[00:02:19] giftpflanze: Okay, das max-age habe ich korrigiert und deployed, bei dem Rest, wenn es momentan funktioniert, fehlt mir Wissen und Zeit. [00:04:09] scfc_de: So, if I use scp where does that file go? [00:07:19] Carlos_Danger: Well, where are you aiming for? :-) What are you using as the second parameter to scp? [02:22:35] how does one pass a output redirect to jsub? [02:56:49] Betacommand: What do you mean? "jsub 'somecommand | somecommand"? That won't work. In most cases, you will have to use a wrapper script. [02:58:17] scfc_de: I use > to save the output in a file [03:23:10] Betacommand: Then you'll have to use a wrapper script (unless you want to capture the whole output of the job because - of course - it's saved to ~/*.out and .err). [06:59:52] !log tools merged a lint patch to the gridengine module. Should be a noop [07:02:00] !log tools merged a lint patch to the gridengine module. Should be a noop [07:02:20] labs-morebots, what gives? [07:02:20] I am a logbot running on tools-exec-07. [07:02:20] Messages are logged to wikitech.wikimedia.org/wiki/Server_Admin_Log. [07:02:20] To log a message, type !log . [07:02:36] grrr [08:27:43] Hi guys, sorry for re-posting: someone knows why I can login correctly on http://wikitech.wikimedia.org/ but not anymore on gerrit? [08:28:49] When you say 'log in' on gerrit… is that a thing you normally do? Via ssh? [08:28:56] Or are you talking about not being able to push things? [08:34:28] ilmerovingio: ? [08:42:10] andrewbogott: login on the web interface [08:43:59] in the last week i did some code review for puppet modules but now i can't login anymore [08:44:46] Ah, I see. Ok, so you aren't even using an ssh key to log in, it's just your username/pass [08:45:40] It may be that gerrit is case-sensitive but wikitech is not? [08:45:48] What is your username? I'll take a look. [08:50:10] ilmerovingio: ? [08:50:55] guido.iaquinti [08:50:57] sorry :) [08:54:02] ilmerovingio, are you loggin in as just 'guido.iaquinti' or as 'guido.iaquinti@gmail.com'? [08:55:26] i tried both [08:55:30] let me check again [08:56:21] yes, it doesn't works. The problem is that I logged in on saturday i think without problem [09:01:45] It seems unlikely that you have a '.' in your username. [09:01:51] I'm looking at the user db now... [09:03:32] Try username with these caps: Guido.iaquinti [09:03:57] yes, i've already tried [09:03:59] also now [09:04:03] Ah, so I see. [09:04:31] OK, lemme look at a log from Saturday... [09:04:37] You mean 8 days ago, right? [09:04:42] Or two days ago? [09:12:39] i think 2 days [09:17:46] OK… there were some network issues between our datacenters, and I think that our account databases may be out of sync. I'm investigating. [09:19:19] andrewbogott: many thanks for your help :) [09:32:48] ilmerovingio: I need to run, but will try to keep you updated. [09:33:06] andrewbogott: don't worry, many thanks for your help [09:42:14] !petan-build [09:42:14] make -j `getconf _NPROCESSORS_ONLN` deb-pkg LOCALVERSION=-custom [12:51:50] !labs [12:51:50] https://labsconsole.wikimedia.org/wiki/$1 [12:51:57] :/ [12:52:18] ther is a doc on wiktech how to use putty with tools? [12:52:40] (i know, putty is evil...) [12:53:03] what do you need to do Steinsplitter [12:53:28] how to config, my configuration dos not work. [12:54:13] "wrong passphrase" [12:54:23] but my passphrase & cert is correct :) [12:56:42] I haven't used putty in like 8 years [12:57:11] :-) [12:57:34] but you have pls a link with the acces data at least (host port etc.) ? [12:57:42] that i do [12:58:17] yourusername@tools-login.wmflabs.org [13:00:22] :D ah, it works. wrong host *g* [13:00:23] thx :) [13:00:37] typo in host :/ lol [13:01:52] good [13:17:54] !putty [13:17:55] official site: http://www.chiark.greenend.org.uk/~sgtatham/putty/ | how to tunnel - http://oldsite.precedence.co.uk/nc/putty.html [13:18:02] !access [13:18:02] https://wikitech.wikimedia.org/wiki/Access#Accessing_public_and_private_instances [13:18:06] Steinsplitter ^ [13:18:22] I don't think putty is evil [13:19:24] well... maby it is the encoding. UTF-8 is the problem. if utf-8 selected i cannot access :P [13:20:00] maby i find som time at home to fix this [13:55:32] ilmerovingio: try now? [14:27:49] hi. It seems that I cannot sudo on either of my instances. I also lost admin menus on wikitech. But, I am still listed as an admin for the project. [14:29:35] in syslog I am getting errors from nslcd on both instances [14:37:30] ChrisJ_WMDE: the ldap was recynced, try now [14:48:38] matanya: still getting errors. nslcd[1111]: [14914e] error writing to client: Broken pipe [14:52:00] I asked in #wikimedia-operations [15:23:47] ChrisJ_WMDE: https://wikitech.wikimedia.org/wiki/LDAP#Client [15:30:00] matanya: thanks, but I cannot debug it because I lost my credentials to sudo [15:31:05] you will need an admin then ChrisJ_WMDE [15:33:15] "Lost credentials?" [15:35:10] i should say admin more often if you replay so fast Coren :) [15:35:36] ChrisJ_WMDE: coren can help you [15:36:12] hi coren [15:36:32] What instance is this, ChrisJ? [15:36:53] The project is scrumbugz, the problem is on all three instances, I cannot sudo [15:38:09] see you [15:41:06] Looking into it. [15:42:13] There doesn't seem to be a sudo policy to get to root in your project. [15:43:27] yes. It disappeared... [15:43:51] You should be able to add it back with the 'Manage Sudo Policies' link in the sidebar. [15:45:11] Do you have access to that section allright? [15:45:47] yes. actually the default policy still exists there [15:46:24] Yeah, but the default allows project member -> project member. Not root for project admins. :-) [15:46:30] Lemme add it for you. [15:47:08] ok thanks [15:48:49] Hm. How odd. [15:49:53] That should do it. Try? [15:50:31] it works! thanks a lot Coren [15:50:44] No worries. [20:16:11] anybody objects if i (try) to reboot nginx-devunt.pmtpa.wmflabs? it's totally unresponsive [20:21:43] coren: nginx-devunt seems to be broken, it cannot create a socket on eth0 (see https://wikitech.wikimedia.org/w/index.php?title=Special:NovaInstance&action=consoleoutput&project=nginx&instanceid=a13c687c-e77d-4f70-8bea-2cef52a8fa3b®ion=pmtpa) [20:21:46] any magic tricks? [20:51:43] drdee: Checking. [20:51:54] ty [20:53:27] drdee: That's IPv6 (which labs doesn't support). [20:54:18] yeah i saw that as well, it's not my instance but don' know what should happen [20:54:20] drdee: And it's just npd whining (you can tweak the config so that it doesn't try); it shouldn't harm anything else. [20:54:27] i can't ssh into that box [20:54:39] i thought it might be related to the eth0 stuff [20:54:46] ssh just hangs [20:54:55] it doesn't close the connection, it does not timeout, it just hangs [20:55:11] Probably gluster being insane again. Yeay gluster. Lemme check. [20:55:42] Hi Coren [20:56:02] drdee: Ah, not even. The instance does seem to be off its rocker. That's after a reboot, you say? [20:56:08] yup [20:56:36] nginx-dev1.pmtpa.wmflabs might have the same problem [20:56:36] I'm on my laptop with a relatively new SSH key which I haven't used with labs before. I added my id_rsa.pub to my openstack preferences on wikitech [20:56:54] explodinate, then rebuild. I can't even get in via root. You /might/ be able to salvage it if you knew what was wrong and puppet runs on reboot, but it's normally not worth the trouble. [20:57:01] But when I try to ssh to tools-login.wmflabs.org it fails with Permission denied (publickey,hostbased)... [20:57:26] Krenair: