[00:03:11] Withoutaname: Okay, I'm blind :-). What's the word in front of the character? [00:06:16] scfc_de: Hey, do we have a fork of meetbot or is it just a configuration file somewhere? [00:07:41] ok, so go to view-source:https://en.wikivoyage.org/w/index.php?diff=2544908 and find where it says ‎ [00:07:56] in between the span is where the hidden char is [00:08:12] scfc_de: And either way would you mind giving me access? The bug about the topic changes is making me itchy. [00:08:19] marktraceur: It's the OpenStack fork (didn't know what to pick; will revert to the main development branch some time) and a configuration file; see https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Tools/meetbot#Administration for the whole deal. [00:08:22] it'll show up when you paste it into google [00:08:55] marktraceur: Denying someone to take work away from me? :-) What's your wikitech username? [00:09:19] scfc_de: Should be marktraceur [00:09:28] wait no sorry [00:09:34] wrong url lemme get the right one [00:09:49] view-source:https://en.wikivoyage.org/w/index.php?diff=2544908&oldid=2544859 [00:10:22] scfc_de: do you see it in that one? [00:10:36] Firefox recognizes that as a protocol? [00:10:37] Wow. [00:11:21] yeah? [00:11:58] marktraceur: Done. I had a look at how to resolve that bug some time ago, and setting the topic is very confusing in MeetBot (two classes that overwrite some method IIRC). Just so you don't get too frustrated early :-). [00:12:22] huh: I had put a lot of that in between File:A...a.svg.png [00:12:28] surprised its not showing up [00:12:44] Oh dear. [00:12:53] scfc_de: Maybe I'll postpone it, then, if it's not fast [00:13:26] Withoutaname: Ah, now I see the empty ! So I'm not blind after all. [00:14:07] So . [00:14:23] it's causing a lot of problems [00:14:26] especially for commons [00:14:56] for example it resolves to legitimate category titles but movescripts dont recognize it [00:15:08] so a lot of errors pop up [00:15:36] I want to know if Commons has a lot of that in file or category names [00:17:19] (U+200E = https://en.wikipedia.org/wiki/Left-to-right_mark) [00:17:47] Withoutaname: You should be able to search for that in the Labs replica DBs. [00:18:25] ok [00:19:02] I'll probably need to find a script for it though because plain old Special:Search doesn't do it properly [00:20:00] Withoutaname: Do you have an example in Commons to test if a query works? [00:21:30] I'm not sure what you mean but I tried uploading https://test.wikipedia.org/wiki/File:Aa.svg.png to see if it worked [00:21:52] I put the filename in Special:Upload as File:Aa.svg.png [00:24:34] test.wikipedia.org isn't replicated to Labs. So I'm looking for a page/file in "real" Commons that has this character so I can test whether a query that lists all pages/files in Commons works (= returns this page/file). [00:25:27] (I don't want to run a query against millions of pages only /not/ to know afterwards if there were no results because there are no pages with that character or because the query is incorrect :-).) [02:44:54] TimStarling: why does bots.wmflabs.org not resolve from my side? [02:46:02] petan^ [02:47:46] hhm now it does [02:54:38] Jasper_Deng: DNS for labs has been flapping all day -- https://bugzilla.wikimedia.org/show_bug.cgi?id=63709 [03:20:00] Ah, I had the same problem. [03:20:07] Thought it was something on my end. [07:16:21] hello all… is the labs dns issue resolved for everyone? [08:02:12] hillo [08:04:35] * Coren waves at hashar [08:58:06] umm [08:58:37] I rebooted language-replag-master.eqiad.wmflabs and I can no longer log in nor do anything about it on wikitech [09:13:58] is there a problem with labs ? I get Firefox can't find the server at tools.wmflabs.org. This is for http://tools.wmflabs.org/reasonator/?q=Q1352906&lang=en [09:14:27] it is now back, it happened several times in the last four minutes [09:14:32] back and gone [09:22:37] also, I have gotten this two times today: [09:22:38] ssh: Could not resolve hostname bastion2.wmflabs.org: Name or service not known [09:36:02] GerardM, Nikerabbit, there was a dns outage earlier, but I thought it was fixed... [09:36:08] not, I take it? [09:36:10] * andrewbogott kicks pdns [09:42:42] andrewbogott: