[00:06] *** tooth has quit IRC (Remote host closed the connection) [14:03] is it poor practice to run ntpd from VPSs ? [14:03] i had it suggested to me it was wasteful [14:10] That's odd. My ntpd is still unable to sync [14:10] maybe arp changed policies [14:10] used to work just fine [14:11] what policy would there be to disallow this? filtering 123 or something? [14:12] I'm checking another arp box [14:12] the non-VPS arp box is peering ntp just fine [14:12] ahh - this other VPS is peering only via v6. [14:12] perhaps a ipv4 port 123 block has been installed for all VPS [14:13] When the big NTP DDOS was going down, ARP firewalled it [14:13] ahh [14:13] I thought up_the_irons had un-blocked it since [14:13] perhaps we can turn it back on then? [14:13] FWIW RandalSchwartz, my NTP is peered v4 and v6 just fine [14:14] maybe just some netblocks then [14:14] (or at least I think it's just fine...) [14:14] RandalSchwartz: gimme server IP and i'll see if it is in the "bad ntp hosts" list [14:14] <-- not an expert on ntpq output [14:15] up_the_irons: any interest in running an arp time server to go with the dns servers? it'd be nice to hit a local time server vs laboring the ntp pool [14:17] is this still accurate? http://support.arpnetworks.com/kb/main/how-to-configure-ipv6-on-freebsd [14:17] I can't find my mail on how to repair my v6. [14:17] yes, i'm interested, just not enough time to do it [14:17] RandalSchwartz: i believe it is accurate up to 9.x [14:18] Nice pun. [14:18] hahahaha [14:18] maybe a bunch of us customers could chip in to bounty fund it? [14:18] i didn't even notice... [14:18] i'd offer $10 toward it [14:18] so if my default route is $x::1, can I still use $x::0 as my IP address? [14:18] i'd consider that :) [14:18] RandalSchwartz: i believe so; no specific ordering on the gateway is required AFAIK [14:19] anyone else be interested in pooling in to get an arp time server or two in place? [14:19] i suppose even 1 would suffice, with a preferred tag on it in ntp.conf [14:21] actually i'm not sure if prefer reduces load to other servers, or if it just bumps the weight-of-value [14:22] I don't think throwing more money at up_the_irons is going to give him the time to set one up. [14:23] *shrug* [14:23] well, it would raise the priority, cuz instead of working XYZ hours on $client work, I can then do the NTP servers [14:23] I stand corrected :) [14:23] PRIORITEHZ! [14:23] how about a twitter broadcast? i'd suggest email blast but that's for up to decide [14:24] could there be an order form set up for 'sponsor arp ntp servers' that customers can easily go to, put in the contribution amount, and hit order? then you can just process using cc on file [14:24] twitter broadcast, yes; email blast, no (the only emails that go to everyone are the sales receipts) [14:24] i'm trying to think of the simplest way to do it [14:24] that would make it even harder ;) [14:25] yea [14:25] cuz i'd need to make the form [14:25] how would you want to structure this? [14:25] and cc processing [14:25] yea [14:25] Simplest thing: Bitcoin wallet. [14:25] hmm, interesting [14:25] i don't have bitcoin unfortunately, but i'd still want to contribute toward this [14:26] darn it... doesn't look like you can reset ipv6 while the box is up... I guess I have to finally reboot. [14:26] no kidding? [14:26] i think there are a handful of a sites that do bounty stuff [14:26] service netif restart; service routing restart ? [14:27] well pick one and i'll be your first to donate server brother [14:27] RandalSchwartz: how much uptime you have on that? [14:27] Uh - I don't want to kill my v4 [14:27] which host? kvr08? [14:27] but it looks like there's no "stop" or "restart" commands for the v6 stuff [14:28] why not just do ifconfig manually ? [14:28] well I'm afraid of screwing up v4. :) [14:29] fortune favors the bold [14:30] Wow, now that's a heckuva reason to run updates on a Friday afternoon :) [14:31] fortune -a favors all. [14:32] is there an easy way to remove all ipv6 routes so it'll build them up again? [14:32] maybe this for bounties: https://www.bountysource.com/ [14:32] dunno if it is only for code though [14:32] RandalSchwartz: ah ok [14:34] route flush inet6 ? [14:35] no... that's still not enough :( [14:35] ... default 2607:f2f8:3080::1 UGS em0 [14:36] but ping6 ip6.me doesn't work [14:37] em0 says inet6 2607:f2f8:3080:: prefixlen 64 [14:37] RandalSchwartz: you firewalling icmp6 by any chance? [14:37] no [14:37] curl ip6.me also does nothing [14:37] can you mtr -6 red.stonehenge.com ? [14:37] http://www.reddit.com/r/cryptodev/comments/24u17d/bountysource_now_supports_bitcoin_for_funding/ [14:37] cryptodev approved :) [14:37] icmp6 is required for neighbour discovery in ipv6. I wasn't asking about the ping :P [14:38] I have no v6 firewall rules that I'm aware of. [14:38] I'll check though [14:38] RandalSchwartz: nope. I also can't ping your ::1 gateway [14:38] so something is still broken there. [14:38] (traceroute/mtr gets me 2607:f2f8:a654::1 and nothing further) [14:38] (coming from ARP btw) [14:38] up_the_irons? need a ticket on it? [14:39] yes, but trust me, nothing has changed [14:39] bbiab [14:39] LIES. As a service provider, all you do is change things and lie about it and make people upset, because that's a solid business strategy :P [14:40] ^ sarcasm, btw [14:40] Hm my ntp just isn't happening... nothing in tcpdump [14:40] RandalSchwartz: You don't have a /48 by any chance? [14:40] yes, I have a /48 [14:41] RandalSchwartz: Ah, so you should be using ff.... for your gateway [14:41] fe80::1 [14:42] RandalSchwartz: and set fe80::2/64 as your interface's ip [14:42] any other addys like 2607:f2f8:3080:: are just aliases [14:42] (The wiki article you cited is only for the default /64 type configuration) [14:43] how do I say that in rc.conf? [14:43] RandalSchwartz: the IP? Just replace 2607:f2f8:3080:: with it. The aliases... fuck if I know :P [14:43] well... I have to have that [14:43] https://forums.freebsd.org/viewtopic.php?&t=26400 [14:44] A nice purple link... [14:45] RandalSchwartz: so it would be like ifconfig_interface_ipv6="inet6 fe80::2 prefixlen 64" ifconfig_interface_aliasX=inet6 2607:f2f8:3080:: prefixlen 64" [14:45] (I like to expand the :: in configs for clarity too, but that's just me) [14:47] yeah - I got that [14:47] ok em0 looks good [14:47] routes look good [14:47] but I still can't get out [14:48] no - fe80::2 is only on lo0 [14:48] that can't possiibly work [14:48] and default also went to lo0 [14:48] also not working well [14:48] unless that lo0 is magic [14:48] lol so your config didn't apply right :P [14:48] ... ipv6_ifconfig_em0_alias0="fec0:0:0:2::80/64" [14:48] .. ipv6_ifconfig_em0_alias0="fec0:0:0:2::80/64" [14:48] RandalSchwartz: fe80:: is a link-local address, so it can be wherever [14:48] oops [14:49] right but there's no fe80::2 on em0 [14:49] Specifically, you put it on your em0 and up_the_irons has his stuff configured to point at you [14:49] RandalSchwartz: well you need to put it there :P ifconfig_interface_ipv6="inet6 fe80::2 prefixlen 64" [14:49] (I assure you this works for many customers, including yours truly, albeit I'm on Debian) [14:50] I don't have anything that said ifconfig_interface_ipv6 before [14:51] do I need to add %em0 to ipv6_defaultrouter ? [14:52] does accept_rtadv need to be 1 or 0 [14:52] AHH [14:52] that did it. ipv6_defaultrouter="fe80::1%em0" [14:52] essential piece missing [14:53] Yes, link-local address needs the link specified [14:53] congrats [14:53] try mtr -6 red.stonehenge.com [14:53] works [14:53] * brycec proceeds to DOS you [14:53] and curl ip6.me shows 2607:f2f8:3080:: properly [14:53] >.> [14:53] all is good now [14:53] <.< [14:53] 18684 packets transmitted, 18683 received, 0% packet loss, time 18762ms [14:53] Not bad [14:54] 1kpps [14:54] er, 100pps [14:55] should my address also be fe80::2%em0 ? [14:55] I don't have that there [14:56] oh wait, that's inherent [14:56] yeah :) [14:56] I'm config'ing _em0 [14:56] Oooh... and ntp is working too [14:56] Lucky you :) [14:58] * brycec should figure his out [15:02] woo grats randal [15:04] Yeah - I haven't had ipv6 since like february [15:04] oooh - that means I can connect to ipv6 freenode again [15:06] now I just need to understand where /48 fits in :) [15:09] Hmm. [15:09] RandalSchwartz: pretty simply, ARP just routes any traffic destined to that /48 to your fe80::2 [15:10] All I did was copy the information from the portal into /etc/hostname.em0. [15:10] Simply put, it's just there. Add aliases as desired. [15:10] caze from #c? [15:10] #c? [15:10] CaZe: Which is great if you're on OpenBSD, however RandalSchwartz is not... Also, has a /48 which gets routed differently from the /64 most users have. [15:11] Well that's just being crazy. [15:11] It is. And that's why up_the_irons has a big warning "if you can't figure this out, I'll just switch you back to a /64" or such [15:14] "Please note, however, that this is an experts only option. That is, we assume you know how to set up your side. We will not provide additional support for your end of the