--- Day changed --- Log opened Wed Jun 18 00:00:02 2003 01:24 * alt is away: bed 02:40 -!- pflanze [~chris@dclient80-218-22-118.hispeed.ch] has quit [Quit: [x]chat] 02:53 -!- mistral [mistral@212.159.71.212] has joined #uml 04:15 -!- pflanze [~chris@cc-linux4.ethz.ch] has joined #uml 04:15 < pflanze> Hello 04:15 < pflanze> What do I have to do to get a uml with mem=512MB working? 04:16 < pflanze> It says: CONFIG_HIGHMEM not enabled - physical memory shrunk to 430964736 bytes 04:16 < caker> run skas and config himem in the UML kernel? 04:16 < pflanze> where the number varies from run to run. 04:16 < ichilton> pflanze: enable the option CONFIG_HIGHMEM :) 04:16 < pflanze> Yes, both skas and highmem 4g enabled in uml kernel. 04:16 < ichilton> oh 04:16 < ichilton> strange 04:16 < caker> dumb question, but running the right kernel? :-) 04:17 < pflanze> Strange thing is also, that dmesg|grep -i mem does not show the above line from inside uml. 04:17 < pflanze> It's just printed to stderr. 04:17 < pflanze> or stdout rather. 04:17 < pflanze> ah shiiit 04:18 < pflanze> of course, caker: I've made a chroot... 04:18 < caker> k 04:18 < caker> it happens :-) 04:20 < pflanze> Hmm, with the right one it won't even boot. 04:20 < pflanze> Kernel panic: Kernel mode fault at addr 0x7c42277, ip 0xa01a58d1 04:21 < pflanze> And it still prints 'Kernel virtual memory size shrunk to 28311552 bytes', though no 'CONFIG_HIGHMEM not enabled' anymore. 04:21 < caker> not sure if this matters, but is highmem enabled on the host? 04:21 < pflanze> shrunk to 28MB ! 04:21 < pflanze> yes it is. 04:21 < caker> hmm 04:22 < pflanze> I have 1 1/4 GB RAM on the host. 04:22 < pflanze> (1292592 k) 04:23 < caker> what uml kernel/umlpatch are you using? 04:24 < pflanze> This is user-mode-linux-2.4.20-6um from debian unstable, rebuilt with CONFIG_HIGHMEM=y 04:25 < pflanze> hmm, I did put CONFIG_HIGHMEM4G=y and CONFIG_HIGHIO=y into the config, too, seems like 'make oldconfig ARCH=um' did throw those out. 04:25 < pflanze> I have those two running in the host (with 2.4.21) 04:26 < pflanze> Do you run with mem >= 512MB? 04:26 < caker> I experimented with that a while ago, and was told skas' highmem limit was around 1GB, and tts was around 380, but was told today that was incorrect 04:27 < caker> otherwise, no, not more than a few hundred 04:27 < pflanze> What's this?: CONFIG_KERNEL_STACK_ORDER=2 04:27 < pflanze> CONFIG_KERNEL_HALF_GIGS=1 04:27 < pflanze> CONFIG_SMP=y 04:27 < green> don't mess with this kernel stack order ;) 04:27 < pflanze> CONFIG_UML_SMP=y 04:28 < pflanze> I don't have smp on this machine. 04:28 < pflanze> Should I change the above? 04:29 < caker> pflanze: check out CONFIG_KERNEL_HALF_GIGS 04:29 < pflanze> (Does uml even run with multiple processors? I thought I've read that this isn't supported with skas.) 04:29 < pflanze> caker: how do you mean check out? comment out? 04:29 < caker> pflanze: mp inside uml only works with tt 04:30 < caker> pflanze: let me paste the help 04:30 < pflanze> ah 04:30 < caker> This determines the amount of address space that UML will allocate for its own, measured in half Gigabyte units. The default is 1. Change this only if you need to boot UML with an unusually large amount of physical memory. 04:30 < pflanze> don't need to 04:30 < caker> change to 2, and get 1gb address space? 04:31 < pflanze> ok I'll try 04:47 -!- yeti [~yeti@p3EE22A19.dip.t-dialin.net] has joined #uml 04:47 < yeti> shak: alive? reading? 05:00 -!- yeti_ [~yeti@p3EE22947.dip.t-dialin.net] has joined #uml 05:00 -!- yeti [~yeti@p3EE22A19.dip.t-dialin.net] has quit [Killed (NickServ (GHOST command used by yeti_))] 05:00 -!- yeti_ is now known as yeti 05:05 -!