Time Nick Message 07:51 boegel can you guys access http://blog.jcuff.net/ ? I can't, and James tweets it's OK for him 09:00 boegel nvm about James' blog, turns out our uni blocks it because some torrent site is hosted at the same IP 12:10 pdurbin boegel: that's hilarious :) 12:38 boegel pdurbin: not sure if it is :) 12:38 pdurbin fair enough :) 12:40 boegel pdurbin: are you saying our network team is overexagerating? or that it's funny that James' blog is hosted alongside torrents/warez? 12:41 pdurbin the latter :) 12:41 pdurbin but i'm kidding 12:41 boegel pdurbin: ah, cause I might agree with the former as well 12:41 pdurbin i'm quite anti-warez. i'm a boy scout 12:41 pdurbin well, not literally... 12:41 boegel pdurbin: it's probably a virtualized server, apparently 2500 sites are hosted on the same IP 12:41 pdurbin interesting 12:42 boegel pdurbin: it's just annoying that I need to connect differently to read James' blog :P 12:42 pdurbin no kidding 12:42 pdurbin you said you let james know already? 12:44 boegel yes, via Twitter 12:44 pdurbin awesome. thanks 12:44 boegel pdurbin: I was complaining first that his blog is down ;-) 12:47 pdurbin crimsonfubot: lucky down for everyone or just me 12:47 crimsonfubot pdurbin: http://www.downforeveryoneorjustme.com/ 12:48 boegel pdurbin: yeah, well, that's down as well :P 12:48 pdurbin is it?!? 12:48 boegel pdurbin: 12:48 boegel pdurbin: at least, it doesn't work from inside our uni network 12:48 boegel pdurbin: btw, https://twitter.com/jamesdotcuff/status/251301469310054400 12:51 pdurbin huh 12:52 pdurbin i hope your uni doesn't block all blogger sites 12:52 pdurbin (though i'm more of a jekyll/github pages guy) 12:56 pdurbin agoddard: dunno if you saw my "High Availability with MySQL" to our internal mailing list, but i'm curious about your thoughts 12:57 pdurbin (i linked back to some chat about Scalr at http://irclog.perlgeek.de/crimsonfu/2012-07-18#i_5824414 which is public, of course) 12:57 Itkovian now now 12:57 Itkovian let me introduce you to hakyll then pdurbin :-p 12:57 pdurbin ventz: were your ears burning? i also talked about the mysql cluster you set up 12:58 pdurbin crimsonfubot: lucky hakyll 12:58 crimsonfubot pdurbin: http://jaspervdj.be/hakyll/ 13:00 pdurbin Itkovian: you're a haskell hacker! sqee! please submit a solution to this kata! http://irclog.perlgeek.de/crimsonfu/2012-09-11#i_5977620 13:02 boegel pdurbin: I'm into Haskell too, when I find the time 13:06 pdurbin double sqee! squee? squee! 13:09 ventz pdurbin: oh yea? :) 13:09 ventz what about it? 13:21 larsks Hey, did anyone ever end up with a successful RT4 repository? 13:47 Itkovian pdurbin: what's the question/exercise? 13:48 Itkovian the autovivication ons? 13:48 Itkovian *one? 14:31 Itkovian pdurbin: challenge accepted 14:52 pdurbin Itkovian: \o/ #soexcited 14:53 pdurbin larsks: yes. http://wiki.greptilian.com/request-tracker/4/ 14:53 Itkovian if it documented anywhere what the JSON should contain, or do I just look at the example? 14:53 pdurbin ventz: i'll forward it to you 14:53 pdurbin Itkovian: it needs to pass `prove` or it doesn't count 14:54 Itkovian sure 14:54 pdurbin so excited 14:57 pdurbin ventz: sent. subject: High Availability with MySQL, was Re: Mysql.rc brutality 14:58 pdurbin if you can talk about your mysql cluster here that would be great 14:58 pdurbin or link to some public puppet manifests, ideally :) 15:52 ventz pdurbin: it's definitely not crazy difficult 15:52 ventz and yes, that's what the cluster achieves 15:52 ventz there are some "gotcha's" when it comes to the data type 15:52 ventz it's not really meant for throwing wordpress against it 15:52 ventz it's designed for custom applications where the dev can modify the table create statement so that it takes advantage of the cluster (2 lines) 16:11 dotplus http://puppetlabs.com/puppetconf-live 16:11 agperson watching! 