[00:04:04] Warning: There is 1 user waiting for shell: Fvue (waiting 161 minutes) [00:04:05] Warning: There is 1 user waiting for access to tools project: Merl (waiting 9 minutes) [00:12:52] !screenfix is script /dev/null [00:12:53] This key already exist - remove it, if you want to change it [00:12:57] !screenfix [00:12:58] script /dev/null [00:12:59] [= [00:17:33] Warning: There is 1 user waiting for shell: Fvue (waiting 175 minutes) [00:17:34] Warning: There is 1 user waiting for access to tools project: Merl (waiting 22 minutes) [00:31:01] Warning: There is 1 user waiting for shell: Fvue (waiting 188 minutes) [00:31:02] Warning: There is 1 user waiting for access to tools project: Merl (waiting 35 minutes) [00:44:30] Warning: There is 1 user waiting for shell: Fvue (waiting 201 minutes) [00:44:31] Warning: There is 1 user waiting for access to tools project: Merl (waiting 49 minutes) [00:44:51] anyone else having an odd "/bin/sh: 1: jsub: not found [00:44:51] " error when trying to run a tool via a cron job [00:45:36] http://pastebin.com/qUGtppu0 [00:48:37] JasonDC: You need to set PATH in your crontab or use /usr/local/bin/jsub (cf. https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help#Scheduling_jobs_at_regular_intervals). [00:57:55] Warning: There is 1 user waiting for shell: Fvue (waiting 215 minutes) [00:57:56] Warning: There is 1 user waiting for access to tools project: Merl (waiting 62 minutes) [01:00:10] alright, thanks [01:02:40] I am trying to get ~/public_html/index.cgi to execute as a cgi from the web directory of my service group. Can I configure that from my own .htaccess? [01:02:51] i.e. is it disabled somehow? [01:07:31] because of course I have a 500 error which I can't diagnose [01:10:34] yeah- a minimal counterexample shows me that it has something to do with having the cgi files in the main directory [01:11:06] http://tools.wmflabs.org/enwp10/cgi-bin/x.cgi versus http://tools.wmflabs.org/enwp10/x.cgi [01:11:24] Warning: There is 1 user waiting for shell: Fvue (waiting 228 minutes) [01:11:25] Warning: There is 1 user waiting for access to tools project: Merl (waiting 76 minutes) [01:17:46] carl-cbm: I'm not sure; I think there is an "ExecCGI" directive but you can't use it from .htaccess. So you're probably limited to /cgi-bin. Don't know if you redirect that transparently. [01:19:39] scfc_de: they apparently do have it configured to run php scripts from public_html ? [01:21:30] carl-cbm: Hmmm. You may be able to convince Coren to enable that for *.cgi as well. [01:22:09] I presume php has an exec() command [01:22:27] carl-cbm: You want to call a CGI from PHP? Sure, why not. [01:22:37] just bootstrap it I mean [01:24:17] carl-cbm: Sure. Is it just aesthetics avoiding /cgi-bin? [01:24:54] Warning: There is 1 user waiting for shell: Fvue (waiting 242 minutes) [01:24:55] Warning: There is 1 user waiting for access to tools project: Merl (waiting 89 minutes) [01:26:56] scfc_de: well, otherwise I need to do directory rewriting, since the page I'm looking at is index.cgi [01:28:09] carl-cbm: Okay. BTW, https://jira.toolserver.org/browse/TS-1652 can be closed? [01:30:44] scfc_de: I can't find the button to close it :( [01:31:41] carl-cbm: You have to login, (and then fast, as JIRA will log you out very soon otherwise :-)), then "Assign To Me", then "Resolve this issue" (or similar). [01:32:54] ok, it is resolved now [01:34:49] carl-cbm: Thanks. [01:37:45] carl-cbm: The "--where" idea is nice lateral thinking :-). [01:38:14] Warning: There is 1 user waiting for shell: Fvue (waiting 255 minutes) [01:38:15] Warning: There is 1 user waiting for access to tools project: Merl (waiting 103 minutes) [01:39:34] thanks. I will have to bug Coren about the cgi. It will work with index.php, but php doesn't seem to have a clean exec() function. It will not work with index.py even though the docs claim that .py scripts will run inside public_html [01:41:17] if I just take a working .py fiel and rename it index.py, it stops working [01:42:40] carl-cbm: Sorry, never used Python for web stuff. And