[00:03:50] init: nss-ldap: do_open: do_start_tls failed:stat=-1 [00:03:50] [00:03:50] init: nss_ldap: could not search LDAP server - Server is unavailable [00:03:50] [00:03:50] Console output to bots-irc2 [00:04:02] RECOVERY host: bots-irc2 is UP address: bots-irc2 PING OK - Packet loss = 0%, RTA = 4.07 ms [00:04:20] !logging :o [00:04:20] To log a message, use the following format: !log [00:04:22] do that [00:04:23] :D [00:04:30] I have no idea what's going on [00:04:43] methecooldude: ^ [00:05:26] !log bots LDAP issue on bots-irc2 > "init: nss-ldap: do_open: do_start_tls failed:stat=-1" and "init: nss_ldap: could not search LDAP server - Server is unavailable" [00:05:27] Logged the message, Master [00:05:50] meh why it exist? [00:05:54] I thought we deleted irc2? [00:06:00] jeremyb: ^ [00:06:59] methecooldude: I think that best solution for now would be to delete it, but I don't know who created it nor why... [00:07:06] ask jeremy [00:07:30] * jeremyb clicks delete finally [00:07:55] :D [00:08:15] it's too bad when you delete then all the details are gone [00:08:31] !log bots deleted irc2 [00:08:32] Logged the message, Master [00:08:41] would be better to just have a semnatic option to indicate it's deleted [00:08:52] s/option/property/ [00:08:52] hm... [00:08:57] that's true [00:09:19] otoh allow everyone (even anons?) to see deleted pages [00:09:27] also a good idea [01:25:32] petan: Would you mind adding me to the bots project? Interested in seeing what's going on and might want to move some of my stuff over [03:17:53] PROBLEM Disk Space is now: WARNING on nova-production1 nova-production1 output: DISK WARNING - free space: / 565 MB (5% inode=86%): [06:01:07] 01/04/2012 - 06:01:07 - Creating a home directory for johnduhart at /export/home/bots/johnduhart [06:02:08] 01/04/2012 - 06:02:08 - Updating keys for johnduhart [06:17:53] PROBLEM Disk Space is now: CRITICAL on nova-production1 nova-production1 output: DISK CRITICAL - free space: / 279 MB (2% inode=86%): [10:44:03] Change on 12mediawiki a page Wikimedia Labs/Create shared sql service for all projects was modified, changed by Petrb link https://www.mediawiki.org/w/index.php?diff=480900 edit summary: [15:55:13] 01/04/2012 - 15:55:13 - Creating a project directory for deployment-prep [15:56:23] shoulda been here already [15:56:23] !labs [15:56:23] https://labsconsole.wikimedia.org/wiki/$1 [15:56:35] !labs Access [15:56:35] https://labsconsole.wikimedia.org/wiki/Access [15:56:49] that's a console [15:57:24] arr, i thought i was on this channel already:) [15:57:27] wrong logfile now [15:57:33] !log added project deployment-prep for hexmode and petan [15:57:34] added is not a valid project. [15:57:41] !logging [15:57:41] To log a message, use the following format: !log [15:57:49] what is the global project? [15:57:54] no idea [15:57:55] for news like "added project":) [15:58:02] !sal [15:58:02] https://labsconsole.wikimedia.org/wiki/Server_Admin_Log see it and you will know all you need [15:58:10] maybe there is a log where Ryan log this [15:58:31] dunno [15:59:32] btw you didn't insert us to project :) [15:59:39] https://labsconsole.wikimedia.org/wiki/Special:NovaProject [15:59:51] hexmode: you know how to use instances? [16:00:07] !socks-proxy [16:00:07] see https://labsconsole.wikimedia.org/wiki/Access#Accessing_public_and_private_instances ssh @bastion.wmflabs.org -D [16:00:32] petan|work: no... This'll be my first time on labs [16:00:36] ok [16:00:41] there is a login server bastion [16:00:52] from there you can ssh to instances [16:01:05] hexmode: so you need a wiki user on labsconsole.wm [16:01:15] you ssh using key, for sudo you use password, it's not so hard [16:01:21] hexmode: then i can add you to the new project [16:01:54] mutante: MarkAHershberger is on there already ;) [16:01:59] ah [16:02:00] o.0 it's busy in here [16:02:05] hey [16:02:06] :) [16:02:43] mutante: my user is Petrb [16:02:44] !log deployment-prep added new project deployment-prep for hexmode and petan [16:02:46] Logged the message, Master [16:02:58] :) thats the way:) [16:03:00] yes that's probably how to log that :) [16:04:13] 01/04/2012 - 16:04:13 - Creating a home directory for mah at /export/home/deployment-prep/mah [16:04:26] :) [16:04:27] hexmode: I think I will create a web server (small instance) and 80 gb sql instance we would use until we get a better server [16:04:42] that would be enough even for full simple wiki [16:04:50] i was about to say "added users", but the bot already tells you:) [16:04:56] petan|work: sure lets do that [16:04:59] ok [16:05:14] 01/04/2012 - 16:05:13 - Updating keys for mah [16:05:14] 01/04/2012 - 16:05:13 - Creating a home directory for petrb at /export/home/deployment-prep/petrb [16:05:53] I need netadmin too :) [16:05:59] to create instances and Mark too [16:06:16] 01/04/2012 - 16:06:16 - Updating keys for petrb [16:06:39] !log deployment-prep added members MarkAHershberger & Petrb - added them to sysadmin and netadmin roles [16:06:40] Logged the message, Master [16:07:59] !log deployment-prep configured firewall for webserver [16:08:00] Logged the message, Master [16:08:16] OrenBochman: did you get tidy yet? [16:08:38] btw OrenBochman import of simple wiki is nearly done, I will fix db now [16:08:54] ok [16:09:04] can you also put in tidy [16:09:26] it has loose html tags - can you install tidy to clean these up - e.g. bottom of /w/index.php/Italic [16:09:59] btw where can you see the import progress ? [16:10:09] !log deployment-prep create instances for apache and mysql [16:10:10] Logged the message, Master [16:10:15] on [[special:statistics]] [16:10:19] out of 300k [16:10:22] pages [16:10:36] petan|work: the first log line is missing on https://labsconsole.wikimedia.org/wiki/Nova_Resource:Deployment-prep/SAL though [16:10:44] i'm doing tidy? [16:10:46] uh [16:10:47] petan|work [16:10:59] jeremyb: yes? [16:11:08] < jeremyb> i'm doing tidy?