16:24 semiosis dotplus: thanks 16:24 semiosis wow there's 830+ people in #puppet now... thats just crazy 16:41 Pax Just cross posted the link to the #splunk room on irc.efnet.org :) 16:52 agperson Pax: any update to this: http://forge.puppetlabs.com/thartmann/pam_access ? 16:53 agperson (sort of assuming that is you) 16:56 Pax yup! 16:57 Pax thats me! snuff and I did some re-writes and posted a newer version under the huit account 16:57 Pax http://forge.puppetlabs.com/huit/pam_access 16:57 Pax I should probably take down the older one 16:57 Pax but there a parts of it I still like, I sort of like the "gold" copy-ness of the file 16:58 agperson thanks 16:59 agperson Pax: I agree, line-by-line management makes things very flexible but lots of opportunity for problems to arise 17:00 agperson also puppet runs a *lot* faster that way... 17:00 agperson (err, the opposite…a lot slower with lots of augeas calls, faster managing full config files) 17:02 Pax I have to admit I am a bit of a template fan boy :) 17:32 larsks pdurbin: Thanks! 17:38 pdurbin larsks: sure 17:38 * pdurbin reads https://docs.rc.fas.harvard.edu/puppetconf2012/ (thanks to sjoeboo) 17:49 pdurbin heh. "thinks we're living in the most exciting time in IT. it sucks, but sucks way less than it used to." 17:50 pdurbin "Unmovable deadlines, corners cut, more shortcuts, more fragile code/applications. = Technical Debt" 17:52 pdurbin sjoeboo: thanks. good stuff 17:59 pdurbin wow. 1.6 terabytes in /mnt/lost+found/ 18:00 pdurbin ventz: not crazy difficult? the mysql clustering? do you recommend it? any docs you can link to? 18:06 agperson pdurbin: it's not very useful for off-the-shelf apps unless they explicitly support it, or you want to do a lot of work 18:07 agperson we tried with wordpress, drupal, mediawiki…no luck 18:07 agperson best for custom-built apps that need those specific capabiilities it provides 18:14 pdurbin agperson: that's very unfortunate. what's the point? 18:15 agperson i gather it is supposed to be highly-available, highly-scalable, and all in-memory…but not NoSQL 18:15 ventz pdurbin: all this time i thought it was my script that fudged URLs 18:16 ventz i found a bug in the tinyurl API call 18:16 ventz pdurbin: about mysql cluster -> more of a in-person discussion :) 18:16 pdurbin agperson: i'll send you the email. if i can find your address 18:17 pdurbin ventz: more of a geeknight thing? 18:21 pdurbin agperson: sent 18:21 agperson cheers 18:21 pdurbin would be great to continue the discussion here 18:22 pdurbin certainly on topic 18:22 * dotplus is planning to build a mysql "cluster" with HA in mind to run our RT systems 18:23 pdurbin see? :) 18:24 pdurbin dotplus: me too. for RT 18:24 pdurbin and whatever 18:24 dotplus so I'd love to see anything others have done recently, especially if it's puppetized. I think that that puppetlabs-mysql isn't going to be flexible enough for me (we're not using vendor rpms for mysql, won't support multimaster and various other settings) 18:24 pdurbin dotplus: have you looked at scalr? i linked it above 18:25 pdurbin well, linked to previous chat about it 18:29 ventz pdurbin: not necessarily that, but just too much info 18:29 ventz and i have a lot of stuff today 18:30 ventz the summary: it works, it requires someone to look after it/maintain it (that is when you add apps and such) 18:30 ventz it's great if you truly need 0% downtime 18:30 ventz it's a HUGE pain for "off the shelf" apps 18:33 dotplus "off the shelf" as in "cannot modify how the app interacts with/queries the db"? 18:34 pdurbin dotplus: please. he's busy ;) 18:34 pdurbin ventz: how many nodes in the cluster? 