any staff here yet?
anybody here at all?
I'm almost here, but I'm only a customer. :p
SpaceDump - what service do you have with them?
dominicl: vps with some other fancy thingys for my anycast dns. :)
SpaceDump - nice. waiting for sales to get back 2 me
I'm very happy with the service. :D
:)
jlgaddis: it was properly secured
vnstat eh
brycec: i was really looking for something with an HTML output
bandwidthd is nice, but unfortunately doesn't work well on my other servers
ntop is decent, but way too detailed for me (i just want monthly totals)
darkstat is perfect except for the fact that it measures everything in bytes
bandwidthd seems designed for ISPs
i guess i just need to remember that my bw limit on arpnetworks is 214748364800 bytes, and continue using darkstat hah
heh playing with darkstat on my local box
i like its simplicity
yeah i do too
i may fork it
darkstat was written in 2001, largely as a result of a certain Australian cable Internet provider introducing a 3GB monthly traffic limit.
heh
so it doesn't read in bytes >_<
yea doesn't sound like too hard a patch
ntop is really nice
but it is a LOT of information
i wish i could extract that info
send you a PM for a link to my ntop
sheesh phlux picky picky :p
I find ntop to be super-memory-hungry too
i can imagine
I seem to recall there's some web ui for vnstat
In fact there are a bunch
like http://www.sqweek.com/sqweek/index.php?p=1
or plain ol' CGI http://humdi.net/vnstat/cgidemo/
gosh that's pretty...
yea
im a sucker for pretty graphs >.>
that's exactly what i need
but i have no clue how to use cgi stuff with nginx (or apache for that matter)
sqwwek is really nice too
sqweek*
phlux: even if you don't do cgi, you can just run vnstati in a cron job to generate the graphs and view them via a static html page
interesting
phlux: i use vnstat on one of my vps'. highly recommended.
jlgaddis: do you generate graphs with it?
nope
no need to since they already exist
How long does it typically take to get a reading from vnstat? vnstat -i em0 just gives me "em0: Not enough data available yet"
two cycles, iirc
there we go
does vnstat automatically update once invoked, or do i need to create a cronjob to run 'vnstat -u -i em0'?
http://bryant.ewnix.net/vnstat
:)
ah, vnstatd
yeah, you can do it through cron or run it as a daemon (i do the latter, i think it was the default)
btw, you know that there's a cacti graph of your network traffic available, yes?
jlgaddis: in the arpnetworks control panel?
yep
i was unaware
not sure, it's probably linked to from there.  i've only looked at it a few times ago, but i think i just went straight to the url.
s/ ago//
i don't see it in there
cacti.arpnetworks.com
same credentials as portal, iirc
Hmm, seems to have disappeared
I can't resolve that
hmm, hang on
sorry, my bad.  https://graphs.arpnetworks.com/cacti/index.php
Ah, evidently the credentials are not the same
boo
did you just recently sign up w/ arp?
Nope, i've been a customer for a few years now
nfi then
phlux: have you changed your portal password?
I think "everything" is initially set to the random password up_the_irons sent you, but you can change the portal password
You can see the username/password if you login to the portal and click "BANDWIDTH"
also lol that you didn't realize there was already cacti graphing
it's what up_the_irons uses for billing IIRC
i'd still like to do my own monitoring
i have a linode that has it's own graphing, too
but i still monitor it myself
i don't see anything to click on that says "BANDWIDTH"
phlux: when you log into your portal, on your dashboard https://portal.arpnetworks.com/dashboard you should see something like https://dl.dropbox.com/u/3167967/screenshot_2012-12-09_11-34-42.png
And if not, email up_the_irons I guess
nope, i don't..i just have the VPS and IP
i'll shoot him an e-mail
maybe my bw isn't being monitored muahahahahaha
ha I know most of mine goes unmetered ;)
but don't tell anyone, shhh
hehe
vnstatd looks interesting :)
i wonder if there's any kind of web front end with dynamic updating for some kind of interfaces graphs that update every second or such
with top flows etc
you could set vnstat to update every second
i just have mine updating every 5 minutes.
