[00:13] mnathani_: The only theoretical limit on q-in-q-N-q is set by the MTU, and any implementation specifics. The "-in-q" just eats into the payload size. Theoretically, given a 1500 MTU (1454 payload) you could have 363 VLAN tags. [00:16] (And jumbo frames MTU=9000 yields 2.238 layers of VLAN. [00:16] ) [00:26] thats a lot of vlans [00:40] *** gluffis has joined #arpnetworks [00:41] mtu is per vlan [00:41] the physical link can support more than 9k mtu sometimes. [00:41] not necessarily much more [00:46] actually usualyl if you can set 9k mtu you can do 9k mtu + vlan, i never 8996 mtu support [00:59] maybe it's time to upgrade my old OpenBSd 4.7 machine to 5.4 :) [01:08] yeah probably, there's a few issues with upgrading [01:08] like old packages won't execute [01:12] [04:27] mercutio: usually is :) something to do this weekend [04:33] yeah it got worse a little bit [04:33] like with the timestamp change [04:34] ok [04:34] it used to be pretty easy [04:34] it's still not really difficult, just annoying :) [04:36] hehe [06:15] I've setup siteXX.tgz and just re-install whenever a new release comes out [06:19] that is one way to do it :D [06:22] this works great except on my Linux boxes where it's never as easy [06:23] we are, it seems, supposed to use some gigantic "configuration management" software that creates more problems than it solves [06:56] systemd, since Lennart knows best :D [06:59] I can't wait until systemd has an IRC client. [06:59] Then I won't have to use screen anymore! [06:59] haha [06:59] 'lets fix what is not broken' [08:09] *** dj_goku_ has quit IRC (Ping timeout: 264 seconds) [08:10] CentOS assigns my workstation a different IP address every time I boot [08:10] apparently it doesn't turn the network on until I log into the console, which makes the whole network feature of this computer useless [08:10] I blame NetworkManager. [08:11] I meant to SSH to it [08:11] and with all the news we hear about lack of IPv4 addresses, I now have 6 public IP addresses assigned to my desktop [08:11] Haha. [08:11] someone is lucky enough to work at a university that got a /16 way back when... [08:12] Yeah, my university had the same situation. [08:13] *** dj_goku has joined #arpnetworks [08:13] *** dj_goku has quit IRC (Changing host) [08:13] *** dj_goku has joined #arpnetworks [08:15] disabling networkmanager did the trick [08:16] Word. [08:16] now if only I could disable systemd [08:16] :( [08:17] they must have made firefox talk to networkmanager because firefox thinks it's not online [08:18] and of course that's more accurate than getaddrinfo("mail.google.com") [08:18] Oh yeah, that's a thing. [08:24] *** _Zodiac has joined #arpnetworks [08:40] networkmanager includes a really awful, half-baked attempt at network location detection [08:40] and those xml configs, yuck [09:13] mkb: If I had to guess, the network stack not being configured until you log in sounds like dependency based startup working perfectly and you not having any network services like sshd starting on boot [09:43] *** _Zodiac has left "Ciao" [10:07] gluffis: why not 5.6? [10:07] (And speaking from experience, upgrading from the ISO is a painfully smooth process :) ) [14:38] brycec: did you have to update all of your packages? [14:38] s/update/uninstall + reinstall/ [14:38] brycec: did you have to uninstall + reinstall all of your packages? [14:38] i was skipping between -currents. [14:38] For 5.6, yes [14:38] (or was it 5.5? I can't remember) [14:38] yeah i can't remember when it was too it was recentish [14:39] then i upgrade my base before kernel [14:39] because the kernel wouldn't compile [14:39] Luckily my OpenBSD servers are mostly base stuff [14:39] and it wouldn't execute anything [14:39] so i had to boot with the snapshot kernel [14:40] wow, you're either doing a lot of things custom, or you messed something up royally :p I did the 5.5 (and 5.6) upgrades completely headless without issue [14:40] thish is 5.7 i think [14:40] 5.7 hasn't been released yet [14:40] and going from -current to -current [14:40] (as "stable") [14:40] not for another 2 months or so [14:41] OpenBSD arp.meh.net.nz 5.7 GENERIC.MP#834 amd64 [14:41] OpenBSD 5.7-beta (GENERIC.MP) #834: Tue Feb 3 18:08:48 MST 2015 [14:41] ok it's in beta [14:42] usually