[00:51:27] i'm getting mails about 3 phabricator instances failing every single day [00:51:58] but those are the ones we cant ssh to ,apparently [00:52:34] nuke from orbit? [00:52:57] no, that destroys all testing for phab [00:53:04] people want the data back [00:53:10] if puppet's not busted I think there's a way to add your keys as root [00:53:23] my key is in root keys but that doesnt help [00:53:27] wait, you're a real root [00:53:40] hmm so really really screwed up [00:53:53] is there a way to get console on the VMs? [00:55:30] i know for ganeti VMs but i have not seen one for openstack VMs [00:57:04] mutante: http://docs.openstack.org/user-guide/cli_access_instance_through_a_console.html ? [01:00:19] thanks, searched "get-vnc-console" does not appear on wikitech [01:00:27] trying that from .. hmm.. i guess labcontrol [01:00:41] You must provide a username or user id via --os-username, --os-user-id, env[OS_USERNAME] or env[OS_USER_ID] [01:02:36] ok, so then it also wants a project name.. [01:03:39] and after that an --os-auth-url [01:07:09] mutante, see silver:~krenair [01:07:18] there's some openstack-env file under there which will be helpful [01:09:04] which are the failing instances bd808? [01:09:20] I see 6 total instances in the phabricator project [01:09:26] dunno. ask mutante [01:09:55] Krenair: thanks, the instance names are: [01:10:16] phab-01, phab-02, phab-03 and harbormaster.phabricator [01:10:24] so pretty much anything phabricator [01:11:15] krenair@silver:~$ nova get-vnc-console phab-01 novnc [01:11:15] ERROR (ClientException): The server has either erred or is incapable of performing the requested operation. (HTTP 500) [01:13:13] same for xvpvnc [01:13:35] apparently we don't get the "spice" option [01:14:04] * Krenair looks into how OSM does it [01:15:06] OSM only gets console output [01:15:43] try phab-02 [01:15:45] with the same command [01:15:47] just because [01:15:55] i have a feeling phab-01 is more broken than the others [01:20:00] no luck. a root sitting on labcontrol may have more luck [01:24:19] Krenair: how do you know the password? [01:24:28] which password? novaadmin? [01:24:30] yes [01:24:36] it's in the wikitech private config [01:24:55] gotcha [01:26:06] haven't been able to find anything interesting-looking in the console outputs [01:26:23] harbormaster1 failed to start NFS [01:26:41] so i ran the -env script where i try it [01:26:56] it still complains about missing os-username [01:27:08] even though that should be exported now..looks [01:27:49] oh, look: [01:27:51] ERROR (CommandError): No server with a name or ID of 'phab-01' exists. [01:28:07] did you change OS_TENANT_NAME? [01:28:17] it's "testlabs" [01:28:18] my default is testlabs but you need phabricator [01:28:30] oh, of course, per project [01:28:38] yeah :/ [01:29:12] now it .. doesnt show an error [01:29:17] but that's all [01:29:33] expected to get that URL it talks about [01:30:09] ERROR (ClientException): The server has either erred or is incapable of performing the requested operation. (HTTP 500) [01:30:13] after timeout [01:30:16] so no luck either [01:32:30] YuviPanda: ping [01:33:33] Krenair: thanks for trying.. i wonder if we really never used those consoles before or we just do it wrong. will ask again next week [01:33:46] Negative24: vacation [01:34:03] ah [01:34:17] Btw, I can get into phab-02 and it still has puppet failing [01:34:28] Well I could get in. Let me try [01:34:58] Negative24: oh? how about phab-01 [01:40:42] mutante: no phab-01 but I'm in phab-02 [01:42:33] mutante, are you able to send commands to the instances via salt? [01:42:57] Krenair: yuvi tried [01:43:00] and they worked [01:43:42] But why can I login an no one else? [01:45:25] Negative24: thanks, so there are 2 issues, last time it was like "no ssh to -01 or -02" [01:45:33] now i can also ssh to -02 [01:45:38] cool [01:45:45] the mails i get are from the puppet fail [01:45:49] Krenair: probably, yes [01:45:51] I cleared out the full 1 MiB /tmp partition [01:46:07] actually... [01:46:08] that may have done stuff [01:46:10] * Krenair goes afk [01:46:15] root@novaproxy-01:~# [01:46:18] that's where i end up [01:46:32] phab-01 gets me to novaproxy-01 [01:46:38] hrm [01:46:49] how are you sshing? [01:47:23] wrong :p [01:47:29] ssh root@phab-02.wmflabs.org [01:47:32] which was