--- Day changed

--- Log opened Mon Jun 09 00:00:01 2003
00:13 -!- ElectricElf [david@elf.noc.oftc.net] has quit [Read error: Connection reset by peer]
00:13 -!- ElectricElf [david@elf.noc.oftc.net] has joined #uml
02:44 -!- green [green@217.76.32.60] has quit [Quit: Leaving]
03:13 -!- radical [~kernet@202.88.184.39] has joined #uml
03:27 -!- _slh__ [~chatzilla@68.154.227.191] has quit [Ping timeout: 488 seconds]
04:08 -!- green [green@linuxhacker.ru] has joined #uml
04:45 -!- avoozzzl is now known as avoozl
05:34 -!- radical [~kernet@202.88.184.39] has quit [Ping timeout: 496 seconds]
05:55 -!- BB [~chris@ns1.8086.net] has joined #uml
06:47 -!- fle [~elf@home04019.cluj.astral.ro] has joined #uml
06:48 < fle> hello
06:54 -!- fle [~elf@home04019.cluj.astral.ro] has quit [Quit: BitchX Official FTP Site -- ftp://ftp.bitchx.com]
07:27 -!- _slh__ [~chatzilla@68.154.227.191] has joined #uml
08:47 -!- radical [~kernet@202.88.184.39] has joined #uml
08:59 -!- radical [~kernet@202.88.184.39] has quit [Remote host closed the connection]
09:00 -!- radical [~kernet@202.88.184.39] has joined #uml
11:03 < david> hi
12:12 < desaster> hey
12:32 -!- smcavoy [~smcavoy@grimlock.drive-megawheels.net] has joined #uml
12:33 < mistik1> weeeeeeeeee
12:33 < mistik1> redhat8.0 up and running
12:48 < david> caker: around?
12:49 < BB> hmm did you know a 1TB sparse file with an ext3 filesystem on ends up as 18GB .. I found that an interesting way to loose 18GB ;)
12:54 < mistik1> hehe
12:54 < mistik1> BB: a little bored were we?
13:05 < avoozl> BB: doesn't seem to take up any noticable space here on xfs: -rw-r--r--    1 root     root     1099512676352 Jun  9 19:05 tmpfile
13:06 < avoozl> du says it takes about a megabyte
13:10 < avoozl> can't seem to grow it over 8.0T though, must be the max filesize :P
13:10 < avoozl> then it takes up about 8megabytes
13:11 < radical> has anyone tried rhl9 with umlbuilder? i am getting errors for bdflush rpm
13:12 < avoozl> BB: finally a good replacement for /dev/zero :)
13:14 < BB1> heh
13:14 < BB1> I'd done a mke2fs -j on it also ;)
13:14 < BB1> then i gave it to a m8 to use with redhat installed
13:18 < BB1> radical i looked couple of days ago, but didnt try as it wasnt in the list of distros
13:19 < green> radical: There is no bdflush rpm in RH9
13:19 < green> BB1: basically totally empty sparse file for ext2/ext3/reiserfs is roughly equal filesize/1000
13:20 < green> s/equal/will take space/
13:21 < radical> green: exactly hmm i actually selected rhl8 in umlbuilder.. so i guess it was givin that error.. will have to hack ard a bit i guess
13:31 < avoozl> green: another reason to use xfs :)
13:47 < green> avoozl: hehe.
13:47 < green> avoozl: or ntfs (from w2k), or reiser4 or whatever else ;)
13:48 < BB1> any idea which is best for loads of small files on a 1.5TB drive?
13:49 < green> xfs sucks on loads of small files that is easily demonstrated by my scrip that creates 59 millions of files on reiserfs in 1 hour (deletes in 1 hour), I tired waiting for completion on xfs after 4 hours (and it progressed only to 1/3 of workload or so). And I never managed to wait till xfs deletes the created files
13:50 < BB1> lol
13:50 -!- radical [~kernet@202.88.184.39] has quit [Remote host closed the connection]
13:51 < BB1> I guess i should prob do some live data tests and just make additional copies
13:51 < green> yes, that's wise idea.
