--- Day changed --- Log opened Wed Jul 02 00:00:05 2003 04:49 -!- revenger_ [revenger@p508B170B.dip.t-dialin.net] has joined #uml 04:51 -!- revenger [revenger@p508B17BF.dip.t-dialin.net] has quit [Ping timeout: 490 seconds] 09:07 -!- litost [~sombitch@131.96.158.138] has joined #uml 10:10 -!- AquaJo [~torbofh@pD9559F4A.dip.t-dialin.net] has joined #uml 10:14 -!- faf [~chatzilla@194.2.91.222] has joined #uml 10:14 < faf> hi 10:15 < faf> Is it possible to use uml to debug a kernel ? 10:16 < AquaJo> It ist, that is/was(?) one of the main purpose 10:16 < AquaJo> but I don't know how to do that 10:16 < faf> my issue is that the kernel is already cutomised, I would like to use it without adding another patch 10:17 < AquaJo> the hostkernel or the kernel which you nlike to debug? 10:18 < faf> the one I d like to debug is heavily custom 10:19 < AquaJo> as far as I know, you'll ned to apply the uml-patch 10:26 < faf> hum 10:44 < faf> ok, i've read a little bit more from the How-to. 10:44 < faf> I try the patch, and hopefully it won't complain too much 10:45 -!- faf [~chatzilla@194.2.91.222] has quit [Quit: ChatZilla 0.8.9 [Mozilla rv:1.1/20020826]] 10:59 -!- snide [~snide@80.14.150.243] has joined #uml 11:05 -!- ajax [~ajax@12-221-0-198.client.insightBB.com] has quit [Quit: BitchX-75p3i-openSSL -- just do it.] 11:15 -!- snide [~snide@80.14.150.243] has quit [Quit: [BX] snide has no reason... just kidding :)] 12:53 -!- ido [~Ido@212.150.75.108] has joined #uml 12:54 < ido> hey people. 12:55 < AquaJo> moin 12:56 < ido> I've got an interesting issue on some server 12:57 < ido> anyhow, i've got a few ip's which are supposed to be bound to it 12:57 < ido> anyhow, i see only ONE of them on ifconfig -a 12:58 < ido> but they do show up on netstat -a's connection list 12:58 < ido> and they even work 13:03 < AquaJo> strange 13:03 < ido> tell me about it. 13:04 < ido> its a colocated server 13:04 < ido> i thought maybe the hosting company might be routing ip's to it 13:04 < ido> but that still doesn't seem it'd work 13:04 < ido> since they still have to be bounded 13:04 < caker> ido: um, did you ifconfig each ip? 13:05 < ido> how could i ifconfig each ip ? 13:05 < david> ido: is the box arping for them? 13:05 < caker> ifconfig eth0:1 netmask up 13:05 < caker> ifconfig eth0:2 netmask up 13:05 < caker> ... 13:06 < ido> they dont exist 13:06 < ido> only eth0 does 13:06 < caker> ido: they're virtual -- try it 13:06 < ido> hrmf, virtually they dont exist 13:06 < ido> otherwise i would have seen them with ifconfig -a 13:06 < ido> david: how could i check that ? 13:07 < david> ido: arp -n 13:07 < david> uh, wait, no 13:07 < caker> ido: you've got to bring up the IPs with the command I showed you 13:08 < david> ido: I'm confused 13:08 < ido> caker: that doesn't solve the weird issue 13:08 < david> ido: you have additional IPs, they're not ifconfiged, they're not routed, but they work? 13:08 < ido> those ip's work ALTHOUGH they aren't bounded as eth0:0, eth0:1, etc. 13:08 < ido> exactly. 13:08 < david> ido: sure they belong to that box? 13:09 < ido> yep 13:09 < david> well 13:09 < caker> define "up" 13:09 < david> someone must be making them arp for the IP 13:09 < caker> er working -- they just pinging? 13:10 < ido> same services work on them 13:10 < caker> bizarre 13:10 < caker> any NAT going on or something strange? 