nonsense [01:47:40] but if i do it the right way [01:47:44] i get denied.. [01:47:51] Permission denied (publickey,keyboard-interactive). [01:48:05] by "right way" i mean: ssh root@phab-02 [01:48:12] phab-02.wmflabs.org A record would point to novaproxy-01 [01:48:29] phab-02.phabricator.eqiad.wmflabs A record should point to phab-02 [01:48:31] yep, and it does [01:48:40] that one doesnt let me login [01:48:52] eh [01:48:53] so where do you end up Negative24? [01:49:45] As me in my directory [01:49:58] what's the hostname? [01:49:58] Pretty much normal [01:50:10] hostname -f ? [01:50:33] Warning: Permanently added 'phab-02.phabricator.eqiad.wmflabs' (ECDSA) to the list of known hosts. [01:50:36] Permission denied (publickey,keyboard-interactive). [01:50:48] phab-02.phabricator.eqiad.wmflabs [01:51:02] well, so i take it back, only you can login :) [01:51:11] Negative24, cat /etc/ssh/userkeys/root [01:51:12] and that bothers me [01:51:54] mutante, wait, what is your root key's public part? [01:51:57] Looks normal with regular admin keys [01:52:07] Want me to pastebin [01:52:56] Krenair: it ends in mutante@vanguard [01:53:22] i'm running out of time [01:53:30] Negative24, is that key in there? [01:53:38] Yep [01:53:49] Ends with "75sDD04r2AyFbYPdbVKJDw7cai7V/w== mutante@vanguard" [01:54:10] does /etc/ssh/sshd_config contain PermitRootLogin yes? [01:54:13] separate quesiton [01:54:18] what is the actual puppet error? [01:55:01] wanna run puppet agent -tv? [01:55:02] PermitRootLogin yes [01:55:17] Puppet: Error: Could not request certificate: getaddrinfo: Name or service not known [01:55:32] Looks like that's from the puppetmaster move [01:55:39] * Negative24 has to go [01:55:44] yea, same here [01:55:47] Negative24, can you add me as a project admin? [01:55:48] we gotta continue later [01:55:53] I'll be back in an hour for more debug [01:55:57] but i dont know what "the master move" was [01:56:05] thanks so far! [01:57:34] Krenair: can't: "The action you have requested is limited to users in the group: cloudadmin." [01:57:42] bah [01:57:45] we broke that earlier [01:57:45] ok [01:57:47] never mind [02:15:31] bd808, could you add me to the phabricator project please? [03:16:18] Krenair: what's this "fatal: unable to access 'https://github.com/jdloft/dotfiles.git/': Failed to connect to url-downloader.wikimedia.org port 8080: Connection timed out" [03:17:22] oh, you won't be able to connect to url-downloader anymore [03:17:42] need to remove that as your proxy config [03:17:52] what was it? [03:17:54] proxy? [03:18:12] I have a feeling phab-02 is really out of date [03:19:13] it still is a proxy [03:19:20] but only for production [03:20:48] ok [03:20:54] Krenair: from auth.log "sshd[21970]: User root not allowed because account is locked" [03:22:12] did mutante use his own account? [03:22:39] no [03:24:34] at least I assume not [03:25:28] auth.log doesn't contain mutante [03:25:43] So why is root locker or not accepting pubkey auth? [03:25:47] *locked [03:26:19] you may need to set a password on it [03:26:52] Alright [03:27:32] can you try logging in as root? [03:27:52] 'password' [03:28:51] shouldn't be possible [03:29:31] yep, not possible [03:29:45] Heh [03:29:50] same error as twentyafterfour [03:29:52] "User root from bastion-01.bastion.eqiad.wmflabs not allowed because not listed in AllowUsers" [03:29:58] but my key is not in there [03:30:29] I can paste it in [03:32:08] I could turn on PasswordAuthentication [03:38:56] I suggest you don't do that while the root pass is 'password' :) [03:39:34] Yeah [03:40:09] is it normal to lock the account when there isn't a root pw? [03:40:37] because then root access would always be impossible [03:59:16] not sure [04:38:05] dunno if anyone is arround still, but i can't get into discourse1001.eqiad.wmflabs, the console in wikitech says: 2016-02-13T04:36:59.949462+00:00 discourse1001 puppet-agent[1025]: Could not request certificate: Connection refused - connect(2) for "" port 8140 [04:38:32] i tried rebooting, but same problem [04:41:08] ebernhardson: I can look, but most likely the best solution is to delete the instance and recreate. It’s brand new, right? [04:42:46] andrewbogott: it's somewhat new, i have a backup of the data on it from last week but not since then. [04:42:59] oh, so it was accessible at some point