I believe petan and Coren as the only people able to look into the logs are sleeping :-). [01:49:02] I think it is just a combination of different apache settings, I can bug Coren about it some time. The rest of the scripts seem to run fine out of cgi-bin [01:51:43] Warning: There is 1 user waiting for shell: Fvue (waiting 269 minutes) [01:51:44] Warning: There is 1 user waiting for access to tools project: Merl (waiting 116 minutes) [01:53:18] scfc_de: are you a volunteer on the wmf labs, or one of the staff? [01:53:48] carl-cbm: Volunteer. [01:53:59] carl-cbm: And you? :-) [01:54:16] thanks for the help. I'm not really involved with the labs apart from migrating things off of toolserver [01:55:46] carl-cbm: I basically just make my own stuff work as well :-). [02:05:09] Warning: There is 1 user waiting for shell: Fvue (waiting 282 minutes) [02:05:10] Warning: There is 1 user waiting for access to tools project: Merl (waiting 130 minutes) [02:18:34] Warning: There is 1 user waiting for shell: Fvue (waiting 296 minutes) [02:18:35] Warning: There is 1 user waiting for access to tools project: Merl (waiting 143 minutes) [02:31:54] Warning: There is 1 user waiting for shell: Fvue (waiting 309 minutes) [02:31:55] Warning: There is 1 user waiting for access to tools project: Merl (waiting 156 minutes) [02:45:15] Warning: There is 1 user waiting for shell: Fvue (waiting 322 minutes) [02:45:16] Warning: There is 1 user waiting for access to tools project: Merl (waiting 170 minutes) [02:58:35] Warning: There is 1 user waiting for shell: Fvue (waiting 336 minutes) [02:58:36] Warning: There is 1 user waiting for access to tools project: Merl (waiting 183 minutes) [03:12:00] Warning: There is 1 user waiting for shell: Fvue (waiting 349 minutes) [03:12:01] Warning: There is 1 user waiting for access to tools project: Merl (waiting 196 minutes) [03:18:52] it looks like the SGE setup is going to need some tweaking [03:20:33] carl-cbm: Why? [03:21:14] qstat -f -u '*' [03:22:37] At the moment, it's just memory, and addshore's task don't last for long, I believe. [03:23:06] (Though hack-xyz isn't a very good description :-).) [03:23:27] most of them were started 3 hours ago - about 175 jobs [03:23:55] a job that is supposed to run once per hour would have no chance [03:24:40] I think that's all this interwiki/wikidata stuff, so hopefully there will be an end sometime :-). [03:25:01] it would be easy enough to tweak the max slots-per-user on sge [03:25:28] Warning: There is 1 user waiting for shell: Fvue (waiting 362 minutes) [03:25:29] Warning: There is 1 user waiting for access to tools project: Merl (waiting 210 minutes) [03:26:26] Probably much nicer to just set up tools-exec-05 :-). WMF isn't really short on hardware. [03:26:50] indeed [03:31:14] carl-cbm: But you're right, you can see the sharp increase in . [03:35:47] I did not know about the ganglia server, thanks. It's good to know this is a new problem that I happened to run into [03:38:43] are the db servers in the ganglia? [03:38:49] Warning: There is 1 user waiting for shell: Fvue (waiting 376 minutes) [03:38:50] Warning: There is 1 user waiting for access to tools project: Merl (waiting 223 minutes) [03:52:10] Warning: There is 1 user waiting for shell: Fvue (waiting 389 minutes) [03:52:11] Warning: There is 1 user waiting for access to tools project: Merl (waiting 237 minutes) [03:56:53] carl-cbm: I haven't them yet. [03:57:33] I was wondering mainly because the load on the exec servers looked low for the number of jobs, so the slowdown must be somewhere else [03:59:44] It's memory; do a "qstat -j $JOBNUMBER" for a job of yours that's pending. [04:00:51] sure, that is what is keeping new jobs from running. I was talking about what is keeping the running jobs from finishing [04:04:05] addshore's programming skills?! :-) When I last saw one of the interwiki/wikidata scripts, it looked optimizable. So probably just a huge task. [04:05:30] Warning: There is 1 user waiting for shell: Fvue (waiting 402 minutes) [04:05:31] Warning: There is 1 user waiting for access