[12:27] <halcyonCorsair> ajmitch: Depends: libdns21 (= 1:9.3.2-2ubuntu1.3) but 1:9.3.2-2ubuntu1.4 is to be installed
[12:27] <halcyonCorsair> there are about four of those messages
[12:30] <ajmitch> great
[12:30] <ajmitch> that's usually security or updates
[12:31] <halcyonCorsair> ajmitch: no, its the original version of the package vs. the one that i built, i'm downgrading by specifying each one to be the older version...
[12:31] <ajmitch> did you bump the version to 1:9.3.2-2ubuntu1.4?
[12:31] <keescook> halcyonCorsair: yeah, that's what I've done in the past.  used "apt-cache madison PKG" to figure out the prior version, and then specified each on on the apt-get install line
[12:32] <ajmitch> apt won't automatically downgrade
[12:32] <ajmitch> hello keescook
[12:32] <keescook> hiya ajmitch!
[12:32] <keescook> how goes it?
[12:32] <ajmitch> good, how are you?
[12:33] <keescook> good, wading through a kernel update.  :)
[12:33] <ajmitch> I see that we finally have a name for 8.04
[12:33] <ajmitch> heh
[12:33] <keescook> Yeah, I was curious if "H" was going to be reused
[12:33] <keescook> I wonder why "C" was skipped.
[12:33] <ajmitch> but noone remembers hoary now :)
[12:33] <halcyonCorsair> i guess those voting for "Hungry Hippo" lost out in the end :)
[12:34] <keescook> that would have rocked.
[12:34] <ajmitch> how's the search for a security guy going?
[12:35] <keescook> ajmitch: good, I think we're almost done.
[12:35] <ajmitch> that's great
[12:36] <ajmitch> I won't bother applying then :)
[12:37] <halcyonCorsair> hmm....how do i remove my custom packages to make sure they don't interfere with anything anymore?
[12:37] <ajmitch> I thought you just did that?
[12:39] <halcyonCorsair> i downgraded them to the original version yeah...so if i do a apt-get upgrade, they're not gonna come up to haunt me?
[12:40] <halcyonCorsair>  ah...it seems they won't, way cool
[12:41] <ajmitch> no, since they're not in an apt repository in your sources.list
[12:41] <ajmitch> so apt forgets all about them
[12:43] <halcyonCorsair> learn something new every day...
[12:47] <keescook> ajmitch: you keep a table of high/med/low changes between debian unstable and ubuntu, is that right?
[12:47] <ajmitch> keescook: rc bugs fixed in debian that aren't in ubuntu
[12:47] <osmosis> can anyone recommend a small boxen...something like a mac mini, but with hardware that doesn't fail.
[12:47] <keescook> I just got burned by an update in debian that was prior to Etch, but after Feisty.  (i.e. fixed in gutsy)
[12:47] <ajmitch> keescook: django.ajmitch.net.nz/rcbugs
[12:48] <keescook> ajmitch: and that's just gutsy, right?
[12:48] <ajmitch> yeah, just gutsy right now
[12:48] <ajmitch> you want it for multiple distros?
[12:49] <keescook> ajmitch: yeah, but I'm trying to figure out how to specify what I'm looking for.
[12:49] <keescook> because it's "changes between a stable release of ubuntu's version and the following stable release of debian"...
[12:50] <keescook> anyway, might not be needed, I guess.  A CVE should have been assigned for this bug.
[12:57] <halcyonCorsair> how can i get apt-get to install config files over the top?
[12:58] <ajmitch> overwrite existing config files?
[12:59] <ajmitch> keescook: if there's something I can help with, just ask - I've got a mirror of debian's BTS which I rsync
[12:59] <keescook> ajmitch: okay, I'll ponder it.  :)
[12:59] <ajmitch> halcyonCorsair: it's generally not a good thing - you can purge & reinstall if you really need it
[01:00] <ajmitch> or remove the config file & reinstall with apt-get -o DPkg::Options::="--force-confmiss"
[01:01] <halcyonCorsair> mmm, yeah i did the purge and reinstall...although it kinda felt like i was taking to it with a hatchet
[01:08] <keescook> ajmitch: any chance I could convince you to add a tiny column to your rcbugs output for main/restricted/universe/multiverse?  like maybe a single character: m, r, u, v ?