in the past? [04:43:01] That’s different :) [04:43:23] andrewbogott: yes it's been accessible, and somewhat used (it's a test for a mailing list replacement) [04:45:19] basically there was a discourse instance in the search project, i backed it up and shut it down on tuesday and started a new instance tuesday in the `discourse` project yuvi created for it. It's not the end of the world to lose the data since tuesday but would be nice to keep [04:45:58] when did it stop working? [04:47:11] andrewbogott: ad noticed ~12 hours ago, [04:47:22] andrewbogott: my best guess would be related to this: 2016-02-13T04:12:18.398131+00:00 discourse1001 diamond[370]: sudo: unable to resolve host discourse1001 [04:47:23] ok, I think I see why it’s down... [04:47:50] it’s my fault, approximately — I was cleaning up duplicate dns entries and must have cleaned up yours and left the zombie behind :( [04:47:51] other instances in labs seem to agree no dns entry exists [04:47:53] I’m looking for details now [04:47:55] ok :) [05:00:29] !log phabricator Added Alex Monk (Krenair) as projectadmin [05:00:32] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Phabricator/SAL, Master [05:00:50] Krenair: ^ [05:01:33] valhallasw`cloud, I'm a little disturbed. My bash history suggests someone else has access my cyberbot account. [05:01:59] There a commands in the history that are unfamiliar to me. [05:02:29] ebernhardson: this is only getting worse… I think you should delete it rather than wait for me to fumble. [05:02:42] Krenair: do you want couldadmin? It seems to me that you have more than enough reason to have it? [05:02:44] Make sure you give it a few minutes after deletion before recreating if you want to reuse the same hostname [05:03:23] * bd808 sees that Krenair has given up cloudadmin before [05:03:24] andrewbogott: actually, it let me log in now so i might be able to pull a backup [05:03:31] oh? [05:03:38] maybe I fixed it better than I thought [05:03:48] andrewbogott: i think you might have, the discourse instance running in docker is now up too [05:03:49] but, yeah, if you don’t mind, grab the data, delete the instance, and then tell me when you’re done [05:03:52] so I can clean up the mess [05:03:54] ok [05:04:18] it’s in a between state now, will leak some resources on deletion which I’d like to take care of while I’m thinking about it [05:06:21] andrewbogott: ok instance backed up and deleted [05:06:26] thanks [05:06:31] sorry about the mess :( [05:06:47] andrewbogott: no worries, it's all worked out :) [05:09:26] ok, all cleaned up. [05:09:27] bedtime [05:09:36] 6Labs: Instance discourse.search.eqiad.wmflabs in SHUTDOWN state - https://phabricator.wikimedia.org/T126191#2024926 (10EBernhardson) Something went wrong with a cleanup of DNS entries in labs and rendered the host into an undeterminate state. Was able to pull a backup and have had to delete the instance, settin... [07:33:00] 10Labs-Other-Projects: Succesful pilot of Discourse on https://discourse.wmflabs.org/ as an alternative to wikimedia-l mailinglist - https://phabricator.wikimedia.org/T124690#2025033 (10EBernhardson) [07:33:02] 6Labs: Instance discourse.search.eqiad.wmflabs in SHUTDOWN state - https://phabricator.wikimedia.org/T126191#2025031 (10EBernhardson) 5Open>3Resolved a:3EBernhardson [07:34:10] 6Labs: Create temporary test mailman mailing list to test synchronization with https://discourse.wmflabs.org/ - https://phabricator.wikimedia.org/T126547#2025035 (10EBernhardson) there is already a labs project, he's asking for a mailing list he can point at it to test syncronization instead of testing with a li... [07:34:23] 6Labs, 10Wikimedia-Mailing-lists: Create temporary test mailman mailing list to test synchronization with https://discourse.wmflabs.org/ - https://phabricator.wikimedia.org/T126547#2025036 (10EBernhardson) [07:37:02] 10Labs-Other-Projects: Succesful pilot of Discourse on https://discourse.wmflabs.org/ as an alternative to wikimedia-l mailinglist - https://phabricator.wikimedia.org/T124690#2025041 (10EBernhardson) [07:40:15] PROBLEM - Free space - all mounts on tools-worker-1001 is CRITICAL: CRITICAL: tools.tools-worker-1001.diskspace.root.byte_percentfree (<10.00%) [07:45:07] ebernhardson: Austin Hair is one of the current wikimedia-l moderators [07:45:20] list admin technically [08:10:09] PROBLEM - Free space - all mounts on tools-worker-1001 is CRITICAL: CRITICAL: tools.tools-worker-1001.diskspace.root.byte_percentfree (<10.00%) [08:12:42] 10Labs-Other-Projects: Succesful pilot of Discourse on https://discourse.wmflabs.org/ as an alternative to wikimedia-l mailinglist - https://phabricator.wikimedia.org/T124690#2025074 (10Steko) Re: Multi lingual interface. Discourse has an internationalised interface, admins can choose the default language but us... [09:26:14] 6Labs, 10whatcanidoforwikimedia.org: Project wcidfwm (What can I do for wikimedia) - https://phabricator.wikimedia.org/T115092#2025104 (10Samtar) @Petrb little bit confused, I worked on and completed wcidfwm quite a while ago.. [09:30:51] 6Labs, 10whatcanidoforwikimedia.org: Project wcidfwm (What can I do for wikimedia) - https://phabricator.wikimedia.org/T115092#2025106 (10Samtar) @Steinsplitter also, whatcanidoforwikimediacommons.org is your project so just thought you should be aware of this ^ [09:39:11] 6Labs, 10whatcanidoforwikimedia.org: Project wcidfwm (What can I do for wikimedia) - https://phabricator.wikimedia.org/T115092#2025115 (10Steinsplitter) >>! In T115092#2023416, @Dzahn wrote: > is this project also for http://whatcanidoforwikimediacommons.org/ ? The project is hosted on wmpl toolserver and the... [10:35:11] CP678: yes, me. [10:35:18] trying to figure out your delted file issue [10:41:36] Hello [10:42:45] Why cannot I see who is here? [10:43:12] I'm not sure what you mean. [10:43:24] The list of users [10:43:56] That sounds like an issue with your client? [10:44:06] Yes. [10:44:16] Mutter on iPad. [10:45:01] check the mutter docs on how to view that list? [10:45:33] What docs ? ^^ [10:52:39] 10Tool-Labs-tools-Database-Queries, 6Phabricator: Archive Tool-Labs-tools-Database-Queries project - https://phabricator.wikimedia.org/T107699#2025204 (10Bugreporter) Yes it is better to request it at mailing list instead of creating Phabricator tasks. [10:57:20] 6Labs, 10Wikimedia-Mailing-lists: Create temporary test mailman mailing list to test synchronization with https://discourse.wmflabs.org/ - https://phabricator.wikimedia.org/T126547#2025206 (10Krenair) Can it not be on a test mailing list in labs? [12:06:51] how does php erro reporting works in labs? :/ [12:07:19] Steinsplitter: they end up in your error.log by default [12:07:42] yes, but error reporting for the file itself does not work? [12:07:53] whiteout editing the .config ? [12:08:13] I'm pretty sure you can tell php to output in the http response as well [12:08:40] so, er, https://stackoverflow.com/questions/1053424/how-do-i-get-php-errors-to-display ? [12:08:48] does not work [12:12:06] Steinsplitter: syntax errors would not be reported, no. [12:12:36] then tai error.log :) thx [12:12:38] *l [12:39:46] 6Labs, 10Wikimedia-Mailing-lists: Create temporary test mailman mailing list to test synchronization with https://discourse.wmflabs.org/ - https://phabricator.wikimedia.org/T126547#2025294 (10AdHuikeshoven) Any test mailman list will do which behaves technically the same as for example wikimedia-l. [12:40:23] 6Labs, 10Tool-Labs, 6Project-Creators: Migrate Tools access request process to Phabricator - https://phabricator.wikimedia.org/T72625#2025297 (10Luke081515) So what is your proposal? A project (type component), where you can put all tasks in? [16:01:18] Can someone tell me how I can correct a translation, please? [16:03:45] ? [16:04:50] Nnemo: translatewiki.net? [16:05:01] it depends on what translation you want to correct, really [16:22:13] The page: tools.wmflabs.org/xtools-ec/index.php?project=az.wikipedia.org&user=Cekli829 [16:23:54] At the end, the page says that the translations are propulsed by translatewiki.net and Intuition. [16:23:58] So? [16:27:10] Who wants a trout? [16:28:01] How can I correct the translation? [16:29:23] Nnemo: https://translatewiki.net/w/i.php?title=Special:MessageGroupStats&language=fr&group=tsint-xtools [16:30:08] Thank you Glaisher. I will look at that. [16:30:11] But I don't think those get updated regularly so you'll have to submit a patch with the new translations or poke someone who can do that/ [16:48:59] Wait... So, before I edit the translation, will it be placed on the page? How? [16:51:35] Glaisher: Would you then upload the translation to the page? [16:52:16] I