- yeti [~yeti@p3EE22947.dip.t-dialin.net] has left #uml [Client exiting] 05:08 -!- DnsInfector [~DnsInfect@ARennes-204-1-10-115.w81-53.abo.wanadoo.fr] has joined #uml 05:17 -!- DnsInfector [~DnsInfect@ARennes-204-1-10-115.w81-53.abo.wanadoo.fr] has quit [Quit: Client exiting] 05:51 < pflanze> ooh 05:51 < pflanze> stilll Kernel virtual memory size shrunk to 418381824 bytes 05:52 < pflanze> even with CONFIG_KERNEL_HALF_GIGS=2 05:54 < pflanze> btw it's still hanging after mounting devfs. 05:54 < pflanze> strace -p shows all of the linux threads waiting in a syscall except one: 05:54 < pflanze> munmap(0xa0a35000, 4096) = 0 05:54 < pflanze> munmap(0xa0a36000, 4096) = 0 05:54 < pflanze> munmap(0xa0a37000, 4096) = 0 05:54 < pflanze> ... 05:54 < pflanze> munmap(0xa0886000, 4096) = 0 05:54 < pflanze> old_mmap(0xa0886000, 4096, PROT_READ|PROT_WRITE|PROT_EXEC, MAP_SHARED|MAP_FIXED, 4, 0xffaf000) = 0xa0886000 05:54 < pflanze> munmap(0xa0887000, 4096) = 0 05:54 < pflanze> ... 05:54 < pflanze> and so on in an endless loop. 05:55 < pflanze> please anyone tell me how they do get uml's work with 512MB memory. 05:57 < pflanze> btw the init=/bin/bash trick I've mentioned yesterday does not work better. 05:57 < pflanze> It's just that about in 9 of 10 cases, uml hangs after mounting devfs, regardless of the init parameter. 05:58 < pflanze> I think it does not hang as frequently if given a small mem parameter. 05:58 < pflanze> What's broken here? 06:14 -!- renate [renate@ppp-62-245-211-114.mnet-online.de] has joined #uml 06:16 < renate> Hi, i got a problem with the installation of RedHat9 06:18 < renate> I followed the instructions at http://linuxhacker.ru/uml/, but the installer doesn't accept the modified stage2 image 06:18 < renate> i tried installation via nfs and ftp 06:18 < renate> the original one is accepted but the installation hangs later 06:19 < renate> can anybody help me? 06:31 -!- DnsInfector [~DnsInfect@ARennes-204-1-10-115.w81-53.abo.wanadoo.fr] has joined #uml 06:57 -!- renate [renate@ppp-62-245-211-114.mnet-online.de] has quit [Quit: Client exiting] 07:14 -!- yeti [~yeti@p3EE22947.dip.t-dialin.net] has joined #uml 07:24 -!- yeti [~yeti@p3EE22947.dip.t-dialin.net] has quit [Quit: Client exiting] 07:26 -!- DnsInfector [~DnsInfect@ARennes-204-1-10-115.w81-53.abo.wanadoo.fr] has quit [Quit: Client exiting] 07:35 -!- ASY- [~someone@66.36.133.68] has joined #uml 07:35 -!- ASY [~someone@66.36.133.68] has quit [Read error: Connection reset by peer] 07:35 -!- ASY- is now known as ASY 07:42 -!- DnsInfector [~DnsInfect@81.53.249.115] has joined #uml 08:54 -!- Dorian [~Dorian@host81-132-237-58.in-addr.btopenworld.com] has joined #uml 08:56 -!- Dorian [~Dorian@host81-132-237-58.in-addr.btopenworld.com] has left #uml [] 09:10 -!- ichilton [~ian@pc3-stoc3-4-cust203.midd.cable.ntl.com] has quit [Quit: leaving] 09:10 -!- ichilton [~ian@pc3-stoc3-4-cust203.midd.cable.ntl.com] has joined #uml 09:24 -!- DnsInfector [~DnsInfect@81.53.249.115] has quit [Quit: Client exiting] 09:26 -!- yeti [~yeti@p3EE22947.dip.t-dialin.net] has joined #uml 10:01 -!- tanelorn [~tanelorn@dragonix.ccsg.de] has joined #uml 10:01 < tanelorn> ho 10:01 -!- sjc [~sjc@80.46.14.25] has quit [Remote host closed the connection] 10:04 < tanelorn> i've got a question concerning skas mode.. anyone tried to patch 2.4.21? 10:15 < shak> yes 10:15 < shak> I've patched 2.4.21 sucessfully. 10:19 < tanelorn> shak: where in make menuconfig is CONFIG_PROC_MM gone...? 10:19 < tanelorn> *doh* 10:20 < shak> tanelorn: let me just check 10:20 < shak> the bottom of General Setup. 10:21 < tanelorn> yeah. now it is there. 