Time  Nick      Message
03:32 ironcamel pdurbin: unicode issues with http://danceb.in should be fixed now
03:33 ironcamel pdurbin: let me know if you still have issues
13:04 pdurbin   ironcamel: the fish i posted to http://danceb.in/FphvM6ym4RGycl7bl9DNYg is gone. http://irclog.perlgeek.de/crimsonfu/2012-05-25#i_5639032 . i guess stuff expires...
13:19 pdurbin   posted a new one: http://danceb.in/EHXZtGux4RGZD62l9DNYg . some weird red boxes. . .
13:23 pdurbin   BIND security problem: Handling of zero length rdata can cause named to terminate unexpectedly | Internet Systems Consortium - http://www.isc.org/software/bind/advisories/cve-2012-1667
13:26 pdurbin   had beers with mattdm yesterday and now i'm looking at https://code.seas.harvard.edu/openstack
13:26 pdurbin   huh. maybe they're actually running devstack... "A tool to rapidly deploy an openStack instance. By default only works on Ubuntu, but we have a branch to get it working on CentOS 6 in progress."
13:28 pdurbin   mattdm had mentioned openstack here: http://irclog.perlgeek.de/crimsonfu/2012-05-03#i_5534205
13:34 ironcamel pdurbin: the default is for posts to expire in 1 week
13:35 ironcamel set DANCEBIN_EXP if you want to change the default https://metacpan.org/module/App::Nopaste::Service::Dancebin
13:35 pdurbin   "Running in EC2, we're using AMI "ami-3ecd1e57", a CentOS 6.2 64-bit image.  Using "OpenStack" security group, with a "m1.small" image size." -- https://code.seas.harvard.edu/openstack/devstack/blobs/master/README-SEAS.txt
13:37 pdurbin   ironcamel: cool. this actually makes me worry less about spammers attacking my dancebin installation (if i were to stand one up). i assume there's a way i could remove pastes manually as well
13:37 ironcamel pdurbin: the red boxes are due to the language detection
13:37 pdurbin   ironcamel: do you see the red boxes too? http://danceb.in/EHXZtGux4RGZD62l9DNYg
13:37 ironcamel if you set the language to text, they go away
13:37 ironcamel nopaste -l text
13:38 pdurbin   DANCEBIN_EXP=never:1  # Never Expire
13:38 ironcamel pdurbin: of course you can remove them. its just stored in a database. if you know sql, then you will be able to remove them, haha
13:39 pdurbin   ah. looks perfect with -l text: http://danceb.in/VEP4rW6x4RGZD62l9DNYg  thanks! good job on the unicode support
13:40 pdurbin   ironcamel: but you'll write a script to make deleting spam super easy, right? :)
13:41 ironcamel sure
13:42 ironcamel throughnothing fixed the unicode thing last night
13:42 pdurbin   good job, throughnothing
13:52 pdurbin   right. and various openstack packages in epel: http://dl.fedoraproject.org/pub/epel/6/x86_64/repoview/letter_o.group.html
13:53 pdurbin   openstack-nova-2012.1-7.el6.noarch . i guess that's newish...
13:53 pdurbin   i bet there's newer stuff at https://admin.fedoraproject.org/pkgdb/acls/name/openstack-nova
13:54 pdurbin   yep. Wed Jun 06 2012 Pádraig Brady <P@draigBrady.com> - 2012.2-0.2.f1 http://pkgs.fedoraproject.org/gitweb/?p=openstack-nova.git;a=blob;f=openstack-nova.spec;hb=HEAD
13:59 pdurbin   huh. i guess 2012.1 is essex https://launchpad.net/nova/essex/2012.1 . and 2011.3 is diablo https://launchpad.net/nova/diablo/2011.3
14:01 pdurbin   "OpenStack releases are numbered using a YYYY.N time-based scheme. For example, the first release of 2012 will have the 2012.1 version number. During the development cycle, the release is identified using a codename." http://wiki.openstack.org/ReleaseNaming
14:04 pdurbin   ok and 2012.2 is folsom: https://launchpad.net/nova/folsom . er. will be, i guess. again, the 2 means the second release of 2012. ok. starting to make sense :)
14:06 pdurbin   and the latest fedora spec file has http://launchpad.net/nova/folsom/folsom-1/+download/nova-2012.2~f1.tar.gz as Source0, which is https://launchpad.net/nova/folsom/folsom-1 . "This is another milestone (folsom-1) on the road to Nova 2012.2"
14:06 * pdurbin wonders which version he should try
14:07 pdurbin   westmaas: any advice? ^^ or are you going to tell me to use ubuntu? :)
14:30 pdurbin   heh. "Status: Dormant" http://distrowatch.com/bulinux
14:38 pdurbin   "Set this option if you would like to use Qpid instead of RabbitMQ. Qpid is set as the default in the Fedora OpenStack packages." http://fedoraproject.org/wiki/OpenStack_devstack
14:39 pdurbin   Apache Qpid™: Open Source AMQP Messaging - http://qpid.apache.org
14:40 sjoeboo   weird, cause i think activemq is apache as well
