**** BEGIN LOGGING AT Sat Jan 25 02:59:59 2014
Jan 25 04:24:07 <builder_> build #504 of atheros is complete: Failure [failed compile_8]  Build details are at http://buildbot.openwrt.org:8010/builders/atheros/builds/504
Jan 25 05:12:34 <builder_> build #506 of orion is complete: Failure [failed compile_8]  Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/506
Jan 25 05:15:06 <builder_> build #438 of sibyte is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/438
Jan 25 07:27:47 <builder_> build #448 of avr32 is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/448
Jan 25 08:27:57 <OpenWrt-Commits> juhosg r39393 trunk/target/linux/generic/patches-3.8/811-pci_disable_usb_common_quirks.patch * kernel/3.8: update pci_disable_usb_common_quirks patch
Jan 25 08:34:37 <OpenWrt-Commits> juhosg r39394 trunk/target/linux/ar71xx/image/Makefile * ar71xx: image: move WNDAP360 macro
Jan 25 11:19:54 <OpenWrt-Commits> juhosg r39395 trunk/target/linux/ar71xx/files/arch/mips/ath79/mach-archer-c7.c * ar71xx: archer c7: define LEDs connected to the AR8327 switch
Jan 25 11:51:46 <builder_> build #494 of cobalt is complete: Exception [exception]  Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/494
Jan 25 13:59:31 <dtaht_nuc> has anyone ever got a stun/turn server working with webrtc on openwrt?
Jan 25 13:59:45 <jow_laptop> webrtc ?
Jan 25 14:01:59 <jow_laptop> I mean "webrtc on openwrt" ?
Jan 25 14:23:24 <Devastator> can I discuss a bit about networking in here?
Jan 25 14:38:43 <dcosnet> sure
Jan 25 14:40:51 <Devastator> if I have two interfaces, eth0 and eth1, when default route is set to eth0 and for some reason eth0 has no connectivity, I have to change default route to eth1 before executing ping -I eth1 <ip>, right?
Jan 25 14:41:08 <Devastator> or I would get destination unreacheable, correct?
Jan 25 14:47:48 <vampo> I suppose it depends on what address you are trying to ping
Jan 25 14:48:22 <vampo> if it is in the eth1 subnet, it should work without route updates, even without the -I
Jan 25 14:48:34 <Devastator> external one, like 8.8.8.8
Jan 25 14:48:56 <vampo> it it's beyond eth1, then eth1 should have a gateway first
Jan 25 14:55:55 <vampo> are you trying to monitor connectivity of the 2 interfaces?
Jan 25 14:56:22 <Devastator> vampo yes, a primary link and a backup link
Jan 25 14:57:30 <Devastator> vampo I have this script: http://codepad.org/w025aJOx it switches fine to backup link, but it doesn't switch back to primary because when backup link goes down, ping -I eth1 -c1 8.8.8.8 returns destination unreacheable and it will only switch back to primary when backup goes back online
Jan 25 15:00:12 <Devastator> vampo I'm thinking something like.. check which gateway_if is set before executing ping, if it is backup_if, ip route del default via 2.3.4.5 dev eth1, ip route add default via 1.2.3.4 dev eth0, then execute ping
Jan 25 15:01:26 <Devastator> and I do the same thing before switching to backup link, because if both are down, then "quit for about 5 minutes"
Jan 25 15:02:07 <dtaht_nuc> jow_laptop: sure. Or better put, webrtc through openwrt, using a local stun/turn server, and perhaps call setup via a local web server
Jan 25 15:03:19 <dtaht_nuc> Someone had got it all working, but was stuck on submitting packages, and I confess I dropped the ball: http://comments.gmane.org/gmane.comp.embedded.cerowrt.devel/1268
Jan 25 15:03:37 <vampo> Devastator: I assume you always want to use the primary if the primary is up?
Jan 25 15:03:50 <Devastator> vampo exactly
Jan 25 15:04:07 <vampo> Devastator: what if you always switch to the primary if the backup doesn't respond?
Jan 25 15:04:17 <vampo> Devastator: even if the primary is down
Jan 25 15:04:53 <vampo> Devastator: worse case, it will cycle one more time
Jan 25 15:05:45 <Devastator> vampo yes, something like, while primary is down, keep checking both, the first one that gets back should be used
Jan 25 15:06:25 <Devastator> I better do this carefully or the routing table cache will become a mess
Jan 25 15:06:36 <Devastator> unless I flush cache during the checks
Jan 25 15:13:31 <vampo> Devastator: I think I see what you mean... the script can't check if the primary has recovered unless you change routes first?
Jan 25 15:13:56 <Devastator> vampo yep
Jan 25 15:14:44 <vampo> how about setting static routes. 4.4.4.4 always goes out eth0, 8.8.8.8 always goes out eth1
Jan 25 15:16:09 <Devastator> it's an option, then I have a fixed high available ip to test from each one
Jan 25 15:16:19 <Devastator> s/from/for
Jan 25 15:18:14 <vampo> it 8.8.8.8 is good enough, 8.8.4.4 should be just as good ;)
Jan 25 15:20:05 <Devastator> a guy just gave me another idea
Jan 25 15:20:43 <Devastator> policy based routing, primary goes out primary gateway, backup goes out backup gateway
Jan 25 15:33:25 <dtaht_nuc> me commits resiprocate from 2012
Jan 25 15:33:34 <dcosnet> hmm
Jan 25 15:33:49 <dcosnet> anyone here ever done load balancing routes?