10:22 < tanelorn> i've converted .config-2.4.20->2.4.21 and then applied the patch. no mm compiled, cause make menuconfig threy CONFIG_PROC_MM away. then untared kernel again and the option was gone. now: unpacking kernel, patching kernel, copying config, make menuconfig and hopefully... 10:22 < tanelorn> got totally confused... 10:23 < yeti> ping shak ;-) 10:23 < yeti> hi! 10:23 < tanelorn> how many umls are you running on your machines? and: what for? 10:24 < shak> hi yeti 10:25 < shak> yeti: my machine is dying :( 10:25 < yeti> why? 10:25 < shak> I think I've sorted it 10:25 < shak> the heat 10:25 < yeti> ok 10:25 < yeti> let'er die 10:25 < shak> heh 10:25 < shak> I'll try and debian on that machine asap 10:25 < yeti> one reasonmore to reinstall 10:25 < shak> I've got to pop out in a moment. 10:25 < yeti> can you set up a small system with debian and give me access via ssh? 10:27 < yeti> or maybe my setup could run with nestin gevel 1 as uml? 10:27 < yeti> but i havent tried up to now 10:27 < shak> yep, I should get some time on Sunday 10:27 < shak> if my machine hadn't died today 10:27 < shak> then I'd have done it 10:27 < shak> I have a spare machine in the corner 10:27 < shak> I'll use that. 10:28 < yeti> i can compile things on my smp box... i just would like to set up a debian on a system you can test... 10:29 < shak> that would be great 10:29 < yeti> i think even k6@300 with 128M ram would be enough 10:29 < yeti> 4g hd would be enough too 10:29 < shak> I think I've got a dual PII w/ 40GB HDD 10:30 < yeti> ok... please make only a small partition for the system, some swap and let the rest stay unpartitioned 10:31 < shak> I'll do that 10:31 < yeti> small = 1/2G .. 4G 10:32 < shak> how much swap do you want? 10:32 < yeti> none ;-) i just thought: swap can go on a logical volume ;-) so it is variable ;-) 10:33 < tanelorn> is it possible to limit a specific uml (or uml processes in general) not eating up all cpu? 10:34 < shak> tanelorn: well, there's a CPU cap thing. 10:34 < yeti> jes, search thru the mailinglist for schedulers and google for cpu-cap patch 10:34 < shak> but I'm having some problems with it. 10:34 < shak> it doesnt apply to .21 very cleanly. 10:34 < yeti> i think cap is not the thingy 10:34 < yeti> if no other task runs, why not give all power to the umls? 10:34 < shak> and I'm not sure how it goes with SMP 10:34 < shak> yeti: you can limit each UML 10:35 < shak> so one runs @ no more than 10% CPU. 10:35 < yeti> i want something like fair time distribution 10:35 < yeti> if only one uml wants time, allow 100% 10:35 < yeti> if n are running, divide fair 10:36 < tanelorn> shak: okay, but plain vanilla kernel does not... i've heard about -ac kernels would do, ever tried? 10:36 < shak> I've not tried with -ac on .21 and patching it. I'm going to play about more this evening. 10:37 < tanelorn> you mean taking -ac and patching skas-mode...? 10:38 < shak> trying to take -ac, patch it with skas and patch it with the cpu cap patches 10:40 < tanelorn> i thought -ac does all that scheduling and limiting-stuff? 10:41 < shak> I'm not sure 10:42 < tanelorn> i dont know whether explicitly "your" cap patches are used, but.. im not that kernel-guy to know which is better suited... 10:42 < shak> this is extremely good for capping CPU usage, Im not sure if the -ac tree has anything as effective. 10:43 < tanelorn> okay... i think ill try it... 10:43 < tanelorn> are you running umls "productive"? is it suited for some kind of virtual hosting? 10:44 < shak> I'm setting mine up for running in a commercial environment yes 10:46 < tanelorn> pretty. im going to reboot. cu later. :) 10:46 -!- tanelorn [~tanelorn@dragonix.ccsg.de] has quit [Quit: leaving] 10:47 < yeti> shak: i mailed a friend in berlin... i wanna restart my linux-on-linux things... 