14:40 sjoeboo   i could be wrong though
14:40 pdurbin   "Nova requires the QPID messaging server to be running." Getting started with OpenStack on Fedora 17 - FedoraProject - http://fedoraproject.org/wiki/Getting_started_with_OpenStack_on_Fedora_17
14:43 pdurbin   hmm, ActiveMQ ( http://en.wikipedia.org/wiki/Apache_ActiveMQ ) implements Java Message Service (JMS): http://en.wikipedia.org/wiki/Java_Message_Service
14:44 SEJeff    sjoeboo, Redhat wrote qpid for their MRG product and wanted qpid to become the _de facto_ standard
14:44 SEJeff    They failed
14:47 pdurbin   reading the conclusion of "Understanding the Differences between AMQP & JMS" http://www.wmrichards.com/amqp.pdf
14:48 pdurbin   looks like MGR means Messaging, Realtime, Grid: http://www.redhat.com/products/mrg/
14:49 SEJeff    The latest activemq with nio is actually insanely fast
14:52 pdurbin   i guess my takeaway is that AMQP is newer and more interoperable than JMS. of course, zeromq is newer still. :)
14:52 SEJeff    pdurbin, Did you watch the video interview about ZeroMQ by the guy who created AMQP for JPMorgan?
14:52 SEJeff    Really informative
14:56 pdurbin   SEJeff: you mean the floss weekly interview, right? i listened to it
14:56 pdurbin   ah, here's the seas openstack readme: https://code.seas.harvard.edu/puppet/puppet/blobs/feature.openstack/modules/openstack/README.md
14:57 SEJeff    yes
14:57 pdurbin   so does anyone have any opinions about rabbitmq vs qpid?
14:57 SEJeff    rabbitmq doesn't do acls
14:58 SEJeff    which SUCKS for a lot of use cases AMQP was written for
14:58 SEJeff    You can use a hack called rabbitcage ontop of rabbitmq, but activemq does them properly
14:58 pdurbin   heh. rabbitcage
14:58 SEJeff    rabbitmq is easy to setup and low memory footprint, but isn't that great
14:58 SEJeff    qpid runs MRG, so you've got redhat fully behind it
14:59 SEJeff    I use rabbitmq for celery (http://celeryproject.org/), and it works great. But I wouldn't use it for much else
14:59 pdurbin   huh. what do you use celery for?
15:00 pdurbin   (i can't believe i just wrote that)
15:03 pdurbin   ah. "What kinds of things should I use Celery for?" http://docs.celeryproject.org/en/latest/faq.html#what-kinds-of-things-should-i-use-celery-for
15:06 pdurbin   'finish the web request as soon as possible, then update the users page incrementally. This gives the user the impression of good performance and “snappiness”, even though the real work might actually take some time.' sure. makes sense. general messaging goodness
15:06 pdurbin   heh. "Celery does not depend on Django anymore" :)
15:11 SEJeff    pdurbin, I use celery as a distributed task queue
15:12 SEJeff    "various things" can http post to a url (django backed api really) and will get a uuid
15:12 SEJeff    they can then poll a url with that UUID and see the job status
15:12 SEJeff    when it is done, they then request the url with the results giving the uuid as a query argument and get the data. The jobs take from 1 to 20 minutes
15:20 pdurbin   SEJeff: interesting. thanks
15:21 * SEJeff  wrote a pexpect based network management framework
15:21 SEJeff    I've used tcl/expect and it is ok, and perl's expect module in the past
15:21 pdurbin   man, maybe i should try pexpect. i was struggling with Expect.pm last week
15:21 SEJeff    Wrote this one to learn python's pexpect and it is the most friendly of them all. Perl's expect module is gross
15:22 SEJeff    Yeah it is the worst expect implementation I've used
15:22 SEJeff    This was my first foray into tcl and expect: http://www.digitalprognosis.com/opensource/scripts/change_pass.exp it was reasonable enough
15:22 pdurbin   cool. any pexpect code?
15:22 SEJeff    It kills me!
15:23 SEJeff    I've worked 3 years on this amazing network management app
15:23 SEJeff    work won't let me open source it
15:23 SEJeff    It is the awesome
15:23 SEJeff    I'll eventually rewrite it all from scratch
15:24 pdurbin   SEJeff: so you *have* slung some perl :)
15:24 SEJeff    I worked at ticketmaster man
15:24 SEJeff    100% perl shop
15:24 SEJeff    Arguably one of the biggest perl shops on the internet
15:24 SEJeff    And they funded mod_perl 2.0 entirely by hiring the mod_perl developer
15:24 pdurbin   ticketmaster is always held up as a perl shop, sure
15:28 SEJeff    How about this
15:28 SEJeff    I am comfortable with storing my credit card and password with ticketmaster
15:28 SEJeff    The reason...
