hahahah
packet inter-leaky
well, my irc client seems to have connected from arp
but can't ssh
DNS is working again from my server. Yay
my machine is up
hazardous: check your machine
back up
Try again
that was annoying
s1.lax went over its 128K route limit and CEF died
shutdown a few peers using out-of-band then re-enabled CEF
[channel suddenly doubles in size]
lol
that's one way to get people to join... ;)
haha
up_the_irons do you actually sleep or is this normal response time at 6 am :p
seems i'm up
wooooo
hazardous: it's normal if something major breaks (like a core router), cuz pagerduty is set to scream like hell (and it does :)
fun
you know, smileys in parentheses always weirds me out
like (:) is incomplete, (:)) is weird
yeah me too
I hear ya (no really, i do :) )
also, my vps is back online. thanks to the arp folks for their work
hope it stays stable
cheers!
What'd I miss?
@up_the_irons │ s1.lax went over its 128K route limit and CEF died
Cool, good to know.
Glad to be back, too.
https://twitter.com/arpnetworks/status/426730720862560256
TWITTER: One LAX core router (s1.lax) hit its CEF limit and went into software routing; time to move peering to the 6500 (s7.lax) (Fri Jan 24 14:58:26 +0000 2014)
<3 BryceBot
Oh right I forgot about Twitter. My only thought was "I gotta get on IRC!"
That's my thought for most of the day, though.
the mentioning of the twitter post reminded me to build a new version of bitlbee to fix twitter. so now i can finaly use twitter again after i nearly forgot that is wasn't working
ouch
i seem to recall some TAC-confirmed bug with CEF on certain switch chassis'
in this case, it's actually a documented phenomenon. if you exhaust the hardware resources, then it likes to switch to software routing. go figure.
yeah
the one i'm thinking of causes it do the equiv of a core dump and totally die :)
friend of mine works for an ISP/MSP up in OR, had a 4500 hit that bug and dump a bunch of stuff offline - backup 4500 then totally died in the same exact way
just remembered it involving CEF is all :)
huh, funny. looks like there was a routing change to one vps at like 0230 MST
that one goes abq > denver > sjc > lax - other one still just goes abq > lax
that sounds like a nasty bug
yeah, didn't sound like much fun
http://arstechnica.com/gadgets/2014/01/modems-warez-and-ansi-art-remembering-bbs-life-at-2400bps/
Ars Technica: "Modems, wArEz, and ANSI art: Remembering BBS life at 2400bps"
2400 was for the cool people... I started at 110 baud. :)
You still got that acoustic coupler in the garage?
IIRC they were built out of whatever they made abrams tanks out of
up_the_irons: what model is the s1.lax ? (Cisco I a assume)
up_the_irons: talked to my buddy - it was actually a similar thing- overflowing CEF table, switched to process switching mode, then died
must be exclusive to whatever ios rev they were on, but they swapped out that gear for new juniper stuff recently
is that only for specific usernames D:
https://twitter.com/officialjaden/status/329768040235413504
TWITTER: How Can Mirrors Be Real If Our Eyes Aren't Real (Thu May 02 01:23:22 +0000 2013, retweeted 17062 times)
oh, any
lol
everybody's hugging the bot, oy
it's a lovable mutt
Does more than just twitter URLs, you know :p Wikipedia, XKCD, Spotify, Amazon, Newegg... all sorts
eg http://xkcd.com/1321/
Cold http://xkcd.com/1321/ (2014-01-24) 'You see the same pattern all over. Take Detroit--' 'Hold on. Why do you know all these statistics offhand?' 'Oh, um, no idea. I definitely spend my evenings hanging out with friends, and not curating a REALLY NEAT database of temperature statistics. Because, pshh, who would want to do that, right? Also, snowfall records.'
it's definitely nice to not need bothering opening twitter
how does it work brycec
do you define individual url regexes and scrape them
or do you use oEmbed or something
All of the above
Depends on the URL
Amazon hits noembed.com
xkcd hits a database of my own
Wikipedia, Newegg hit their own API
(Spotify is noembed too)
fun
spotify has its own api btw
that returns a lot of interesting things
hazardous: which is what noembed uses :P
At least I assume it does
Actually come to think of it, they just scrape the page
it looks like it just returns scraping the
element
which I'm okay with
the actual spotify api returns things like
this song can be played in: US CA DE UK
length, a lot of other funky stuff
Requires oauth doesn't it?