10:47 < yeti> but i concentrate on the shell level for now... 10:47 < yeti> web interfache laaaaaaater 10:47 < shak> ;) 10:48 < yeti> the guy is a former opennuke developer 10:48 < shak> ;) 10:48 < yeti> i hope he has some fun making the portal, website and some docs... 10:49 < yeti> i want to build something more universal than hosting-only... 10:49 < shak> yeah, it should be cool. 10:49 < shak> me too, hosting only is very limited. 10:49 < yeti> i'll use uml as abstaration layer 10:50 < yeti> the real hardware will only get as less software as needed, as much as possible shall run in VMs 10:52 < shak> Im trying to set up a router UML 10:52 < shak> so that it has the external connection to the internet 10:52 < shak> and then the users SSH into the external UML 10:52 < shak> and access their internal one 10:53 < yeti> how do you limit the access to the external interface to the uml? 11:06 < david> hello 11:06 < yeti> hi! 11:22 -!- smcavoy [~smcavoy@207.61.160.163] has joined #uml 11:39 -!- smcavoy [~smcavoy@207.61.160.163] has quit [Remote host closed the connection] 11:40 -!- smcavoy [~smcavoy@grimlock.drive-megawheels.net] has joined #uml 12:28 -!- chase [~chase@schnuffel.beschoner.net] has joined #uml 12:28 < chase> howdy 12:29 < chase> shak: how was this performance-limiting patch called? i cannot remember... %-) 12:31 -!- chase is now known as tanelorn 12:31 -!- tanelorn [~chase@schnuffel.beschoner.net] has quit [Client Quit] 12:32 -!- tanelorn [~chase@schnuffel.beschoner.net] has joined #uml 12:32 < tanelorn> re 12:32 -!- desaster [~desaster@80.222.254.34] has quit [Quit: Internet] 12:59 < shak> tanelorn: er, I can't quite remember the application is called "cap" though 12:59 < shak> if you search the mailing list 13:02 < tanelorn> im search thru kernelnewbies.org, perhaps there ill find something 13:12 < yeti> shak: you have kinda scheme/convention/definition where to put the umls's homes? 13:13 < shak> /mnt/uml/users/ and then the umlid 13:13 < yeti> i shak i really am thinking about NOT makeing a difference between users and umls... ]:-) 13:14 < yeti> parts of the code (tap, lvs, ...) may be usefull for users and other emulators/vms too... 13:14 < yeti> e.g. hercules... s390 emu 13:14 < yeti> i want it as "orthoginal" as posssible 13:15 < shak> ;) 13:15 < shak> I'm just getting this machine back on its feet 13:15 < shak> then I'm going to have another go at uml_switch 13:15 < yeti> i dont use hardcoded paths, i use ~or $HOME ;-) 13:16 < yeti> i hink i'll introduce a /etc/taptab ;-) 13:16 < yeti> or some equivalent structure 13:16 < shak> ;) 13:17 < yeti> up to now noone could tell me how to define an interface without an ipaddress in /etc/network/interfaces 13:17 < yeti> (debian specific) 13:17 < yeti> i'd love to keep them in that file but "it doesnt fit" 13:24 < david> yeti: er,give it an IP of 0.0.0.0 13:25 < yeti> hmmm... think i tried but it complained... maybe the mask? i'll retry... 13:27 < david> if the IP is 0.0.0.0, who cares about hte netmask 13:29 < shak> right, I fixed that. 13:29 < shak> vim can be a hazard 13:29 < shak> if you forget to press escape before :wq! then you end up with problems while editing kernel files. 13:30 < yeti> david: doesnt work 13:30 < david> yeti: no idea then - google for it 13:30 < yeti> david: no netmask doesnt fit debians interfaces definition 13:31 < yeti> david: 0.0.0.0 and 255.255.255.255 give errors 13:32 < yeti> i'll look at that later 13:45 -!- Dorian [~Dorian@81.132.93.246] has joined #uml 13:47 -!- Dorian [~Dorian@81.132.93.246] has left #uml [] 14:02 < alt> UML rocks!!!! < 14:03 < shak> UML does rock 14:03 < alt> it's making testing 100 times easier :) 14:03 < alt> no... 1000 times! 