[05:40:47] @seen petan [05:40:47] zhuyifei1999: Last time I saw petan they were quitting the network with reason: Remote host closed the connection N/A at 7/10/2013 9:22:49 PM (08:17:58.5923890 ago) [06:59:00] ls [07:59:36] !log bots petrb: had to restart bsql01 OOM [08:00:50] petan: :< [08:01:11] addshore: I need you to move your bot from bots soonish [08:01:17] I mean from bsql [08:01:29] funny you should say that :P [08:01:33] it should be happening today ;p [08:01:39] cool [08:01:42] ^^ [08:01:42] you need any help? [08:01:47] like moving the db to tools etc [08:01:50] petan: Where's labs-morebots? [08:01:55] zhuyifei1999_ dead [08:02:00] I am afraid I can't fix it :P [08:02:03] nah, i will be totally scrapping the old db and recreating :) [08:02:03] let me check [08:02:07] ok [08:02:15] (but not quite yet) ;p [08:02:58] petan: so !log bots petrb: had to restart bsql01 OOM is not working [08:03:04] indeed [08:03:32] sec [08:21:37] petan: I will infact ask for some adcive with my db in a sec :> [08:49:45] so petan still here? :> [08:50:13] addshore: yes [08:50:20] :O [08:50:45] hm [08:58:20] addshore? [08:58:41] pm :> [09:14:13] @notify petan [09:14:14] This user is now online in #huggle. I'll let you know when they show some activity (talk, etc.) [09:31:35] petan: i want to run a slightly intensive task but i want to monitor its output live [09:31:59] jsubs output to the .out file just seems stupidly slow.. [09:33:10] * addshore thinks a tools-dev-exec would be great [09:33:36] or just a bigger -dev [09:34:23] !screenfix [09:34:24] script /dev/null [09:42:33] addshore for now you can use unused bots-bnr1 [09:42:39] mhhhm [09:42:42] i may have to [09:42:44] addshore or you can use unused toolsbeta-exec-01 [09:42:50] true! [09:42:56] i think i just killed tools-dev... [09:42:59] http://ganglia.wmflabs.org/latest/?r=20min&cs=&ce=&c=tools&h=tools-dev&tab=m&vn=&mc=2&z=medium&metric_group=ALLGROUPS [09:43:00] really? [09:43:02] I don't think so [09:43:03] :P [09:43:12] everything stopped ;p [09:43:18] o.o [09:43:20] meh [09:43:32] well [09:43:33] it's somewhat running just slowly [09:43:37] *kills* [09:43:42] petan: , its not running at all :P [09:43:49] it is I am sshed there [09:43:53] I can like move my cursor :P [09:43:56] :< [09:43:59] that's all [09:44:01] brb 2h [09:44:05] see ya [09:52:31] mhhm, cant run on tools beta xD [10:34:00] hello, I cant access bots-login and tools-login, it always raises an "Access denied (publickey)" error [10:55:40] Pyfisch: details? [10:56:02] zhuyifei1999: which details do you need? [10:58:49] Pyfisch: made https://wikitech.wikimedia.org/wiki/Special:Preferences#mw-prefsection-openstack ? [10:59:21] @notify petan [10:59:21] This user is now online in #huggle. I'll let you know when they show some activity (talk, etc.) [11:01:11] zhuyifei1999: yes there is a key in [11:01:33] that's strange. Where's petan? [11:01:57] @seen petan [11:01:57] zhuyifei1999: petan is in here, right now [11:02:42] Pyfisch: How about google it? [11:03:29] I am running http://meta.wikimedia.org/wiki/WM-Bot version wikimedia bot v. 1.10.12.0 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 [11:04:16] @labs-info Pyfisch [11:04:16] I don't know this instance, sorry, try browsing the list by hand, but I can guarantee there is no such instance matching this name, host or Nova ID unless it was created less than 47 seconds ago [11:04:45] @labs-user Pyfisch [11:04:45] Pyfisch is member of 3 projects: Bastion, Bots, Tools, [11:09:38] zhuyifei1999: google tryed, nothing found [11:11:26] petan: are you here? [11:14:04] Pyfisch: can you try "Permission denied (publickey)"? Seems that most of the google search results are that. [11:15:22] will try [11:16:26] petan! [11:33:18] Pyfisch: Is it working? [11:33:33] not yet [11:54:32] [bz] (8NEW - created by: 2Quim Gil, priority: 4Normal - 6enhancement) [Bug 51050] Connecting wikitech.wikimedia.org user profiles with community