to tools project: Merl (waiting 250 minutes) [04:18:51] Warning: There is 1 user waiting for shell: Fvue (waiting 416 minutes) [04:18:52] Warning: There is 1 user waiting for access to tools project: Merl (waiting 263 minutes) [04:32:13] Warning: There is 1 user waiting for shell: Fvue (waiting 429 minutes) [04:32:14] Warning: There is 1 user waiting for access to tools project: Merl (waiting 277 minutes) [04:45:32] Warning: There is 1 user waiting for shell: Fvue (waiting 443 minutes) [04:45:33] Warning: There is 1 user waiting for access to tools project: Merl (waiting 290 minutes) [04:58:54] Warning: There is 1 user waiting for shell: Fvue (waiting 456 minutes) [04:58:55] Warning: There is 1 user waiting for access to tools project: Merl (waiting 303 minutes) [05:12:15] Warning: There is 1 user waiting for shell: Fvue (waiting 469 minutes) [05:12:16] Warning: There is 1 user waiting for access to tools project: Merl (waiting 317 minutes) [05:14:49] is there some similar thing like https://wiki.toolserver.org/view/Toolserver_database available on labs database? [05:25:39] Warning: There is 1 user waiting for shell: Fvue (waiting 483 minutes) [05:25:40] Warning: There is 1 user waiting for access to tools project: Merl (waiting 330 minutes) [05:39:05] Warning: There is 1 user waiting for shell: Fvue (waiting 496 minutes) [05:39:06] Warning: There is 1 user waiting for access to tools project: Merl (waiting 344 minutes) [05:52:25] Warning: There is 1 user waiting for shell: Fvue (waiting 509 minutes) [05:52:26] Warning: There is 1 user waiting for access to tools project: Merl (waiting 357 minutes) [06:05:46] Warning: There is 1 user waiting for shell: Fvue (waiting 523 minutes) [06:05:47] Warning: There is 1 user waiting for access to tools project: Merl (waiting 370 minutes) [06:19:12] Warning: There is 1 user waiting for shell: Fvue (waiting 536 minutes) [06:19:13] Warning: There is 1 user waiting for access to tools project: Merl (waiting 384 minutes) [06:32:36] Warning: There is 1 user waiting for shell: Fvue (waiting 550 minutes) [06:32:37] Warning: There is 1 user waiting for access to tools project: Merl (waiting 397 minutes) [06:46:01] Warning: There is 1 user waiting for shell: Fvue (waiting 563 minutes) [06:46:02] Warning: There is 1 user waiting for access to tools project: Merl (waiting 410 minutes) [06:59:31] Warning: There is 1 user waiting for shell: Fvue (waiting 576 minutes) [06:59:32] Warning: There is 1 user waiting for access to tools project: Merl (waiting 424 minutes) [07:12:56] Warning: There is 1 user waiting for shell: Fvue (waiting 590 minutes) [07:12:57] Warning: There is 1 user waiting for access to tools project: Merl (waiting 437 minutes) [07:26:21] Warning: There is 1 user waiting for shell: Fvue (waiting 603 minutes) [07:26:22] Warning: There is 1 user waiting for access to tools project: Merl (waiting 451 minutes) [07:29:56] !screenfix [07:29:56] script /dev/null [07:39:49] Warning: There is 1 user waiting for access to tools project: Merl (waiting 464 minutes) [07:53:23] Warning: There are 2 users waiting for access to tools project, displaying last 2: Merl (waiting 478 minutes) Sumanah-testing (waiting 12 minutes) [08:06:52] Warning: There are 2 users waiting for access to tools project, displaying last 2: Merl (waiting 491 minutes) Sumanah-testing (waiting 25 minutes) [08:15:51] !tr Merl [08:15:52] request page: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Access_Request/Merl?action=edit talk page: https://wikitech.wikimedia.org/wiki/User_talk:Merl?action=edit§ion=new&preload=Template:ToolsGranted link: https://wikitech.wikimedia.org/w/index.php?title=Special:NovaProject&action=addmember&projectname=tools [08:16:39] !tr Sumanah-testing [08:16:39] request page: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Access_Request/Sumanah-testing?action=edit talk page: https://wikitech.wikimedia.org/wiki/User_talk:Sumanah-testing?action=edit§ion=new&preload=Template:ToolsGranted link: https://wikitech.wikimedia.org/w/index.php?title=Special:NovaProject&action=addmember&projectname=tools [08:16:50] @seen sumana [08:16:51] petan: I have never seen sumana [08:16:56] @seen sumanah [08:16:56] petan: Last time I saw sumanah they were quitting the network with reason: Quit: Leaving N/A at 5/22/2013 1:56:33 PM (1.18:20:23.4102570 ago) [08:17:06] hmmm [08:17:16] is there thought of having a centralized logging server? [08:17:19] for tools labs at least? [08:17:52] unified logging from multiple processes / jobs seems hard to do right now... [08:18:24] it could also be easily used to find out status of tools [08:18:27] thoughts, petan? [08:18:32] @seen Moogsi [08:18:33] YuviPanda: Last time I saw Moogsi they were quitting the network with reason: Ping timeout: 248 seconds N/A at 5/18/2013 4:51:49 PM (5.15:26:43.1745520 ago) [08:18:51] YuviPanda logging of what [08:19:02] of log output from tools? [08:19:06] oh [08:19:10] in my case, i've the GitHub -> Gerrit Bot [08:19:11] well, that is a good idea [08:19:14] there's a web part, the 'receiver' [08:19:25] that recevies web hook pushes, and then just starts a job [08:19:33] now the log for the receiver and the job will have to be different [08:19:35] and hence complicated [08:19:42] if we implement something like a central scribe server or something [08:19:52] we can also expose nice UI for non-tech people to gawk at logs :) [08:20:09] hmm [08:20:22] Warning: There is 1 user waiting for access to tools project: Mauro742 (waiting 10 minutes) [08:20:48] petan: it'll also help with 'is this tool up?' questions [08:20:57] so how would you design it? how the logs would be recorded from a tool? [08:20:59] also if logging is easy and nice more people will use logging [08:21:25] petan: something like https://github.com/facebook/scribe [08:21:33] we shouldn't reinvent the wheel, there are plenty of these around [08:21:58] scribe is not the only thing there are plenty of other options too [08:22:03] thx petan [08:22:10] yw [08:23:47] hey chrismcmahon [08:24:12] lol addshore [08:24:38] are you trying to kill the cluster? [08:26:06] Coren|Sleep I am going to create some more exec nodes [08:26:47] petan: is this (logging server) something tools labs would be interested in? in the future, at least... [08:27:01] ? [08:28:51] !log tools created 2 more exec nodes, setting up now... [08:28:52] Logged the message, Master [08:29:02] YuviPanda I think yes [08:29:15] alright, powercut in about 3 minutes [08:29:16] :) [08:29:20] :P [08:31:21] addshore is there any place near that table you are at [08:31:35] oh dammit, you are all at the hackathon [08:31:35] grr [08:33:29] YuviPanda: You didn't come? [08:33:35] anomie: exams :( [08:33:47] worst part is that when the schedule was announced, I have a 10 day window right during the hackathon [08:33:57] but of course I wouldn't know that when this was all being planned [08:34:41] now it's time for my daily power cut [08:34:42] bye [08:36:27] Coren|Sleep I think we should create a bigger swap because now the instances are 100% when it comes to operating memory but only up to 3gb of real memory is used on them [08:37:16] petan: When the new exec nodes up and running, let me know and I'll update http://tools.wmflabs.org/anomiebot/packages.html [08:37:58] ok [08:38:25] anomie are these in puppet or not? [08:38:32] if not it makes sense they aren't on all of them [08:38:35] Well, if I can get a job through addshore's 200-some [08:38:58] petan: The packages? No idea. I just came up with a way to find the inconsistencies. [08:39:06] ok [08:39:58] petan: is there any possibility to tell grid engine on tool lab to not execute two jobs at the same time? (like user_slot on ts) [08:40:14] anomie: The queue is free :) [08:41:00] addshore: Oh, good! What was all that, anyway? [08:41:20] trying to clear up a 700,000 article backlog as quickly as possible :) [08:41:49] only 250,000 left now ;p [08:43:08] !log setting up lvm on new exec nodes because it is more flexible and allows us to change the size of volumes on the fly [08:43:09] setting is not a valid project. [08:43:12] addshore: if you run the same job mutiple times you