just occasionally [09:43:46] lemme see if the cache is still too small... [12:26:59] does the login node block irc connections or something? [12:58:18] gry: what u mean [13:24:08] petan, connection to irc times out when I try to run an irc bot on login node as a test. [13:24:40] times out or get lost? [13:25:06] if it just get ping timeout, then you aren't sending ping / responding to ping in regular intervals, which apparently freenode requires [14:01:10] gry: Login node does not block IRC, but freenode wants identd to bypass user limits and only the exec nodes have that. [14:01:24] But also, you shouldn't be using IRC bots from the login nodes [14:03:45] !log project-proxy replaces SSL certificate for the general labs webproxy [14:03:46] Logged the message, Master [14:04:41] Coren: I'll look into running interactive jobs in a bit then, as I apparently need to debug. Thanks. [14:06:55] maybe try qrsh? [14:07:42] oh, no, the rules … [14:17:07] gry: No, using qrsh for debugging is okay. :-) [14:17:26] gry: That's why the message says "you shouldn't be here except if you're debugging" :-) [14:18:24] gry: Just don't leave something running unattended through qrsh [14:19:43] Coren: so not debugging any other process but only the one running is ok? [14:20:58] No, that's okay. It's just not for general work or starting stuff by hand for normal use; but when you're debugging it's perfectly okay. [14:21:51] ah, ok [14:22:37] I just want to avoid people running stuff outside grid control as a rule since that messes up resource allocation. [14:39:37] hi, I've a problem I want to run a code in labs that it retrieve a url in labs but it seems it can't open it from inside, What do you suggest? [14:40:10] I was thinking about a proxy [14:40:29] but I don't have one and I can't trust any open proxy [14:48:55] Amir1_: You can access the URL by the instance's name that is serving it; i. e. for tools.wmflabs.org, use tools-webproxy. [14:49:14] scfc_de: let me check [14:49:18] (If I understood the problem correctly.) [14:50:39] scfc_de: I change this to what http://wikistats.wmflabs.org/api.php?action=dump&table=%s&format=xml ? [14:52:08] Coren, hi [14:52:45] Amir1_: Moment, please. [14:52:55] sure :) [14:53:24] why so complicated? [14:55:18] Amir1_: Usually I would recommend determining the IP for the URL (here: 208.80.155.156), then going to wikitech, listing all instances with their public IPs, and then searching for the IP there. *But*: It isn't listed there. Odd. [14:56:16] so there is option except an open proxy [14:57:10] Amir1_: https://wikitech.wikimedia.org/wiki/Nova_Resource:Wikistats claims to handle that URL, and it has only one instance. Could you try wikistats-live, please? [14:57:46] let me check [14:58:45] It works! [14:58:48] thank you :) [14:59:19] np [15:10:24] Coren, ping [15:12:45] can someone check why language-replag-master is no longer accessible after I rebooted it? [16:55:36] Hi, where can I find the fingerprint of the host public key for: tools-login.wmflabs.org [16:59:14] morray: https://wikitech.wikimedia.org/wiki/Help:SSH_Fingerprints/tools-login.wmflabs.org [16:59:57] thx se4598 [17:28:20] can someone check why language-replag-master is no longer accessible after I rebooted it? [18:11:12] Hi; I need to transfer a MMP from the toolserver to labs. I thought I requested a place to put that a year ago, but no-one acted upon it back then. It seems to be totally gone now. I have a labs account, but there still doesn't seem to be a actual tool made for it. What is the link to request a place to put this mmp? [18:14:03] Sumurai8: Do you mean "Tools" ("Tool Labs") by "labs"? Or do you need a complete Labs project? What is the Toolserver MMP's name? [18:14:50] it's a simple irc bot; I guess it is just a tool (I have an account on tools) [18:16:19] So, I found https://wikitech.wikimedia.org/w/index.php?title=Special:NovaServiceGroup&action=addservicegroup&projectname=tools - where I created a new tool - which now says I don't have access to?! [18:16:34] if I do "become wikilinkbot" [18:16:36] Sumurai8: What is the tool's name? [18:16:46] wikilinkbot @ scfc_de [18:17:05] http://toolserver.org/~wikilinkbot/help.htm [18:17:18] Sumurai8: you need to be