[01:08] <ajmitch> keescook: yeah, though this list is only for universe right now :)
[01:09] <ajmitch> I'll have to change some code around, should only be a few lines
[01:09] <keescook> oh! heh. then can I convince you to add main?  :)
[01:09] <ajmitch> it needs burnt down & rewritten anyway
[01:09] <keescook> hehe
[01:10] <ajmitch> it's seriously crap code that I can hardly understand
[01:10] <ajmitch> must have been some late nights there
[01:10] <keescook> heh
[01:38] <halcyonCorsair> hey, can anyone tell me why my feisty server doesn't have colours when I ssh in? (i don't think it had colours with local logon either)
[01:47] <halcyonCorsair> i mean, .bashrc says it'll be fancy and colourful if it detects xterm-color, but dapper seems to ignore that and be colourful anyway....
[01:47] <halcyonCorsair> why not feisty?
[01:50] <halcyonCorsair> hmm...ok, its not that part...its the ls alias somehow...
[01:50] <halcyonCorsair> not, thats wrong
[01:52] <halcyonCorsair> vims not displaying colours...thats whats wrong
[01:55] <halcyonCorsair> *sigh* i'm talking to myself here...
[01:56] <Kamping_Kaiser> hehe
[01:57] <ajmitch> halcyonCorsair: ls displays colour, vim doesn't?
[01:57] <ajmitch> if so, there were changes from vim 6 to 7, check /etc/vim/vimrc
[01:58] <halcyonCorsair> yes....it seems feisty comes fully loaded with vim-tiny which could be the problem...
[01:58] <ajmitch> quite
[01:58] <ajmitch> that was done for a reason, of course :)
[01:59] <halcyonCorsair> oh?
[01:59] <ajmitch> disk space for the desktop cd
[01:59] <Kamping_Kaiser> bleh @ the excuse
[02:00] <ajmitch> I presume that noone cared enough to reseed the full vim for the server
[02:00] <ajmitch> Kamping_Kaiser: you'll live
[02:00] <halcyonCorsair> best way to get vim in colour would be to update to full blown vim, or try and readd colour conf to /etc/vim/vimrc?
[02:00] <Kamping_Kaiser> ajmitch, potentially
[02:00] <ajmitch> ok, lunchtime, back later
[02:01] <Kamping_Kaiser> you nead real vim to get colouring
[02:16] <osmosis> can anyone tell my why feisty has different xen packages for desktop and server ?
[02:19] <osmosis> what is PAE ?
[02:19] <halcyonCorsair> Kamping_Kaiser: what do i need for real vim to start colorizing?
[02:22] <Kamping_Kaiser> halcyonCorsair, a ~/.vimrc or /etc/vim/vimrc
[02:22] <Kamping_Kaiser> afaik
[02:23] <halcyonCorsair> crap....so i'm gonna have to do "THAT"!
[02:24] <Kamping_Kaiser> hm?
[02:27] <halcyonCorsair> i know nothing about configuring vimrc
[02:30] <halcyonCorsair> oh....that was easy...hehe!@
[02:41] <CharlieSu> Does anyone know of a linux program that is basically like WinSCP that uses an ncurses interface?
[02:59] <sommer> hey all is it bad form to compare PEAR to CPAN and RubyGems?
[04:13] <pschulz01> halcyonCorsair: Howdy..
[04:13] <halcyonCorsair> pschulz01: sup
[04:13] <pschulz01> halcyonCorsair: Did you find an answer to the 'files' problem?