metrics - https://bugzilla.wikimedia.org/show_bug.cgi?id=51050 [11:57:10] zhuyifei1999 yes [11:57:48] petan: finally. Can you solve Pyfisch's problem? [11:58:30] which problem [11:59:10] (10:34:01) hello, I cant access bots-login and tools-login, it always raises an "Access denied (publickey)" error [12:00:07] ok [12:00:14] Pyfisch are you here [12:01:59] @notify Pyfisch [12:01:59] This user is now online in #wikimedia-labs. I'll let you know when they show some activity (talk, etc.) [12:26:51] petan: now, I am here [12:28:18] hello [12:29:58] can you describe what is wrong [12:30:19] Pyfisch? [12:31:51] * Pyfisch runs with his head against a wall [12:32:26] petan: problem solved [12:32:58] I just forgot to start PAGEANT before using PUTTY. [12:34:46] Pyfisch: Why don't you use OpenSSH? [12:35:35] zhuyifei1999: why I should not use putty? [12:36:13] nvm [12:36:35] zhuyifei1999: https://en.wikipedia.org/wiki/OpenSSH#Microsoft_Windows_support for this reason [12:36:44] I am using option 2 :-p [12:36:59] zhuyifei1999: what is "nvm"? [12:37:13] never mind [12:38:00] petan started using that before me [12:38:56] Pyfisch: And why don't you use Linux? [12:40:02] this question is more difficult to answer ;-) [12:40:31] I always used Windows and I only use Linux for this server stuff now [12:40:57] :P [12:41:40] yeah I know some people here arround will kill me for still using windows *runs away* [12:43:44] I use windows as well [12:43:46] I have to [12:44:02] petan: Why? [12:44:28] because this company I work for provides only laptops with windows and I am not allowed to reinstall them [12:44:38] so my office laptop is windows [12:45:02] petan: I thought you're working for wikimedia [12:45:14] if only I could :P [12:45:42] * zhuyifei1999 is going to kill petan's company [12:45:50] it's pretty big one [12:45:58] 200 000+ employees [12:47:08] zhuyifei1999: I is enough if you kill the managment ;-) [12:48:25] * zhuyifei1999 is going to kill the managment unless he allows petan to use Linux [12:49:40] linux isn't very popular in large bussinesses [12:49:53] our servers are running HP-UX [12:50:13] big companies needs professional support [12:51:04] most big companies run solaris, AIX etc [12:52:00] tbh I can't imagine some of our servers, especially these with nearly 1000 days of uptime running linux :P [12:52:18] Hi, I'm trying to migrate a tool from the toolserver to labs [12:52:18] wrought1 hey [12:53:07] hi petan, I signed up submitted my key, but when I try to add a new tool here http://tools.wmflabs.org/ I get this response: "No Nova credentials found for your account." [12:53:19] logout, login [12:53:22] ;) [12:53:24] petan: Imagine a server running windows :-P [12:53:30] I know some [12:53:35] they barely have uptime more than 6 days [12:56:46] Anyway, windows is hard to use [13:03:15] petan: That's not my experience. I've worked at several very very large companies and Linux dominates. [13:03:43] petan: Solaris used to be common but started being dethroned at least 5 years ago; more in some places. [13:03:46] hmm... it may depend on level of support they require [13:03:59] h-pox is... marginal. [13:04:10] linux is definitely very good for web servers and like, but I didn't see many linux boxes running oracle etc [13:04:23] except for my testing boxes :) [13:04:30] petan: Hell, Oracle even has its own distro now. :-) [13:04:38] yes... now :) [13:04:52] petan: Has been the case for what... 6, 7 years? [13:05:06] but in times when these servers were installed, linux 2.2 was being released as a hot new stuff [13:06:03] yes I know about oracle linux, it's a cool thing, but in fact I didn't see that redhat fork much in action... [13:06:29] I installed 2 oracle servers running oracle-tux for some development environ [13:06:45] but on production they had solarix anyway (running SPARC) [13:07:30] the main reason they had different system on development than on production, was that