13:51 < green> all benchmarks are lying
13:52 < BB1> yeah, most of our stuff will (hopefully) be hard links
13:52  * green gave uml 100 megabytes more RAM only to find out that it does not use those. hm.
13:53 -!- radical [~kernet@202.88.184.39] has joined #uml
13:54 < BB1> heh its just playing hard to please
14:28 < mistik1> can you make a cow of a swapfs?
14:42 < green> why would you even want to do that? ;)
14:57 < mistik1> dont make much sense now that I think about it
14:58 < BB1> i guess u could rm ; dd; mkswap it everytime you booted to keep it trim
14:58 < mistik1> radical: I found that with umlbuilder the gui is a curse
14:59 < mistik1> the commandline with the switched worked like a charm while the very same options did not work in the gui
14:59 < mistik1> s/switched/switches/
15:02 -!- amrut [~amrut@203.124.158.218] has quit [Quit: Client Exiting]
15:13 < radical> mistik1: okk.. cool.. i'll give it a shot.. 
15:20 < mistik1> would you like to see the line I used?
15:22 < radical> mistik1: yup
15:25 < mistik1> umlbuilder --distro Redhat8.0 --modules base,base-x --hostname UMNode.geeksinthe
15:25 < mistik1> hood.net --ipaddr 192.168.1.209 --rpmdir /var/distros/redhat/8.0 --dir ~/UML --m
15:25 < mistik1> emsize 256 --fs /,root_fs,512,ext3,0 --fs /usr,usr_fs,1024,ext3,1 --fs /var,var_
15:25 < mistik1> fs,1024,ext3,2 --initrd false
15:25 < mistik1> bah
15:26 < mistik1> it broke up the paste
15:26 < radical> long one heh np 
15:26 < mistik1> I just install base and base-x then I installed apt-get and get what else I needed from online source as I normally would
15:27 < radical> hmm okie.. 
15:28 < radical> oh well, same bdflush error! i'll have to remove the reference to bdflush.. wherever that might be.. as soon as i get time!!
15:39 < green> hehe, better install native rh9 ;)
15:39 < mistik1> true enough
15:40 < mistik1> I guess i'll stop being lazy and download the ISOs
15:40 < mistik1> i've not touched 9.0 yet
15:45 < green> hehe
16:32 -!- radical [~kernet@202.88.184.39] has quit [Ping timeout: 490 seconds]
16:36 -!- radical [~kernet@202.88.184.39] has joined #uml
17:02 -!- Ryback_ [~ra993482@malazarte.lsd.ic.unicamp.br] has joined #uml
17:25 -!- fle [~elf@home04019.cluj.astral.ro] has joined #uml
17:26 < fle> hello
17:34 -!- fle [~elf@home04019.cluj.astral.ro] has quit [Quit: [BX] Eat, drink and be merry...for tomorrow we die]
17:46 -!- zlb [blizzard@stealth.leenoox.org] has joined #uml
17:48 < zlb> does UML have problems running on host which has grsecurity patch?
17:49 < mistik1> no clue
17:50 < mistik1> I could see the posability though
17:50 < mistik1> if ACLs are tight
17:50 < zlb> I keep getting segfault as soon as I run it, and is not the 2g/2g problem
17:50 -!- Ryback_ [~ra993482@malazarte.lsd.ic.unicamp.br] has quit [Quit: Client Exiting]
17:50 < mistik1> donno
17:51 < mistik1> I dont use grsec, dont much believe in it
18:04 < david> zlb: do you allow processes to be ptraced?