(Which would be more overhead than it's worht)
only the authetnicated add to playlist one does i think
yeah, metadata doesn't
2009
ah cool, good to know
AD AR AT AU BE BG BO BR CH CL CO CR CY CZ DE DK DO EC EE ES FI GB GT HK HN HU IE IS IT LI LT LU LV MC MT MY NI NL NO NZ PA PE PH PL PT PY RO SG SI SK SV TR TW
289.450000
(ew xml)
there's .xml .json .xsl and some other weird things and jsonp
xsl? Excel Spreadsheet?
wait, who needs six digit decimal accuracy on track length?
oh wait that's xls
xsl is some weird xml thing
http://ws.spotify.com/search/1/track.json?q=mad+world
yeah, xslt stuff
Anyways, BryceBot supports most everything on http://noembed.com/providers with a few of my own override noembed
yeah, their spotify.pm just scrapes
i kinda want to submit, but i have never used perl before
and it vaguely scares me
lol, it's pretty easy actually. https://github.com/leedo/noembed/pull/24
I just copied another module and changed the names and URLs
"I have not tested this code as I do not have an environment to run noembed within." lol
I'm not about to setup perl-wsgi wor whatever
But it worked :)
haha
yeah, i would probably end up doing like
perl xxx argv to test
and just making it dump to stdout or something
i actually tried to learn clojure basics last week
the hardest part was syntax, i guess
and operator arg arg
I embarked to learn Perl once, many many years ago
It's fairly readable as-is at least
see https://github.com/toddfries/Finance-CoinBase/blob/master/cb for some fun perl code that should be readable ;-)
Which part is readable?
the word "Github"
mike-burns: RandalSchwartz can say 'readable or not' the rest of you wimps I wouldn't trust to judge my perl codez ;-)
RandalSchwartz: curious if you've played with SPORE before
the EA game?
I wrote Perl for five years! (True facts: I've met RandalSchwartz on multiple occasions.)
brycec: github.com/spore .. the multi language spec that lets you write a single .json file and after loading it .. you get a magic class that you can do function calls on that were described in the .json file
$ wc -l api-description/services/btc-e.json \
> Finance-btce/lib/Finance/btce.pm
65 api-description/services/btc-e.json
621 Finance-btce/lib/Finance/btce.pm
oh heh
65 lines of json to do what I do (almost) in 621 lines of raw perl code
(In all caps, I thought it was the game)
SOAP but with JSON?
$randomsitesapi but with JSON
I have a json file I'm working on with one function call to reset my cablemodem. soon as I learn all about SPORE and html form submission
144 lines of json to describe all you'd ever want to do interacting with twitter for another example
see https://github.com/toddfries/api-description/blob/master/services/twitter.json for that
then in my app I could do: use Net::HTTP::Spore; my $api = Net::HTTP::Spore->new_from_spec("twitter.json"); $api->enable('Net::HTTP::Spore::Middleware::Format::JSON'); my $res = $api->retweeted_by_me({ format => 'json', count => 2 });
and handle the return bits in the $res var
m0unds: yeah, i imagine if it was software switching that much traffic, it would eventually melt the RP ;)
mnathani: 4500 with SupIV
w00t! Newegg info is working once again! http://www.newegg.com/Product/Product.aspx?Item=N82E16813131981 Sorry it was non-working for awhile.
ASUS Z87-A LGA 1150 Intel Z87 HDMI SATA 6Gb/s USB 3.0 ATX Intel Motherboard -- $139.99
up_the_irons: you know something's wrong when your switch chassis becomes a puddle
of cement ;)
(referring to the UK Metro's recent issues)
brycec: tnx for fixing newegg info :)
m0unds: yep
http://www.newegg.com/Product/Product.aspx?Item=N82E16813131982
ASUS Z87-C LGA 1150 Intel Z87 HDMI SATA 6Gb/s USB 3.0 ATX Intel Motherboard -- $129.49
potentially a strange question, but does anyone know how well freebsd runs on low memory machines (like 256/512 or less)
do i have to stick to pkg_add -r or can i do a lot of ports stuff
i'm using pkgng on a 768
it's alright except pkgs are still wonky on freebsd
but much more bearable than ports
256 could be quite different to 768
i imagine it's probably fine as long as you don't try using zfs
i'm more or less brand new to bsd
i usually do debian or something
i used to run openbsd on 64mb of ram
i kinda wish linux could have bsd's see_other_uids thing be part of core
hazardous: wasn't ther eplans to?
along with some much bigger frameowrk
i dunno how anyone is meant to keep up with linux
i'm not even using cgroups yet
memory use isn't even remotely a problem on a 768; the box uses less than 300MB on a regular basis
but it's also not running a ton of stuff
it's sitting at 206MB right now
haha
hazardous: what were you gonna do with it? just learning freebsd?
more or less
should be fine
i actually have a bsd box in "prod" right now except i have no idea how to use it
lol
all i know is i managed to set up a jail with an application server
i still have a Sparc laptop with OpenBSD 2.1. I think it has like 32MB of RAM. I used to use it for IRC; ran fvwm2 great! :)
whoa, haha
Sorry, no "anycast" jobs were found that match your criteria
http://www.workopolis.com/jobsearch/job/14847946?uc=E8
strong voip component
ahh they specialise in voip
you shooting for that job mnathani ?