version of sw they needed was only for SPARC on solaris and they didn't have any testing server with this platform [13:07:49] * solaris [13:12:07] petan: I actually know a Windows Server 2003 that hardly ever shut down. [13:12:28] was it doing something in fact? :P [13:13:30] I don't know. It's not mine [13:14:28] I know where it is, but I'm not allowed there. [13:17:27] that is worst thing for me [13:17:36] when I know something but I am not allowed to do that :P [13:17:46] and I have to wait for others to do that for me [13:17:58] I really hate that [13:18:21] btw Coren what is the plan for that new tools sql server? [13:19:03] petan: The plan is simple: make new DB. Point users at new db. Decommission old db after $interval. :-) [13:19:17] do you think you could configure it secure enough so that me and scfc could have some kind of access there, so that we can help with maintenance? unlike replicas it's not going to contain any really dangerous data... [13:20:07] petan: That DB isn't going to have only Tool Labs stuff, but we'll see. [13:20:33] yes, addshor want to move one db to tools-db when you expect it to exist [13:20:55] if it was soon he could do that on final db and we could save some moving of data [13:21:09] his db is going to be huge I guess, now on bots it has some 30gb [13:23:02] petan: I'm doing that now (the new DB) [13:23:12] nice :0 [13:28:00] er [13:28:23] damn [13:28:35] @optools-permanent-on [13:28:35] Operator tools are now in permanent mode on this channel [13:29:09] let's try them out :o [13:29:10] @jb blah [13:29:19] !ping [13:29:19] pong [13:29:20] pong [13:29:30] @unjb blah [13:29:31] heh [13:29:32] :P [13:29:45] damn [13:29:45] this needs fix [13:30:35] petan: @optools-permanent-on not in [[meta:wm-bot]] [13:30:50] I just implement it it's not even working :D [13:31:45] and wm-bot is still op [13:38:05] that is correct [13:38:08] that's what the function is about [13:38:19] it produces less spam when the bot need to op / deop all the time [13:38:23] and when chanserv crash like yesterday it suck [13:38:43] @jb blah [13:38:48] @unjb blah [13:38:52] here we go [13:39:02] petan: do we have an smtp server on labs? [13:39:15] I think we do but this is undocumented [13:39:17] !labsadmin [13:39:27] how does it do the .forward thing? [13:39:37] every instance has local MTA [13:39:46] I suppose that thing is sufficient for this [13:39:53] @search admin [13:39:53] Results (Found 3): morebots, labs-morebots, hyperon, [13:40:03] !toolsadmin [13:40:03] https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Documentation/Admin [13:40:16] https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Documentation/Admin#Mail [13:40:32] YuviPanda it say "there is tools-mail which only Coren knows what it is for" :P [13:40:44] undocumented... [13:41:31] petan: hehe [13:41:31] ok [13:41:32] a [13:51:34] petan: The local MTAs aren't, strictly speaking. tools-mail is the MTA, the other hosts just blindly forward any locally delivered mail there (and don't accept incoming mail at all) [13:53:17] ok [13:53:27] first piece of documentation yay [13:54:21] Coren: hmm, so unlike toolserver we don't have toolname@tools.wmflabs.org forwarding [13:54:52] YuviPanda: That's pending on Luis. :-) [13:54:56] aaah [13:54:58] right [13:55:19] Strictly speaking, we /do/ have it up and ready, just needs the final okay. [13:56:27] hey all, I came in earlier asking about migrating a tool from the toolserver to labs. I added an FAQ here adding the step of asking for Nova access via IRC: https://wikitech.wikimedia.org/wiki/Help:Access#Access_FAQ [13:56:38] Let me know if there is some better location or phrasing, or please simply improve [13:57:21] wrought1: Strictly speaking, asking on IRC just speeds the process -- it's not required. We hack at the list of pending requests at regular interval even unprompted. :-) [13:57:52] ah okay, it was unclear to me the request list was the right