13:10 < ido> not that i can tell 13:12 < david> ido: ok, start with the basics 13:13 < david> ido: ping one of the IPs then do 'arp -n' if the IP is listed, then it's not local 13:13 < ido> nope, its not listed 13:19 < ido> whats the next step ? 13:22 < david> ido: no idea - Seems weird 13:22 < david> ido: 100% it's not bound to something? 13:22 < ido> any other way to check ? /proc ? 13:23 < caker> what's your subnet, and how many IPs is this affecting? 13:24 < caker> What other interfaces show when doing "ifconfig -a" ? 13:24 < ido> 9 ip's 13:24 < ido> welp, eth1 is down (isn't used) 13:24 < ido> and the lo'calinterface. 13:24 < ido> thats it. 13:24 < caker> using bridges or anything else? 13:24 < caker> hmm 13:25 < ido> as i said, its a colocated box, and its not mine, so i don't quite have answers to those questions 13:25 < ido> i've mailed their support about it. hoping to get answers. 13:25 < caker> not using a broadcast address or anything strange, right? what's your netmask 13:26 < ido> 255.255.255.0 14:39 -!- shak [~shak@pc1-hudd4-5-cust78.hudd.cable.ntl.com] has joined #uml 14:43 < shak> hrm 14:43 < shak> does anyone use any kind of bandwith limiting on their UML machines? 14:43 < david> shak: yes 14:43 < shak> how do you do it? I've had a look and it looks like you can use iptables, just curious how other people do it though 14:44 < david> using 'tc' and the QoS stuff in the kernel 14:44 < shak> is QoS true limiting or just prioritising? 14:44 < david> both 14:45 < shak> hrm 14:45 * shak goes to google 14:46 < david> lartc.org 14:46 < shak> thanks :) 14:46 < david> np 14:47 < ido> hmmpf. 14:47 < ido> anyone with experience recovering broken tgz files ? 15:06 < shak> device uml-bridge is a bridge device itself; can't enslave a bridge device to a bridge device. 15:06 < shak> why?! 15:06 * shak pokes it 15:08 < david> shak: er, you can't add a bridge to a bridge 15:08 < shak> I noticed 15:08 < david> shak: you want to add a regular interface to the bridge 15:08 < shak> what I was trying to do 15:08 < shak> was to create a step between the physical interface 15:08 < shak> and the bridge 15:08 < shak> wherby I could do some routing 15:09 < shak> but I seem to have been foiled 15:09 < david> you don't route between interface and bridge 15:09 < david> makes the whole use of a bridge somewhat useless :-) 15:09 < shak> yup 15:10 < shak> I need some form of virtual switch or something 15:10 < david> virtual switch = bridge 15:10 < david> a bridge will do that for you 15:10 < david> why do you need to route? 15:11 < shak> I have a nifty app 15:11 < shak> that is a router that sort of emulates some of the cisco stuff 15:11 < shak> that I wanted to test 15:11 < Pahan> Oooh, a usable Cisco emulator? 15:11 < shak> not really usable 15:11 < david> zebra kinda uses the same commands as Cisco stuff 15:11 < shak> is some testing code 15:12 < shak> I'm unclear as to whether zebra can do level three stuff? 15:12 < david> shak: layer 3? 15:12 < shak> i.e. via IP 15:12 < shak> instead of needing to work via Mac 15:12 < david> zebra is an IP routeing tool 15:13 < david> it does RIP, OSPF, static route and some other stuff 15:13 < david> zebra doesn't have much of anything to do with layer 2 15:13 < shak> are there actually any decent docs for it now, when I last tried it it was hopeless in terms of docs 15:13 < david> shak: not really - Google might help 15:13 < david> zebra.org sort of stagnated 15:13 < shak> hrm, I was told by a friend that it was almost exclusively layer 2 15:13 < david> http://zebra.dishone.st/ 