you people surely have heard of traceroute6
yes
toddf: Yeah :P
ipv6 configuration confuses me :(
what about it
im running debian, i enabled ipv6 forwarding and installed radvd
kitkatbar: don't worry about mpls... nothing particularly to do with ipv6...
:s
supposedly i have a /48 block of ipv6, just trying to figure out how to use them
kitkatbar: what confuses you
you shouldn't need radvd
oh
... i would be very surprised if arp's routers would listen to your advertisement packets
kitkatbar: you dont need any extra software
you just need route and ifconfig
no wonder,  i have been going in circles
ok, i just arped arp
and i think i broke the internets
lol
kitkatbar: have you requested your /48 block from support yet
no i havent, but it shows up under my control panel
are you sure its not a /64
normally you must request to have the /48 routed to you over link local
ok i will open a support ticket
but im able to use myip:v6ya:know::2
:)
kitkatbar, i dont know if its helpful, but i made a wiki entry about setting up the /48
yea but wasn't it for fbsd?
Its on OpenBSD
http://wiki.arpnetworks.com/wiki/48%20IPv6%20on%20OpenBSD
but same principals apply
to any other os
ill take a look, thx
basically you enable packet fowarding
and use ifconfig to create a link local address
and add the link local address on the other side as your ipv6 gateway
then you can just add whatever ipv6 addresses you want
You fe80::2 <---------> fe80::1 Arp Networks
i see
basically for your subnet they add a route sending it to your ipv6 link local that you add
so they route the traffic to you, and you route it back to their link local
link local = local address that can only talk to things on the other side of the whatever it is connected to
so say you had two computers directly connected with an ethernet cable
they could talk over link local addresses
but could not talk to computers more than 1 level 2 hop away
interesting
yes, but its very simple once you understand it
:)
yea i been reading tons of articles about it, not thoroughly though...  but i see what your saying...  thx  :)
vcs will i need to request reverse delegation to be able to bind ipv6 addresses to domain names in bind & reverse?
yes
you will need to setup a dns server to serve your reverse zone
and of course request reverse delegation
you will need to provide a hostname of your dns server (have it valid for both ipv6 and ipv4)
ya i already run my own dns server but it is only valid for ipv4 atm
bhha
want my vps :(
kitkatbar: http://pastebin.com/XdcUv105
:P
example of a reverse dns zone file
for ipv6
nice
you > google
yeah it really is not difficult to set this up
its just there is very little info on ipv6 over link local out there :P
automatic reverse zone builder for the truely lazy - http://www.fpsn.net/?pg=tools&tool=ipv6-inaddr
LT: yeah, hehe
but by default it includes absurd TTLs
like 3 days
hmm.... I wouldn't think that matters too much on a v6 reverse, not like you have so few addresses you'd want to reuse one often?
well, true
but sitll, i leave all mine at 1 hour
why not :P
yea i thought i was going insane, reading nearly every ipv6 setup guide out there
guess the process of getting it over the local link was unknown to me :P
the thing most people don't understand is that you need to setup a link local address on your end
so they can route the packets through to you :P
its not just as easy as setting up their link local as a gateway
so i guess i should wait on the response email before i proceed with your guide, eh?
yes
ipv6 really is not that difficult to understand, there are just a few new things about it compared to ipv4 that trip most people up
and once they get that, they get ipv6
vcs: what are those things?
link local address for example of course
also the new ipv6 autoconfiguration
although dhcp still works fine i imagine
you imagine wrong... dhcp can't hand out the default route
also slaac can't out the dns server address
which is all very ... useful
ahh
when it comes to autoconfig I think the difficult part to understand is... what were they smoking
hahaha
the same thing they were smoking when they decided to give ipv4 32bits of address space
it wasn't supposed to be what it is now. ;-)
yeah i know i have read into it
it was an "experiment"
maybe if they WERE smoking something. we might have a bigger addresss space
hahaha
"dude man. can you see it? a working net of all.... devices and shit.... connected... treated as equals..."