[04:14] <halcyonCorsair> pschulz01: yes, two solutions, one that is totally evil, and one that is pretty good
[04:14] <halcyonCorsair> i still have to test them though
[04:14] <pschulz01> :-)
[04:15] <halcyonCorsair> pschulz01: soln1: alter /usr/include/linux/posix_types.h and /usr/include/bits/typesizes.h to set FD_SETSIZE to be larger than 1024 (EVIL!!!)
[04:15] <pschulz01> halcyonCorsair: Yup.
[04:16] <halcyonCorsair> soln2: add a listen-to directive to named.conf.options for each physical interface, and then have iptables allow access to that address for dhcp
[04:16] <ajmitch> 2) still sounds a bit painful
[04:17] <halcyonCorsair> ajmitch: not really
[04:17] <halcyonCorsair> ajmitch: can you envision anything less painful?
[04:17] <pschulz01> What does FD_SETSIZE do?
[04:18] <pschulz01> Is it just used as an option? can it be over written within named?
[04:18] <ajmitch> halcyonCorsair: listening on 1 interface & relying solely on the networking stack to get the packets to the right place
[04:19] <ajmitch> which has its own problems
[04:19] <halcyonCorsair> ajmitch: sure, I could do that, but my way means i have to modify fewer scripts :)
[04:19] <ajmitch> it depends whether bind needs to open a socket for every interface it explicitly listens on
[04:20] <halcyonCorsair> i think it does
[04:20] <halcyonCorsair> either way, listening on 5 interfaces is a lot better than listening on 1000
[04:21] <ajmitch> right, I was thinking that you meant  listening on every vlan interface
[04:22] <ajmitch> but if you're listening on just 5, then it sounds much better
[04:22] <halcyonCorsair> no, just the PHYSICALL interfaces
[04:22] <halcyonCorsair> ie. eth2 covers eth2.2 - eth2.254, etc
[04:22] <ajmitch> right, I missed that when reading it earlier
[04:23] <pschulz01> halcyonCorsair: Option3. hack bind
[04:23] <halcyonCorsair> when i first made the change, i forgot to listen on 127.0.0.1 :) wondered why dns suddenly didn't work on the server
[04:23] <ajmitch> ls -la
[04:23] <ajmitch> hm
[04:23] <ajmitch> (ssh is very lagged)
[04:23] <pschulz01> halcyonCorsair: port/*/include/fd_setsize.h <-- from the CHANGES  file.. I'm looking now,
[04:24] <ajmitch> pschulz01: hacking bind or rebuilding should be the last option
[04:24] <halcyonCorsair> pschulz01: eh? not really....what our system was doing in the first place was kinda stupid
[04:24] <ajmitch> since it means an unsupportable package
[04:25] <pschulz01> ajmitch: first step in the  'get stuff fixed' process -> ..create patch.. submit patch..
[04:25] <ajmitch> pschulz01: only if it's necessary
[04:26] <pschulz01> halcyonCorsair: I can't see the file in the bind source anyway for 'linux' so.. hmm..
[04:27] <halcyonCorsair> pschulz01: the files i modified aren't part of bind
[04:27] <pschulz01> halcyonCorsair: That is the main problem! :-)
[04:27] <halcyonCorsair> ajmitch: what package is /usr/include/linux/posix_types.h from?
[04:27] <pschulz01> halcyonCorsair: You entire system is now unsupported.
[04:28] <halcyonCorsair> pschulz01: i'm not sure you understand the nature of what i'm talking about here...
[04:28] <halcyonCorsair> either that, or your making a joke that i missed
[04:28] <pschulz01> halcyonCorsair: dpkg -S /usr/include/linux/posix_types.h
[04:28] <pschulz01> linux-libc-dev
[04:28] <pschulz01> halcyonCorsair: (Just a little one)
[04:29] <pschulz01> halcyonCorsair: (joke.. that is)
[04:29] <halcyonCorsair> right :)
[04:30] <pschulz01> If bind is not able to do what you want it to do, then there are proably going to be other people in the same situation.. so rather than a work around, fixing it properly is probably a 'nice thing to do'.
