[15:00:08] OHHHH [15:00:20] no standup right now [15:00:20] doh [15:09:32] * average is waiting for that RAM memory delivery [15:20:14] somehow I doubt there is story tasking meeting today [15:30:49] Me hoo. [15:30:56] s/hoo/too/ [15:31:07] I am in the Batcave felling lonely :-) [15:31:13] s/felling/feeling/ [15:31:27] * qchris needs some spell-auto-correction :-( [16:18:36] hey ottomata [16:18:41] hey qchris [16:18:46] Hi average [16:18:56] 16G of RAM !!! damnit [16:19:04] works [16:19:10] hiii! [16:19:14] one last question, to top everything up [16:19:25] i feel like I am going down a long rabbithole of python logster ganglianess [16:19:33] doing a lot of cool work, but for very little gain [16:19:35] ottomata: is an SSD running on a SATA3 controller slower/faster than a ramdisk ? [16:19:46] psshs, iunno? [16:20:10] I'm asking because if an SSD < ramdisk [16:20:20] then I could just get lots of RAM and throw VMs in it [16:20:27] and have them run there instead of the SSD [16:20:41] someone in #vagrant says a ramdisk would be faster [16:20:48] I'm not sure [16:22:03] sounds likeley, i don't know much about SATA 3 speeds or whatever, but generally ram is faster than SSD [16:22:10] Typically, RAM will give the fastest access path. [16:23:07] SATA3 is 6Gbps/s . If that doesn't tell you much(I know it doesn't tell me much) I can tell you that I've read an SSD can do 450MB/s write and 550MB/s read on SATA3. And, also, the same SSD, will do 220MB/s write and 250MB/s read on SATA2. [16:23:21] the guys in #vagrant are saying that a ramdisk would be 10-20 times faster than an SSD on SATA3 [16:25:01] so gargantuan amounts of RAM => make ramdisk => mount ramdisk => throw VMs in ramdisk => huge speed improvement [16:26:02] :-) [17:06:14] average: standup :-) [17:34:41] missed it [17:34:48] oh well.. was trying to fix some stuff [17:52:50] oops, sorry, forgot my IRC [17:52:56] I'm a mess when I'm out here :) [19:02:06] changing locs, back in a bit [19:08:31] average, milimetric: I guess as no one shows up, we're not having the Sprint Planning Meeting. Right? [19:29:37] milimetric: you guys doing meeting? [19:29:59] qchris, ottomata: I think Toby rescheduled the retrospective for right now [19:30:09] calendar sayss spring planning meeting [19:30:13] i know! [19:30:24] calendar is currently being changed [19:30:25] i just want to talk about logster stuff :p [19:30:30] I know that too :( [19:30:55] average, milimetric: I guess as no one shows up, we're not having the Sprint Planning Meeting. Right? [19:31:41] ^ (forgot time stamp ... that was 30 minutes ago) [19:32:54] yes qchris, sprint planning has been moved to monday it looks like [19:33:11] I will just say this now: [19:33:28] I am going to *completely* screw up any meeting / scheduling / timing thing for the next few days [19:33:38] :-D [19:33:43] it's *very* hard for me to adjust to people having access to my physical dimension [19:34:00] Where are we having the meeting now. Batcave looks empty? [19:34:09] yea, batcave [19:34:35] ottomata, nuria, average: batcave for retrospective [19:34:46] Wait ... retrospective? [19:59:08] Folks, question: Is it possible to get node.js on stat1? [20:01:34] Also, is there an established place to stick data-crunching scripts? [20:02:04] And finally, is there a nice-ish way in Limn to show a distribution graph? [20:03:31] marktraceur: 1) yes, you can make a change in gerrit for puppet [20:03:52] Smexy. /me does [20:05:11] marktraceur: 2) wikistats for example sits in /a/wikistats_git/ . But I don't know know of any established place for that. But consider looking into the puppet manifests and maybe ask ottomata [20:05:27] marktraceur: 3) you can contact milimetric about that [20:05:42] marktraceur: 1, maybe! [20:05:44] what do you ned it for? [20:06:08] ottomata: I'm writing a script to crunch through some EventLogging data on stat1, then produce CSVs for limn [20:06:33] I can submit the gerrit change for