18:39 ventz 10 18:39 dotplus scalr is an alternative to openstack, opennebula, cloudstack, etc.? 18:39 ventz you ahve 3 "types" 18:39 ventz you need at minimum TWO different "servers" -- not recomended 18:39 ventz the minimum "recommended" model is THREE "servers" 18:40 ventz 1.) API/SQL node (front access only, 2.) STORAGE node (attached to lots of disk), and 3.) Management server 18:40 ventz in an ideal setup: 18:40 ventz 10 nodes = "2 site" redundancy, with each site: 2 SQL nodes, 2 DN (data) nodes, and 1 "management server" 18:40 ventz out of 10 servers, you can lose 8 -- everything except 1 SQL node and 1 DN node 18:40 ventz and still be 100% up 18:45 pdurbin dotplus: no. scalr for "orchestration" 18:45 pdurbin ventz: thank you! :) 18:52 pdurbin dotplus: i'm a little confused about what "off the shelf" means myself. wordpress, sounds like 18:52 pdurbin not sure if RT qualifies 18:53 pdurbin ventz: did you say there was some book you read? 18:53 dotplus isn orchestration the cool new name for provisioning? 18:53 ventz :nod: 18:53 pdurbin dotplus: nope 18:53 ventz pdurbin: but now it's about 80% useless because they released 2 major releases 18:53 ventz that was the difficult part. The book gave me a lot of the basic concepts, but some of them had drastically changed by the time I started building 18:54 pdurbin agoddard: didn't we talk about provisioning vs. orchestration once? 18:54 ventz they went from v5.1,2,3 to v7 18:54 pdurbin ventz: hmm. ok. still worth linking to the book? 18:54 ventz i have the copy 18:54 ventz you are more than welcome to borrow it 18:54 ventz i even have all the sections labled 18:56 pdurbin crimsonfubot: google mysql cluster book 18:56 crimsonfubot pdurbin: MySQL Clustering: Alex Davies,Harrison Fisk: 9780672328558 ...: <http://www.amazon.com/MySQL-Clustering-Alex-Davies/dp/0672328550>; MySQL Clustering - Alex Davies, Harrison Fisk - Google Books: <http://books.google.com/books/about/MySQL_Clustering.html?id=_3nERYD9xqcC>; MySQL :: MySQL Documentation: MySQL Reference Manuals: <http://dev.mysql.com/doc/>; MySQL Cluster Management : Part 2 | (2 more messages) 18:57 pdurbin "Publisher: MySQL Press; 1 edition (March 20, 2006)" that one by Alex Davies must be it 18:59 pdurbin ventz: we have a guy who might want to borrow it. not sure this is my project. but thanks 19:05 GitHub95 [crimsonfu.github.com] pdurbin pushed 1 new commit to master: https://github.com/crimsonfu/crimsonfu.github.com/commit/0c0f6864844ffa424a8532ae4a317bd5b12647d5 19:05 GitHub95 [crimsonfu.github.com/master] mysql cluster fu from ventz - Philip Durbin 19:06 ventz does anyone know of a quick/light CLI based todo list? 19:06 pdurbin Todo.txt: Future-proof task tracking in a file you control - http://todotxt.com 19:07 ventz i am finding out that as the day goes on, stuff comes in from here and there, and eventually i either 1.) lose track of small things I have to do or 2.) there's a giant list and more and more gets added -- where i need to prioritize 19:07 semiosis cool! 19:07 ventz ohhh 19:07 * semiosis uses jira to keep track of things 19:07 ventz pdurbin++ 19:07 ventz that's fricken awesome 19:07 * semiosis is not thrilled about it though 19:08 pdurbin dotplus: it sounds like ventz needs RT ;) 19:09 dotplus "quick/light" :) 19:09 pdurbin semiosis: i keep hearing about jira. is it free and open source? 19:09 semiosis sadly not 19:09 pdurbin pass 19:09 semiosis yeah i'm looking at redmine as a replacement 19:10 boegel semiosis: we use Trac, and I really like it 19:10 semiosis boss wants pivotal tracker b/c it looks good on his iphone :) 19:10 pdurbin semiosis: i set up https://redmine.hmdc.harvard.edu 19:10 semiosis i love trac too but it's not a great fit for my current needs 19:11 pdurbin whorka: and now it's your job to keep it alive :) 19:11 semiosis actually, i should re-evaluate trac, it may not be like what i remember 19:11 pdurbin dotplus: yes, RT not so light :) 19:11 boegel semiosis: what are your needs? 