[04:32] <halcyonCorsair> pschulz01: its not a problem that should be encountered if bind is configured in a more "correct" way
[04:33] <pschulz01> (Please correct me if I'm wrong ajmitch) Unless  FD_SETSIZE is a system limit, in which case there should be a kernel configuration parameter to change it. If it is problem with bind, then.. well.. that was where I was heading.
[04:34] <pschulz01> halcyonCorsair: Just rephasing your problem...
[04:35] <pschulz01> halcyonCorsair: You have a single 'bind' listening on lots of interfaces (1000+) and you want bind to open a socket to listen individually for each one?
[04:36] <halcyonCorsair> pschulz01: thats what the system was doing, however its not really very desirable behaviour when you think about it
[04:36] <halcyonCorsair> i just hadn't questioned the logic of what it was doing is all
[04:41] <halcyonCorsair> how can i check the version of a running module?
[04:46] <ajmitch> halcyonCorsair: kernel module?
[04:48] <halcyonCorsair> ajmitch: yes, i've just installed an updated intel e1000 driver and want to check that the new one is running
[04:48] <ajmitch> hm, I think modinfo just tells you details of the one on disk
[04:48] <ajmitch> so dmesg or syslog will be your best bet
[04:50] <halcyonCorsair> *sigh* it didn't fix the problem....
[04:51] <halcyonCorsair> damn HP for not supporting ubuntu
[04:51] <halcyonCorsair> ....luckily i don't think its a showstopper
[04:51] <ajmitch> what's the problem with it?
[04:51] <halcyonCorsair> e1000: eth3: e1000_request_irq: Unable to allocate MSI interrupt Error: -22
[04:52] <halcyonCorsair> i guess i'll find out shortly though
[04:52] <ajmitch> does it cause issues, or just give an error?
[04:52] <halcyonCorsair> just give an error so far, i'm about to start testing
[04:59] <ph1zzle> hey guys, I am looking for a wan failover software or mechanism, basically designed for a machine with two nics, each hooked up to a seperate wan, and when one wan goes down, it fails over ti the other, any thoughts?
[05:26] <halcyonCorsair> ph1zzle: what sort of NICs? (not that it really matters, just curious)
[05:28] <halcyonCorsair> ph1zzle: look up under your kernel source, look up Documentation/networking/bonding.txt and then the "Configuring Bonding for High Availability"
[05:33] <ph1zzle> sure, I actuall. just now found something easier, and I have been looking at this for a while
[05:34] <ph1zzle> basically what was stated was define two default routes, each with the different network cards, if one route fails, the second will be used, there for if each nic is on a different wan
[05:35] <halcyonCorsair> oh true :) i was thinking more enterprisey :)
[05:35] <ph1zzle> and you have your primary link defined as the link on wan 1 and your secondary on wan 2 then when the router on wan 1 is no good, then wan 2 should suffice
[05:35] <ph1zzle> yeah, well at this point I just want a solution, lol
[05:36] <ph1zzle> this office only contains 20 people, we have a adsl link that seems to fail constantly, and a spriny ev-do link as a backup and frankly... I just want it to work
[05:41] <halcyonCorsair> http://www.epr.ch/brb/linux/backroute.php
[05:41] <halcyonCorsair> that might be another option, i've never used it though
[05:43] <ph1zzle> huh, that actually looks damned cool, thanks halcyonCorsair
[06:06] <halcyonCorsair> ph1zzle: hope it works
[08:15] <kraut> moin
[08:16] <nealmcb> kraut: any idea where people get the idea of adding spaces before punctuation??
[08:16] <kraut> because they want to annoy other people with it?
[08:21] <nealmcb> hmm - looks like it is the default in french punctuation - I wonder when that started and why.  not much info here:  http://en.wikipedia.org/wiki/Punctuation_in_French
[07:29] <radicall> hi all
[07:29] <ScottK> Hellol
[07:32] <radicall> someone has been installed Sun Application Server on Ubuntu Server 6.06?