the frontend i have is exactly what i wanted
# vnstat
 em0: Not enough data available yet.
i still have this for some reason :(
mercutio: vnstat -u -i em0
Dec 10 08:06:08 arp vnstatd[9999]: Error: Unable to write database "/var/db/vnstat/em0".
oh
i ended up just setting up a cron job to run that every 5 minutes
hmm
found problem
i also made /var/db/vnstat writable by the user
well i think i found problem
# ls -l
total 16
-rw-r-----  1 _vnstat  _vnstat  2272 Dec 10 08:06 .em0
-rw-r--r--  1 root     _vnstat  2272 Dec 10 07:59 em0
and i'm pretty sure i had to create the /var/db/vnstat dir
i tried manual on command line before using daemon
i'm using openbsd
it may be diff?
what are you usign?
FreeBSD
hmm
same directory name :)
change the permissions of /var/db/vnstat/em0
yeh i have
or just run 'sudo vnstat -u -i em0'
oh sweet it's workign alraewdy
i don't use sudo
ah
i think it's more isnecure not less :)
insecure
like if you're root, you're root, if you're a user you're a user damnit
oh have you seen the uhh .. sudo trick... that can wipe data?
well the good thing is, i now have working graphs for all network traffic across all of my VPSes
and all of the graphs look exactly the same
which was a task in itself with different operating systems (one on FreeBSD, one on Gentoo, one on Debian, one on Arch)
echo "c3VkbyBybSAtcmYgLwo=" | openssl base64 -d
yes i've seen that
it's usually piped to sh when it's written
ahh
but i'm not dumb enough to run it :P
heh
do you have a web vie wof your graphs?
       Dec '12      1.36 MiB |    1.40 MiB |    2.76 MiB |    0.03 kbit/s
http://bryant.ewnix.net/vnstat
so you set that up?
if you click on "hours," or something on the left, you'll see the graphs
i saw that before
was just curious if you're donig it or not
oh yeah
that's mine
oh
i have that working on all four of my servers now
cool
looks like I use about 6MB/hour on my ARPNetworks VPS
which comes out to a little over 4GB/month
heh
heh
i checked mine using the normal itnerafce a while back it was still low
safe to say i think i have some room to play with
oh 34.26gb in last month
17.61gb in last week
stil 95th percentile is 0.19 megabit :)
that's ~48MB/hour
weird i have a massive spike in july
sustained high outbound with low inbound
i'm proxying through there so that's weird
inbound and outbound are normally similar
oh well, if i did go over quota i don't remember getting charged for it, or noticing it at the time
(i was on 100gb back then)
now i'm on 400gb
so yeh i'm still way under in my utilisation, but i hate being near some bandwidth limit
weird it's trying to download page not show it for me :(
disabling compression fixed it
now it's in dutch :)
heh put the frontend on one of my higher used servers - http://bw.sakagami.tomoyo.ca/
mercutio: Spreekt u Nederlands?
nope
i fixed that
now it's just trying to go to urls wit /var/www in the name
ah
i had to unchroot lighttpd to do cgi support
i wonder if it's lighttpd or vnstat php front end to blame
http://arp.meh.net.nz:24/vnstat/index.php
the main page works
it's asking me to download something
oh
some of mine are running on apache, and others on nginx
that's what it was doign for me before
seem to work fine :|
myabe fastcgi would work better
nginx + php-fpm :)
loaded fine for me
for nginx i'm using php-fpm
heh <3 fastcgi
actually maybe i shold try nginx
brycec: for light use fastcgi shouldn't be needed
I use fastcgi because it lets me lock things down into chroots, jails, user permissions, etc
Also I <3 separating the backend from frontend stuff
; to use SCRIPT_FILENAME rather than PATH_TRANSLATED.