15:14 * shak wonders if the friend got confused 15:14 < shak> thanks 15:14 < david> shak: me thinks your friend doesn't know what zebra is 15:15 < shak> I agree there 15:15 < shak> I only started reading about it about, err, 15 minutes ago 15:15 < shak> and when I last looked it was with the same friend 15:16 < david> shak: zebra is basically a nice IP routeing tool - Why it'd be useful for layer 2 is beyond me 15:18 < shak> Im beginning to understand that its really not useful for layer 2 15:18 < shak> do you think you could do 15:19 < shak> ------ethernet physical connection-------> zebra -----> bridge -----------> UML tuntap devices? 15:22 < shak> you'd be able to use some of the nicer commands from zebra 15:25 < david> uuuh 15:25 < david> yes 15:25 < david> you can do that 15:25 < david> it's just a simple route 15:26 < david> have a netblock for the UMLs & the bridge, another for the outside 15:26 < david> that is what I use 15:26 < david> although you don't need zebra for that, just 'route' 15:26 < shak> meaning that the UMLs dont have real IPs? 15:34 < david> shak: yeah, they have real IPs 15:34 < shak> ah 15:34 < david> shak: public or private IPs is only a concern when using NAT 15:35 < shak> so do your UMLs have two eth devices? 15:35 < shak> one internal and one external? 15:41 < david> yes 15:42 < shak> cool 15:42 < shak> thanks for this, enduring my questioning 15:42 < shak> i really appreciate it 15:43 < david> np 15:55 -!- shak [~shak@pc1-hudd4-5-cust78.hudd.cable.ntl.com] has quit [Quit: Client exiting] 16:08 < BB> evening 16:20 * BB is bored waiting for raid to create *sigh* 16:25 -!- ElectricElf [david@elf.noc.oftc.net] has quit [Ping timeout: 244 seconds] 16:27 -!- ElectricElf [david@elf.noc.oftc.net] has joined #uml 16:28 < AquaJo> BB: take a smaler raid, or more server :-) 16:35 < BB> AquaJo heh 16:36 -!- shak [~shak@80.4.84.78] has joined #uml 16:37 < BB> i could go to pub and come back at last orders... but i think i might be a tad wasted then 16:38 < shak> :D 16:38 < shak> I did that last night 16:38 < shak> thinking that I would be home by midnight, damn late licence pubs 16:38 < BB> heh 16:38 < BB> well i have an hour to kill 16:41 < shak> Im just about to make a patch that applies cleanly to 2.4.21 of the CPU cap patch 16:41 < shak> is it bad etiquette to distribute it? 16:41 < shak> (I think it is, I should really mail it to the author) 16:52 < AquaJo> good night 16:53 < shak> night 16:53 < shak> sleep well 16:56 < AquaJo> thx 16:56 -!- AquaJo [~torbofh@pD9559F4A.dip.t-dialin.net] has quit [Quit: Ich komme wieder, sicher?!] 17:26 -!- Lathiat [lathiat@seven.sixlabs.org] has quit [Remote host closed the connection] 17:26 -!- Lathiat [lathiat@seven.sixlabs.org] has joined #uml 17:39 -!- litost [~sombitch@131.96.158.138] has quit [Quit: fuck this patch reject bullshit!] 18:00 < BB> nice quit msg 18:01 < shak> lol 18:21 -!- ido [~Ido@212.150.75.108] has quit [uranium.oftc.net unununium.oftc.net] 18:21 -!- mistral [mistral@jstevenson.plus.com] has quit [uranium.oftc.net unununium.oftc.net] 18:21 -!- albino [~albino@neuf.dorms.usu.edu] has quit [uranium.oftc.net unununium.oftc.net] 18:21 -!- sssssssh______ [~chatzilla@68.19.239.102] has quit [uranium.oftc.net unununium.oftc.net] 18:21 -!- mistik1 [rasta@ool-44c02046.dyn.optonline.net] has quit [uranium.oftc.net unununium.oftc.net] 18:21 -!- caker [~null@pcp507591pcs.nash01.tn.comcast.net] has quit [uranium.oftc.net unununium.oftc.net] 18:21 -!