ops/puppet no problem [20:06:50] have oyu seen limnpy? [20:06:52] Is it the analytics role or the statistics role? Hm... [20:07:05] https://github.com/wikimedia/limnpy [20:07:08] ottomata: I mean, I've written it basically already [20:07:48] My tutorial from DarTar basically went "generate CSVs, I have Bash scripts, you can do whatever" [20:08:44] I may use this to produce dashboard files though [20:08:52] milimetric_: might have more thoughts in a bit, but we are in a meeting right now [20:08:53] Because we'll (hopefully) have per-wiki dashboards very soon [20:08:59] i *think* installing node.js on stat1 should be fine [20:09:53] i would ask another opsen to be sure, but I'm ok with it [20:10:18] 'kay [20:11:26] make a class in misc/statistics.pp [20:11:29] misc::statistics::nodejs [20:11:30] or whatever [20:11:33] include whatever packages you need [20:11:43] and then include that new class in role/statistics.pp role::statistics::cruncher class [22:05:30] milimetric: looks like the analytics project is in the new mingle instance! everyhting go OK? [22:18:16] awjr: all's good except I'm not sure about Bingle [22:18:31] milimetric: do you have an account on tools labs? [22:18:38] yes, I do [22:18:43] username? [22:20:00] milimetric [22:20:02] :) [22:20:38] that is easy :) [22:25:49] Grrr ganglia! [22:25:50] laters all! [22:26:59] milimetric: ok, you are added to the bingle group [22:27:13] cool [22:27:13] did you get my email yesterday with instructions on what to change in the bingle config? [22:27:30] yes but was flying to sf [22:27:37] ok good [22:27:45] will look now [22:27:48] ty! [22:27:53] np :) [22:28:00] i'll ping with any problems [22:28:05] so if you get to the tools labs bastion, 'become bingle' will get you in [22:28:37] the analytics bingle config is in /data/project/bingle/var/bingle-analytics/bingle.ini [22:28:48] there's some bingle documentation on wikitech: https://wikitech.wikimedia.org/wiki/Bingle [22:28:53] holler if you need any help :) [22:36:58] oh awjr, one other thing: nuria just joined and I don't know how to add her to mingle [22:38:08] where is our "mingle"? [22:38:10] milimetric: you are an administrator of the overall mingle account so you have sufficient privileges for this [22:38:20] milimetric: to do it, go do to the main apge of the mingle instance (https://wikimedia.mingle.thoughtworks.com/) [22:38:21] Did it get migrated? [22:38:22] oh really? ok [22:38:35] then admin -> users [22:38:41] she's not in the system at all as far as I can tell but if I have rights to the whole system, I'll poke around [22:38:42] then click 'new user' [22:38:44] oh ok [22:39:09] after you've added her there, you'll have to had her user to the analytics projects [22:39:11] would you hate me if I made an "analytics bingle" user? [22:39:35] projects -> analytics -> project admin -> team members [22:39:45] milimetric: no, but there's already a bingle user for mingle [22:39:50] do you need another one? [22:39:54] right now it's just set to diederik's account [22:40:01] OH [22:40:11] milimetric: take a look in /data/projects/bingle/var/bingle.ini [22:40:26] copy the username/pass for the mingle bingle user [22:40:31] ok, cool [22:40:32] thx [22:40:36] you'll also need to add the user to the analytics projects [22:40:38] yep :) [22:40:45] AND make sure to make it a project administrator [22:41:13] it's a quirk (flaw) in the mingle api - in order for a user to create/update stuff via the api, they have to have administrator access [22:41:45] yep, I think I remember that's why diederik just used his account [22:41:47] thanks awjr! [22:42:44] np milimetric [22:50:56] hm, awjr, does something need to be restarted? It doesn't seem to be picking up my test bug [22:51:18] milimetric: did you run it manually or are you waiting for the cronjob to run? [22:51:41] just waiting [22:51:57] oh it's 5 minutes [22:52:01] k, sorry :) [22:52:30] np :) [22:52:34] you can try running it manually too