--- Day changed

--- Log opened Mon Jun 23 00:00:04 2003
01:18 -!- ASY [~someone@216.221.51.240] has joined #uml
01:56 -!- rob [rob@213.230.203.222] has joined #uml
01:56 -!- rob is now known as shak
02:08 -!- charles [charles@on-tor-blr-a58-03-733.look.ca] has joined #uml
02:36 -!- ido [~Ido@bzq-218-140-160.cablep.bezeqint.net] has quit [Ping timeout: 496 seconds]
03:02 -!- ido [~Ido@bzq-218-140-160.cablep.bezeqint.net] has joined #uml
03:17 -!- ASY- [~someone@216.221.51.240] has joined #uml
03:17 -!- ASY [~someone@216.221.51.240] has quit [Read error: Connection reset by peer]
03:17 -!- ASY- is now known as ASY
03:19 < ido> :)
03:20 < ido> anyone here ?
03:28 < caker> HELLO
03:28 < caker> oops
03:33 < ido> :)
03:35 < ido> the UML seems to run stuff really
03:35 < ido> slow
03:36 < ido> although what might be causing this is the fact that i've got it running on about 40mb ram and swap
03:36 < ido> although its an athlonXP 1700 processor 
03:37 < caker> Two most common things to improve performance (besides add more ram) is run in skas mode, and use a "tmpfs" filesystem for your /tmp dir
03:37 < caker> Next would be to make sure you didn't compile with "ubd always performs sync I/O" in your UML's kernel config
03:38 < ido> tmpfs is being used..
03:38 < ido> hrmf.
03:38 < ido> if i'll strip the UML kernel, would it take up less space in memory while loaded ?
03:39 < caker> how big is your kernel?
03:39 < ido> 11mb
03:40 < caker> dammn
03:40 < caker> do you have any of the kernel debugging kernel options enabled?
03:40 < ido> nope
03:40 < caker> wow - that's big imho
03:43 < ido> :)
03:44 < BB> mines 33MB before i strip it
03:44 < ido> WOW
03:44 < ido> with debugging ?
03:54 < BB> yeah it has some debug on
03:58 < ido> :)
03:58 < ido> thats why
04:03 < ido> hmm.
04:04 < ido> i've ssh'd into my box, and now i see what i type in the UML window
04:04 < ido> since i have con=pty con0=fd:0,fd:1
04:04 < ido> should i change it to pts instead of pty ?
04:10 < ido> apparently it didn't solve it
04:23 < ido> hmm
04:24 < ido> and fake_ide and fakehd don't seem towork
04:50 -!- mistral [mistral@212.159.71.212] has quit [Ping timeout: 488 seconds]
04:52 < ido> anyone here ?
04:52 < ido> actually they do work
04:53 < ido> but i wonder if i could reproduce a /dev/hda0 with mknod
04:58 -!- mistral [mistral@jstevenson.plus.com] has joined #uml
05:02 < ido> seems to work
05:54 < desaster> yeah, but should that be hda1 ?
06:13 -!- yeti [~yeti@p3EE22A3D.dip.t-dialin.net] has joined #uml
07:07 -!- ichilton [~ian@pc3-stoc3-4-cust203.midd.cable.ntl.com] has quit [Ping timeout: 490 seconds]
07:37 -!- ichilton [~ian@80.6.255.203] has joined #uml
07:49 -!- desaster [~desaster@80.222.254.34] has quit [Quit: EPIC Panic: EPIC4-1.1.12 (524):BE SURE TO INCLUDE THE ABOVE IMPORTANT INFORMATION! -- new_free()'s magic check failed from [window.c/4622].]
07:51 -!- desaster [~desaster@dsl-jklgw1o22.dial.inet.fi] has joined #uml
07:58 -!- sssssssh______ [~chatzilla@adsl-19-228-6.jax.bellsouth.net] has joined #uml
08:09 -!- radical [~kernet@210.7.75.194] has joined #uml
08:42 -!- radical [~kernet@210.7.75.194] has quit [Read error: No route to host]
08:48 -!- litost [~mondain@phynp6.phy-astr.gsu.edu] has joined #uml
09:18 < ido> could i remove the 'UML User Mode Linux GNU/Linux' part to show up from uname -a ?