18:04 < mistik1> hi david 
18:05 < mistik1> you said if it looks good you would post my script on the site
18:05 < mistik1> I think i'm satisfied with it for now
18:06 < mistik1> http://music.geeksinthehood.net:81/~mistik1/ifup-bridge-0.2.sh
18:06 < mistik1> I have now gotten a change to test all the modes
18:07 < mistik1> maybe you can add something to it or just flame me so I get my act togather
18:07 < zlb> david: no
18:07 < mistik1> that would be needed for uml to work
18:22 -!- smcavoy [~smcavoy@grimlock.drive-megawheels.net] has quit [Remote host closed the connection]
18:45 -!- gump [~gump@81.5.136.90] has joined #uml
18:50 < david> zlb: that would be a problemthen
19:07 -!- BB1 [~chris@217.79.122.210] has quit [Ping timeout: 492 seconds]
19:35 -!- radical [~kernet@202.88.184.39] has quit [Quit: Client Exiting]
19:44 < mistik1> darnit
19:44 < mistik1> You cant move a sparce file can you?
19:45 < ljlane> dunno, but cp has a sparse option
19:45 < mistik1> too bad I dont see one for scp
19:45 < mistik1> hmm
19:46 < mistik1> but if I mount the remote via nfs the sparse copy should work
19:46 < ljlane> tar has one too
19:47 < ljlane> perhaps a variation of the ssh/tar copy: tar cf - blah/ | ssh you@remote tar xvf -
19:47 < mistik1> yah
19:48 < david> mistik1: I'd just bzip2 it, bunzip2 it
19:48 < david> then cp --sparse=always in out
19:48 < david> to make it 'sparse' again
19:48 < mistik1> thanks
19:50 < mistik1> that works fine
20:59 -!- BB1 [~chris@217.79.122.210] has joined #uml
21:22 < caker> Hey all
21:25 < mistik1> hi caker 
21:25 < caker> hello mistik
21:26 < mistik1> there should be a law against this much fun/connectivity/power caker 
21:26 < caker> heh - with what?
21:27 < mistik1> linux in general UML especially
21:27 < caker> I say give it to the masses
21:28 < mistik1> did you ever get your network issues sorted?
21:28 < caker> nope - only implemented the work-around
21:28 < mistik1> hmm
21:29 < caker> cron job that pings the IPs every minute, and set the MAC address table timeouts on the switch to 4 hrs
21:29 < mistik1> I was trying to catch you yesterday while I had a similar setup going
21:29 < mistik1> but that site will be down for the next few weeks at least
21:29 < caker> screw it - i'm *guessing* something is weird with the cisco switch's and bridging ?
21:29 < mistik1> I too think its the switch
21:30 < caker> it already ate three days or so of my time, so .. I'm over it ... 
21:30 < caker> I did see your messages, thanks for the follow up
21:30 < caker> next time I'll get the el cheapo switch
21:30 < caker> it will probably work :-)
21:31 < mistik1> It was the first chance I got to test my script with routable UMLs
21:31 < mistik1> worked like a champ but I can see where I would need ebtables
21:31 < caker> I've got another Cisco switch here at home (a 5005 cabinet model), so if I ever care to test, I could recreate the environment here and really identify the switch as the problem
21:32 < caker> I use the bridge-netfilter patches and do my IP limiting in iptable rules
21:32 < mistik1> oh?
21:32 < mistik1> got a URL for me
21:32 < caker> also I have some rules for MAC limiting too (didn't have ANYTHING to do with the ARP issues, I tried with zero rules)
21:32 < caker> http://ebtables.sourceforge.net/
21:33 < mistik1> I use a very extensive iptables script and see i'll prolly have to hack some extentions into it for working with bridges
21:33 < caker> not really - i just use the bridge-nf patch, and then use source-device filtering in my chains
21:33 < mistik1> I love the script because we use sed/awk/grep to process a acl file
21:34 < caker> but - i'm kind of a ipchains idiot, so there are probably better ways to do what I did
21:34 < mistik1> very cisco`esk
21:34 < caker> cool
21:34 < mistik1> it also will restart the ACL portion of the rulesets without bringing down the whole firewall for a second
21:35 < mistik1> it's really quite nice
21:35 < caker> sounds like it
21:35 < mistik1> lemme show you a sample ACL
21:35 < caker> k
21:36 < mistik1> PORTFW IFACE eth1 63.230.20.129 tcp TOPORT 3957  TOHOST 192.168.1.31 HOSTPORT 3957 ACCEPT
21:36 < caker> so you start off with everything dropped
21:36 < caker> and poke holes for everything?
