[13:21:20] hi there! it seems that replica.my.cnf on tools is automatically reset from time to time. is this an intended behaviour? [13:39:51] ireas: When/which was reset? There seems to have been a mass change on November 14th, but since then I see no changes. [13:42:10] scfc_de, ah, okay. I remember that I fixed this for one project back then [13:42:38] scfc_de, I did not use the other one since then, so it’s probably the same reson [13:42:47] yes, last modification on Nov 14th [13:43:00] okay, thank you! [13:44:51] speaking of which, how do I know if there's something wrong with my account? $ sql itwiki_p [13:44:55] ERROR 1045 (28000): Access denied for user 'u2179'@'10.4.0.220' (using password: YES) [13:48:39] "Alternatively you can rename the credentials file from replica.my.cnf to .my.cnf " didn't help [13:51:21] Nemo_bis: That appears to be a bug, but as auth for LabsDB is done by Coren and Sean you need to file a bug for that. [13:56:36] sob ok [13:58:15] Nemo_bis: Do you get this only with itwiki or with all databases? And what tool is that from? [14:05:47] (03CR) 10Tim Landscheidt: [C: 04-1] "Needs debian/changelog entry." [labs/toollabs] - 10https://gerrit.wikimedia.org/r/106648 (owner: 10Tim Landscheidt) [14:06:04] (03CR) 10Tim Landscheidt: [C: 04-1] "Needs debian/changelog entry." [labs/toollabs] - 10https://gerrit.wikimedia.org/r/106283 (owner: 10Tim Landscheidt) [14:06:19] (03CR) 10Tim Landscheidt: [C: 04-1] "Needs debian/changelog entry." [labs/toollabs] - 10https://gerrit.wikimedia.org/r/106281 (owner: 10Tim Landscheidt) [14:07:25] (03CR) 10Tim Landscheidt: [C: 04-1] "@hashar, after the Jenkins job is now running well, could you please abandon this change? Merci!" [labs/toollabs] - 10https://gerrit.wikimedia.org/r/104917 (owner: 10Hashar) [14:10:25] (03Abandoned) 10Hashar: Jenkins job validation (DO NOT SUBMIT) [labs/toollabs] - 10https://gerrit.wikimedia.org/r/104917 (owner: 10Hashar) [14:15:31] Coren: all i tried, just as nemobis on shell [14:15:42] for a whle [14:15:57] * Coren checks. [14:33:28] hello, I have a situation with puppet overwriting my instance config, can someone tell me how to stop it? [14:34:51] ChrisJ_WMDE: As a rule, you can't. That's what puppet /does/. The right thing to do is to adapt the puppet config to match what your instance need. [14:37:22] ok. how would I stop it from removing postfix? [14:40:11] ChrisJ_WMDE: /removing/ postfix? It'm not sure why it'd do that, perhaps it's trying to replace it with the MTA we normally use? (exim) [14:42:18] yes. postfix removes exim [14:44:31] ChrisJ_WMDE: You mean the other way around, I suppose? If so, you could adapt Puppet for this particular host to not enable exim - but: Why? Is there some Postfix functionality that exim doesn't offer? [14:44:39] Well, you need to change your puppet class(es) to not install exim. Or you need to use exim and not postfix. :-) [14:45:17] (I'd recommend the latter; we use exim in general and having a different MTA in some places is just begging for trouble down the road) [14:45:40] ok, thanks I will just change to exim [15:23:20] hashar: if you have time, could look into https://bugzilla.wikimedia.org/show_bug.cgi?id=49846 ? it has VE down completely on beta labs [15:28:42] chrismcmahon: that is handled by James [15:28:54] chrismcmahon: the issue is in Gerrit because there are two repositories named VisualEditor [15:28:59] chrismcmahon: they are going to rename one [15:29:05] I think it happens today [15:29:13] hashar: oh, sorry, I thought we'd done that already, thanks [15:32:22] I have created an exim4 conf file, but will puppet overwrite it with the default? [15:34:48] hashar: ah, nice to know it should potentially be fixed today :) [15:40:15] chrismcmahon: I have let other folks to find out a way to fix it up [15:40:32] chrismcmahon: sorry it lacked feedback on bug report :D [15:46:12] ChrisJ_WMDE: Probably. Are you aware of https://wikitech.wikimedia.org/wiki/Help:Self-hosted_puppetmaster? Also, I think Silke_WMDE has some experience with the general idea on how to test changes. [15:50:24] Coren: scfc_de I can't access tools.wmflabs.org (via HTTP) from toollabs, no? What do I use instead? [15:50:27] just tools-webproxy? [15:50:36] * Coren nods. [15:50:43] Coren: alright! [15:50:49] Openstack networking limitation from Hell. [15:50:56] heh, indeed [15:51:05] Coren: why can't we just add a hosts entry? we have those for the labsdb [15:53:19] YuviPanda: There's one at least on tools-login. [15:53:26] scfc_de: thanks this is what I will do [15:53:44] scfc_de: hmm? curl tools.wmflabs.org on -login doesn't do anything [15:54:46] YuviPanda: Works for me? [15:55:14] scfc_de: ah, hmm, works for me too [15:55:22] scfc_de: I... was doing curl to tools-login.wmflabs.org [15:55:27] but I guess won't work from the exec nodes [15:56:00] YuviPanda: It's on -exec-04, let's see what the other nodes say ... [15:56:10] should be on puppet, hopefully [15:57:12] YuviPanda: Is present on all but tools-proxy, tools-syslog, tools-proxy-test, tools-redis and tools-webproxy. [15:57:19] hmm [15:57:22] where's that coming from? [15:58:10] YuviPanda: /data/project/.system/hosts manually copied there (not by me). Puppetizing is WIP at https://gerrit.wikimedia.org/r/#/c/107103/. [15:58:17] ah, ok [15:58:22] hmm, maybe I could just use it then [15:58:39] thanks scfc_de [15:58:42] brb [16:00:20] virt12 is saturating its uplink [16:00:23] virt2 as well [16:00:35] Coren: ^^ [16:56:40] Hey Guys [16:59:05] My Friends when connect a server with putty https://fbcdn-sphotos-h-a.akamaihd.net/hphotos-ak-prn2/v/1532013_646623838735201_798048869_n.jpg?oh=fe5cf8fbe6f38150e51c744856997a87&oe=52D661D6&__gda__=1389783994_f5ba1197b1d435f380a2711d400447a4 [17:14:38] Selim: Is this the same problem as yesterday? [17:14:57] Nope. [17:15:09] My Account is working fine :) [17:15:31] But somone need help in ar wiki and tell me this proplem [17:18:11] Selim: I don't see a user with a shell name "AboHeidi", but one with "aboheidi2". Perhaps tell them to try that? [17:19:07] ok wait some minutes :) [17:23:49] : Sa,e [17:23:54] Same* [17:24:13] Works now or still failing? [17:25:18] Still Failing [17:25:37] https://fbcdn-sphotos-h-a.akamaihd.net/hphotos-ak-prn1/v/1010732_646639388733646_1308250546_n.jpg?oh=2e0258272d3fae46de2216c89f1f0a7e&oe=52D64F1A&__gda__=1389778536_e7daca056146d89e19c6411c81f5ba65 [17:26:13] I'll make new account for him ok? [17:28:50] That won't change anything. He just needs to check that his public key is properly uploaded at https://wikitech.wikimedia.org/wiki/Special:Preferences#mw-prefsection-openstack, then login will work. [17:31:12] ok i'll tell him [17:32:32] Public Key here: https://fbcdn-sphotos-h-a.akamaihd.net/hphotos-ak-prn2/v/1490624_646644285399823_233620417_o.jpg?oh=8bf8624adab42a0769a46ce94edb7310&oe=52D62046&__gda__=1389748458_507b8cef602d240bc1fdfc037a4980d9 [17:35:06] Selim: Oh, I just noticed the first line in the Putty screenshot: "Unable to use key file ... (SSH-1 private key)". So maybe he needs to generate a new key? (And upload the new public key then again.) [17:35:27] ok [17:35:46] Usually, if Putty or other SSH clients don't find a key file, they try logging in with username and password, and this will fail always on Labs/Tools. [17:40:51] Sry i Don't know what ur talking about it [17:41:59] He Still There :D [17:42:10] What are Doin now? [17:43:10] There are (at least) two ways you can identify yourself to a host with SSH: a) By your private/public key, and b) by traditional username/password. I assume if Putty can't read the key file, it tries username/password instead. This is not allowed on Labs (there are no passwords), so this fails. [17:43:38] ahh! It's Working fine Now [17:43:42] Thank you [17:55:26] scfc_de: weist du wie man "_" in fachsprache nennt? [17:55:32] kriege die kriese [18:01:15] Steinsplitter: Underscore? Unterstrich? [18:14:36] thx [18:14:44] gibts also nicht :D [18:19:54] Hey, Sorry For my Questions But i Can use putty on blackberry? [18:22:39] Selim: https://www.google.co.in/search?q=ssh+blackberry perhaps? [18:23:31] But he need user and pass? [18:23:59] haven't ever used a blackberry phone, so unsure [21:00:31] hi, what is the better way to use pagecount dumps in tool labs? the folder /public/pagecounts is empty, there is other folder where I can get the pagecont dumps of the last month? [21:14:54] danilo: There is a bug with pagecounts dumps atm that will require a restart of much of the tool labs infrastructure. Unless I get a better timeline for the migration to eqiad really soon I'll have to schedule that for the middle of next week or so. [21:15:46] i want to setup one of our labs instances with a permenantly lagged slave(via pt-slave-delay). Before i start doing this all by hand is there anything already setup for doing this in labs? [21:16:09] actualy, "by hand" i will probably end up making a vagrant role [21:34:04] ebernhardson: I was about to say that a vagrant role would seem to fit the bill. :-) [21:48:55] Coren: I need a 1GB pagecounts dump, can I download it in tool labs? [21:49:18] danilo: Sure. Try to not keep to many of them around though. [21:49:38] ok, thanks [21:50:11] I was wondering, what hypervisor does our OpenStack use? [21:50:24] https://en.wikipedia.org/wiki/OpenStack says KVM and XenServer are both supported. [21:53:16] superm401: We use KVM [21:53:30] Thanks, Coren