*puff*cough* yeahdude.
lol
heh
rtsol/rtadvd can do dns, its just an extension that is 'newer' than the origin of v6
vcs:  I read your comment about absurd TTL's.   in general  TTL's are best set to the longest time you can reasonably get away with.  considering most ISP's would give you a few days (or weeks) of notice before re-IP'ing your netblock..  AND, you likely know in advance when you are going to move a server/service,   1d is not an 'absurd' TTL.  in fact,  it a really reasonable TTL.
with the obvious understanding that,  when you know a change is pending, you lower it accordingly (temporarily)
yeah
i understand how it works
and some some isps dont like it less than one day especially for reverse dns
and will add you to spam blacklasts for doing it
well,  that and.. many DNS servers will set an aritrary TTL,  instead of honoring those that are extremely low.
yeah
so you might set it for 1 hour,  but it's very possible that by doing so, you're actually telling many NS's to cache it for 4, 6, 12, or whatever that admin chose.  (personally,  I think that is an absurd practice.  the caching NS should honor whatever TTL the zone admin dictates,  but thats another conversation.)
setting an arbitrary TTL violates rfc, best common practice, and worse common sense. yes people do it, but I make it clear to any client who is moving domains around etc that if people are having issues beyond the TTL lag should permit, someone is doing something wrong, so they get what they deserve.
toddf: I agree wholeheartedly.  but,  if someone else doing something wrong impacts your ability to offer a service..   even though it's not "your issue"  it's still percieved by consumers who aren't as technically oriented..  as "your problem"
yeah, doing things last minute per customer demand vs planning things to take the time to permit idiot arbitrary ttl seettings is another discussion entirely though; I was focusing on the 'last minute per customer demand' scenarios
oh, misunderstood.  Yeah,  I refuse to be held accountable when someone wants a last second change, without proper preparation.  heh
you want cowboy style admin?  thats fine.  but don't bitch at me when you can't hang on to that bucking bull.
how offen are this up_the_irons on ?
coobra: a couple times a day usually
jpalmer: ;-)
ok
evening folks
mornin
morning
coobra: i will make your VPS tonight; sorry for the delay, i was at a going away party last night, drank a lot, and then slept ;)
heh nice :D
up_the_irons: did you get some  :p
coobra: haha, no i'm married
hahhaa
:D
lol
ohh nice
happy for you
She'll be out of school by then,  and we're already got the entire wedding and honeymoon paid for,  save for any expenses we incur while on the honeymoon. (room service, dining out, etc.)
lol
oh i suck so much at webdesign :-/
time for a new hobby
RandalSchwartz: http://i.imgur.com/1OgT0.jpg
is that a model M?
:D
YOUR MOM IS A MODEL M
(heavy, noisy, and generally built like a tank)
you'd still do her
tooth: yea, like zfs on root
it sounded good in my head
any schedule on the provisioning?
richardquisumbin: i'm doing a round tonight
thanks
np
lets sleep some  :D
anyone play w ubiquiti networks gear before?
wow. after an hour or so, it is really nice stuff
Anyone on kvr12?
Anyone experiencing connectivity issues?
My vps lost connectivity between roughly 10:00 - 10:15 pm local
Looks like it's back up now.
jazz57: i see an 800 mbps burst to hit kvr12
the vm hosts have 100 mbps links, so that definitely saturated it
Possibly a DDOS attack?
Any idea where that came from?
jazz57: possible.  if it was, it would surely be spoofed, so finding out the src IP(s) is not useful
Well I'm glad it didn't last longer than it did.
same here :)
up_the_irons: While I have you here, any updates on your carriers?  I noticed XeeX is out and I believe I read you were thinking of changing out MZIMA for Level3.  Is that still the plan?
jazz57: PacketExchange (formerly Mzima) seem to have got their act together, so I am keeping them for the time being.
Ping times are great, so I have no complaints.
jazz57: I'm going to be joining PacketExchange's private peering network, so I'll grab like 120 peers all at once.  Probably won't matter much if you live in the US, but I'm excited about this b/c one peer I will get is China Telecom, which is huge.  I can't peer with them directly, despite my best efforts (i'm too small for them to care)
time to provision some VPS'..  man i'm tired though
That's really cool.  You have a very nice network.  My other VPS provider is single homed with TWTC and they route most of their traffic through SJC.
Good luck and thanks for the info!
Thanks!
Glad you like it
have you guys used openvpn with ip6