21:37 < mistik1> yup
21:37 < caker> I'm giving my users enough rope to hang themselves
21:37 < mistik1> heh
21:37 < caker> next thing for me is to work on a limiting/shaping rule
21:37 < mistik1> me too
21:38 < caker> for when they go over their bandwidth, I'd like to limit their throughput, rather than charge them overages 
21:38 < mistik1> I've not done much there but this is OSS
21:38 < caker> I think that is nicer
21:38 < mistik1> *nod*
21:38 < mistik1> throttling is better than extra fee imho also
21:39 < caker> I agree - and if they don't like it throttled until the end of the month, then they can just upgrade their account, and i'll drop the rule
21:41 < mistik1> wait a sec
21:41 < mistik1> are you using 2.2.x
21:41 < caker> kernel?
21:41 < mistik1> yea
21:41 < caker> no
21:41 < caker> 2.4.21-rc7 on host
21:41 < caker> plus some patches
21:41 < mistik1> then why the bridge-nf patch?
21:41 < caker> 2.5 has it included
21:41 < caker> 2.4 you need ebtables, then bridge-nf
21:41 < mistik1> I think I missed something in my reading
21:42 < caker> let me find you the versions you need, their site is confusing
21:42 < mistik1> yup I certainly did
21:42 < caker> ebtables-v2.0.003_vs_2.4.20.diff  and bridge-nf-0.0.10-against-2.4.20.diff are latest
21:42 < mistik1> these patch cleanly over skas3?
21:43 < caker> do them first, then skas3, but yes will apply fine (with fuzz/offsets)
21:43 < mistik1> ack!
21:43 < mistik1> ok
21:43 < mistik1> np, I just have to start with a clean tree
21:44 < caker> I'm up to the 14th version of my host kernel, and think it's finally dialed in
21:44 < caker> so, i hear ya
21:45 < mistik1> heh
21:45 -!- sangr [~sangr@ip68-13-126-189.om.om.cox.net] has joined #uml
21:52 < mistik1> that went welll
21:53 < mistik1> very clean patches indeed
21:54 < mistik1> caker: I assume I need to use the ebtables userspace app to make use of these patches?
21:55 < caker> if you are going to use ebtables rules, yeah .. otherwise standard iptables works for me
21:55 < mistik1> ok
21:55 < mistik1> maybe you'll give me a look at your ruleset one day
21:56 < mistik1> this way if something is needed in killerwall I can patch it in
21:56 < caker> mine are painfully simple
21:56 < mistik1> I'd like to be able to use my same script
21:56 < caker> drop anything that's source isn't !MAC of UML device
21:56 < caker> allow rule for each IP
21:56 < mistik1> I understand but iptables can be finicky
21:56 < caker> log/limit rule
21:57 < caker> then drop anything else
21:59 < mistik1>  IN IFACE eth0 FROM MAC ! DE:AD:00:BE:EF:00 tcp TOPORT 80 DROP
22:00 < mistik1> an ACL something like that in my current script should work
22:00 < caker> if you're going to use bridge-nf/iptables, the source iface will be the tap device
22:00 < mistik1> fair enough
22:01 < caker> you probably know more about it than I do, so :-)
22:01 < ASY> guys, i just installed new system with tun/tap enabled in the kernel...  however I just realized that i don't even have /dev/net not to mention /dev/net/tun...  any ideas?