09:58 -!- charles [charles@on-tor-blr-a58-03-733.look.ca] has quit [Ping timeout: 496 seconds]
10:14 -!- tito [~tito@tito.dit.upm.es] has joined #uml
10:15 < tito> hi all
10:26 -!- charles [charles@216.154.6.130] has joined #uml
10:26 < tito> can anyone help me, I am trying to use uml with a testing scenario and it hangs at point "Initializing stdio console driver"
10:35 -!- tito [~tito@tito.dit.upm.es] has quit [Ping timeout: 496 seconds]
10:37 -!- tito [~tito@138.4.7.96] has joined #uml
10:37 < tito>  hi all
10:37 < tito>  can anyone help me, I am trying to use uml with a testing scenario and it hangs at point "Initializing stdio console driver
10:37 < tito> thanks in advance
10:53 -!- Lathiat [lathiat@seven.sixlabs.org] has quit [Quit: brb]
11:01 -!- Lathiat [lathiat@seven.sixlabs.org] has joined #uml
11:07 < caker> tito: add this to your command line: con=null con0=fd:0,fd:1
11:11 < tito> ok just a moment
11:14 < tito> pass that breakpoint
11:14 < tito> now I see another breakpoint
11:18 < tito> caker: what is the meaning of "con=null con0=fd:0,fd:1"
11:18 < tito> caker: what is the meaning of "con=null con0=fd:0,fd:1"?
11:19 < caker> con=null sets all the conX's to null, but not con0, set that to stdin:stdout
11:23 < tito> caker: maybe is a silly question(but I'm really a newbie), what is the meaning of "conX"?
11:24 < caker> it's your virtual-linux machine's consoles - where to attach them
11:24 < caker> con0, con1, con2, etc
11:24 < caker> translate into tty0, tty1, tty2, etc inside your UML instance
11:25 < tito> ok
11:25 < caker> You should probably modify the UML's /etc/inittab and make an mgetty listen on con0, if it isn't already
11:25 < caker> err, tty0
11:27 < tito> caker: thanks a lot, I will read more about uml
11:28 < tito> caker: I'm a little bit lost
11:28 < caker> tito: no problem
12:09 -!- mistral [mistral@jstevenson.plus.com] has quit [Ping timeout: 492 seconds]
12:30 -!- charles [charles@216.154.6.130] has quit [Remote host closed the connection]
12:33 < yeti> shak?
13:00 -!- radical [~kernet@210.7.75.194] has joined #uml
13:00 -!- radical [~kernet@210.7.75.194] has quit [Client Quit]
13:16 -!- tito [~tito@138.4.7.96] has quit [Quit: Aplicación Saliendo]
13:36 -!- mistral [mistral@jstevenson.plus.com] has joined #uml
13:49 -!- rawl [~rawl@cable200-75-79-20.epm.net.co] has joined #uml
13:51 < rawl> i am new in uml, could someone please tell me where can i donwload recent root filesystems...
13:52 < yeti> i roll my own...
13:53 < yeti> are you looking for a special linux? or just some food to play with?
13:53 < caker> rawl: what distro are you looking for
13:54 < yeti> rawl: look for deboostrap... this will run on non-debians too and build a basic debian rootfs
13:55 < yeti> i suggest NOT to use some specialized uml builder... at least not now... make some roots on your own.
13:55 < ido> its not that hard either
13:55 < yeti> the easiest uml is host-fs-in the root of the host...  ;-)
13:56 < rawl> yeti to play with
13:56 < rawl> then later make it miself deeper
13:57 < rawl> ok thanks going to check out deboostrap
13:58 < yeti> ;-)
13:59 < yeti> if slow speed doesnt matter, i can give you URLs of tarballs of debootstrap output of all 4 debian flavours
13:59 < rawl> even woodie?
13:59 < rawl> sorry
14:00 < yeti> i just tarred the output of debootstrap runs...
14:00 < david> hello
14:01 < BB> hi david 
14:01 < rawl> hi
14:02 < caker> hello david
14:04 < david> caker: hey, how's linode going?
14:04 < caker> david: very well - things have been going great, and UML is incredibly stable and quick, too ..
14:04 < david> nice :-)
14:05 < caker> david: I was testing out game servers inside a test-linode, with the host machine going full steam with the other users ..
14:05 < caker> david: it was rock-stable, and fast, too
14:05 -!- mistral [mistral@jstevenson.plus.com] has quit [Read error: Connection reset by peer]
14:06 -!- mistral [mistral@jstevenson.plus.com] has joined #uml
14:06 < david> caker: that is my experience also :-)
14:07 < ido> how do i compile a host kernel with SKAS support ?
14:07 < ido> or is SKAS only compiled in the uml
14:07 < yeti> patch in skas mode and activate /proc/mm support
14:07 < caker> ido: skas is on the host kernel
14:07 < ido> hmm, k, and how do i compile it ?