14:03 < yeti> yep 14:04 < yeti> i'm trashing and reinstalling debian umls in below a minute 14:04 < yeti> ;-) 14:04 < alt> hee hee 14:04 < alt> I keep a reference 14:04 < yeti> without cow files! 14:04 < alt> yeah, I should use those 14:04 < yeti> no 14:04 < alt> no? 14:04 < yeti> at least not always 14:04 < alt> okay 14:05 < yeti> i didnt try to cow a real device 14:05 < alt> I've just been doing "cp -r UML-REF NEW-UML" 14:05 < yeti> i prefer my umls' filesystems on logical volumes of "the carrier system" 14:06 < alt> I could try that too 14:06 < yeti> one pass less thru fs code 14:06 < yeti> easy to admin 14:06 < alt> I wonder if I have a spare drive to do that on 14:07 < alt> geez... I have a 20GB drive on that machine I'm not using 8) 14:08 < yeti> i'm thinking of putting the hosts fses and all umls fses in lvs 14:08 < yeti> google for "partition alias" 14:08 < alt> hmm.. and I thought I had data on that drive too :\ 14:09 < yeti> a bit tricky but that would make the host and his umls more equivalent... maybe later i even could boot one of the systems as real one and the others as umls... independently of which one is the real one ;-) 14:10 < alt> ooh.. don't say that :) 14:10 * alt drools 14:10 < yeti> the missing part is booting logical volumes without ramdisk... 14:10 -!- desaster [~desaster@80.222.254.34] has joined #uml 14:11 < yeti> "partition alias" does the trick 14:11 < yeti> a faked(?) partition entry for a LV 14:12 < alt> hmm... I don't think DMA is enabled on that machine either 14:12 < alt> Disk I/O seems slow 14:40 -!- pirlouit [~peter@64.162.195.202] has quit [Remote host closed the connection] 14:46 -!- pirlouit [~peter@64.162.195.202] has joined #uml 14:46 < tanelorn> yeti: im not sure whether i understood you right, you mean "partition alias" is a better solution than cow filesystems? 14:52 < yeti> no 14:53 < yeti> i like uml fs-es on logical volumes 14:53 < yeti> there i have snapshot volumes if needed... so i dont see why i should use a cow file 14:54 < yeti> the "partition alias" comes into the game from the idea of having all fs-es, even the host ones, so uncluding his boot and root on lvs too... 14:55 < yeti> that partition alias trick will make a ramdisk to boot an entirely lv based system unnecesary 14:55 < tanelorn> i see... 14:56 < tanelorn> i think logical volumes are very nice for umls... are ext3/reiser/xfs resizeable? 14:58 < green> ext3 and reiser are resizeable. reiserfs can even be grown online 14:58 < green> not sure about xfs 14:58 < BB> green hmm online growing sounds nice 14:58 < green> the problem with reiserfs is it is not recommended to have reiserfs images on reiserfs filesystem 15:00 < shak> hrm 15:00 < shak> when I've made my bridge in UML, when I come to ifconfig the eth0 inside the UML I get SIOCSIFFLAGS: Device or resource busy 15:01 < shak> I mean, when I've made the bridge outside of the UML 15:01 < david> shak: maybe something is using the tap device already 15:01 < yeti> i lost hundrets of Gigs brcause of reiserfs 15:01 < shak> good point, I'll make a new tap device, 15:01 < shak> now I see why you have the nicely named tap devices 15:01 < shak> makes sense 15:01 < BB> http://www.charleslamb.co.uk/cgi-bin/details.cgi?propid=115 15:01 < green> shak: that might happen with any fs, and even if you do not use any fs at all 15:01 < BB> doh 15:02 < shak> was that aimed at me green ? 15:02 < shak> or yeti? 15:02 < green> shak: nope, it was foer yeti ;) 15:02 < shak> I guessed as much 15:03 < shak> david: am I meant to boot my UMLs with eth0=tuntap,tap1 for this? 15:03 -!