i can get by with extracting tarballs from snapshots [14:42] and doing my cvs upgrade on kernel [14:42] http://www.openbsd.org/faq/upgrade56.html#upgrade "Upgrading without install kernel" Fun times :) [14:43] heh [14:43] the 5.6 was more annoying yes [14:43] because i changed root shell :/ [14:44] so it was the 5.4 -> 5.5 upgrade that required all packages to be removed/reinstalled [14:45] oh [14:45] ok 5.5 was more annoying :) [14:50] *** dj_goku has quit IRC (Ping timeout: 265 seconds) [14:53] *** dj_goku has joined #arpnetworks [14:53] *** dj_goku has quit IRC (Changing host) [14:53] *** dj_goku has joined #arpnetworks [15:33] *** novae has joined #arpnetworks [15:33] *** toddf has quit IRC (Ping timeout: 245 seconds) [15:34] *** toddf has joined #arpnetworks [15:34] *** ChanServ sets mode: +o toddf [15:35] *** novae has left [17:50] *** dj_goku has quit IRC (Read error: Connection reset by peer) [17:50] *** dj_goku_ has joined #arpnetworks [17:56] *** dj_goku_ has quit IRC (Ping timeout: 240 seconds) [18:01] *** dj_goku has joined #arpnetworks [18:01] *** dj_goku has quit IRC (Changing host) [18:01] *** dj_goku has joined #arpnetworks [19:11] *** dj_goku has quit IRC (Ping timeout: 255 seconds) [19:57] *** dj_goku has joined #arpnetworks [20:04] *** dj_goku has quit IRC (Ping timeout: 264 seconds) [20:37] *** dj_goku has joined #arpnetworks [20:37] *** dj_goku has quit IRC (Changing host) [20:37] *** dj_goku has joined #arpnetworks [21:20] Are BSD's a type of Unix? [21:20] freebsd, netbsd, openbsd etc [21:20] and why are they better suited for firewall type workloads? [21:22] *BSD is derived from AT&T Unix, yes [21:22] (Where as Linux was a work-alike cloen written from scratch) [21:23] *clone [21:23] The answer to your second question is much broader, and it depends on the flavour more than "BSD in general" [21:23] But boils down to, not because it's BSD, but because their developers worked really hard at goal X [21:23] OpenBSD's goal is security at every layer [21:25] thanks brycec [21:25] np [21:25] do the 3 BSDs I mentioned all have ports / packages [21:26] Yes [21:26] (As does PC-BSD and DragonflyBSD, two other popular flavours) [21:30] openbsd is really the only bsd that's got good firewall afiak [21:30] afaik [21:30] freebsd has really old version of pf [21:30] i suppose it's "okay" [21:30] openbsd's working pretty actively on their network stuff. but i think it's mostly only a few intelligent programmers. [21:30] whereas linux has way more people working on it.. [21:30] It's okay, and not "insecure" necessarily, but it lacks improvements/features added since then. [21:31] ^ re: FreeBSD's pf [21:31] freebsd actually forked pf [21:31] and multithreaded it. [21:31] so it's hard for them to sync back [21:32] the only really big problem with openbsd is performance. [21:32] they're only just starting to add kernel level smp. [21:32] Yeah, performance is "good" but it's rarely their target. [21:32] but for most people they have way more performance than they need anyway. [21:32] like using a 3 ghz cpu to run word :/ [21:33] but yeah i wouldn't really go with openbsd for 10 gigabit. [21:33] but gigabit is fine. [21:33] Only performance issues I've ever noticed on OpenBSD were disk i/o. Still, quite good, but not 6Gbps :p [21:33] and even then, 10 gigabit should work, it just wont' scale as well for many packets etc. [21:34] brycec: on virtual or real? [21:35] i used to use openbsd as a desktop. [21:35] On virtual - virtio makes an enormous difference. On real, there's not much else to do :p [21:35] it's pretty stable generally. things work or don't work. [21:35] it's not "random" [21:35] whereas linux seems to get random weird obscure problems. [21:36] i kind of got turned off linux years ago when ethernet reordering kept happening [21:36] then people started triyng to "fix" it but it meant if you changed ethernet cards it'd increment to a new higher name. [21:37] that said, linux had way more problems back then. [21:37] it's curious how it's become "linux or bsd" [21:37] where people seem to lump all of the bsd's together. [21:40] And yet it's Linux that is splintered into "distros" and each BSD is just itself. [21:40] heh [21:40] i don't like any of the linux distros [21:41] i use ubuntu and