[15:11:15] <average_drifter>	 hashar: hey Antoine
[15:11:29] <average_drifter>	 hashar: can you help me get access on stat1001 please ?
[15:11:40] <hashar>	 I have no way to do that
[15:11:48] <average_drifter>	 hashar: who should I talk to ?
[15:11:49] <hashar>	 you want to ask in #wikimedia-operations :-D
[15:12:03] <hashar>	 mark paravoid and apergos would be the one to talk to
[15:12:58] <average_drifter>	 hashar: I wrote to them
[15:13:29] <hashar>	 nice :-]
[15:15:15] <drdee>	 morning average_drifter, milimetric, ottomata
[15:15:24] <milimetric>	 morning drdee
[15:15:33] <drdee>	 aight
[15:15:44] <drdee>	 a little bit behind schedule but hey it's my first attempt
[15:15:45] <average_drifter>	 hey drdee ! :)
[15:15:49] <milimetric>	 how's the tiny one
[15:15:50] <milimetric>	 :)
[15:16:06] <drdee>	 a happy pumpkin as she was just fed
[15:16:37] <drdee>	 average_drifter: did you and ottomata deploy webstatscollector?
[15:16:59] <average_drifter>	 drdee: ottomata informed me yesterday that webstatscollector had not crashed and it produced data properly
[15:17:12] <average_drifter>	 drdee: but I'm not sure as if he deployed it into production
[15:17:22] <drdee>	 k
[15:17:30] <drdee>	 sounds like a good start
[15:19:30] <average_drifter>	 :)
[15:19:43] <average_drifter>	 I'm looking for ottomata for access on stat1001 to deploy editor wikistats
[15:19:58] <average_drifter>	 I contacted #wikimedia-operations in the meantime
[15:20:17] <drdee>	 why not run that on stat1 just like wikistats?
[15:21:09] <average_drifter>	 drdee: that is possible, but the page itself stats.wikimedia.org sits on stat1001 ( as I was informed by Erik )
[15:22:16] <average_drifter>	 drdee: Erik says there's an rsync taking data from stat1 => stat1001
[15:22:32] <drdee>	 yes so it should be hosted on stat1001
[15:22:38] <drdee>	 but the crunching needs to happen on stat1
[15:22:49] <drdee>	 so ottomata can just add the editors content to that rsync
[15:23:11] <average_drifter>	 drdee: Erik says he's gonna show me an example of the rsync
[15:23:15] <drdee>	 k
[15:23:33] <drdee>	 is ottomata joining us today or is he already enjoying his long weekend?
[15:23:43] <average_drifter>	 drdee: I'm not yet 100% whether the flow is   stat1 => stat1001   or  viceverse (if the latter, I will need access to stat1001 which I'm talking about right now with mark on #wikimedia-operations )
[15:24:02] <drdee>	 the flow is stat01 --> stat1001
[15:33:37] <average_drifter>	 drdee: ok
[15:33:53] <average_drifter>	 drdee: I talked more with Erik and it seems I will need access to stat1001 to be able to run the scripts that do rsync
[15:34:06] <average_drifter>	 drdee: mark on #wikimedia-operations asked me to create ticket
[15:34:13] <average_drifter>	 drdee: I sent an e-mail which creates the ticket
[15:34:19] <average_drifter>	 drdee: I will need your approval please for this
[15:34:20] <drdee>	 let's talk with ottomata first
[15:34:29] <drdee>	 got your email
[15:34:31] <drdee>	 thx
[15:55:46] <drdee>	 average_drifter: ping
[16:04:09] <average_drifter>	 drdee: here
[16:04:37] <average_drifter>	 talking with Erik on skype
[16:04:40] <drdee>	 okay, have there been any pushes to wikistats
[16:04:41] <drdee>	 ?
[16:04:43] <drdee>	 oh okay
[16:04:45] <drdee>	 shall i join?
[16:05:14] <average_drifter>	 drdee: there's a review in progress with some things that need corrected
[16:06:04] <drdee>	 k
[16:50:17] <average_drifter>	 brb 30m
[16:54:42] <drdee>	 k
[17:58:43] <milimetric>	 https://plus.google.com/hangouts/_/2e8127ccf7baae1df74153f25553c443bd351e90
[18:00:38] <milimetric>	 drdee, erosen, average_drifter, ^^
[18:00:50] <drdee>	 still talking to robla, coming soon
[18:03:01] <drdee>	 coming
[18:32:51] <DarTar>	 folks, I have to skip today's analyst scrum - (hopefully final) follow-up doctors appointment for my wrist in 20 mins
[19:12:37] <dschoon>	 yawn.
[19:12:47] <dschoon>	 don't we have days off this week at some point?
[19:15:45] <drdee>	 tomorrow and friday
[19:15:52] <dschoon>	 we have friday off?
[19:15:53] <dschoon>	 huh
[19:15:55] <dschoon>	 cool.
[19:16:02] <drdee>	 so you can shop
[19:16:11] <dschoon>	 oh! good point.
[19:16:37] <dschoon>	 i definitely need some senseless consumerism to fill this hollow void in my life where my heart should be
[19:16:45] <dschoon>	 maybe i could get some shoes~
[19:16:46] <drdee>	 or in your case struggle with comcast :D
[19:27:32] <dschoon>	 sigh.
[19:38:43] <binasher>	 yo, has andrew otto been around at all today?
[19:55:28] <drdee>	 no, he is taking a day off, can i help?
[19:55:32] <drdee>	 binasher ^^
[19:57:44] <binasher>	 drdee: hey, i wanted to ask him about if he'd like having the /event udplog traffic sent directly to analytics1001
[19:58:04] <binasher>	 but i should probably wait til monday to make any changes with that anyways
[19:58:29] <drdee>	 i think that's okay, dschoon what's your opinion?
[20:27:56] <dschoon>	 my opinion is that lunch was great.
[20:27:59] <dschoon>	 (reading)
[20:33:33] <dschoon>	 i think the answer is that we should wait to answer until otto is back
[20:34:29] <dschoon>	 our public IP points at a cisco, and they have ~3.6T in storage
[20:35:00] <dschoon>	 we want to run the kafka brokers on dells because they have ~24T
[20:35:15] <dschoon>	 that gives us enough headroom to keep a sliding window of a week or so in logs
[20:35:56] <dschoon>	 so pointing the whole stream at a cisco would either mean routing it internally afterward, which hurts throughput overall
[20:36:18] <dschoon>	 or moving the public ip
[20:37:10] <dschoon>	 though i'm not all THAT worried about net, it would really, really suck to have to upgrade
[20:37:21] <dschoon>	 (switches, interfaces, routers)
[20:37:36] <dschoon>	 so i lean toward moving the ip, and that means we should wait for otto
[20:37:40] <dschoon>	 ^^ binasher
[20:49:56] <binasher>	 dschoon: there would still be internal routing of all the traffic though, wouldn't there? or a significant portion
[20:50:15] <dschoon>	 of course.
[20:50:28] <dschoon>	 but hadoop is notorious for gobbling up net
[20:50:37] <dschoon>	 just trying to be conservative about a known-scarce resource