\ [16:11:15] you mean rebuildall.php? [16:11:26] because that's what we have to do now :) [16:11:31] no, wgUseTidy [16:11:33] I will check if import is finished [16:11:36] ah [16:11:40] OrenBochman asked for it [16:11:48] i just wanted to make sure you weren't doing [16:11:48] of course np [16:11:49] it [16:11:59] but I don't know if it's finished yet [16:12:05] import [16:12:55] ok hexmode we have the instances created I will now install all software there and start import of simple wiki and hiwiki latest rev, ok [16:13:02] huh, https://www.mediawiki.org/wiki/Manual:$wgAlwaysUseTidy [16:14:33] PROBLEM Current Users is now: CRITICAL on deployment-web deployment-web output: Connection refused by host [16:14:39] hexmode: https://labsconsole.wikimedia.org/wiki/Nova_Resource:Deployment-prep [16:15:06] mutante: petan|work: tyvm :) :) [16:15:13] PROBLEM Disk Space is now: CRITICAL on deployment-web deployment-web output: Connection refused by host [16:15:45] petan|work: can we do a full simple? [16:15:59] yes [16:16:03] PROBLEM Free ram is now: CRITICAL on deployment-web deployment-web output: Connection refused by host [16:16:23] latest hi would be today [16:16:30] full simple will probably take a day or two [16:16:44] sure, np [16:16:48] ok [16:17:22] hexmode: you can insert more people who are involved to project [16:17:23] if i get other people who want different wikis on there how hard will it be for me to add them? [16:17:23] PROBLEM Total Processes is now: CRITICAL on deployment-web deployment-web output: CHECK_NRPE: Error - Could not complete SSL handshake. [16:17:34] easy :) [16:17:44] awesome [16:17:50] https://labsconsole.wikimedia.org/wiki/Special:NovaProject [16:18:06] there is add a member, you type a wiki name and it would insert them to project [16:18:13] PROBLEM dpkg-check is now: CRITICAL on deployment-web deployment-web output: CHECK_NRPE: Error - Could not complete SSL handshake. [16:18:14] * hexmode imagines a button to "create a dupe of X wiki with X codebase" [16:18:31] ah, that's not so easy yet [16:18:43] but there is a proposal to create it [16:18:46] petan|work: make it so? :) [16:18:53] PROBLEM Current Load is now: CRITICAL on deployment-web deployment-web output: CHECK_NRPE: Error - Could not complete SSL handshake. [16:19:13] http://www.mediawiki.org/wiki/Wikimedia_Labs [16:19:26] that's a summary of all what we are doing right now, and what is going to be done [16:19:58] in proposal on first line is a proposal to create automatic system where dev just pick db, branch and press enter and wiki is up [16:23:52] PROBLEM Current Load is now: CRITICAL on deployment-sql deployment-sql output: CHECK_NRPE: Error - Could not complete SSL handshake. [16:24:32] PROBLEM Current Users is now: CRITICAL on deployment-sql deployment-sql output: CHECK_NRPE: Error - Could not complete SSL handshake. [16:25:12] PROBLEM Disk Space is now: CRITICAL on deployment-sql deployment-sql output: CHECK_NRPE: Error - Could not complete SSL handshake. [16:26:02] PROBLEM Free ram is now: CRITICAL on deployment-sql deployment-sql output: CHECK_NRPE: Error - Could not complete SSL handshake. [16:27:22] PROBLEM Total Processes is now: CRITICAL on deployment-sql deployment-sql output: CHECK_NRPE: Error - Could not complete SSL handshake. [16:28:12] PROBLEM dpkg-check is now: CRITICAL on deployment-sql deployment-sql output: CHECK_NRPE: Error - Could not complete SSL handshake. [16:29:04] OSM needs to be updated [16:29:55] !log sesarch search-test: added Alias /wiki /var/www/w/index.php [16:29:56] sesarch is not a valid project. [16:30:01] !log search search-test: added Alias /wiki /var/www/w/index.php [16:30:02] Logged the message, Master [16:30:58] RECOVERY Free ram is now: OK on deployment-web deployment-web output: OK: 92% free memory [16:32:28] RECOVERY Total Processes is now: OK on deployment-web deployment-web output: PROCS OK: 87 processes [16:33:08] RECOVERY dpkg-check is now: OK on deployment-web deployment-web output: All packages OK [16:33:48] RECOVERY Current Load is now: OK on deployment-web deployment-web output: OK - load average: 0.10, 0.28, 0.26 [16:34:38] RECOVERY Current Users is now: OK on deployment-web deployment-web output: USERS OK - 1 users currently logged in [16:35:18] RECOVERY Disk Space is now: OK on deployment-web deployment-web output: DISK OK [16:35:20] !log search search-test: set $wgArticlePath = "/wiki/$1"; $wgUsePathInfo = true; [16:35:21] Logged the message, Master [16:38:03] !log search search-test: set $wgUseTidy = true; [16:38:04] Logged the message, Master [16:38:07] OrenBochman: ping? [16:38:26] hey [16:38:41] so, when i flip the switch i see a difference in the page [16:38:52] so i think it's done. but let me know what you think [16:39:00] i have to run very shortly [16:39:13] it is good now [16:39:22] thanks! [16:39:36] sure [16:39:40] my vacation is over tommorow [16:39:51] I'll be in touch as time allows [16:43:48] sure. i'll work on it some too [16:44:13] is there some other way that's easier to get you besides IRC? [16:45:38] email [16:46:05] my gmail OrenBochman at Gmail.com [16:47:03] petan|work: What's this deployment-prep stuff we have here? :) [16:47:23] hexmode wants to try 1.19 there on production db [16:48:26] johnduhart: If you have ideas on how to follow up on that wikitech-l post I made about this, I'm all ears! [16:48:42] oh [16:48:51] Didn't get the email, one sec [16:49:13] I go to school in a underground bunker that doens't get cell signal [16:49:38] johnduhart: sent it a couple of days ago [16:50:00] oh that one [16:50:23] well, I thought I did... [16:50:31] * hexmode doesn't see it right off [16:50:37] sounds reasonable until testlabs is working [16:51:05] http://thread.gmane.org/gmane.science.linguistics.wikipedia.technical/57950/focus=57969 [16:51:43] Your organization has chosen to limit viewing of this site (http://thread.gmane.org/gmane.science.linguistics.wikipedia.technical/57950/focus=57969), due to the rating of its content (sharewares download,hobbies/interest,web host). [16:51:46] fsck [16:51:53] hehe [16:52:05] I'm literallly loling [16:52:15] hahahahaha [16:52:19] who dat? [16:52:21] heh [16:52:21] that is pretty lame for mailing list archives [16:52:28] johnduhart: ssh proxy? :P [16:52:52] I have firewalled everything apart of port I need [16:52:54] petan|work: too lazy :p [16:53:04] and that one port is enough for me to get everywhere :D [16:53:21] in work [16:55:13] oh crap I have a CCNA test to do this period [17:01:16] oh, that must have been at school. [17:01:19] * jeremyb runs away [17:42:58] back [18:17:49] hexmode: I will wait for Ryan with import of simple wiki, I need more space [18:18:05] I need temporary 200 gb to extract dump there [18:19:52] moar space! [18:20:19] you know if I can import .18 db there? [18:20:26] I don't know if it's compatible or not [18:20:40] maybe I should install 1.18 and update it? [18:22:30] !rt is http://rt.wikimedia.org/Ticket/Display.html?id=$1 [18:22:30] Key was added! [18:28:50] can $1 be in the middle of the string as well? [18:28:56] yes [18:29:02] do i need to escape it then? [18:29:16] !pxe is http://wikitech.wikimedia.org/view/$1#PXE_booting [18:29:16] Key was added! [18:29:25] !pxe Dell_PowerEdge_R410 [18:29:25] http://wikitech.wikimedia.org/view/Dell_PowerEdge_R410#PXE_booting [18:29:25] !ghsh is $1 $2 gsdfgs$34ghrsh $4gsh _$5 [18:29:25] Key was added! [18:29:32] :) nice, thx [18:29:34] !ghsh gsdgh hs gfhj trjrtj trj [18:29:34] gsdgh hs gsdfgsgfhj4ghrsh trjrtjgsh _trj [18:29:39] !ghsh A B C D [18:29:39] A B gsdfgsC4ghrsh Dgsh _$5 [18:29:41] !ghsh A B C D E [18:29:42] A B gsdfgsC4ghrsh Dgsh _E [18:29:50] ;) [18:31:28] hi petan, [18:31:33] hey [18:31:47] I tried your solution for gaining ssh access to reportcard1 [18:31:52] but it doesn't work :( [18:32:06] right, did you insert it to wiki? [18:32:17] you need to copy the public key [18:32:30] manage your ssh keys [18:32:34] in console [18:32:53] ok [18:32:55] thx [18:33:31] let me know if you didn't know where it is [18:34:05] 01/04/2012 - 18:34:05 - Updating keys for diederik [18:34:08] 01/04/2012 - 18:34:08 - Updating keys for diederik [18:34:10] 01/04/2012 - 18:34:09 - Updating keys for diederik [18:34:11] 01/04/2012 - 18:34:11 - Updating keys for diederik [18:47:12] PROBLEM HTTP is now: CRITICAL on deployment-web deployment-web output: Connection refused [18:52:12] RECOVERY HTTP is now: OK on deployment-web deployment-web output: HTTP OK: HTTP/1.1 200 OK - 452 bytes in 0.003 second response time [19:01:12] PROBLEM Current Load is now: WARNING on bots-cb bots-cb output: WARNING - load average: 0.58, 7.22, 5.70 [19:06:12] RECOVERY Current Load is now: OK on bots-cb bots-cb output: OK - load average: 0.37, 2.97, 4.28 [19:15:33] hexmode: I changed my mind I will start a slow import so I won't need so many mb [19:18:17] maxsem@bastion1:~$ ssh mobile-feeds.pmtpa.wmflabs [19:18:17] Permission denied (publickey). [19:18:22] WTF? [19:18:44] are you using proper key? [19:20:22] PROBLEM Disk Space is now: CRITICAL on pageviews pageviews output: DISK CRITICAL - free space: / 0 MB (0% inode=93%): [19:22:00] lo [19:22:03] l [19:22:16] Reedy: is .19 compatible with 18 dump? [19:22:20] mediawiki [19:22:34] should be [19:22:36] I want to run importDump [19:22:38] ok [19:22:41] i don't think much has changed [19:23:09] 01/04/2012 - 19:23:09 - Updating keys for maxsem [19:25:22] RECOVERY Disk Space is now: OK on pageviews pageviews output: DISK OK [19:27:08] 01/04/2012 - 19:27:08 - Updating keys for maxsem [19:27:53] hexmode: done [19:28:02] import is running for both, but site is working [19:28:08] w00! [19:28:10] 01/04/2012 - 19:28:09 - Updating keys for maxsem [19:28:38] mhm, I tried deleting my key and then reimporting it [19:28:50] MaxSem: do you have a private key on bastion? [19:29:01] or forward it? [19:29:03] now I can't even login into bastion [19:29:08] no [19:29:09] ah [19:29:20] you will need to wait for Ryan then [19:33:39] btw OrenBochman import is done [19:45:03] hexmode: I deployed some extension we use in production and it seems to work as for now :o [19:45:08] * extensions [19:45:15] :) [19:45:21] want to test it? [19:45:31] url? [19:45:37] you need to connect to proxy [19:45:43] !socks-proxy [19:45:44] see https://labsconsole.wikimedia.org/wiki/Access#Accessing_public_and_private_instances ssh @bastion.wmflabs.org -D [19:47:21] hrm... need to find out the proper user [19:47:30] mah [19:47:49] and now... my public key [19:48:00] let me make sure rsa and dsa are both loaded [19:48:07] ok [19:48:12] yep [19:48:17] ok... hrm [19:48:33] ryan gave me a place to put them [19:48:42] need to find it.... [19:49:52] hi wiki is a big piece of vandalism :D [19:59:14] petan: ok, looks like my key matches they key up there... any ideas? [19:59:25] works? [19:59:33] no [19:59:36] ah [19:59:43] ssh mah@b... doesn't work [19:59:50] what does it say [19:59:58] Ryan "improved" it [20:00:08] so that it's more safe but no one can log in :) [20:00:11] heh [20:00:12] "Permission denied (publickey)." [20:00:18] typical message [20:00:19] let me do verbose [20:00:21] :) [20:00:41] do you have a private key in .ssh/ [20:00:42] ? [20:00:47] that's important [20:00:54] also is your key in console hm? [20:01:07] https://labsconsole.wikimedia.org/wiki/Special:NovaKey [20:01:11] console hm? [20:01:11] yes [20:01:13] this [20:01:35] that's weird unfortunatelly I can't do much, Ryan can [20:01:35] without going through every character, it looks like a match [20:01:39] k [20:01:49] isn't he in office [20:01:56] no idea [20:01:57] what time is in SF [20:02:02] you are there? [20:02:13] noon in sf... I'm on the east coast [20:02:16] ah [20:02:27] I thought you are from sf too :D [20:02:52] is actually someone working there XD [20:03:01] no, I'm in Amish country [20:03:06] heh [20:03:10] :) [20:03:13] yes, they are when I visit [20:03:22] heh [20:03:53] Heh, you're right, none of the SF people have been active here in the past few hours [20:04:14] I guess Ryan would normally be here but I think he's traveling today, right? [20:04:19] no [20:04:33] he doesn't [20:04:48] he returned yesterday and said that should be today... [20:04:51] Oh [20:04:59] or I think so... [20:06:10] just confirmed that my pubkey is the right one char by char [20:06:27] * hexmode goes to do something else till Ryan shows [20:06:32] ok [20:07:38] hexmode: another possibility is to ask someone with admin rights to give us a public ip [20:07:43] :o [20:07:55] so that we could open deployment.wmflabs.org [20:08:08] mutante: can you do this? [20:08:11] petan: we'll need that eventually, I think so ok [20:08:36] I can register dns but IP's are restricted [20:08:56] Sara can do this too but I didn't see her a long time [20:09:40] * petan goes to prepare apache for wild public [20:09:56] petan: i can, but how many IPs do you need and please provide a reason why you need them [20:10:06] sure, hexmode^ [20:10:10] we need only one ip [20:10:24] petan: its like RIPE rules, because v4 IPs are getting rare [20:10:29] I don't really know what is this project for but hex does ;) [20:10:33] mutante: one IP needed to test deployment of 1.19 [20:10:58] need something so we can put this on a public ip w/ port 80 open [20:11:31] yep, ok, give me a couple minutes, checking some other puppet change right now [20:11:48] and i need to lookup the commands, but can do in a bit [20:11:55] petan: why doesn't my cable company just switch to IPV6? then we could all just forget about scarcity [20:12:07] heh [20:15:30] !sal [20:15:31] https://labsconsole.wikimedia.org/wiki/Server_Admin_Log see it and you will know all you need [20:15:51] hey Ryan! [20:15:55] howdy [20:15:59] :) [20:16:04] need to talk to you [20:16:16] sup? [20:16:38] first (briefly) there are two wiki clone's one Oren's for search project and second Mark's deployment project [20:16:42] both use huge db [20:16:48] so I wanted to talk with you about db and space [20:17:15] I made a simple proposal of mariadb, feel free to rewrite it http://www.mediawiki.org/wiki/Wikimedia_Labs/Create_shared_sql_service_for_all_projects [20:17:25] why do they need huge databases? [20:17:31] you said you wanted to move sql outside of labs [20:17:35] yep [20:17:38] would like to [20:17:43] because one is clone of simple wiki and second is clone of two wikis [20:17:54] one is 20gb second is 80gb sql [20:17:59] hm. root is hard [20:18:00] re:) [20:18:03] for shared database [20:18:16] hm... maybe, but it would be shared between devs [20:18:22] I don't think we would fight heh [20:18:24] hi Ryan, i created a new project for hexmode [20:18:35] I don't think testing db would contain private data [20:18:46] ideally there'd be an api that would let people create as many databases as they'd like [20:18:47] and now i was about to raise the IP limit and add one public IP [20:18:55] mutante: cool [20:19:01] creation of db is ok, but creation of users isn't [20:19:15] you need a service users for bots etc [20:19:18] deployment-prep? [20:19:27] yes [20:19:28] Ryan_Lane: hexmode tell you more [20:19:30] petan: yep. that would be ideal [20:19:49] regarding deployment [20:20:02] creation of dbs, ownership of their own dbs, giving permissions to the dbs to ldap users [20:20:11] creating users [20:20:14] what about non ldap users [20:20:15] ok [20:20:15] < hexmode> mutante: one IP needed to test deployment of 1.19 [20:20:15] Ryan_Lane: fwiw, I can't get to bastion w/ mah@... and my key [20:20:16] *local users [20:20:20] key looks right [20:20:35] are you in the project? [20:20:38] yes [20:20:44] he's in bastion and even in deployment [20:21:19] actually I don't know if he's on bastion :p [20:21:19] he's not in bastion project [20:21:23] heh [20:21:24] as far as I can tell [20:21:25] true [20:21:27] that's it [20:21:34] I thought everyone is there [20:21:51] hexmode: what's your wiki name again? [20:22:00] Mark*Hersh* [20:22:03] ah. right [20:22:03] MarkAHershberger [20:22:03] MarkAHershberger [20:22:06] that :) [20:22:12] easy to type ;) [20:22:36] !log deployment-prep raised floating IP quota to 1 [20:22:37] Logged the message, Master [20:23:05] hexmode: which dn you want? [20:23:17] deployment.wmflabs.? [20:23:23] is good [20:23:29] right [20:23:50] setting up [20:24:23] !log deployment-prep allocated ip208.80.153.215 [20:24:24] Logged the message, Master [20:24:25] labs-home-wm: ...? [20:24:36] labs-home-wm: shouldn't you be adding mah's key? [20:25:00] petan: ah,ok, so that part you could do already, nice [20:25:40] !log mobile upped the quota for floating ips to 1 [20:25:41] Logged the message, Master [20:26:01] !log deployment-prep registered deployment.wmflabs.org [20:26:02] Logged the message, Master [20:26:10] nice to watch:) [20:27:03] !log mobile allocated IP 208.80.153.216 [20:27:04] Logged the message, Master [20:27:10] !log deployment-prep opened port 80 for wide net [20:27:11] Logged the message, Master [20:27:44] preilly: ok. allocated an ip for your project [20:28:00] preilly: you can associate it, then add a DNS entry for it [20:28:05] http://deployment.wmflabs.org/ [20:28:08] hexmode: ^ [20:28:24] http://deployment.wmflabs.org/simple_wiki/w/index.php/Main_Page [20:29:01] w00! [20:29:04] tyvm [20:29:10] preilly: please log actions via !log mobile [20:30:09] 01/04/2012 - 20:30:09 - Updating keys for brion [20:30:46] !access | hashar [20:30:46] hashar: https://labsconsole.wikimedia.org/wiki/Access#Accessing_public_and_private_instances [20:30:57] yeah that [20:31:01] Ryan_Lane, is there any delay on key updates? bastion.wmflabs.org is rejecting my key [20:31:02] Ryan_Lane: minor bug i think: if you create a new project and then log the very first message to that project name the bot says it logged it but does not appear on the wiki page (because it first needs to create the page) [20:31:03] !log mobile associated mobile-feeds host name on wmflabs.org domain [20:31:05] Logged the message, Master [20:31:23] hashar: you can either 1. forward your agent, or 2. set up proxycommand [20:31:34] brion: a slight one. maybe a minute or two [20:31:35] Ryan_Lane: perfect! That force me to actually update my keys and automatize all that ssh key loading / connection [20:31:50] hexmode: Special:Statistics [20:31:54] there you see progress [20:32:04] brion: lemme make sure you are in the project [20:32:13] ah handy [20:32:20] Krinkle: Ryan provided that link : https://labsconsole.wikimedia.org/wiki/Access#Accessing_public_and_private_instances [20:32:28] Krinkle: ssh password is disabled :-D [20:32:31] hmm. you're in the project [20:32:34] heh [20:32:35] lemme check your home directory [20:33:03] hexmode: http://deployment.wmflabs.org/hi_wikipedia/w [20:33:06] that's the second one [20:33:11] awright i'll fiddle with this later [20:33:13] thanks Ryan_Lane :D [20:33:19] hmm. you don't have a home directory [20:33:24] * Ryan_Lane goes to check the scripts [20:36:12] 01/04/2012 - 20:36:12 - Updating keys for mah [20:37:53] brion: ok. you should be good now [20:38:22] hexmode: are the extensions ok or you also want to test head? [20:38:31] I took the revision we have in production [20:38:35] or I think so [20:38:51] only mediawiki is head [20:39:13] !access | MaxSem [20:39:13] MaxSem: https://labsconsole.wikimedia.org/wiki/Access#Accessing_public_and_private_instances [20:39:25] that page will show you how to forward your key, or use proxycommand [20:41:21] hmm, my connection settings had not changed [20:41:30] lemme dig it for a while [20:43:47] Ryan_Lane: got it!!! You can a add a prelogin banner : http://dpaste.org/vhVzW/ [20:44:11] I know I had setup that one day [20:44:24] ah. cool [20:50:54] New patchset: Ryan Lane; "Adding in a pre-login banner for labs" [operations/puppet] (test) - https://gerrit.wikimedia.org/r/1782 [20:51:25] New review: Ryan Lane; "(no comment)" [operations/puppet] (test); V: 0 C: 2; - https://gerrit.wikimedia.org/r/1782 [20:51:25] Change merged: Ryan Lane; [operations/puppet] (test) - https://gerrit.wikimedia.org/r/1782 [20:52:37] btw Ryan_Lane what is current status of storage usage we have for labs? [20:52:50] I mean physical storage [20:52:56] not what I see on wiki [20:52:57] kind of crap :) [20:53:01] heh [20:53:09] hardware order was just approved [20:53:11] maybe I could kill some instances on bots [20:53:15] oh [20:53:21] well, there's two kinds of storage [20:53:25] volume and instance [20:53:42] we'll want to move most of the storage to volume storage at some point [20:53:50] ok [20:53:53] volume storage is what we are ordering [20:54:02] that's the /mnt [20:54:05] on instances [20:54:12] instance storage is /mnt and / [20:54:18] yes [20:54:23] volume storage will be per-project [20:54:26] ah [20:54:27] ok [20:56:39] Ryan_Lane: if we had server for sql I could kill all sql instances in bots [20:56:56] that would free 6gb of ram and 120 gb of storage [20:57:06] yeah [20:57:15] a server for SQL is going to take a while [20:57:17] well, kind of [20:57:23] we can manually set up databases... [20:57:25] also I would kill other sql's from wiki, that would free 4 gb of ram too [20:57:54] maybe we could create a project on labs and start setting it up there, once you get a hardware we just move it there [20:58:06] yeah. that sounds good [20:58:06] so there would be one big instance which would contain all databases we use now [20:58:12] writing the service is going to be the hard part [20:58:18] we could start now [20:58:22] yeah [20:58:29] ideally it would be an openstack style project [20:58:38] using eventlet, with an API, written in python [20:58:45] ok [20:59:03] I think others are working on a database as a service product [20:59:09] red dwarf is one [20:59:19] they are reusing most of nova for it, though.... [20:59:24] I'm not a huge fan of that [20:59:40] maybe we should fix that proposal now, to make it clear what are we going to prepare [20:59:51] others would be probably interested in that [21:00:00] we could write a proposal up on the wiki [21:00:00] petan: will return in a bit... was out and now have to eat... c u tmw if you aren't here when I'm back? [21:00:07] work on it there, then publish a blueprint to openstack [21:00:16] sure that's what I meant [21:00:30] I'd like to make sure we have our ideas together before publishing the blueprint [21:00:31] hexmode: ok [21:00:36] since it would be a competing project [21:00:41] right [21:13:20] damn, the banner shows up even when login succeeds [21:13:25] oh well, better than nothing [21:15:10] New patchset: Ryan Lane; "Clean up banner content" [operations/puppet] (test) - https://gerrit.wikimedia.org/r/1784 [21:15:34] New review: Ryan Lane; "(no comment)" [operations/puppet] (test); V: 0 C: 2; - https://gerrit.wikimedia.org/r/1784 [21:15:35] Change merged: Ryan Lane; [operations/puppet] (test) - https://gerrit.wikimedia.org/r/1784 [21:20:26] Change on 12mediawiki a page Wikimedia Labs was modified, changed by Ryan lane link https://www.mediawiki.org/w/index.php?diff=481082 edit summary: /* Proposals */ [21:21:18] oh. wow. cool [21:27:15] heh [21:27:23] hopefully someone notice that pages are being updated [21:27:49] hee ryan, best wishes for 2012! [21:27:57] yay [21:29:51] drdee: thanks :) [21:30:05] just deployed new version of OpenStackManager and LdapAuthentication on labsconsole, btw [21:30:08] check out the new interface :) [21:30:14] ok [21:30:25] projects are collapsible if you click their project name [21:30:41] also, and this one is awesome… you can manage your own puppet classes and variables [21:30:49] per-project [21:30:53] !group is https://labsconsole.wikimedia.org/wiki/Special:NovaPuppetGroup [21:30:53] Key was added! [21:30:53] if you have the sysadmin role [21:30:57] cool [21:31:28] is it possible to export it to others project [21:31:49] I guess that's what All is for :) [21:32:03] to have a global option... [21:32:25] I wanted to make interface for nagios so that everyone can check what services they want to monitor [21:32:50] nagios is getting a config from wiki [21:33:50] so, it's easy to use it [21:40:29] well, you must be a member of sysadmin in the project to modify its puppet group [21:40:32] *groups [21:41:15] yay! I got onto bastion finally [21:42:54] the home directory script was broken :( [21:43:04] it's fixed now [21:43:10] hexmode: cool [21:43:13] but, I have a bug in the script that I need to fix [21:43:28] hexmode: you can test anything on wiki but you should avoid changing db layout [21:43:42] because it's still loading data to it [21:43:56] np [21:44:01] hexmode: also create a better index page :P [21:44:14] http://deployment.wmflabs.org/index.html [21:44:54] oh, that one is perfect ;) [21:44:57] hehe [21:45:08] although, maybe a small wiki for that... [21:45:12] anyway [21:45:20] sure, no problem [21:45:54] hexmode: so, is this going to have full copies of these databases, or small recursive exports? [21:46:00] so I'm on bastion, but can't ssh elsewhere? trying ssh deployment, deployment.pmtpa ... [21:46:06] is this a replacement for prototype? [21:46:15] atm it's running full import of pages on simple [21:46:26] no other tables, etc [21:46:27] Ryan_Lane: I think we'd like to get it to be a replacment, yes [21:46:42] on prototype we have small imports of the content [21:46:43] not full [21:46:47] so it'll help you guys automate labs setup :) [21:46:52] sounds good [21:46:59] hm... hex wanted a full import, I think :o [21:47:08] i'm still not getting into bastion :( [21:47:15] just a warning: 1.2T 210G 877G 20% /var/lib/nova/instances [21:47:15] Ryan_Lane, any tricks for usernames with spaces? [21:47:24] yay [21:47:55] hexmode: if you don't need a full import you can just kill the running import, it would be ok [21:47:58] petan: I did, but just because it was going to be used for testing and didn't know what people would want [21:48:04] ok [21:48:17] I don't think it's a problem we can delete db later... [21:48:20] how do I ssh to deployment? [21:48:26] so, if /var/lib/nova/instances hits 100% we're fucked [21:48:29] ssh deployment-web / sql [21:48:36] two instances [21:48:39] brion: your shell name is different than wiki name [21:48:42] on web is running apache and import [21:48:42] k [21:48:46] on sql only db [21:49:13] 01/04/2012 - 21:49:13 - Creating a project directory for embed-sandbox [21:49:14] 01/04/2012 - 21:49:13 - Creating a home directory for brion at /export/home/embed-sandbox/brion [21:50:13] 01/04/2012 - 21:50:13 - Updating keys for brion [21:52:15] 01/04/2012 - 21:52:14 - Creating a project directory for ryanlandsucks [21:53:34] huh [21:53:45] ryanlandsucks? [21:53:47] :D [21:54:04] heh [21:54:13] people can create their own projects now [21:54:23] not sure if I'm going to keep it that way or not. [21:54:26] we are all prisoners in ryanland [21:54:41] cool [21:54:48] I'm thinking I'm gonna take it away.... [21:55:22] so... is there a place I should go for "first steps" or the like [21:55:36] a faq for "WHY CAN'T I USE SUDO???" [21:55:41] you can [21:55:42] you can use