venue for that request, is it the proper location for all requests? [13:59:46] 'directions unclear, hair on fire'? :) [14:00:28] wrought1: There's actually an autogenerated request for general shell access upon wikitech account creation; https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools has a link to the access request form to Tool Labs specifically. [14:04:02] hmm, you don't actually need an SMTP server to *send* email do you [14:04:33] but it'll probably get rejected for the most part [14:04:33] :( [14:04:47] Coren: is port 25 blocked? [14:05:20] YuviPanda: See above, in re "Pending on Luis". :-P [14:05:29] Coren: sending email too? [14:05:34] oooooh [14:05:34] right [14:05:37] nevermind me. [14:05:44] YuviPanda: No, sending should work. [14:05:52] assuming it's not blacklisted [14:06:01] on the receiver's side [14:06:17] YuviPanda: It shouldn't be. No SPF record yet, but I doubt our public IPs have bad karma. [14:06:23] right [14:22:51] hi [14:23:03] I'm Kenrick from Indonesian Wikipedia, I have set up a service at tools.wmflabs.org (at http://tools.wmflabs.org/idwiki-gpu ) but it returned 500 Internal Server Error when I access it through my web browser. [14:23:19] I put my files at "/data/project/idwiki-gpu/public_html" via WinSCP using the configuration taught at https://wikitech.wikimedia.org/wiki/Help:Access_to_ToolLabs_instances_with_PuTTY_and_WinSCP#How_to_set_up_WiNSCP_for_direct_access_to_your_Tool_Labs_account and I'm aware of https://wikitech.wikimedia.org/wiki/User:Magnus_Manske/Migrating_from_toolserver#PHP [14:23:45] I have done taking ownership (using command "take") of all those files using putty and have verified it through winSCP, but I still got 500 Internal Server Error. Any idea why? [14:23:49] Thanks [14:24:22] kenrick95 hi [14:24:24] let me check [14:26:19] [Thu Jul 11 14:07:45 2013] [alert] [client 10.4.1.89] /data/project/idwiki-gpu/public_html/.htaccess: allow not allowed here [14:26:42] kenrick95 I think there is something wrong in that .htaccess [14:26:45] unfortunately log isn't clear on that [14:27:21] maybe I should delete it, my site can live without the .htaccess though [14:30:12] okay, it was resolved after I deleted that file [14:30:18] Thank you for helping [14:30:21] yw [14:43:53] Pyfisch do you need any help moving the bot's db? [14:44:19] Damianz: how's your moving of cluebot to tools project? :D [14:44:46] petan: It would be great if you could help me moving it [14:44:51] ok [14:45:06] once it's not being used, it shouldn't really take a long time to move it [14:45:14] you just need to tell me which db's you need to move [14:45:19] and from which server [14:48:39] petan: server: bots-bsql01 db: wikidata_bot [14:49:22] HAHA @ password for mysql root on bots project :D [14:49:29] ok [14:50:28] * Pyfisch dont understands [14:50:39] the password for root is just funny [14:50:51] unfortunatelly I can't tell you that yet [14:54:15] Pyfisch ok, and what is tool-user you want to access it as [14:55:14] petan: um I have no tools account yet for this [14:55:30] you should create one first :P [14:55:49] when you need to backup stuff somewhere you need to get a backup storage first [14:56:00] petan: it needs to be a service account I guess? not my normal acc [14:56:05] yes [15:02:32] petan: local-fischbot is the group [15:02:52] right [15:03:00] can I start it now? [15:03:05] is bot currently down? [15:03:19] petan: bot is down [15:03:24] right [15:03:27] you can start [15:17:06] [bz] (8NEW - created by: 2Chris McMahon, priority: 4Unprioritized - 6major) [Bug 50622] Special:NewPagesFeed error - https://bugzilla.wikimedia.org/show_bug.cgi?id=50622 [15:17:41] ok Pyfisch it's there [15:17:49] tools-db [15:17:57] petan: thx [15:17:59] just use sql local [15:18:32] petan: one more question: how to move all my files from bots to my computer? [15:18:41] use scp [15:19:19] but you definitely need to use tunneling for that :P [15:19:32] OR [15:19:40] if they are not private, tarball them and move them to public_html [15:24:19] petan: Not started [15:24:29] meh [15:24:37] do u need some help w that? [15:25:04] Mostly need to decide how it's going to work in regards to auto discovery as it's multiple parts [15:25:14] Just busy at work currently [15:25:17] ok [15:25:23] basically you can keep the bot where it is [15:25:33] but database should be probably moved... meh [15:25:42] but the tools-db isn't accessible from bots [15:25:44] damn [15:25:48] well, you need to move the bot as well [15:25:54] no problem [15:25:57] we can do that later then [15:26:19] you can actually do this part by part [15:26:31] first move the bot, it can still access bots-bsql01 from tools [15:27:10] It needs to be able to access more than just sql [15:27:19] There's at least 2 other ports it uses [15:27:29] ah [15:28:02] Damianz given how specific your bot is, maybe we should keep it on own instance for now, maybe we can find a way to access other db servers from bots [15:28:11] I can't move the web interface because tools doesn't support domains afaik, can't move core because I need some form of discovery of where it's running, can't move relay before need discovery... [15:28:18] Moving to tools is doable.. but not straight forward [15:28:24] it's not really problem for it to be on own instance but database servers on bots needs to go out [15:28:57] Ryan has no problem with keeping bots with few small instances [15:29:04] but db servers are huge [15:29:14] Pyfisch what are the processes on bnr3 [15:29:41] Pyfisch I was thinking your bot is off? [15:29:47] I suppose I could have a job to grab the grid info for processes and save it to a file someone... but I kinda need more real time than that [15:29:58] petan: I was also thinking that :-O [15:30:10] Damianz: again, you don't need to move the bot... we just need to sort out the database :P [15:30:32] But the web interface needs the db and the relay and the bot needs the db and the relay and the core [15:30:45] petan: killed it ;-) [15:30:49] ok [15:30:53] Long term I'd like to move it all, but it's intertwined - probably going to have a copy run in test mode with no changes for a while [15:31:35] Damianz ok you can do whatever you want, but we really need to plan some database switch soonish [15:32:52] I'll think about the easiest way... but I'm going out for dinner first as it's been a hell of a two weeks [15:33:12] :0 [15:33:13] k [15:33:16] andrewbogott ping [15:33:22] andrewbogott: you still need these instances in bots? [15:33:35] @labs-project-instances bots [15:33:35] Following instances are in this project: bots-cb, bots-sql2, bots-4, bots-labs, bots-dev, bots-salebot, bots-apache01, bots-analytics, bots-bnr1, bots-bsql01, bots-secure, bots-gs, bots-bnr3, bots-login, bots-dev1, [15:33:41] bots-analytics etc [15:42:34] Coren: Hi [15:42:51] Krinkle: Hoy. [15:43:18] Coren: There's a bot I run in labs right now (http://tools.wmflabs.org/visualeditor/dirtydiffs/) and I'd like it to run every 12 hours [15:43:40] The script needs .bashrc (for custom PATH) and writes to the home directory. How should I run/schedule it? [15:44:07] Krinkle: cron is your friend. Do you know how to fiddle your crontab? [15:44:20] Sure [15:44:25] But shouldn't this use the grid? [15:44:48] Currently I run it from tools-login (takes about an hour to run) [15:44:52] Krinkle: It should; which is why what should be in your crontab is a jsub of your job, and not the job itself. :-) [15:45:07] Ah, you haven't pushed it to the grid at all yet. [15:45:25] yeah, I wasn't sure how to handle it for this case yet. [15:45:57] The very easiest way is to have a shell script that does everything needed to set the environment for the job, and start /that/ with jsub. [15:47:18] In which case your crontab invokation is going to be as trivial as "jsub -once your-shell-script" [15:47:34] There is a shell script around it (home=/data/project/visualeditor; /src/ve-dirtydiffbot/scripts/run.sh ) however it doesn't read .bashrc itself, I was hoping that wouldn't be needed since the script is generic and in git. [15:47:45] do grid jobs not start with the same enviroment? [15:48:09] (actually, bash_profile, not bashrc) [15:48:38] Krinkle: They start with a reduced environment owing to the scheduler not being able to know that the exec nodes are, fundamentally, the same as the submit host. [15:49:27] Coren: Interesting, I didn't know about "jsub -once" yet. Can I also give it name that will cause the jsub to be rejected if that name is already used? (just in case the script runs longer than the cron interval) [15:49:29] Krinkle: Which is why it's generally a better idea to not rely on the environment in a job beyond "PATH includes at least /bin and /usr/bin and HOME is set to the right place" [15:49:53] Krinkle: You can make the job name explicit with -N somename; that defaults to the name of the script sans extension. [15:49:57] Coren: OK, in that case I"ll add a shell script to $HOME that reads bash_profile and then executes the script [15:50:03] Krinkle you can wrap that script to another script [15:50:10] so that you don't need to change it in .git [15:50:11] :o [15:50:15] yeah [15:50:16] * git [15:50:59] Krinkle: Efficiency hint: Have your "wrapper" be just ". ~/.bash_profile && exec the_other_script" [15:51:06] Coren: So using cron is fine? Toolserver had its own thing (cronie?) that would supposedly be more loose about the time in favour of interval (e.g. not everything has to be at midnight, just once a day) [15:51:36] Coren: I'd like that, I tend to avoid that due to bad experience with multiple programs in 1 argument being interpreted incorrectly [15:51:44] cron is fine, though I generally recommend picking a 'random' minute rather than 0 if you don't want to get stuck in a traffic jam. :-) [15:51:53] Alrighty [15:52:28] petan: Nope, those bots run in tools now, you can delete those instances. [15:52:29] Thanks! [15:52:36] ok [15:52:43] andrewbogott so which all I can delete? [15:52:46] even -dev? [15:52:53] @labs-project-instances bots [15:52:53] Following instances are in this project: bots-cb, bots-sql2, bots-4, bots-labs, bots-dev, bots-salebot, bots-apache01, bots-analytics, bots-bnr1, bots-bsql01, bots-secure, bots-gs, bots-login, bots-dev1, [15:53:16] petan, I'll do it, stay tuned [15:53:33] andrewbogott: you can delete dev* and analytics [15:53:40] I don't need these 3 [15:53:50] others should stay for now [15:54:02] 'k [15:55:12] Coren: Where can I find complete jsub doc? jsub --help/man jsub don't include -N for example for job name. [15:55:45] Krinkle: Ah, true, the jsub manual presumes you are familiar with qsub whose options it accepts. [21:18:39] I am running http://meta.wikimedia.org/wiki/WM-Bot version wikimedia bot v. 1.10.16.0 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 [21:18:42] http://bots.wmflabs.org/~wm-bot/dump/%23wikimedia-labs.htm [21:20:31] @optools-permanent-on [21:20:31] Operator tools are now in permanent mode on this channel [21:20:33] Laggy... [21:20:37] LOL [21:20:55] I installed it on new server and now it eats 800mb of ram instead of 300 wtf [21:21:22] this thing will never live on grid... [21:21:31] it eats 90mb res 800 mb virt [21:21:31] Might want to figure that one out. [21:21:47] @jb blabla [21:21:57] seems fast now [21:22:00] @jb blabla [21:22:02] @unjb blabla [21:22:09] !ping [21:22:09] pong [21:22:10] pong [21:22:31] May I have root access and a big shiny restart button on meta:wm-bot? [21:22:38] :D [21:22:56] I will consider [21:23:20] !myaccess [21:23:21] Technical_13!~T13@wikipedia/Technical-13 has the access level: Superuser [21:24:47] Not sure why Helpmebot is in here though.. [21:25:44] retsreklawts: why was Helpmebot put in here? Curious... :) [21:26:00] you can always @kick Helpmebot [21:26:01] :P [21:26:06] @kick Helpmebot [21:26:06]