- ElectricElf [david@elf.noc.oftc.net] has quit [oxygen.oftc.net ionosphere.oftc.net] 15:03 < tanelorn> btw. are your working with tun or tap? 15:04 < shak> oh, I have it ;) 15:04 < shak> superb, thanks david 15:04 < shak> now onwards and upwards 15:04 < shak> a quick script so I dont ever have to do that again 15:07 < yeti> shak: spent some time on not so quick scripts ;-) 15:08 < yeti> d 15:08 < shak> there's no need, I just need a script that sets up the bridge 15:08 < shak> I have a startuml one for starting the UMLs, a startnetwork inside my UMLs (temporary) and a mkbridge one for starting the bridge 15:09 < yeti> too complicated 15:10 < yeti> but that will be easier when you set up a system to concentrate on this stuff... then it can start the bridge at boot... and so on 15:10 < shak> :) 15:10 < shak> Ive got to find a hard disk for that system, someone has nicked the one that I thought was there 15:10 < yeti> hmmmm 15:12 < yeti> drop me a line if you cant set up that box... then i'll permute some hardware here... 15:12 < shak> heh, where are you? England? 15:12 < yeti> .de 15:12 < shak> ah, .uk here 15:13 < yeti> i can set up something here 15:13 < shak> 7I should be fine, I just need to nick the HDD back 15:14 < yeti> but i only have at max a k6-2@500 with only 4g HD AND 192 m ram left. all other boxes carry umls too but have additional tasks 15:14 < shak> I'll make it up over this week/weekend 15:14 * BB says hi to the fellow uk'er ;) 15:14 < shak> ello BB 15:14 < shak> s/ello/hello 15:14 < shak> you anywhere near York? 15:14 < BB> newcastle 15:15 < shak> because there's that talk this weekend 15:15 < shak> at the brewery 15:15 < shak> in York 15:15 < BB> yeah i know 15:15 < shak> I can't make it 15:15 < shak> I'm very annoyed 15:15 < BB> was thinking about it, but i've got a job on all day sat 15:15 < shak> me too, I'm in London (telehouse docklands) for the day 15:15 < BB> so basicaly if i go, i have to drive which means no beer 15:15 < shak> :( 15:15 < shak> are you going to LinuxUserExpo next week? 15:16 * tanelorn ist from .de, too :) 15:16 < shak> :) 15:19 < tanelorn> is swap necessary for umls? i dont think so, the host cares for this, doesnt it? 15:21 < shak> I haven't got any 15:23 < tanelorn> i think it doesnt make sense here... :) 15:24 < BB> shak i'm not going, mid week isnt a good time with work 15:25 < shak> I'm taking a day off 15:25 < shak> we're going on Thursday 15:25 < shak> to hear Alan Cox 15:25 < BB> heh 15:26 -!- shak [shak@shell0.uh-hosting.co.uk] has quit [Quit: leaving] 15:29 -!- shak [shak@shell0.uh-hosting.co.uk] has joined #uml 15:29 < shak> gah, I messed up there 15:29 < shak> closed my screen instead of detaching it 15:36 < tanelorn> what was the difference between tun and tap devices? 15:38 < david> tun is IP only 15:39 < david> tap is ethernet 15:39 < tanelorn> what does tun/tap abbreviate? i _always_ mix them. _always_. 15:39 < tanelorn> are there any mnemonics for? %-) 15:40 < tanelorn> okay. if ip is all i need, there is no reason to choose tap..? 15:40 < david> UML doesn't support tun 15:40 < david> so you have to use tap 15:41 < shak> I think I might set up an SNMP daemon now 15:41 < tanelorn> sure...? okay... 15:41 < david> tanelorn: er, yes 15:41 * shak goes to find some coke first 15:57 < tanelorn> david: how to tun and tap devs differ? if i configure a tunctl -t tun0 or tunctl -t tap0, are they different kinds of devs...? 15:57 < tanelorn> or just different names for tap-devs? 16:00 < BB> just different names, you could do tunctl -t boat0 16:01 < tanelorn> hehe 16:01 < tanelorn> how do i get tun-devs? 16:02 < BB> hrm no idea ;) 16:03 < tanelorn> why is this tool called tunctl when it produces tap-devices? =) 16:04 < BB> heh 16:04 < BB> i have no idea 16:05 < tanelorn> .oO( boat0 is a pretty idea. i could use customer-identifications ;) 16:17 < tanelorn> is it ay to get this message in my uml in dmesg "dev_ip_addr - device not assigned an IP address" ? 