It may take some time before keys are pushed (minutes though, not hours). [20:57:30] Coren: right, but it's not my instance :) my amazing labs-migration-assistant "detected" the problem [20:57:44] Oh, ah. [20:57:50] Well, clearly it found a problem. :-) [20:57:51] !log started orgcharts instance back up [20:57:51] started is not a valid project. [20:57:57] !log orgcharts started orgcharts instance back up [20:57:58] I've been sitting here for a while... not hours though [20:58:20] matanya, Gloria, andrewbogott_afk ^^ [20:58:41] Apparently it crashed [20:58:47] Oh, nice. [20:59:10] Krenair: I'm seeing a key whose comment reads 'ubuntu laptop', I'm guessing that should be it? [20:59:56] thanks marktraceur [21:00:23] Coren, with my email address, yes [21:00:26] Krenair: You know, I'm not sure how well ssh deals with a space in the key comment. Also, there seems to be a stray blank line in your keys (though that I am pretty sure is harmless) [21:00:35] * Coren looks at logs. [21:00:53] That key has been working with Gerrit, so... [21:01:17] Yeah, likely. What host are you trying to connect to? [21:02:11] Krenair: Also, what is your shell account name? [21:02:17] Wait, it's krenair isn't it? [21:02:20] yes [21:02:32] I see no attempt from you on tools-login. At all. [21:03:07] Woops. [21:03:23] I use a different username on my machine. Wasn't specifying krenair@... [21:03:27] Thanks Coren [21:03:37] D'oh. :-) [21:04:15] Coren, you're a remote contractor for WMF right? [21:05:08] Krenair: Well, I'm a permanent staffer now, but I was. Why? [21:05:33] Coren, oh I'm just sitting in the WMF office at the moment with a few people. was wondering if I might meet you at some point this week [21:05:53] Krenair: Aw, sorry no. I'll be at the office Feb 10 and 11 next. [21:06:54] Also, at FOSDEM if you're going. [21:08:01] Coren: speaking of which, how much overlap is there between FT and indef/permanent and at what conditions can one be either from outside USA [21:09:51] Coren: I would need your helping hands again. Could you please move the database p50380g50769__wvs2ds on s7 to another (fast) storage as you did once before? So I can free up some space on s7. The database is new and -except one table definition- empty. [21:10:01] Nemo_bis: There is no real distinction about being outside the US for status; though non-US staffers get more money rather than employee benefits. [21:10:30] hedonil: Sure, give me a couple minutes? [21:10:47] Coren: 'k. super. [21:11:13] * Gloria NP: "American Pie" by Don McLean from "American Pie" [21:11:17] Nemo_bis: As to the staffer/contractor distinction it's mostly a question of whether you've got a budget earmarked or you're coming out of the general contractor budger. There are probably other subtleties, but HR would know. [21:12:40] That's the source of the distinction, not the distinction itself [21:13:03] But thanks anyway [21:13:33] Coren: Sorry it's not s7 it's on s6 [21:16:27] hedonil: {{done}} [21:16:58] Coren: thx. great. [21:25:38] Nemo_bis: I'm not sure I understand the difference you see, sorry. [21:46:34] Krenair: It seems to be the most common mistake to leave out the username if it is different between your own machine and Labs; I wanted to add a bigger note to [[Help:Access]], but didn't find the right spot. If you have a good idea, be bold! :-) [21:47:34] are usernames public? in that case, it might be an option to get sshd to respond 'hey, we don't know that user here' [21:55:17] valhallasw: I don't know if they can be queried externally. (Inside of Labs, you can always look in /public/keys.) [21:57:04] If someone gives as reason for a Tools access request: "bulk insertion" - why do I find that slightly troubling? :-) [23:33:27] Coren: are you around?