other instance (piramido) but my files are all empty (or at least they look to me empty). [12:23:53] Any hint? [12:23:54] nfs issues again? :/ [12:24:29] no idea ... I only know it does not work as it used to [12:24:29] -bash: test: Read-only file system [12:25:09] !ops [12:25:15] bah, that doesn't work :-( [12:25:26] Mpaa: there were issues with storage yesterday -- I think this might be related. [12:25:46] beta also had some issues [12:26:34] all I know is that all my files appear with 0 size and my .ssh is empty [12:27:53] e.g. rm: cannot remove `catlib.py': Read-only file system [12:28:23] is it possible that I am not logging correctly? [12:28:46] no, the issue is the filesystem [12:29:16] valhallasw, is there anything I can do other than wait? [12:29:36] poke Coren, possibly [12:29:47] or other labs ops [12:30:27] when i change my bugzilla e-mail address the "my comments" search doesn't get updated, is that a bug? [12:31:16] Coren, can you please help. I have file system problems. [12:36:57] valhallasw, can you work on bastion? [12:37:44] i can't even edit that search nor "forget" it [12:37:51] no, I have the same file system issues [12:37:59] giftpflanze: have you tried logging out and in again? [12:38:10] might be a caching issue [12:38:12] that always helps, hm? [12:38:14] or just a bugzilla bug, of course [12:38:31] well, the email address also is your user name, after all [12:38:55] relogging doesn't help [12:39:35] valhallasw, ok .... well Nok but at least is not specific to me ... what is the way-forward in such cases? Wait and see ...? [12:39:44] Mpaa: yes.. basically. [12:40:17] valhallasw, I wanted to fix your comments on archive, you can go ahead wit it if you like [12:40:20] or filing a bug [12:40:57] Mpaa: I'll take a look at it later today [12:41:17] the Widar tool is broken as well [12:41:19] :( [12:41:56] aaaw, my god, changing searches is so obfuscated … i found the final step now [12:44:41] valhallasw, OK, bye [12:56:52] valhallasw: should i close the bug again as fixed or as worksforme or leave it open? [12:57:24] giftpflanze: the 'can't change email search' bug? [12:57:49] probably best to post the workaround, but keeping it open [12:58:00] oh wrong channel [12:58:11] i mean the beta cluster is fucked up again bug [13:57:18] Hello. I'm trying to jsub a python file from my tools account. I have tried "jsub cat_refs_rc.py" in the command line but there is an output error in the .err file which says "[Tue Dec 24 08:57:19 2013] /data/project/arnaubot/cat_refs_rc.py: not an executable file" [13:57:42] Anyone has an idea what's happening? [13:58:53] Arnaugir, try jsub python cat_refs_rc.py [13:59:32] Cyberpower678: cool, that was it. thank you very much [13:59:49] i'm finding my way through python and the command line :P [13:59:56] not that easy [14:00:34] Arnaugir, you always need to specify what language it's programmed in before specifying the file. [14:02:13] or you can add #!/usr/bin/env python to the top of the file, and chmod +x it [14:02:17] Cyberpower678: No you don't, if the file is +x and has a proper #! line at the top. [14:02:33] * valhallasw high-fives anomie [14:02:39] anomie, right, right. I forgot [14:03:16] Arnaugir, ^ [14:03:16] Cyberpower678: And really, you're not specifying which ''language'' it's in, you're specifying the name of the interpreter to run [14:04:07] anomie, wharves. :p [14:05:13] Cyberpower678: Docks have nothing to do with it. It's better people have correct information. [14:06:01] anomie, it's easier for me to remember "specify what language it's written in" as compared to "specify which interpreter to run" [14:07:13] Cyberpower678: That'll bite you when you write something in node.js, where the interpreter is named either "node" or "nodejs" depending on the system you're on. ;) [14:07:53] anomie, alright you win, as usual. :p [14:08:47] anomie: well, that's something that is going to bite anyway [14:09:10] valhallasw: One of the several reasons I'm not terribly fond of nodejs. [14:09:23] * Cyberpower678 goes back to handling ACC requests. [14:10:02] just like it bites that some linux distros (I'm looking at you, Arch) thought it was a good idea to point /usr/bin/python to /usr/bin/python3 [14:10:04] 10 hours until Christmas eve. [14:11:05] Cyberpower678: You're in a UTC-24 timezone? [14:11:26] anomie, :D [14:11:44] Cyberpower678, I just see Christmas eve as the night before Christmas. [14:11:52] It's not night yet. :p [14:14:39] anomie, do you think I should give Cyberbots I and II off for Christmas. :p [14:15:26] Cyberpower678: Ask them whether they'd rather have off or just keep working? [14:15:40] anomie, good idea. [14:17:22] :D [14:18:27] * anomie looks at http://www.timeanddate.com/worldclock/sunearth.html?n=0&month=12&day=25&year=2013&hour=0&min=12&sec=0 to work out where it will be dusk in 10 hours from Cyberpower678's earlier statement [14:20:46] anomie, try 8.5 hours instead to be more specfic. [14:22:52] anomie, that should make it fairly easy for you to work out. [14:27:11] 15:26 [freenode] CTCP TIME reply from Cyberpower678: Tue Dec 24 09:27:03 [14:27:12] :Y) [14:27:30] the magic of IRC clients [14:27:41] * Cyberpower678 stabs valhallasw with a sword. [14:33:58] Coren: when you are here, please answer to me if you can upgrade the grid engine to allow for the massive weblink checking in magnitudes we reasoned about ealier. i finally decided that threading doesn't work reliably and want to integrate with tools [14:34:21] +if possible [14:35:32] giftpflanze: Probably by giving that project its own grid node like I have for some Cyberpower tools; that allows you to manage resources differently than with vmem ceilings. [14:36:29] are job number limitations lifted for that? as you probably know i want to do an array job [14:37:21] do you remember the number of threads running at the same time you thought would be reasonable? [14:37:31] s/threads/jobs/ [14:38:50] @coren [14:41:13] giftpflanze, you can run a maximum of 15 or 16 threads before the new ones just sit there queued up. [14:41:32] i vaguely remember that [14:41:50] That's why I have a cyberbot node where I can now run 20 jobs at the same time. :D [14:42:29] lucky you :D [14:43:00] i assume that that are different tasks/tools? [14:43:18] Yea. All Cyberbot though. [14:43:50] giftpflanze, and your own node becomes somewhat resistant to labs failures. :p [14:44:08] great! … [14:44:38] i think the canonical solution would be raising job limits for heavy tools [14:44:49] It's kind of funny when I look at tools in Ganglia, how everything but my node is marked red. ;p [14:45:12] ah, so why don't we add more nodes? [14:45:55] because personal nodes can only be accessed by very specific users. [14:46:19] Meaning while everyone's bots are frying the exec nodes, yours is just idling in yours. [14:46:20] as well as more web nodes [14:46:44] Cyberpower678: that is the answer to what? [14:47:12] adding more nodes won't fix things when labs fails. [14:47:21] hm, ok [14:47:28] I'm probably talking a bunch of bs though. :p [14:47:45] but i think there are too few grid resources [14:47:56] I just noticed that personal nodes don't seem to fry when labs fails. [16:00:40] I am trying to connect to piramido from bastion, but I have publickey issues. [16:00:49] I used to be able to ssh piramido [16:00:53] what can be wrong? [16:01:19] it could be everything [16:04:21] giftpflanze, any hint to fault finding? [16:04:45] Have you been able to before? [16:06:54] yes [16:07:09] i have no knowledge of that stuff, mpaa [16:08:07] Reedy, I suceeded, ot was probably something wrong with ssh-agent [16:08:10] thanks nayhow [16:09:08] Not forwarding your key? [16:09:15] yes [19:51:52] Can anybody tell my why https://tools.wmflabs.org/krdbot/cgi-bin/Dateientlinkerlog.pl does run into timeout? Are there still NFS problems? [19:55:44] i have some questions about http://tools.wmflabs.org/?status [19:56:27] wait, nm. [20:24:58] krd: The webserver is occasionally overloaded when a crawler hits it, that's why the new per-tool web scheme is recommended rather than rely on the shared server. [20:25:21] krd: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help/NewWeb [20:27:43] Coren, where is the sources for the tool labs indexes? [20:27:52] I'm going to make this into a google code-in task. [20:28:33] johang: https://git.wikimedia.org/tree/labs%2Ftoollabs/HEAD/www [20:28:44] jorm: ^^ [20:28:52] *sigh* Even more complicated? [20:29:51] krd: For the most part, it's much simpler. [20:30:03] Since you have complete control over the webserver. [20:30:11] jorm: there are a few bugs on the topic btw [20:30:25] Unless you have .htaccess files, it's no more complicated than just typing 'webservice start' as your tool. [20:30:39] amazingly, i am fixing some of these. [20:31:45] or, maybe not. [20:32:08] hrm. is there any way to test changes to this without deploying to production? [20:32:35] jorm: You can easily make a tool labs project to do exactly that. [20:32:49] jorm: Since that tool would have the same webserver et. al. [20:32:58] oh, hi coren [20:33:00] Coren: I also had to