16:44 * BB moo's 16:44 < shak> does anyone have an snmp daemon runnng? 16:44 < shak> moo BB. 16:45 < shak> Im just wondering which one to install 16:46 < BB> I have ucd-snmp 16:47 < shak> :) 16:47 < shak> is sourceforge down for you? 16:58 -!- DnsInfector [~DnsInfect@ARennes-204-1-5-4.w80-15.abo.wanadoo.fr] has joined #uml 17:18 < tanelorn> is chosing tuntap as network a wise decision? it seems... strange to me. ive got one uml running and i am not able to nmap my uml.... 17:19 < DnsInfector> you should get timing error 17:19 < DnsInfector> but beeing able to nmap your host 17:19 < DnsInfector> which nmap are U using ? 17:20 < tanelorn> mmmh. maybee. waiting as nmaping a 33.6k modem user... 17:20 < tanelorn> nmap-3.27-r1 17:20 < DnsInfector> namap -v -v -v 17:20 < DnsInfector> you should get results faster 17:20 < DnsInfector> at least in a more interactive way 17:20 < tanelorn> shortly after stoping nmap i couldnt even ping that machine. i had to do a ping from the uml first. then it worked. 17:21 < DnsInfector> i used to test ip-personnality with UML and had not a single pb to nmap it 17:21 < DnsInfector> have you check your routing 17:21 < DnsInfector> ? 17:21 < tanelorn> "RTTVAR has grown to over 2.3 seconds, decreasing to 2.0" 17:21 < tanelorn> i have. :) 17:21 < DnsInfector> yes i have this kind of message too 17:22 < DnsInfector> but nmaping is correct isn't it ? 17:24 < tanelorn> okay. i dont know. stopped running and started it over now. 17:26 < tanelorn> ...still..running... 17:30 < tanelorn> finally finished. 17:30 < tanelorn> scanned in 306.273 seconds 17:30 < tanelorn> results seem to be okay 17:31 < tanelorn> superior. :( 17:32 < tanelorn> from a nearby linux i start pinging my uml, which is okay. then i nmap my uml from the host and the pings get corrupted. 17:34 < david> tanelorn: you don't make a tun device - UML doesn't use tun devices 17:35 -!- DnsInfector_ [~DnsInfect@ARennes-204-1-3-69.w193-252.abo.wanadoo.fr] has joined #uml 17:38 < shak> hrm 17:38 < shak> mrtg isnt liking this 17:39 < shak> or snmpd isnt set up properly 17:39 * shak goes back to poke it 17:40 * DnsInfector_ is fighting with zebra 17:41 * shak gets out the garlic 17:41 < shak> zebra is _evil_ 17:41 < tanelorn> shak: why..? 17:41 < DnsInfector_> that's a big animal 17:41 < shak> tanelorn: it just is, far too complex unless you have loads of time. 17:42 < tanelorn> shak: what do you want to achieve? 17:42 < DnsInfector_> and it's poorly documented 17:42 < shak> very poorly. 17:42 < DnsInfector_> and has few howtos 17:42 < shak> tanelorn: nothing, I ended up using something else 17:42 * tanelorn nods 17:42 < shak> no Im trying to poke snmpd 17:42 < tanelorn> shak: hehe... 17:42 < tanelorn> shak: okay. what did you want to do? 17:43 < shak> well, some dynamic routing for a client 17:43 < DnsInfector_> me that's still my ipv6 pb 17:43 -!- DnsInfector [~DnsInfect@ARennes-204-1-5-4.w80-15.abo.wanadoo.fr] has quit [Ping timeout: 492 seconds] 17:43 < DnsInfector_> with the UML that doesn't take the right prefix 17:44 < tanelorn> shak: oh no. you tried _to route dynamically_ using zebra!? 17:44 < shak> tanelorn: yes. 17:45 < DnsInfector_> tanelorn: ipv4 or ipv6 ? 17:45 < DnsInfector_> tanelorn: i'm interested too 17:49 < tanelorn> tanelorn: ipv4 17:50 < tanelorn> shak: dont say it too loud, perhaps the spanish inquisition is near.. *psccht* 17:50 < shak> oh no, not the spanish inquisition 17:50 * shak loves Monty Python 17:52 < alt> Python!!!! 