15:29 SEJeff    One of my first projects when joining the company
15:29 SEJeff    They have this thing called an Ingrian which is essentially a hardware credit card encrypter.http://www.secureconsulting.net/2009/05/ingrian_data_secure_lives_on_a.html
15:30 SEJeff    They have these archives of datafiles that contain customer information. Your credit card is never once written to disk _ever_ before being encrypted with a key that is retrieved via xmlrpc from that Ingrian datasecure
15:30 SEJeff    that is required by PCI
15:30 SEJeff    I then wrote a script which gets a separate key (that rotates every XX time) and encrypts the entire archive. That is above and beyond PCI
15:31 SEJeff    It is a perl script which does this that I wrote so I know with a certainty the data is safe and doubly encrypted
15:31 SEJeff    So I might not be some perl expert like ironcamel, but I certainly am comfortable with perl
15:32 pdurbin   awesome
15:33 pdurbin   oh oh. speaking of TCL. at the bar last night one of our network engineers was telling me that cisco's ios (how annoying that i have to differentiate that these days) supports some sort of embedded TCL
15:33 SEJeff    The systems engineering team wanted us to encrypt that data because they know what they're doing. PCI is more checking of boxes than real security
15:33 SEJeff    pdurbin, Have you seen the stuff the Arista switches support? You can manage hundreds of switches with...
15:33 SEJeff    wait for it...
15:33 SEJeff    XMPP! aka jabber
15:33 pdurbin   Cisco IOS Scripting with Tcl - Cisco Systems - http://www.cisco.com/en/US/docs/ios/12_3t/12_3t2/feature/guide/gt_tcl.html
15:34 SEJeff    https://eos.aristanetworks.com/2011/08/management-over-xmpp/
15:34 pdurbin   i know what xmpp is. sheesh ;)
15:34 SEJeff    Yeah doesn't mean everyone else does
15:34 SEJeff    or googles since this is all logged
15:34 pdurbin   yeah yeah. and if it were me i'd probably link to the wikipedia page or something
15:35 pdurbin   westmaas used to call me "linky" when we worked together
15:36 SEJeff    I bet you write great docs though
15:37 westmaas  haha
17:09 pdurbin   i'm trying to put docs here but it's a bit of a mess: http://wiki.greptilian.com ... a work in progress
17:41 pdurbin   bleh. it's either install something newer than perl 5.8.8 on our cluster or re-write `unless ($repo ~~ @projects)`
18:13 mattdm    http://actech.seas.harvard.edu/2012/06/science.html
18:15 pdurbin   science!
18:21 pdurbin   coffee!
18:21 pdurbin   decaf?!
18:23 pdurbin   hmm, Academic Computing Technical Blog: Virtualization with KVM - http://actech.seas.harvard.edu/2012/05/virtualization-with-kvm.html
18:23 pdurbin   mmm, future directions.... 'We can offer "Platform-as-a-Service" (PaaS) using Red Hat's OpenShift project.'
18:24 pdurbin   oh, cool, you're on github: https://github.com/seasac-ops
18:26 pdurbin   hmm, workshops on python and git: http://acops.seas.harvard.edu/workshops
18:29 mattdm    :)
18:30 SEJeff    python is evil
18:30 pdurbin   i guess you could suggest the new github windows tool instead of msysGit at http://acops.seas.harvard.edu/git-workshop/workshop.s5.html
18:30 SEJeff    just ask #crimsonfu
18:30 * SEJeff  ducks
18:30 mattdm    I am not involved in the windows software. :)
18:32 mattdm    I think the choice of msysgit was largely for our ssh instructions, which currently are a putty-centered mess for windows, and "just do this" for mac and linux.
18:33 mattdm    with msysgit, it becomes "install msysgit; now, see the mac and linux instructions"
18:33 pdurbin   SEJeff: i think our libwww-perl is broken on the cluster. :( i'm about ready to rewrite http://git.greptilian.com/?p=scripts.git;a=blob;f=grgit.pl;hb=HEAD in python :(
18:33 pdurbin   mattdm: sure. command line instructions ftw
18:34 SEJeff    Y NO LWP?
18:34 SEJeff    Is lwp old school?
18:34 pdurbin   KISS :) LWP::Simple! :)
18:35 SEJeff    Yeah thats what I've used
18:36 SEJeff    is libwww-perl different than LWP?