don't know the procedure for how that tool updates its translations but I might look into it if I have some time [16:52:30] (better if you ping a maintainer for the tool) [16:56:18] Who ? [16:56:35] it's intuition-based, so I think it should be updated automatically [16:59:01] 6Labs, 10Tool-Labs, 6Project-Creators: Migrate Tools access request process to Phabricator - https://phabricator.wikimedia.org/T72625#2025612 (10valhallasw) The proposal would be a form which creates a Maniphest task, which would require the same fields as the access request on wikitech. If we can require a... [17:21:48] andrewbogott: around ? [17:22:05] matanya: briefly. What’s up? [17:22:46] I am failing to create a m1.gigantic instance [17:23:00] the video project needs more horse power [17:23:31] so andrewbogott i don't know weather it is capacity issue or something else [17:23:58] ok, looking [17:24:04] you checked your quotas already? [17:25:08] haven't [17:25:28] (how do i do that?) [17:25:39] there’s a link on the manage projects page [17:25:49] my wikitech login is a bit messsed up, so if you can check and tell me what increase you need... [17:26:21] looking [17:26:36] Cores: 17/30 [17:26:37] RAM: 18432/51200 [17:26:37] Floating IPs: 0/1 [17:26:37] Instances: 2/10 [17:26:38] Security Groups: 1/10 [17:26:54] so probably more cores [17:27:12] how many cores does a .gigantic have? [17:27:56] 16 [17:28:02] ah, that’d do it. [17:28:41] try now? [17:29:16] wored, thanks andrewbogott [17:29:21] *worked [17:29:25] cool [17:36:51] andrewbogott: one more question please, is it safe to apt-get upgrade an instance ? [17:58:12] matanya: it’s hard to predict. It is usually safe, but I’d do a dry run first. [17:58:26] Any time someone has tried to dist-upgrade (e.g. from precise to trusty) it has killed the instance. [17:58:40] i'm going 8.0 - 8.2 [17:58:46] so i hope it is safr [17:58:46] e [17:58:56] I’d expect it to work, but no guarantees :) [17:59:56] heh, thanks much [20:59:19] 10Quarry: Include query execution time - https://phabricator.wikimedia.org/T126888#2025894 (10Slaporte) 3NEW [20:59:56] 6Labs, 10Labs-Infrastructure, 10DBA, 6operations: db1069 is running low on space - https://phabricator.wikimedia.org/T124464#2025901 (10jcrespo) 5Open>3Resolved 73% -more could be done, but resolving for now. [21:07:58] 10Quarry: Include query execution time - https://phabricator.wikimedia.org/T126888#2025903 (10mahmoud) +1 ! :) [21:12:42] !log tools.stashbot Post restart test {{T126419}} [21:12:45] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.stashbot/SAL, Master [21:28:33] !log tools.stashbot Post `restart` test {T126419} [21:28:36] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.stashbot/SAL, Master [21:43:37] I have edited several translations. Now, can you upload them to the page, please? [21:44:33] Nnemo: that should be automatic, but it can take some time [21:45:48] So we don't need to do anything? [21:47:32] Nnemo, you just have to wait now. [22:30:01] !log phabricator Fixed puppet on phab-03 and harbormaster1 by signing the certs on deploy [22:30:04] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Phabricator/SAL, Master [22:34:31] no luck with phab-01 and phab-02 [22:54:27] Hi, I have a question about using X!'s page history tool. Can someone help? [22:55:33] ari___: that depends on the question. [22:57:43] I'm looking at stats for the page history tool. Is there any documentation that explains the output? [22:59:42] * White_Master . [22:59:52] I don't know, but probably not? It looks fairly self-explanatory to me :/ [23:00:26] ari___: which part is unclear? [23:02:31] "Links from this page:" is unclear. Does this mean links embedded on the page to other wikipedia pages? There's also "External Links" which I think means links on the queried page to other websites. [23:03:05] ari___: yes, links on that page, linking to other wikipedia pages. [23:04:06] okay thank you! [23:04:38] Is there a way to measure links outside of Wikipedia leading to a specific page on Wikipedia? [23:05:20] there's no way to index that, but google allows you to use linkto: or something like that, I think [23:07:14] thank you, I'll try investigating that! [23:45:37] 6Labs, 10Labs-Infrastructure, 7LDAP: Remove shell user "80686" - https://phabricator.wikimedia.org/T63967#2026026 (10Krenair)