[08:31] any eta on getting more dedis? [10:24] Wasn't the ETA yesterday? :P [10:26] up_the_irons? toeshred? [10:29] they are no longer on hold as far as i know. but i would still check with up_the_irons to confirm. [10:44] Thanks toeshred :) [10:45] I figured the guy to [probably] rack them would know ;) [11:14] rackin n stakin! [11:14] pepwepwpp [12:06] yeah the sold out sign is no longer there [12:29] qbit: ^ [12:29] oh sweet [12:29] i was waiting for a button to show up [12:29] ?:D [12:29] s/?// [15:10] question for you guys. I'm giving a presentation on HAProxy to a local group in october. I kinda want to have a little bit in there for everyone, basic to advanced. the time slot is flecible between 45 - 90 minutes. what kinds of things would YOU guys be interesting in knowing? [15:12] for basic stuff, I was planning on going over stats, load balancing, how it does "HA" and health checks for backend servers. for more advanced, I'd like to cover throttling, DDoS, network/latency diagnostics, load balancing things like mysql, redis, postgres, openvpn, etc. eliminating haproxy as a spof.. those kinds of things.. [15:16] oh, ssl termination, and http vs tcp mode. that'll be towards the end of the beginner part. along with the different (and most common) LB algorithms. [15:17] what about intermittent problems, fault-finding [15:17] caveats etc. [15:18] usually the simple getting things to work stuff is more obvious [15:18] well, the biggest troubleshooting section was going to be on network issues/latency, and how to use haproxy to determine where the fault it. is that what you were thinking, or something else? [15:19] well haproxy is going to multiple servers, right? [15:19] so servers could be hitting issues that the other server isn't hit [15:19] in theory, yes [15:19] with pages half-loading sometimes etc. [15:19] certainly [15:19] aand health checks might not notice such properly. [15:19] so knowing where a problem lays.. [15:20] well, the part about health checks, I'm actually going to cover "full stack" checks. [15:20] i assume it adds something to the headers saying what server it hit [15:20] it does, in the logs [15:20] i haven't used haproxy. [15:21] i like the idea of ssl termination with distant locations though [15:21] preferably with spdy/http2 [15:21] so yeah, sounds like you'de be interested in the health checks portion. that is where I'm intending to mention.. you COULD just have it check if a port is open and responding. or.. lets say in a lamp environment, you write a small php script to query a DB, and return a result. if the result isn't what you expect, the chack fails and the server is removed from the pool [15:22] yeah i prefer the php script to query db [15:22] is there some kind of failover server too? [15:23] for the advanced stuff, would you be at all interested in SSL/TLS stuff, like.. SHA1/SHA2 certs, wildcards, various ciphersuites? etc [15:23] so if web site has too many problems in general can point to down for maintenance page temporarily and alert. [15:23] multi-site ssl [15:24] cuz part of me wants to cover that, and part of me thinks it's outside the scope of an haproxy talk. to just mention at high level that haproxy can handle the SSL termination, alloiwing you to manage it on a small number of frontends, instead of the dozens or hundreds of backends that you have [15:24] yaeh i think it's outside of scope [15:24] to go into detail [15:24] can it do multisite ssl? [15:25] ooh, yes. I should absolutely cover a way to let users know all backends are down or something, and how to fail over to backup servers, or return an "under maintenance" page. good call. [15:25] yes, multisite SSL [15:25] can it do name-only multisite ssl? [15:25] or do they need extra ip's? [15:25] it does SNI [15:25] cool. [15:26] * jpalmer is updating his asana tasks with the advanced health checks, high level SSL stuff, and using a backup server or under maintenance page. [15:26] i thikn generally for talks for me it's about seeing what's possible [15:26] not about how to implement/do things. [15:27] so for me i see it was a way to get a nice overview and get me thinking of other ways of doing things. [15:27] yeah, I think for the basic part, I'll do a demonstration showing the very basics. for the advanced stuff, I'll turn off the slides and just talk. let them know what the possibilities ARE, rather than how to specifically do them. [15:28] and when people go into details too much from questions, or during their talk i kidn of tune out, UNLESS it's something of biting significance [15:28] say like BGP route limits. [15:29] i wonder if you should also mention otehr solutions and what made you pick haproxy over them [15:29] hehe that is almost always how I start a presentation ;) [15:29] cool [15:30] "this isn't the only tool, and I'm certainly not saying it's the best tool. However it's the tool I chose, for my use-case, and here are a few reasons why" [15:30] yip [15:30] yeah i've never seen any of your talks so i don't have anything to go on :) [15:31] (in this, haproxy won out over using nginx or apache, for 2 primary reasons. #1) TCP mode, and #2) sheer performance. [15:31] i like trafficserver myself [15:31] no worries, I appreciate the suggestions regarldess :) [15:31] it can do http2 termination now too :) [15:31] I used trafficserver years ago, haven't touched it in a while [15:32] brb, I need to feed some horses. thanks for the input. [15:32] all good. [15:51] http://blogs.windows.com/buildingapps/2015/08/06/windows-bridge-for-ios-lets-open-this-up/ [16:26] *** neish_ has quit IRC (Read error: Connection reset by peer) [16:26] *** neish has joined #arpnetworks