18:36 pdurbin   nope http://search.cpan.org/~gaas/libwww-perl/lib/LWP.pm
18:36 SEJeff    Ah ok
18:36 SEJeff    Yeah I like that
18:36 SEJeff    I really like the xmlrpc stuff in perl. It is equally as simple as the python version, but makes more sense in my head
18:38 pdurbin   is that my cue to plug this again? https://github.com/crimsonfu/code/blob/master/cobbler.pl
18:40 SEJeff    Or mine to plug this :P  http://www.digitalprognosis.com/opensource/scripts/nhs
18:44 pdurbin   onall?
18:47 SEJeff    http://www.digitalprognosis.com/opensource/onall which is a lightly patched version of: http://code.google.com/p/onall/downloads/detail?name=onall-2.11.tar.bz2
18:47 SEJeff    onall is the best parallel ssh implementation that exists period. Oh and it is written in perl
18:47 pdurbin   ah. you need to work on your seo
18:47 SEJeff    The director of systems engineering @ ticketmaster wrote it
18:47 SEJeff    It isn't my project
18:47 pdurbin   ok :)
18:47 SEJeff    and they took code.ticketmaster.com down when 1/2 of us left the company
18:48 SEJeff    but it is still open source :)
18:48 SEJeff    I've been meaning to start a github project for onall actually. Would you be interested in it?
18:48 pdurbin   um. have you heard of salt? ;)
18:49 SEJeff    Different use case
18:49 SEJeff    ssh is still useful
18:50 pdurbin   sure. gsh is the one i'm most familiar with: http://outflux.net/unix/software/gsh/
18:50 SEJeff    why does that url sound familar?
18:50 SEJeff    Is that kees cook's domain?
18:50 pdurbin   yep
18:51 SEJeff    ha! Eidetic memory ftw
18:51 SEJeff    onall is written by a unix admin for unix admins
18:52 SEJeff    I use it something like nhs ... args to filter | grep | foo | onall -qy -t 5 "uptime" for instance
18:52 SEJeff    or more often just nhs -args | onall ...
18:53 pdurbin   hmm, no mention of onall here: http://stackoverflow.com/questions/243750/how-can-i-automate-running-commands-remotely-over-ssh/245114#245114
18:53 SEJeff    Feel free to add it if you want
18:53 SEJeff    I recommend you trying it out
18:53 SEJeff    You'll like it given your leanings
18:53 SEJeff    The perl ssh libs are crap on old perl versions so it forks ssh in threads
18:53 pdurbin   ok. we use pdsh here
18:53 SEJeff    We had to make it support rhel3
18:54 SEJeff    Yeah onall is just a single perl script too. Simple to hack on
18:55 pdurbin   cool
19:53 SEJeff    pdurbin, I stand corrected, this is better: https://github.com/ticketmaster/onall
19:53 SEJeff    didn't realized they moved it all to github
19:54 pdurbin   "Seriously, who puts underscores in command line options?" it says :)
19:54 SEJeff    Yeah thats from Ryan
19:54 SEJeff    he is hilarious
19:54 SEJeff    That guy was employee #22 @ redhat
19:54 SEJeff    helped build akamai before I even knew what Linux was
20:00 pdurbin   so he's in boston?
20:22 SEJeff    He is in CA right now
20:22 SEJeff    works for Mozilla as a developer
20:22 pdurbin   ok
20:22 SEJeff    https://github.com/rtilder
20:22 pdurbin   oh, SEJeff, you should listen to dave winer talk about xmlrpc on the stack exchange podcast
20:23 SEJeff    link?
20:23 SEJeff    I prefer rest strongly over xmlrpc
20:23 SEJeff    but don't mind using it when I have to
20:23 pdurbin   http://blog.stackoverflow.com/2011/11/se-podcast-27-dave-winer/
20:23 pdurbin   we need a bot
20:24 pdurbin   look at me googling and copying and pasting like a sucker
20:24 SEJeff    Our bot is named pdurbin
20:25 pdurbin   not anymore. someone write us a bot
20:25 SEJeff    supybot
20:26 SEJeff    http://www.theswellesleyreport.com/2012/06/wellesley-high-grads-told-youre-not-special wow
20:26 pdurbin   this one? http://sourceforge.net/projects/supybot/
20:27 SEJeff    Yeah it is one of the more commonly used bots
20:27 SEJeff    google: github supybot and look at all of the plugins
20:27 pdurbin   ok. i'm still an irc noob
20:38 pdurbin   https://github.com/code4lib/supybot-plugins http://supybot.com
20:40 pdurbin   Supybook - the Supybot handbook http://supybook.fealdia.org/
20:46 pdurbin   yeah, all i want is `google lucky` http://supybot.fr.cr/doc/use/plugins/google.html#google-lucky-snippet-search
21:00 SEJeff    Oh that has google calc nice!
21:00 * SEJeff  uses google calc every single day
21:01 SEJeff    Love to convert from nano/micro/and milliseconds