Jan 25 15:34:07 <dcosnet> i assume iptables + some script will be needed
Jan 25 15:34:47 <vampo> help from the ISP is needed as well
Jan 25 15:35:21 <vampo> or the script has to be clever enough to maintain sessions on each interface
Jan 25 15:35:25 <dcosnet> or just 2 diferent ISPs no?
Jan 25 15:35:31 <dcosnet> yea
Jan 25 15:35:44 <dcosnet> especially for things using cookies, or ssl
Jan 25 15:36:54 <dcosnet> i can do HA and load balancing with nginx, just not at the network layer yet
Jan 25 15:36:58 <Devastator> and ftp, which is bad without load balancing
Jan 25 15:37:04 <dcosnet> also never tried so idk
Jan 25 15:37:09 <dcosnet> o i dont use ftp
Jan 25 15:37:14 <vampo> there are discussions out there about this on ipv6
Jan 25 15:37:50 <vampo> with twin ipv6 links outgoing trafic will go out of either line
Jan 25 15:38:05 <dcosnet> hmm
Jan 25 15:38:22 <dcosnet> my isp does support ipv6
Jan 25 15:38:25 <vampo> but return trafic will be dropped by the isp if they are not the subnet destination
Jan 25 15:38:29 <dcosnet> i just have it disabled currently
Jan 25 15:38:41 <dcosnet> mainly because i havent coded a ipv6 iptables yet
Jan 25 15:38:48 <vampo> I assume something similar applies to ipv4
Jan 25 15:39:35 <dcosnet> is there any good scripts to convert ipv4 rules to ipv6 for iptables? now that i think of it? heh
Jan 25 15:39:38 <vampo> or something along those lines... it's been years since I was playing with those tunnels :)
Jan 25 15:39:53 <dcosnet> ok
Jan 25 15:41:14 <vampo> there are some good initial scripts out there for iptables
Jan 25 15:41:31 <dcosnet> yea i coded my own though based on line 20 others
Jan 25 15:41:34 <vampo> and the rules are much simpler since there is no NAT involved
Jan 25 15:41:35 <dcosnet> :)
Jan 25 15:41:40 <dcosnet> ok
Jan 25 15:41:51 <dcosnet> s/line/like/
Jan 25 15:42:23 <dcosnet> i suppose dont need to mess with forwarding rules on ipv6 because of that
Jan 25 15:42:31 <dcosnet> ya nice
Jan 25 15:42:37 <vampo> basically, keep ping open and allow only the ports you need (equivalent of port forwarding)
Jan 25 15:43:00 <dcosnet> ah i block ping on ipv4
Jan 25 15:43:12 <dcosnet> pinging dcos.net should fail
Jan 25 15:43:50 <vampo> you can block in on ipv6 as well but "the rules" say that it should always be allowed
Jan 25 15:44:05 <dcosnet> ok, interesting
Jan 25 15:44:33 <dcosnet> i got an early tcp/ip book that covers ipv6 but its from around 2003?
Jan 25 15:44:34 <vampo> I remember there was some practical reason as well for that
Jan 25 15:44:35 <dcosnet> lol
Jan 25 15:44:50 <Devastator> correct, you can limit how many pings you want to reply, but it's good practice to reply pings afaik
Jan 25 15:44:51 <vampo> yes, that counts as early :D
Jan 25 15:44:56 <dcosnet> maybe as early as 1998
Jan 25 15:44:59 * dcosnet looks
Jan 25 15:45:24 <dcosnet> omg!
Jan 25 15:45:26 <dcosnet> 1996
Jan 25 15:45:31 <dcosnet> lol
Jan 25 15:45:56 <dcosnet> paid 1$ at a fleamarket
Jan 25 15:46:01 <dcosnet> couldn't say no
Jan 25 15:46:31 <dcosnet> and yes it has ipv6 in it
Jan 25 15:47:12 <dcosnet> addison/wesley tcp/ip
Jan 25 15:47:28 <dcosnet> UCN
Jan 25 15:47:54 <dcosnet> i suppose i should get a modern more practical one
Jan 25 15:49:37 <vampo> or google
Jan 25 15:49:52 <vampo> the HE site has some great info as well
Jan 25 15:59:15 <dcosnet> ya
Jan 25 16:26:21 * dtaht_nuc finishes procd-ing almost his entire universe
Jan 25 18:15:01 <OpenWrt-Commits> juhosg r39396 trunk/tools/firmware-utils/src/mktplinkfw.c * firmware-utils/mktplinkfw: allow to reserve space in the generated images
Jan 25 18:15:05 <OpenWrt-Commits> juhosg r39397 trunk/target/linux/ar71xx/image/Makefile * ar71xx: image: reserve 4 erase blocks on the TP-Link devices
Jan 26 01:11:33 <OpenWrt-Commits> luka r39398 packages/utils/lxc/Makefile * [packages] lxc: lxc-lua depends on luafilesystem
**** ENDING LOGGING AT Sun Jan 26 02:59:59 2014