can use one job having multiple taks [08:43:46] how? :) [08:44:10] !log tools setting up lvm on new exec nodes because it is more flexible and allows us to change the size of volumes on the fly [08:44:11] Logged the message, Master [08:46:23] addshore: -t 1-500 then environment variable $SGE_TASK_ID contains the task number if you have to know it in your script [08:47:19] hmmm, currently I need to submit a seperate parameter with each task :/ [08:48:57] you can create one file and use the tasknumber as line number to grab the needed parameter [08:50:23] addshore: e.g. http://wiki.gridengine.info/wiki/index.php/Simple-Job-Array-Howto#Pulling_data_from_the_ith_line_of_a_file [08:50:27] [= [08:50:36] will check it out ina second :) [08:51:22] now they're just messing with me. scheduled power cut didn't happen today. grr [08:52:05] or..... it just hasn't happened yet ;p [08:55:38] Coren|Sleep the new boxes aren't same as previous, they don't have so huge /tmp but the fs is living on lvm so if we needed to extend swap, or /tmp or any other fs, we can do that online :P [08:56:55] erm [08:58:58] addshore: yes, that's the frustrating part [08:59:16] also how can I programmatically retreive job id of something that I just submitted? [08:59:32] currently second line 3rd word of the output of jsub is jobid [08:59:36] but i don't know if I can rely on that [08:59:49] can I, Coren|Sleep / petan / anyone else? [09:00:05] hi addshore [09:00:05] YuviPanda: I think that's pretty safe - that's just the output of qsub [09:00:27] what's your way to move tool/bot from bots to tools? [09:00:33] including 'prog?: /bin/echo'? [09:00:42] (I was testing jsub with an echo) [09:03:05] Yuvi: in a min, at hackaton [09:04:12] * YuviPanda hacks Coren a ton [09:05:37] hey liangent ! [09:06:34] coren where? :P [09:06:59] petan: Where are _you_? [09:07:37] addshore: what's your way to move tool/bot from bots to tools? [09:08:08] liangent: what sort of tool? [09:08:18] (what des it need)? [09:08:55] *does [09:09:07] * Coren is sitting at the bottom, centet, behind the speaker. Brion is about 3 feet from me at the end of the table. [09:09:26] whee :D [09:09:34] addshore: primarily files [09:09:39] maybe databases too [09:09:50] replicated dbs? :) [09:09:56] or just a userdb? :) [09:10:20] userdb [09:10:25] lovely :) [09:10:36] btw ssh from bots-login to tools-login seems impossible [09:10:51] thats correct :P [09:11:01] it seems we need some host-based authentication [09:11:11] liangent: Something odd on the toolserver side we havn't tracked down. It works the other way 'round though. [09:12:11] hmmm, liangent actually this works for me [09:12:18] are you forwarding your ssh key? [09:12:39] addshore: You can ssh from the labs to the toolserver? [09:12:46] addshore: note that forwarding your ssh key means you effectively share it with all roots [09:12:52] oh, to the toolserver? Coren Ill check [09:13:00] it's better to create a specific key for internal labs use [09:13:05] Coren: there was some talk yesterday where this thing was mentioned [09:13:14] addshore: Didn't work two day ago, something at the TS side is blocking traffic from labs [09:13:28] Coren: time tweaked to use utc: (2:53:12 PM) Damianz: SSH from labs to production is restricted, for some reason the toolserver ip range is in the range blocked [09:14:16] I was taking about ssh from bots-login to tools-login [09:14:24] Aha. So it'd be on our end? I need to talk to Leslie about it. I'll track her down, she's around here [09:15:06] addshore: ProxyCommand ssh -a -W %h:%p bastion1.pmtpa.wmflabs [09:15:14] so I guess I had it disabled... [09:16:57] addshore: I often see cautions telling me not to enable forwarding [09:17:00] is it safe here? [09:17:01] Coren: Should one of us get up and invite to the workshops? [09:17:12] Coren: ... and would you do that? [09:17:39] liangent: basically: do you trust the roots to your private ssh key [09:18:05] valhallasw: as my ssh key is labs / wikimedia specific I see no real reason not to [09:18:51] mine is not [09:19:08] make a labs specific one ? ;p [09:20:26] addshore: I've never tried using two keypairs simultaneously [09:20:38] any guide about it? [09:21:08] liangent: https://wikitech.wikimedia.org/wiki/User:Magnus_Manske/Migrating_from_toolserver under connecting to different servers [09:25:39] Warning: There is 1 user waiting for access to tools project: Hydriz (waiting 0 minutes) [09:27:17] Coren is there a free place near you with power? [09:27:31] Seat between me and Brion is empty [09:29:05] Coren: So according to platonides's email on toolserver-l, there's db access on labs now? [09:29:18] ye [09:29:19] legoktm: yes =] [09:29:22] ooh [09:29:25] how do i get access?? [09:29:26] legoktm: I've been travelling so I haven't yet caught up on email; but yes [09:29:29] althoguht I dont think all langs dbs are availible yet [09:29:37] @search wikitech [09:29:38] Results (Found 10): pxe, wikitech, mobile-cache, tooldocs, sal, rq, tr, proxy, toolsrequest, replicateddb, [09:29:48] !replicateddb | legoktm see this [09:29:48] legoktm see this: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help#Database_access [09:29:51] legoktm: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help#Production_replicas [09:29:57] :O [09:31:03] * legoktm hugs petan addshore and wm-bot [09:31:37] xD [09:31:44] we are all so helpfull ;p [09:34:00] any chance we can get metawiki added to the list? [09:34:09] that would be helpful for cross-wiki stuff [09:34:12] ^^ Coren [09:35:02] addshore: valhallasw: is this config file correct http://pastebin.com/GV2ebAyW to use a specialized keypair for labs http://pastebin.com/PV7sZFkh ? [09:35:07] the latter [09:35:23] actually the only difference is -a and 0A [09:35:24] -A [09:35:42] Im afraid I cant help you as I am on windows and use pagent :/ [09:35:45] ugh, I never proxy, I always just ssh in and then ssh further [09:35:52] Users can create databases with names starting with username__ (where username is your credentials' user, which can be found in your ~/replica.my.cnf followed by two underscores) <-- Can I add things after the two underscores? and why 2 underscores? [09:36:18] two underscores? [09:37:00] https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help#Creating_databases | addshore [09:37:09] I havn't read that bit yet ;p [09:37:35] interesting with two underscores :O [09:38:02] legoktm: You can, and two underscores to make the namespace unique (no username can have them) [09:38:16] ah, ok. [09:38:38] thanks :) [09:38:59] now i need to make my scripts TS and labs compatible! [09:39:06] :D [09:39:08] Warning: There is 1 user waiting for access to tools project: Hydriz (waiting 13 minutes) [09:39:26] it seems removal of pubkeys from wikitech doesn't take effect instantly [09:39:31] this sounds bad [09:39:44] liangent: nfs is not instant [09:39:53] !tr Hydriz [09:39:53] request page: https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Access_Request/Hydriz?action=edit talk page: https://wikitech.wikimedia.org/wiki/User_talk:Hydriz?action=edit§ion=new&preload=Template:ToolsGranted link: https://wikitech.wikimedia.org/w/index.php?title=Special:NovaProject&action=addmember&projectname=tools [09:40:10] It's not NFS that puts the delay; the scripts that update the keys rune IIRC every 60s [09:40:18] run* [09:40:20] ah, ok [09:46:25] laz rulz ;p [09:46:33] *labz >.< [09:48:08] addshore: Coren rules ;-) [09:48:39] Dunno about "rules", but I work hard on that tool labs. :-) [09:49:01] * addshore hopes Coren doesn't ask questions about the spike here >> http://ganglia.wmflabs.org/latest/stacked.php?m=load_one&c=tools&r=day&st=1369388912&host_regex= [09:49:09] hehe [09:49:22] I think he is already aware of that hence the explanation why 2 new nodes [09:49:30] hehe :P [09:50:24] Yeah, I [09:50:31] I'll help finish their setup today [09:54:03] Did anyone see what limit I hit when I was running all of my jobs earlier? [09:54:25] Memory or slots? I didn't check at the time... [09:54:33] memory limit [09:54:36] ahhh [09:55:00] wait a moment I am not entirely sure [09:55:08] hmmm, lets see :) [09:55:11] it might be