logged in to wikitech for that link [18:17:35] Sumurai8: it's a seperate account from your normal WMF account [18:18:08] I am, and I created a group wikilinkbot. It said it succeeded, but when I tried to do "become wikilinkbot" on ssh, it says "sudo: sorry, a password is required to run sudo" [18:18:21] whatever that may mean [18:18:31] I log into ssh with a key, not a password [18:18:36] Sumurai8: log out, then log in again [18:18:38] Sumurai8: Did you create it just now? Then you need to log out and log in again. [18:19:02] (We should put that as a notice in the "New tool created" message.) [18:19:22] Ugh -_-' [18:19:25] Yeah, that worked [18:19:36] scfc_de: yeah, or fix linux to update groups on demand ;-) [18:19:55] I hate configuration stuff with a deep passion :S [18:20:49] scfc_de: apparently running it in a new shell should work, so we could just change become to start a new shell, and *then* run sudo [18:21:36] valhallasw: But doesn't that lose the previous environment & Co.? Well, in most cases you don't need that for sudo anyway ... Hmmm. [18:51:12] Coren: around? [18:57:15] I am noticing that the "Add" and "Remove" maintainer links on http://tools.wmflabs.org/ link nowhere sensible. The correct url for both seems to be https://wikitech.wikimedia.org/w/index.php?title=Special:NovaServiceGroup&action=managemembers&projectname=tools&servicegroupname=local-{project-name} - to whoever can fix that/knows how to fix that [19:04:37] Sumurai8: that would help yes, now you have to lookup your project again [19:18:02] Filed https://bugzilla.wikimedia.org/63741 for that. [19:47:07] die had ik al gevonden akoopal ;) [20:48:21] !ping [20:48:22] !pong [20:48:24] !python del [20:48:24] Successfully removed python [20:48:28] ok [20:50:34] You don't want to stop there, do you? :-) [20:50:55] for the moment :P [21:39:22] !log deployment-prep Cherry-picked I8f77e0c into puppet and forced puppet run on deployment-bastion [21:39:25] Logged the message, Master [22:12:29] How can I prevent lighttpd from executing my .py file, and instead force it to let the user download it? [22:23:03] Sumurai8: I assume you need to set (not append) cgi.assign in ~/.lighttpd.conf (cf. https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/Help), but I don't have a tested configuration ready. [22:23:29] it complains about duplicate configuration when I set it :| [22:23:36] I can't use -= [22:23:52] and I am unfamiliar with lighttpd, so I don't even know if it is possible [22:24:34] Sumurai8: Neither me unforuntately :-). [22:25:23] I just wrapped the entire file in a printed heredoc and replaced all instances of \n with \\n, so it will simply print the contents, but there should be a better way :| [22:26:31] Might ask it on serverfault tomorrow. Gonna have some sleep. [22:56:34] !log integration Restarting integration-slave1002 [22:56:35] Logged the message, Master [23:04:23] !log deployment-prep Re-enabled puppet on deployment-apache02 and forced a puppet run [23:04:25] Logged the message, Master [23:12:43] Wouldn't heartbleed potentially have leaked *anything* that could be in memory on a public facing server using openssl on an open port? [23:14:12] potentially, yes [23:15:40] paravoid: So anything that's meant to be private on those machines should technically be considered compromised? [23:16:14] wait [23:16:21] not in memory in general [23:16:29] in memory of the process that handles SSL [23:16:49] for instance, in memory of an apache/php process [23:17:39] Oh, that's actually worse, because it's a smaller space to find the SSL key, right? [23:18:23] kind of [23:18:39] it's not worse, in the sense that not everything is exposed [23:18:53] * a930913 just realised why cookies are compromised too. [23:19:16] and as for the SSL key, there's no evidence (yet) it's extractable in this way [23:19:30] on Linux, at least [23:19:44] paravoid: Didn't they demonstrate on their own machine that they could extract it? [23:19:50] no [23:20:06] Or was it just they could extract 63 kbytes? [23:20:39] *64-1 [23:20:47] 64K of memory per request [23:21:31] Don't you have to send a nominal one byte? [23:21:59] Or was that someone explaining it badly? [23:22:43] to do what? [23:35:25] paravoid: Nvm.