[09:36] <raky> net boot client set in bios to boot from network, but doesn't http://pastebin.com/d36396a38 this is my dhcpd.conf file
[09:44] <sommer> raky: I've actually done a net boot, but don't you need to add an entry for your tftp server?
[09:45] <raky> i'm following the hottoforge guide here http://www.howtoforge.com/ubuntu_pxe_install_server
[09:47] <sommer> raky: I think you need a next-server entry to point to your tftp server.  Check the last highlight box in the guide.
[09:47] <osmosis> what is the difference between xen i386 and xen PAE ?
[09:48] <necrite_> hi all
[09:48] <necrite_> i have too many "dummy internal connection" on my webserver
[09:48] <necrite_> any idea why apache make this connections?
[10:25] <osmosis> how do I make a xen initrd ?
[10:52] <ajmitch> good morning
[10:56] <raky> i keep getting a message that eth0 subnet is not configured, here's the paste:  http://pastebin.com/d36396a38
[11:10] <soren> raky: Pointing out that it's a dhcp server config file might help..
[11:10] <soren> raky: What's the IP of your eth0 ?
[11:10] <mralphabet> soren: took me a minute too
[11:11] <soren> mralphabet: Well, it's pretty obvious when I actually clicked the link, but until then..
[11:11] <ajmitch> hello soren
[11:12] <soren> ajmitch: Hey, Andrew.
[11:12] <raky> 169.254.4.178 / 192.168.1.2 mralphabet
[11:12] <soren> raky: Um... Which one is it?
[11:13] <soren> raky: It can only have one.
[11:13] <raky> 169.254.4.178
[11:14] <ajmitch> sounds like a link-local address
[11:14] <ajmitch> ie it's not been configured
[11:14] <mralphabet> raky: pastebin your /etc/network/interfaces
[11:14] <ajmitch> you had a mention of eth0:avahi in the dhcp configuration?
[11:15] <soren> raky: If you have an error about eth0 not begin configured properly, it often helps to provide information about said interface.
[11:15] <raky> this is when configing dhpcd server
[11:23] <raky> /etc/network/interfaces http://pastebin.com/d3c0769e1
[11:26] <soren> raky: That config will not work.
[11:26] <soren> raky: Your gateway needs to be on the same subnet as you.
[11:26] <soren> Er..
[11:26] <soren> Oh, it says dhcp.
[11:26] <soren> That's a pretty strange interfaces file :)
[11:27] <raky> lol, i need to understand more, antoher person said the same thing  was wrong
[11:27] <ajmitch> I hope this isn't on the same box that you're running a dhcp server on?
[11:28] <ajmitch> the dhcp server needs to have a configured interface (statically assigned IP address) to hand out leases
[11:29] <ajmitch> and you shouldn't assign 169.254 addresses :)
[11:48] <mralphabet> raky: normal dhcp client http://pastebin.com/m740da274
[11:50] <mralphabet> raky: are you expecting the DHCP server to assign the IP to its own interface?
[11:50] <soren> ajmitch: Strictly speaking, the dhcp server can get an address from dhcp. It's a bit strange, but doable.
[11:50] <ajmitch> soren: from itself?
[11:51] <soren> ajmitch: No, from something else.
[11:51] <soren> ajmitch: ...which will only hand out and IP to that machine.
[11:51] <soren> ajmitch: It's probably not what he wants, but it's possible :)
[11:51] <ajmitch> right, but in the common case it's not done :)
[11:52] <soren> Indeed :)
[11:52] <ajmitch> and assigning link-local addresses is still bad :)
[11:53] <mralphabet> very
[11:53] <soren> It doesn't get assigned as it's a dhcp interface.
[11:58] <ajmitch> soren: no, but that was the apparant intent, at least
[11:58] <soren> Right. I think we should stop trying to make sense of that interfaces file.
[12:02] <ajmitch> maybe if I look at authtool I'll get an email from the soc student who made some changes :)