sudo [21:55:46] just not on bastion [21:55:49] to create a new instance in my project, would i need additional permissions? i assume yes as i can't find 'create a ne winstance' on the 'manage instances' page :) [21:55:53] !bastion [21:55:53] ssh user@bastion.wmflabs.org [21:55:56] ah [21:56:15] brion: you need to be net admin [21:56:19] !access [21:56:19] https://labsconsole.wikimedia.org/wiki/Access#Accessing_public_and_private_instances [21:56:39] https://labsconsole.wikimedia.org/wiki/Special:NovaProject#embed-sandbox says i am the netadmin [21:57:15] in that case you should be able to do that [21:57:23] maybe try to logout and back [21:57:29] volume creation has a link but tells me i can't do it as i must be a member of the sysadmin role: https://labsconsole.wikimedia.org/w/index.php?title=Special:NovaVolume&action=create&project=embed-sandbox [21:57:34] already tried logging out [21:57:43] hm... [21:57:43] I need to remove the volume link... [21:57:49] oh [21:57:50] I know [21:57:58] fucking nova and its stupid intersecting groupos [21:58:18] brion: I see the link next to name of project, before it was under [21:58:25] brion: ok. you're good to go now [21:58:29] hm... [21:58:48] i see it now! thx [21:58:50] we aren't using volumes in the normal ec2/openstack way, so that interface won't work [21:59:10] * Ryan_Lane removed it from the sidebar [22:01:13] so... what is my sudo password? My wiki one doesn't seem to work [22:01:20] wiki one [22:05:53] probably too hard [22:05:56] !password [22:05:57] gfgjoagaewhgAW#YAU_#Y$*U*U^*^%Q#Tqyhe [22:06:01] that's recommended [22:06:08] hahahaha [22:07:29] woohoo! i can log into my embed-sandbox instance :D [22:07:32] no more project creation without having cloudadmin roll :) [22:07:37] *role [22:07:43] bag [22:07:46] * bah [22:07:49] that was a bad oversight. heh [22:07:55] well, I thought about it, and it's a security hole [22:07:57] * petan has no roll :P [22:08:02] heh [22:08:24] security hole? [22:08:25] since anyone can create their own project, then give themselves sysadmin and create instances [22:08:35] I thought it was mentioned for that [22:08:50] what it would be for if they couldn't do that [22:09:16] just cause you have an account doesn't mean you should be able to create instances :) [22:09:21] now all i need's a public ip ;) [22:09:27] at some point I'd like anyone to be able to create an account [22:09:33] !rt does it work yet? [22:09:33] http://rt.wikimedia.org/Ticket/Display.html?id=does [22:09:38] for labs? [22:09:45] :o [22:09:58] we really need to set up a proxy :D [22:10:08] I can't login there [22:10:18] sudo works... had to figure out my web password though.... [22:10:25] heh [22:10:37] you can save it to bot :) [22:10:48] but I know more secure ways too [22:13:28] petan: thanks! I'm making a similar setup on one of my windows boxes to allow cross platform testing. [22:13:42] ok [22:14:27] brion: ok, you can allocate an ip, associate it with your instance, then add a dns address for the address via: https://labsconsole.wikimedia.org/wiki/Special:NovaAddress [22:14:33] PROBLEM Total Processes is now: CRITICAL on embed-sandbox embed-sandbox output: CHECK_NRPE: Error - Could not complete SSL handshake. [22:15:23] PROBLEM dpkg-check is now: CRITICAL on embed-sandbox embed-sandbox output: CHECK_NRPE: Error - Could not complete SSL handshake. [22:15:29] I may need a couple of small tasks in the next few days [22:15:40] right [22:15:51] I think that jeremy is also working on search [22:16:00] OrenBochman: if you needed to install anything there let me know [22:16:03] PROBLEM Current Load is now: CRITICAL on embed-sandbox embed-sandbox output: CHECK_NRPE: Error - Could not complete SSL handshake. [22:16:13] but I think its ready for my work [22:16:23] check the extensions etc. [22:16:30] will do [22:16:43] PROBLEM Current Users is now: CRITICAL on embed-sandbox embed-sandbox output: CHECK_NRPE: Error - Could not complete SSL handshake. [22:17:23] PROBLEM Disk Space is now: CRITICAL on embed-sandbox embed-sandbox output: Connection refused by host [22:18:13] PROBLEM Free ram is now: CRITICAL on embed-sandbox embed-sandbox output: CHECK_NRPE: Error - Could not complete SSL handshake. [22:26:43] RECOVERY Current Users is now: OK on embed-sandbox embed-sandbox output: USERS OK - 1 users currently logged in [22:27:23] RECOVERY Disk Space is now: OK on embed-sandbox embed-sandbox output: DISK OK [22:27:40] http://embed-sandbox.wmflabs.org/ sweeeeeet [22:28:13] RECOVERY Free ram is now: OK on embed-sandbox embed-sandbox output: OK: 91% free memory [22:29:33] RECOVERY Total Processes is now: OK on embed-sandbox embed-sandbox output: PROCS OK: 86 processes [22:30:23] RECOVERY dpkg-check is now: OK on embed-sandbox embed-sandbox output: All packages OK [22:31:03] RECOVERY Current Load is now: OK on embed-sandbox embed-sandbox output: OK - load average: 0.00, 0.07, 0.07 [22:34:03] PROBLEM Free ram is now: WARNING on deployment-web deployment-web output: Warning: 17% free memory [22:35:37] * jeremyb looks up [22:35:53] OrenBochman_: how goes it? [22:39:59] jeremyb: it's fully operational now [22:40:10] but service is still not running there [22:40:13] java [22:41:59] * jeremyb was confused by having the bots-apache1 IP be exactly 1 bit difference from upload.pmtpa.wikimedia.org :P [22:42:12] heh [22:42:17] that's possible [22:42:26] $ for i in 2 3; do ipcalc 208.80.15$i.211 | head -n 1; done [22:42:27] Address: 208.80.152.211 11010000.01010000.10011000. 