17:02:14 #startmeeting 17:02:14 Meeting started Tue Sep 24 17:02:14 2013 UTC. The chair is TheSnide. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:02:14 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:02:28 hi alll 17:03:14 today will be a loose meeting 17:03:21 hm. 17:03:59 is there something I should know about this meeting? I never bothered about the MeetBot pings before 17:04:01 ;-D 17:05:44 nah, just attend if you want. speak if you feel to. 17:06:05 #topic 2.2 17:06:27 sql schema is mostly completed 17:07:06 i added an extra url table to avoid complex runtime url decoding 17:07:43 i'm also rewriting cgi html part from scrtach 17:08:05 ... but i still use html::template. 17:08:52 so far the startup cost of a request is far below 10ms 17:09:39 we'll see how it fares when sql queries will pile up to construct the model, but it looks very promising 17:10:58 #info sql branch perf-test box is a RPi. 17:11:22 e'll see how *that* scales. 17:11:25 :-D 17:11:53 scale horizontally with more RPi's? :) 17:12:02 * TheSnide has the habit of having an anemic perf-test box at work. 17:12:08 TheSnide, is there a migration service from rrd -> sql ? 17:12:13 *waves* 17:12:34 #info 2.2 will *not* scale horiz yet. 17:13:26 Um, some background please. sql will replace data storage done by rrd so far? 17:13:28 #info sql in 2.2 is _only_ for *metadata* (storables) 17:13:39 Ah 17:14:04 the stuff needed for munin-html ... and munin-limits? 17:14:09 that means that upward migration from 2.x -> 2.2 is possible, and the nodes can talk to a 2.2 master? 17:14:16 #info rrd _will_ stay. (no pb so far) 17:14:56 fenris02: yup, protol will not change 17:14:57 with SQL for metadata, it will be simpler to scale RRD updates and graphing horizontally 17:15:38 yes, but 2.2 will be sqlite-only 17:15:55 #info sql in 2.2 will be sqlite-only 17:16:01 TheSnide, only the master is writing to the sql db correct? not multiple writers ? 17:16:29 as the db is rewritten wholly each m-u run 17:16:51 yeap. it is a drop-in for the storable architecture 17:17:09 2.4 will bring MoM 17:18:02 Methods of Mayhem? 17:18:57 but let's focus on simple,manageable tasks 17:19:29 mom = master of masters 17:20:00 aka horz scaling 17:20:45 yeah, but that abbrev was completely unknown to me 17:21:20 #info yes, but 2.2 will be sqlite-only 17:21:23 TheSnide, which version replaces m-n with async? 17:21:51 not 2.2 17:21:56 * fenris02 nods 17:22:11 dunno yet 17:22:42 will probably call it 3.0 when that happens 17:25:55 ... as the protocol might get enhanced. 17:28:39 and i'm tempted to add plugin-exec abilities to asyncd, and let m-n die out of attrition. 17:33:38 ok. so no more questions. 17:34:07 #topic bugs 17:34:28 aany particular bug ? 17:35:52 Besides my personal favorites and the things that are really hard to fix? 17:35:53 ok, anything else ? 17:36:19 oh 17:36:42 what are your fav ? 17:37:10 A request, could somebody(tm) triage the list at and give an idea how to handle at least the more important ones? 17:37:24 #1205 shouldn't be to difficult to fix 17:38:02 #1382 (that was me) will be fixed in rrd once I've managed to fork projects on github 17:38:21 GrumpyFux: how about we fix at least 5 each before next meeting? ;) 17:38:22 (Tobi already told he'll pick it) 17:40:05 kjetilho: Quite an idea, however there are a lot that a incomplete or require re-design of the way munin operates. 17:40:21 But there are too many to do a triage during that meeting here 17:41:24 until next meeting we can probably find 5 closable bugs each. it would be a start. 17:43:53 find? Also propose how to deal with it, too 17:44:44 by the way, are there now pointers in trac to all reports in the Debian BTS? 17:51:31 * TheSnide got disconnected... 17:53:37 * TheSnide still has a flaky connection... 17:56:39 * TheSnide does not think upstream url are done 17:58:02 wtf 17:58:15 .kickban m-r-b 17:59:40 xSmurf: ? 18:00:36 unless my bouncer fucked up he was spamming like crazy 18:00:54 oh wait it's a linked chan bot isn't it? 18:01:11 yup 18:02:57 so, any other point ? 18:03:08 #topic anything else 18:04:15 Well, IMHO there's a long-time project "documentation" in Trac. The current state is terrible 18:04:54 #topic doc 18:06:08 everybody hides? 18:07:17 #info 2.2 will have a enhanced embedded doc/ 18:07:31 (the munin guide) 18:08:16 Then just a statement: I'm inclined to aim for a complete re-write of the doc in Trac. But that's a huge task and none I could handle. 18:08:29 i'm slowly taking info from the various trac pages to integrate them in the guide. 18:08:48 My time ressources are very limited, and not predictable, I'm afraid 18:08:48 Why focus on the trac? Doc should IMHO be moved to the guide. 18:09:00 doc in trac will be more a wiki. 18:09:25 Whereever. More important it's well organized so information can be found. 18:09:26 more a living bag-of-hacks. 18:09:45 my focus is purely the guide for now. 18:12:39 more a living bag-of-hacks. 18:13:24 #topic branches 18:14:24 #info the branches are in my own git repo on github. 18:15:24 #info it is steveschnepp/munin, branches "sql" and "guide" if you'd like to have a peak. 18:16:28 but beware, i do rebase from time to time 18:16:48 so, anything else ? 18:17:21 (more than 1h elapsed. time to close) 18:17:40 so, see you next week for your weekly dose of munin. 18:17:45 #endmeeting