arch, but both are frustrating [21:42] basically combintaitons of rolling release and stable core are necessary i reckon [21:42] which is what freebsd-stable is meant to be [21:42] but the complexity of both can be kind of overwhelming [21:43] It's how I view Arch for that matter (leaning towards rolling-release, but they do still have stable/testing tiers) [21:48] they keep updating the kernel [21:48] and then your modules don't work [21:49] that's my biggest issue with it atm, and my custom kernel seems to have issues on arch atm [21:49] i think i need some option enabled [21:49] but it does waiting on enp0s25 or something [21:50] Why do you run a custom kernel? [21:51] so that i can do system updates without having to reboot :/ [21:51] mercutio: just don't install the updated kernel until you're ready to reboot [21:52] but i generally run custom kernels in most places anyway. [21:52] well the same custom kernel [21:52] yeah there may be a way to hold it back [21:53] there is... Ignore=linux [21:53] ahh [21:53] Or --ignore=linux if you feel like typing it every time [21:53] i'll probably just end up figuring why my network and serial config don't load with my custom kernel [21:53] And anything that depends on the kernel upgrade (eg: virtualbox host modules) will also be held back. [21:54] i also compile zfs from soruce [21:54] instead of using one the existing packages? [21:54] well don't use a package [21:55] yeah [21:55] i think packages usually compile from source too actually [21:55] ubuntu's packages were mental for zfs. [21:55] mercutio: AUR are compiled from sources on fetch/makepkg, yes [21:55] are you using arch? [21:56] But pacman-installed stuff is binary. (I mean, sure, someone compiled it :P) [21:56] yes [21:56] It's my Linux desktop of choice. [21:56] % systemd-analyze [21:56] Startup finished in 9.022s (firmware) + 1min 14.878s (loader) + 2.816s (kernel) + 6.326s (userspace) = 1min 33.044s [21:56] i reckon that's pretty cool. [21:56] Startup finished in 2.865s (kernel) + 7.775s (userspace) = 10.640s [21:56] weird you don't get firmware [21:56] oh netiher does this computer [21:56] # systemd-analyze [21:57] Startup finished in 5.007s (kernel) + 1min 33.719s (userspace) = 1min 38.727s [21:57] Still, 90+ seconds seems rather slow [21:57] that has an extra 90 seconds from the waiting for ethernet interface issue [21:57] (at least in this day and age, with systemd, etc) [21:57] cos it's using custom kernel [21:57] you can shorten that time [21:57] ahh [21:57] the kernel time is slow too though :/ [21:58] yeah i like arch the most for play systems [21:58] and ubuntu for "dumb" systems. [21:59] although i am thinking of moving more stuff to arch anyway [21:59] i kind of wish more aur stuff shifted to base [21:59] but it is kind of nifty getting such regular updates [22:00] * mercutio does update, only logrotate is updated, hmm.. [22:00] how are you finding systemd? [22:00] nifty/annoying, depending on the day :p [22:00] Just fine. [22:02] i keep hoping radeonsi update comes through that fixes notion slowness [22:02] there was an issue with drawing primitives ages ago, which still doesn't seem to be fixed. [22:02] and notion uses some "legacy" stuff that's unaccelerated and insanely slow. [22:03] so like if you drag a window from one desktop space to another it's really slow [22:03] and that's just to show the window title background [22:04] it reminds me of sun3s :) [22:04] with their unaccelerated framebuffers. [22:04] but even on the cpu it should be "fast enough" [22:07] heh [23:07] how would I get tmux to open panes / windows and execute a certain command in each one? [23:07] for instance pane 1: telnet localhost 2501, pane 2: telnet localhost 2502 etc [23:08] @google tmux multiple panes [23:08] Google API failure :( [23:15] *** gluffis has quit IRC (Ping timeout: 250 seconds) [23:17] *** gluffis has joined #arpnetworks [23:21] *** Seju has joined #arpnetworks [23:29] *** Seji has quit IRC (*.net *.split) [23:29] *** toeshred has quit IRC (*.net *.split) [23:36] *** toeshred has joined #arpnetworks [23:50] *** dj_goku has quit IRC (Ping timeout: 246 seconds) [23:54] *** dj_goku has joined #arpnetworks [23:54] *** dj_goku has quit IRC (Changing host) [23:54] *** dj_goku has joined #arpnetworks