22:01 < mistik1>  IN IFACE mistik1-2786 FROM MAC ! DE:AD:00:BE:EF:00 tcp TOPORT 22-65399 DROP
22:02 < mistik1> moprobe tun
22:02 < mistik1> err
22:02 < mistik1> modprobe tun
22:02 < ASY> i made it built-in.  not a module.   
22:03 < ASY> hm...  i guess i will try rebuilding the kernel...   will check dmesg first.
22:03 < caker> running the new kernel?  do you have to mknod the devices (don't think so) .. ?
22:03 < mistik1> have you tried just making a tap
22:03 < mistik1> not with devfs
22:03 < caker> is he running devfs?
22:04 < mistik1> to think of it
22:04 < mistik1> I dont use devfs on the box i play uml on
22:04 < mistik1> I never had to make the node
22:04 < caker> same
22:04 < ASY> not sure about the host box... 
22:05 < mistik1> try just making a tap
22:05 < mistik1> tunctl -u <username>
22:05 < ASY> Universal TUN/TAP device driver 1.5 (C)1999-2002 Maxim Krasnyansky
22:05 < ASY> debby:/dev# tunctl
22:05 < ASY> Failed to open '/dev/net/tun' : No such file or directory
22:05 < Pahan> har har har
22:05 < Pahan> devfs hates you?
22:05 < mistik1> then I bet /dev/net will show up
22:06 < mistik1> hrmmm
22:06 < mistik1> wierd
22:06 < Pahan> Oh, wait, the stupid thing attempts /dev/misc/net/tun when devfs is detected.
22:06 < Pahan> Bleh.
22:06 < mistik1> ahhhhhh
22:07 < mistik1> ln -s /dev/misc/net /dev/net ?
22:07 < ASY> i just had to do like 4 consequent kernel builds to establish proper config...  maybe i should try to do make clean and a complete rebuild...  or should i try to boot with devfs=mount in lilo?
22:07 < ASY> no /dev/misc either
22:07 < caker> ASY: just do this: mkdir /dev/net ; mknod /dev/net/tun c 10 200
22:07 < caker> then remember to set proper file permissions on it
22:08 < ASY> ok
22:08 < mistik1> that is wierd
22:08 < ASY> caker: this worked.  thanks.
22:08 < caker> ASY: np
22:08 < mistik1> I've done this now on 5 different distros and never say that
22:08 < mistik1> with and without devfs
22:08 < ASY> my question would be "but why"...
22:09 < caker> silly distros - who knows what you get with them
22:09 < mistik1> heh
22:09 < mistik1> I smell an LFS user
22:09 < mistik1> or debianista
22:09 < caker> nope :-) just learned not to trust them
22:09 < caker> or rely on them, rather
22:09 < mistik1> I hear that
22:09 < ASY> caker: any idea what was suppose to create this in the first place
22:09 < caker> you
22:10 < mistik1> I use a redhat base system
22:10 < caker> or the filesystem your distro installed
22:10 < mistik1> the latest distro to get on my nerves is mandrake
22:10 < ASY> hm...  you might be right.  (woody here).  but yes, you might be right.  my previous setup was done with a friend, he could have done this without me realizing
22:10 < mistik1> what a hoser
22:11 < caker> I still need some more beta testers .. anyone care to have a free linux machine on a fast network for a week?
22:11 < mistik1> caker: I've been so deep in my testing that I still have not gone back to signup
22:11 < caker> mistik1: it's all good
22:11 < mistik1> I will though
22:12 < mistik1> right now i'm working on setting up a linux terminal server out of UML
22:12 < mistik1> then i'm gonna make it a liveCD
22:13 < caker> I guess I'll email my local LUG, but those guys are a**holes most of the time .. bunch of prima donnas
22:13 < mistik1> lol
23:59 -!- sangr [~sangr@ip68-13-126-189.om.om.cox.net] has quit []
--- Log closed Tue Jun 10 00:00:00 2003