remove the old access.log which was owned by root [20:33:09] i wonder if it wouldn't be just simpler for me to do this myself. [20:33:12] and my tool shows 404 now. [20:33:12] this is not a code-in task. [20:33:22] krd: Ah, yes, I need to put that in a FAQ somewhere. Lemme do that now before I forget. [20:33:41] krd: What url? [20:34:15] tools.wmflabs.org/krdbot/cgi-bin/Catwatch.pl [20:35:15] Coren: i had some questions earlier that you didn't answer yet [20:36:15] krd: Hm. .htaccess or cgi-bin. :-) Look at the very bottom of the web page, there is the short configuration stanza to add to your .lighttpd.conf to have cgi-bin functionality. [20:36:37] giftpflanze: Could you repeat them please? I'm on my iPad and scrollback access sucks from it. [20:36:44] np [20:37:20] Is this the recommended approach? Or shall I move it to something else? [20:38:09] krd: Either works; but since your scripts are named .pl you can simply put them directly in public_html to shorten the URL to tools.wmflabs.org/krdbot/Catwatch.pl [20:38:39] great, thank you. [20:39:44] Coren: may i give you a paste? [20:39:48] giftpflanze: Sure. [20:41:01] http://bpaste.net/raw/I27m2GK3vgZaPQh24eQO/ [20:42:35] giftpflanze: Yes, if I set aside a node for your tool you have no limitation beyond its actual memory and CPU; you can schedule as many jobs as you need on it. [20:42:55] good [20:43:26] Coren: I still get 404 [20:43:27] giftpflanze: Although to do a true array job you need to use MPI in your code. [20:43:36] krd: Lemme see what's up. [20:44:07] yeah, we discussed earlier that it isn't a true array job [20:45:22] krd: Oh, like the note says at the bottom, you have to move your cgi-bin inside your public_html with lighttpd. I can do it for you if you want. [20:45:27] Coren: i don't suppose it would be possible for me to test changes locally? That seems like a crazy hassle. [20:45:44] Coren: I just saw it myself. [20:45:47] I'll try it. [20:45:53] jorm: Not if you want to test anything that speaks to ldap or checks local groups. Do you want me to set a project up for you to do tests? [20:46:04] that would be awesome. [20:46:24] i can't imagine testing this code in the live instance; breaking toollabs seems. . . ill advised. [20:46:27] coren: so, do you remember what number of jobs would be needed for the task? [20:46:59] giftpflanze: No, but that's immaterial. As long as you keep your memory within what's actually available. [20:47:30] Coren: i thought it would be necessary to determine the size of the node [20:48:26] Memory is the important factor. [20:50:56] jorm: Err, you don't seem to have a user account on Wikitech? [20:51:50] Coren: um, how do we proceed? [20:52:16] giftpflanze: Open a bz, I'll have the opportunity to create your node between xmas and new years [20:52:26] ok [20:52:56] jorm: Nevermind, you just weren't added to tools. [20:54:39] sec. i'm running to get some food. [20:57:49] jorm: https://tools.wmflabs.org/betaweb/ Caveat: since the default web pages have absolute links in them they fall back to the default site. [20:58:14] jorm: Log in to tools-login and you can 'become betaweb' to access its home. The site is in public_html [21:02:22] jorm: Nevermind the caveat, I've changed the source in that copy to use all relative URLs. [21:05:11] Coren: Do you have a minute for 2 cpan installs? [21:05:42] krd: That takes a bit more than a minute, I have to make debians out of them if they don't already exist. What do you need? [21:06:13] There re 2 open bugs [21:06:42] I would tell you the links, but my computer seems to be just crashing. [21:06:53] Hm. I suppose I can do a round of install now. Lemme see. [21:07:04] thx. [21:07:14] I'll have to reboot in the meantime. [21:08:55] Coren: btw https://blog.wikimedia.org/ <- the privacy thing [21:12:31] PiRSquared: Didn't know it would end up on the blog, but I know about the substance of this for a while; I participated in the discussion that led to it. :-) [21:14:25] Which one was that? [21:15:48] Which what? Discussion? Internally with LCA; while we discussed the reponse we would have to make to the wikide's sorta-petition. [21:17:05] https://de.wikipedia.org/wiki/Wikipedia_Diskussion:Kurier/Archiv/2013/09#Deep_user_inspector [21:17:28] it seems mostly dewiki cares [21:22:13] Coren: With lighthttpd a script that don't compile will show no error message but an empty file download. How can this be improved? [21:22:26] Coren: what do we do about the dns caching when the bot runs in tools? [21:25:41] krd: The error messages end up in your ~/error.log [21:27:03]