19:12 semiosis multi-project, multi-person, issue tracking, wiki documentation, scm integration, supports agile/scrum/kanban type of workflows, looks good on iphone 19:13 pdurbin mmm, kanban 19:13 semiosis oh and i forgot to mention, zero administrative overhead :) 19:14 pdurbin and a pony 19:14 semiosis you know it 19:16 semiosis anyone have an opinion on redmine vs. trac? 19:16 pdurbin haven't used trac in ... 7 years 19:19 ventz trac is too "cluggy" 19:19 ventz we used it for changemanegment at mozy 19:20 pdurbin heh. cluggy 19:21 semiosis cluggy? is that like kludgy? 19:26 pdurbin semiosis: is it just that you're trying to save money? doesn't jira do all that? 19:26 semiosis not trying to save money actually 19:27 semiosis jira does all that but it's kinda hard to use 19:27 semiosis and when it doesnt work/do what i want, it's really hard to fix 19:31 semiosis this is sort of related to the discussion... years ago i took over administration of a bunch of servers where there was no change/configuration management, all setup was done interactively on the servers & not much was documented or even recorded about past changes 19:31 semiosis so first thing i did was set up a trac and wrote a nifty (imho) script to bring order to the chaos, the result of which is this cheesy & poorly written (but incredibly effective) script: https://gist.github.com/3795948 19:32 semiosis that allowed me to make very effective use of trac's scm integration, wiki, and ticketing system 19:32 semiosis it was the missing link 19:36 pdurbin "savn -- brings version controlled order to chaotic system configurations" 19:36 semiosis savn -- like save + svn... get it? 19:36 pdurbin semiosis: it was savn you time 19:36 semiosis wow! 19:36 semiosis and sanity 19:36 semiosis yup 19:37 pdurbin savn - be born again 19:37 semiosis eh? 19:38 pdurbin have you been saved? 19:38 pdurbin don't answer that. off topic 19:38 semiosis haha 19:58 pdurbin semiosis: hard to fix jira in what way? 19:59 pdurbin i feel like i've only heard good things about jira 19:59 semiosis well we're using the self-hosted jira 19:59 semiosis and i dont have a lot of time to spend on it 19:59 semiosis a couple things that didn't work as expected are 20:00 semiosis SCM integration -- i could never get jira & fisheye to work together, even after engaging support. when they asked for remote access to the server & wanted me to schedule an appointment with their engineers i gave up 20:00 semiosis it wasnt worth all the trouble 20:01 semiosis Formatting in issues/comments/etc -- supposedly there's ways to do code blocks, section headings, things like that, to format the text 20:01 semiosis this has never worked 20:01 semiosis even after i followed the instructions to "enable" the supposedly built-in feature 20:02 semiosis maybe once i upgrade to jira 5 things would work better 20:02 pdurbin semiosis: they don't use markdown in issues/comments? 20:03 semiosis i've not been able to get any formatting to work 20:03 semiosis i dont remember if it was markdown or some other thing, but i do remember the markup syntax just appearing as plaintext everywhere i tried it :( 20:04 semiosis with trac otoh everything that was supposed to work out of the box did, and everything i had to enable/configure myself worked exactly like the instructions said it would 20:08 pdurbin huh 20:08 pdurbin i could ask the guys if they like redmine 20:08 pdurbin whorka: or you could. they're down the hall for you 20:08 pdurbin i think they like it 20:09 semiosis ah, dont worry about it, but thx 20:09 whorka they like it well enough to keep using it 20:09 semiosis thats about as much as i like jira lol 20:13 pdurbin time for ice cream. see ya