17:53 < shak> right 17:53 < alt> We're the knights of the round table... 17:53 < alt> hee hee :) 17:53 < shak> mrtg appears to be working 17:53 < shak> ;) 17:53 < alt> NI! 17:53 < alt> okay guys... I need help 17:53 < alt> I have the following choices for filesystem 17:53 < alt> ext2, JFS and Reiser 17:53 < alt> I'm running a Mailserver using Maildir 17:53 < alt> any suggestions? 17:53 < shak> personally, I'm a reiser an 17:53 < shak> s/an/fan 17:55 < tanelorn> :) 17:55 < alt> okay, that's one for reiser and "ext3, but you don't have it, so reiser" 17:57 < alt> 2 for ""ext3, but you don't have it, so reiser" 18:01 < alt> I'm getting a lot of votes for ext3 in #cola 18:01 < shak> I dont really like ext3 18:08 -!- smcavoy [~smcavoy@grimlock.drive-megawheels.net] has quit [Quit: ACK!] 18:17 < silug> ext3 has the distinct advantage of being hammered on by 1e666 people 18:20 < alt> hee hee 18:56 -!- pflanze [~chris@cc-linux4.ethz.ch] has quit [Ping timeout: 492 seconds] 19:07 -!- tanelorn [~chase@schnuffel.beschoner.net] has left #uml [] 19:46 -!- pflanze [~chris@217.162.115.201] has joined #uml 21:29 -!- pflanze [~chris@217.162.115.201] has quit [Ping timeout: 488 seconds] 21:33 < alt> just so everyone knows, the skas3 patch doesn't seem to cause problems in 2.4.21 21:33 < alt> it's probably safe to use 21:33 < yeti> seems so here too 21:33 < alt> (haven't tested it enough) 21:34 < alt> yeti: glad someone agrees with me :) 21:36 < yeti> btw... tail merging would be fine for a maildir mailserver... 21:36 < alt> tail merging? 21:36 < yeti> so reiser has a space advantage over ext(2|3) 21:36 < alt> oh 21:36 < alt> hmmm 21:36 < alt> well, nothing is set in stone yet anyhoe 21:37 < alt> anyhow 21:37 < yeti> but ext2/3 is extremely stable 21:37 < yeti> and you can convert ext 2<->3 21:37 < alt> I have ext3 installed on it ATM 21:37 < alt> I'm thinking ReiserFS still tho 21:38 < yeti> i donnow about tail-merging in jfs, xfs... 21:38 < alt> hmmm... I get the feeling that the Compaq1600 and 1200 are pretty similar servers 21:38 < yeti> tail-merging is using the space behind the end-of-file up to the end of block for other files 21:38 < alt> okay 21:39 < alt> that's a nice trick 21:39 < yeti> maybe xfs/jfs can do that too... and maybe there is a patch for ext2/3... 21:40 < yeti> i'm not that far... 21:40 < yeti> such questions dont come up in my "play net at home" 21:40 < alt> http://cgi.ebay.ca/ws/eBayISAPI.dll?ViewItem&item=3031123190&category=11212 21:41 < yeti> and @work i always favour stability over other aspects, i am verrrrrrry conservative therte 21:41 < alt> well, this is a work box 21:41 < alt> so it needs to be stable 21:43 < yeti> i need to find something to male the bridge less verbose... some of my logfiles are >100m now... 21:43 < yeti> but it is 3:42am here... think this is an after sleep item ;-) 21:43 < alt> hee hee 21:43 < alt> ttyl then 21:44 < yeti> i'll be back soon... 21:44 < yeti> have fun! 21:44 -!- yeti [~yeti@p3EE22947.dip.t-dialin.net] has quit [Quit: Client exiting] 21:53 -!- caker [~null@68.52.196.167] has quit [Read error: Connection reset by peer] 22:31 -!- DnsInfector_ [~DnsInfect@ARennes-204-1-3-69.w193-252.abo.wanadoo.fr] has quit [Quit: Client exiting] 22:33 -!- caker [~null@pcp507591pcs.nash01.tn.comcast.net] has joined #uml 23:13 -!- kayl [~jada@6532176hfc248.tampabay.rr.com] has joined #uml 23:14 -!- kayl [~jada@6532176hfc248.tampabay.rr.com] has quit [Client Quit] 23:53 -!- tarcan [~tarcan@200.97.184.70] has joined #uml 23:54 < david> hello 23:56 < alt> hi 23:56 < tarcan> hi --- Log closed Thu Jun 19 00:00:00 2003