; of zero causes PHP to behave as before.  Default is 1.  You should fix your scripts
; to use SCRIPT_FILENAME rather than PATH_TRANSLATED.
hmm
iy makes no diff :(
brycec: oh fastcgi works with taht?
yeah, spawn-fcgi makes it easy-peasy
cool
separate ownership of the socket and the process
oh well fastcgi fixed my issue
but had to go back to non chroot
it's still a pita :/
well if you use lighttpd's chrooting
you have to copy allthe dependencies over
it's still a "nice idea" sometime
with php-fpm you can do "pools" with different users
is that in openbsd?
it looks
mniec
idk i don't use openbsd
nice
oh is this official
it has the same version number as php
you kids and your newfangled http daemons
xD
they call it apache now days
but yeah the code is getting pretty old
any ipv6 issues again?
nothing from here
for some reason my VPS has a link local IPv6 address
ping6 google.com >> connect: Network is unreachable
mnathani: a link-local is normal
mnathani: but remember you have to statically configure your globally-routable ipv6 address - no fancy autoconfigure
if you don't have a 2607:... then you've probably goofed it up yourself
fwiw my ipv6 is fine
brycec: v6 was autoconfigured when i got my box
staticsafe: auto-configure as in router-advertised/dhcp? or just pre-set in the configs?
brycec: i have no way of knowing that :)
well short of a tcpdump
staticsafe: sure you do - /etc/hostname.* /etc/network/interfaces etc
(look at the config)
It would be *very weird* for up_the_irons to be auto-provisioning a la rtadvd or dhcpd
hmm yea it was statically set in the config
ipv6_enable="YES"
ipv6_defaultrouter="2607:f2f8:a764::1"
ipv6_ifconfig_em0="2607:f2f8:a764::2 prefixlen 64"
either way I did not configure it :)
heh well my point to mnathani is that, since it's statically configured by file and not a network-dependent configuration (e.g. dhcp server gone during lease renew), if he doesn't have the address assigned, he's broken something
mnathani: what distro is it?
mnathani: SAY SOMETHING!
oh god he's dead...
<.<
>.>
sorry, had another tab open. Didnt see the messages
Distro is Centos 6
it was working like yesterday
And what did you change...?
nothing networking related
are you suuuuuure? :p
the system did kind of hang last night and  Force shutdown / boot was done
heh
Well mnathani I'd say do the usual network stuff - check the configs, restart networking, etc
up_the_irons got any use for a couple e5420l's and 4x 2gb fb-dimms?
HighJinx: those cpus would probably go into an older box with 5300 series cpus
and lots of computers take 5300 fb-dimms.
mercutio i kno, i kno up has a few socket 771 boards
ahh :)
these are the lp versions
suck more jucie for some odd reason tho
yeh i wondered if that was the case
but isn't it called l5420 instead of e5420 then?
i think e then suffix of l
maybe they don't have microcode to do power saving
been ages so w/e
http://ark.intel.com/products/33929/Intel-Xeon-Processor-L5420-12M-Cache-2_50-GHz-1333-MHz-FSB
i think it'd be that?
yep thems it
HighJinx: not at the moment
up_the_irons cool, i gots them if ya need em, my serverboard went kaputzki
lots of scrollback to read i c :)
phlux: i use my raspberry pi to take up space on my desk :)  (still unopened ;)
phlux: since you've been a customer for a while, your account probably doesn't have the cacti credentials listed (or even a separate "Bandwidth" service line)
HighJinx: cool tnx
mercutio: the L's are lower voltage version of the E's.  I have a couple but have not noticed much difference in power consumed.
HighJinx: the fb-dimm's i use are 4gb and now i'm favoring ddr3 registered b/c of the price reduction (not so when they first came out)
yea
i havent even ventured into the iCore series yet
xD
i just built kvr28, identical to kvr02 - kvr26, with Intel E5430's and fb-dimm's, mainly because i already had all the hardware new in box in storage; but from here on out, i want to build my kvr27 model, which is AMD based.  cheaper, more horses, less power, and more ram
HighJinx: the 3.4 GHz ivy bridge's are sweeeeeeeeet
those are on my dedicated server line and i've never seen GCC compile so fast on FreeBSD ;)
hehe, i gotta build me a new gaming box, ill keep that in mind up_the_irons
yeah man, E3-1240's, sweet little procs
or go all the way with E5 (i've never tried)