- dg [dgl@dsl-217-155-7-197.zen.co.uk] has quit [uranium.oftc.net unununium.oftc.net] 18:23 -!- mistik1 [rasta@68.192.32.70] has joined #uml 18:25 -!- dg [dgl@217.155.7.197] has joined #uml 18:26 -!- mistral [mistral@212.159.71.212] has joined #uml 18:28 -!- sid_ [~sid@81.168.9.234] has joined #uml 18:28 -!- sid_ [~sid@81.168.9.234] has left #uml [] 19:14 -!- ido [~Ido@212.150.75.108] has joined #uml 19:14 < ido> mount: none already mounted on /proc 19:14 < ido> not mounted anything 19:14 < ido> INIT: Entering runlevel: 4 19:14 < ido> INIT: Id "1" respawning too fast: disabled for 5 minutes 19:14 < ido> INIT: Id "2" respawning too fast: disabled for 5 minutes 19:14 < ido> INIT: Id "3" respawning too fast: disabled for 5 minutes 19:14 < ido> INIT: no more processes left in this runlevel 19:14 < ido> what could have caused this on the UML ? 19:27 < desaster> wrong tty entries in /etc/inittab 19:28 < desaster> the problem is that you're not using devfs while the inittab is 19:28 < desaster> OR 19:28 < desaster> the other way around 19:44 < ido> no no, see, the problem is, i haven't changed the fs at all 19:44 < ido> actually i've even tried a couple of pre-made FS's downloaded off the net 19:44 < ido> all give me the same problem 19:45 < ido> http://archive.linuxfromscratch.org/mail-archives/blfs-support/2002/07/1384.html 19:45 < ido> exactly as described in this thread 20:03 < desaster> ido: you don't need to have changed the fs at all, 20:03 < desaster> i think most of the pre-made images are devfs enabled 20:03 < desaster> but you aren't running uml with devfs enabled 20:04 < desaster> this is very common question, i wish there was a faq entry for it.. 20:04 < ido> i'm pretty sure i am.. 20:04 < ido> unless i've recompiled and forgot to enable it 20:04 < desaster> you could boot to the uml in single mode, and fix /etc/inittabn 20:05 < desaster> there's something like ttys/0 20:05 < desaster> and you should change those to tty0, tty1, tty2, so on.. 20:05 < ido> CONFIG_DEVFS_FS=y 20:05 < ido> # CONFIG_DEVFS_MOUNT is not set 20:05 < ido> nope, that seems ok 20:10 < ido> any other idea's ? 20:13 -!- albino [~albino@neuf.dorms.usu.edu] has joined #uml 20:14 < ido> i'm pretty sure it has to do with the host's kernel 20:14 < ido> or inittba 20:15 -!- caker [~null@pcp507591pcs.nash01.tn.comcast.net] has joined #uml 20:15 < ido> hey caker, morning 21:54 -!- walters [walters@verbum.org] has joined #uml 21:54 < walters> anyone here gotten 2.5.69+uml to work? For me it hangs right after "Initializing stdio console driver" 21:56 < walters> 2.4.21+uml works fine with the same rootfs 22:06 -!- Pahan [pahan@syr-24-59-98-110.twcny.rr.com] has quit [Remote host closed the connection] 22:07 -!- Pahan [pahan@syr-24-59-98-110.twcny.rr.com] has joined #uml 22:14 < walters> ah, nevermind, i forgot to add CONFIG_SSL and stuff. 23:24 -!- walters [walters@verbum.org] has quit [Quit: Client exiting] 23:44 -!- ljlane [ljlane@aphrael.net] has quit [Remote host closed the connection] 23:52 -!- ljlane [ljlane@aphrael.net] has joined #uml 23:56 < Pahan> Grrr. 23:56 < Pahan> Linux is seriously fucked in the head. 23:56 < david> ljlane: heya 23:56 < david> Pahan: works fine for me 23:56 < Pahan> TUNSETIFF ioctl on tuntap devices oh-so-cleverly cannot clear IFF_NO_PI on existing devices. 23:57 < Pahan> It can set the flag just fine. 23:57 < Pahan> So ridiculously inept :( --- Log closed Thu Jul 03 00:00:00 2003