slots because there were exactly 64 jobs on each box [09:57:07] ahhh, then my next question was, should we introduce a limited ammount of jobs per user/project? [09:57:21] addshore: http://pastebin.com/RV4QMGz1 this is my final version, and hopping between bots and tools works [09:57:28] this was people like me couldnt flood the queue and cause other people not to submit jobs.. [09:57:38] it would just be my jobs in qw and there would still be space on execs? [09:58:12] I think we should set up the cluster large enough to handle all of jobs [09:58:19] hmmm [09:58:51] !screenfix [09:58:52] script /dev/null [09:59:10] I guess now I can just scp files [10:10:54] addshore the boxes are still not in cluster :P yet you are attempting to kill it again :D [10:11:13] but still a lot of free space [10:11:17] hah, im starting less processes this time :) [10:11:30] I just want to clear this silly db backlog [10:11:35] np [10:12:14] if it goes wrong I can just qdel -u local-addshore-dev them all ;p [10:12:24] hence the -dev project ;p [10:12:28] aha [10:12:29] cool [10:12:42] it should stop after around 250 this time [10:25:06] hmm i hit that limit again ;p [10:50:20] All sorted :) just over 100 left running :) [10:53:24] Okay: Workshop on Toolserver migration to Tool Labs: 13.30, "Borneozaal 1" (facing the Hackathon reception, this is in the corridor left). [10:58:54] shall I come or stay here and code? :P [10:59:16] :) [10:59:21] Here is an idea for a missing tool: Rewrite the URLs in all wikis pointing to a migrated tool / to Tool Labs instead of Toolserver. [11:00:56] that is a good plan :) [11:08:18] Silke_WMDE: that tool already exists: MerlLinkBot [11:10:02] http://de.wikipedia.org/wiki/Benutzer:MerlLinkBot [11:17:43] addshore: stats! stats! [11:17:48] /nick addstats [11:19:55] Silke_WMDE how much full is it? if you needed some help I can come [11:20:04] Labs workshop starting RSN [11:20:57] Coren where is it? is it full yet? [11:21:24] petan: I think this one is full if you hadn't already signed up, but we're doing it again later [11:21:33] no prob [11:22:29] Sunday 9:30 I am told [11:24:49] im signed up for thsi one but still eating :/ [11:25:38] Merlissimo: Oh cool, good to know! [11:26:40] petan: Not full yet, but many more registered so I think they'll come. [11:26:56] petan: We are in "Borneozaal 1" [11:29:01] addshore: you were right, the power went out 40 minutes late. [11:29:03] was even more annoying [11:31:59] Warning: There is 1 user waiting for shell: TWillemsen (waiting 0 minutes) [11:33:57] !rq TWillemsen [11:33:58] https://wikitech.wikimedia.org/wiki/Shell_Request/TWillemsen?action=edit https://wikitech.wikimedia.org/wiki/User_talk:TWillemsen?action=edit§ion=new&preload=Template:ShellGranted https://wikitech.wikimedia.org/wiki/Special:UserRights/TWillemsen [11:36:15] bad luck YuviPanda [11:45:24] Warning: There is 1 user waiting for shell: TMg (waiting 4 minutes) [11:45:39] !rq TMg [11:45:39] https://wikitech.wikimedia.org/wiki/Shell_Request/TMg?action=edit https://wikitech.wikimedia.org/wiki/User_talk:TMg?action=edit§ion=new&preload=Template:ShellGranted https://wikitech.wikimedia.org/wiki/Special:UserRights/TMg [11:51:14] is anyone else getting kicked out of ssh like every 10 minutes [11:51:33] petan: the wifi is just a bit flaky, I think. Try using mosh instead of ssh [11:51:48] mosh? [11:52:31] http://mosh.mit.edu/ [11:52:51] (it does predictive local echo to make the lag less irritating) [11:55:10] heh petan, you just got a mention ;p [11:56:04] addshore mention like "there is that idiot with root who ocassionaly removes wm-bot, you need to live with that" or "there is a cool guy petan over there, go and bring him a beer" [11:56:25] petan: "there is a cool guy petan, but I have no clue where he is" [11:56:27] HAHA! [11:56:30] yep [11:56:31] lol [11:56:37] I can come but I was told it's full [11:56:46] there is room [11:56:47] it's not [11:56:50] not even close to full [11:56:56] ok I will try to get there [11:56:58] Coren s intro is being recorded. You'll be able to watch it later on. [11:57:37]