11010011 [22:42:27] Address: 208.80.153.211 11010000.01010000.10011001. 11010011 [22:43:02] same network [22:43:13] but it's fun [22:43:42] same network? [22:44:25] different subnet but same [22:44:49] my bot was loading 100 pages in a second from wikipedia on bots-1 [22:44:55] it's a /22 allocation [22:47:15] Ryan_Lane: Hey, got a question about OSM [22:47:39] Why are you creating empty classes that extend HTMLForm? [22:47:59] petan: can you setup the wiki to keep all the pages in the cache for ever ? [22:49:29] petan: I'm going adding indexing of html to the indexer and would like to run dump to fill out the cache [22:50:35] OrenBochman: in which cache? [22:50:47] I want to deploy memcached there but dunno if it's in puppet [22:51:04] browser cache or cache on server? [22:51:14] petan - I'm not sure myself [22:51:15] on production we have squid and memcached [22:51:17] server [22:51:28] squid is just a proxy cache [22:51:36] I think memcached since it a single server [22:51:37] memcached is a feature used by php [22:51:52] I can set it up but I think puppet is better for that [22:52:05] jeremyb: do you know if memcached is puppetized? [22:53:20] see what you can do - dev won't take long but rendering all that HTML will take about week [22:53:45] I guess I'll need the dumpHtml extentions [22:55:37] petan: Yes, memcached is puppetized [22:55:44] manifests/memcached.pp in the puppet repo [22:56:17] petan: it is i think [22:56:49] OrenBochman: can you check it? [22:56:55] memcached is certainly *not* forever [22:56:56] open instance set up and check memcached [22:57:10] why not just use nginx/varnish? [22:57:12] jeremyb: I set it up to be a year [22:57:17] no [22:57:24] on my windows machine [22:57:26] memcached flushes when the process stops [22:57:39] you guys don't want to wait a year [22:57:42] :D [22:58:01] I'll have a look [22:58:19] you start with a cold cache every time. no persistence [22:59:27] both memcache and squid are in puppet [22:59:46] jeremyb: I'm not sure you are correct [22:59:59] I am sure [23:00:13] no doubts [23:00:18] Indeed [23:00:26] restart memcached, it'll be empty [23:00:35] it has no persistance whatsoever [23:00:35] Reedy is sure too :) [23:01:16] where is it configured ? [23:01:32] there's also pcache which is in mysql in prod and is therefore persistent [23:01:37] Depends on your os [23:01:50] it's not configured to be empty on start [23:01:55] you don't have a choice [23:02:05] http://en.wikipedia.org/wiki/Memcached [23:02:34] "A Memcached-protocol compatible product known as MemcacheDB provides persistent storage." [23:02:40] but i think what you want is not memcached anyway. better to go with nginx/varnish (or squid) [23:02:47] ready I need a cache that would store a copy of the rendered html for unit and regression testing [23:03:05] memcached is more of an object store [23:03:07] * jeremyb keeps repeating himself [23:03:10] it could work [23:03:18] you don't want memcached. you want nginx [23:03:24] but jeremyb is right, something like varnish/squid/nginx etc makes more sense [23:03:48] so lets use squid - its in puppet [23:03:57] and its in production [23:03:59] nginx is also in puppet. we can copy from ms6 [23:04:05] as is varnish [23:04:43] but you could just store them as a file if it's for unit testing [23:04:51] dudes: I was informed that varnish is for bineries [23:04:59] .. [23:05:09] bits.wikimedia.org doesn't serve binaries [23:05:28] http://en.wikipedia.org/wiki/Varnish_(software) [23:05:29] * jeremyb has to hit the road for a it [23:05:31] bit* [23:05:52] let me know if you make up your mind (or i can just make it up for you if you prefer) [23:05:55] bbiab [23:06:36] anyhow I'd like to converge towards the cloud setup - I belive it's squid for non logged in pages [23:06:43] right [23:07:12] so I'll prefer squid - since that's what will be used in production [23:07:37] is it though? [23:07:53] does search hit squids or bypasses somehow? [23:08:11] or does it set a cache-control header to never use the cache? [23:08:12] I think you should install memcached anyway just to get production environment [23:08:19] sure [23:08:27] you should have all caches we have there [23:08:35] but in that case we should let puppet install apache not do it manually [23:08:46] anyway, really bbiab [23:08:50] me too [23:10:01] dudes: we can have both I just don't want to run out of memory or hardrive space - since there will be (multiple ) indexes being created [23:11:10] Reedy: how would yoou suggest to stash all the files to a file for unit testing ? [23:12:41] OrenBochman: memcached can be on another instance [23:12:54] so no problem with resources [23:13:00] If they're static, have them as files [23:13:15] storing them in some non persistant cache seems strange and a needless dependancy [23:16:47] Reedy: I would rather get them off the server but MediaWiki is too slow - cache is where the rendered pages exist in production [23:18:44] I'll be doing both the first time anyhow - perhaps I will stash the html into the xml dump. [23:19:43] it would not be so simple under windows but on linux it could be ok [23:30:17] johnduhart: I have no clue :) [23:30:34] Ah, mind if I change that? :() [23:30:37] :) * [23:30:38] johnduhart: if there's a better/easier way of doing that, fix it [23:30:42] sure [23:30:55] No need to make a new class, just make a new HTMLForm instance [23:30:58] btw johnduhart I inserted you to bots [23:31:00] ah [23:31:05] petan: I saw, thanks