link; rather, we will revert it back to the default /64 setup if you can't get it working." [15:14] http://support.arpnetworks.com/kb/main/what-is-the-difference-between-my-ipv6-64-assigned-space-and-48-allocated-space [15:40] so does anyone else run tinyproxy on their vps? [15:41] phlux: as it happens, I do [15:41] never use it... but it's there if I need it [15:41] oh... so shouldn't both of my fe80:: be /48 instead of /64 ? [15:42] no wait, that'd be silly [15:42] Wow, answered yourself before I could :p [15:42] Yeah the /64 applies to the fe80:: subnet [15:42] instead... I carve out the next /64, and it could be on a separate segment [15:42] repeat 65534 more times :) [15:42] Bingo. [15:43] do I turn on rtadv at some point? [15:43] or is this more or less automatic now? [15:43] No need. ARP doesn't do rtadv or anything [15:43] rtadv/slaac/etc has no place here [15:43] Everything you just did is statically configured [15:46] I mean if I start building a topology of /64's [15:47] *** RandalSchwartz has quit IRC (Quit: RandalSchwartz) [15:47] *** RandalSchwartz has joined #arpnetworks [15:47] *** RandalSchwartz has quit IRC (Changing host) [15:47] *** RandalSchwartz has joined #arpnetworks [15:48] wb [15:48] there. on IPv6 on IRC again [15:48] just like the old days [15:48] I suppose you could, if you wanted to... But I would stick with statically configuring things. [15:48] (depends what you're using the /64s for as well) [15:50] I'd like to allocate a segment for openvpn, and a segment for the other end [15:50] I think openvpn is compatible now... [15:51] I know there's something you can do with ipv6 and openvpn... but I haven't messed with it. [15:51] I only just started setting up my own ipv6 tunnels [15:53] *** mus1cb0x has left "WeeChat 0.4.2" [16:14] ... http://openvpn.net/index.php/open-source/faq/77-server/287-is-ipv6-support-plannedin-the-works.html [16:29] so i moved my network over to inspircd due to a majority vote [16:29] and now i feel like a loser [16:35] considering my speeds were roughly 1/5 of this a few weeks ago, I'm happy with this: http://www.speedtest.net/my-result/3487294027 [16:35] also lol@carrier/server listings [16:56] ew inspircd [17:01] eww IRC, oh wait... [17:02] could be worse... we could be on EFNET. :) [17:04] *** jlgaddis has joined #arpnetworks [17:04] *** jlgaddis has quit IRC (Changing host) [17:04] *** jlgaddis has joined #arpnetworks [17:17] wow IRC had 10 million peak users in 2003, but now averages 400k users at peak [17:19] I wonder if they counted AOL chat in that [17:20] Yes! Figured out my boneheaded ntp problems. It's no wonder NTP never even tried to reach out to my peers, I had "interface listen lo". You'd think that only meant listen, but apparently it tried to use that for outbound connections to peers too. [17:23] "listen lo" is like some navel gazing thingy [17:27] Thank you #arpnetworks for motivating me to figure out my issue :) [17:43] *** hive-mind has quit IRC (Remote host closed the connection) [17:44] *** hive-mind has joined #arpnetworks [18:13] up_the_irons is at Starbucks... again. :) [18:14] How does RandalSchwartz know this? [18:15] Foursquare checkins [18:16] Same way he knows I'm at Busby's West. :) [18:23] aol chat was irc...? [18:29] I don't think it was [18:38] I was told by an AOL insider that they used the ircd code [18:39] it never "appeared" to be IRC [18:39] but it had most of the same basic infrastructre for scaling [18:43] RandalSchwartz: LOL [18:45] irc.aol.com used to link to efnet iirc [18:47] netcom "chat" was irc. [18:48] * RandalSchwartz pulls up up_the_irons current lat/lon [18:49] whoa can u do that from FourSquare? ;) [18:49] you're facing, northwest right now? :) [18:54] *** mus1cb0x has joined #arpnetworks [18:54] VPSs don't have 2 NICs correct? (one for public network traffic, the other for private/vlan) [19:01] I think those both show up as em0 [19:01] (freebsd) [19:02] I'm surprised to see that irc.aol.com is a thing [19:02] And by thing, I mean DNS A record [19:02] both? [19:02] i wonder if aol 'rooms' were backed by irc servers in the past [19:02] mus1cb0x: there's just a single interface. [19:02] ok ty [20:05] *** mus1cb0x has left "WeeChat 0.4.2" [20:46] heh aol [20:46] ME TOO!!!!!!!!!11 [21:41] % telnet irc.aol.com 6667 [21:41] Trying 205.188.149.22... [21:42] so it doesn't seem to work [21:42] or maybe you have to be on aol to use it [21:42] does aol still exist? [21:43] http://get.aol.com/plans/dial-up-internet.php?regtype=upsell&ncid=crosssellusaolp00000044 [21:43] seems they do