14:07 < david> ido: read the kernel-howto
14:07 < ido> i need a Skas patch, right ?
14:08 < david> right
14:08 < yeti> for 2.4 yes, for 2.5 i donnow...
14:08  * yeti didnt touch 2.5 up to now
14:08 < david> 2.5 doesn't have a skas patch yet
14:08 < ido> me neither.
14:08 < ido> workign with 2.4
14:08 < ido> 2.4.21
14:09 < caker> off-topic question .. how do I generate a better/more verbose list of installed RPMs than what "rpm -qa" gives me?
14:11 < ido> is the skas patch called  mod-skas3-20030107.tgz ?
14:11 < david> ido: no
14:11 < david> ido: it would be host-skas3.patch
14:11 < david> caker: -qav? No idea :-)
14:11 < caker> david: heh -qvai ended up being it
14:12 < david> oh, well, I was close
14:13 < ido> isn't mod-skas3-... a module instead of patching the kernel itself ?
14:13 < david> ido: right, but it only works with RH8.0 I think
14:13 < ido> ah
14:14 < ido> so the file i need is host-skas3.ptrace.patch.bz2 ?
14:15 < david> no
14:15 < david> go to user-mode-linux.sf.net/dl-sf.html
14:16 < david> download host-skas3.patch at the bottom
14:16 < ido> isn't it quite outdated over there ?
14:16 < david> no
14:16 < ido> i mean, nov 2002 ?
14:16 < david> that is current
14:16 < david> nothing changed with the host-skas patch since then
14:16 < ido> oh
14:16 < ido> ok
14:16 < ido> thats why i haven't downloaded it yet
14:20 < ido> do i need to just recompile the kernel, or do i need to run menuconfig first and select SKAS support
14:21 < david> you need to enable CONFIG_PROC_MM
14:21 < ido> oh
14:22 < ido> anything else before i compile ?
14:22 < rawl> ido i found this easier to read tutorial
14:22 < rawl> http://archive.linuxfromscratch.org/mail-archives/hints/2001/12/att-0016/01-uml.txt
14:22 < rawl> read it
14:22 < ido> wrong page, got object not found error
14:23 < ido> its http://archive.linuxfromscratch.org/mail-archives/hints/2001/12/att-0016/uml.txt
14:23 < ido> seems that all is working atm
14:23 < ido> except this slowness issue i'm having
14:23 < ido> the host is a LOT faster
14:24 < david> rawl: except that howto was written about a year before skas was on the go
14:24 < david> ido: UML is slow in tt mode compared to skas
14:24 < ido> hoping the SKAS thingy will solve it
14:24 < david> ido: you have /tmp mounted tmpfs, right?
14:24 < ido> yep
14:24 < david> ok, just checking
14:24 < david> :-)
14:24 < ido> eh.
14:24 < rawl> david its because the question about the host kernel
14:24 < ido> what does that have to matter with speed ?
14:25 < caker> skas changes the method of UML's process and memory management
14:25 < rawl> david it help in part, he must see another tutorials, but this one good for begginers
14:25 < ido> isn't it just acting like a better 'swapdrive' ?
14:26 < caker> it's because every memory call in UML in TT mode was making corresponding block reads/writes (guessing)
14:29 < rawl> now scrolling i see no rational
14:29 < rawl> lity
14:34 < david> ido: er, you have a /tmp mounted with tmpfs because RAM is quicker than your HD
14:34 < silug> i had an old drive turn up bad sectors in /tmp a couple of weeks ago
14:35 < silug> i was only running one uml on the box, so i never bothered to point TMPDIR to tmpfs
14:35 < silug> i can't help but wonder if that had anything to do with it...
14:37 < ido> david: oh, but ofcourse
14:38 < ido> but that doesn't affect the whole system performance. i mean, it does, depends on the use.
14:38 < ido> /usage
14:38 < caker> everything makes memory calls
14:38 < caker> **everything**
14:39 < ido> hmm.
14:39 < ido> any iptable freaks ?
14:39 < ido> - over here that is
14:40 < BB> you might find a user, dont know about a freak
14:41 < ido> > i want to limit my small brother's game playing hours. which means, for example, to allow upto 2 hours of use on a specific port, when it reaches a limit, block the port. and viola, he won't be able to play. question is how can that be done...
14:42 < david> ido: iptables has a module to do that
14:42 < david> ipt_time, I believe
14:42 < ido> really ?
14:42 < ido> sweet
14:42 < caker> does it work on udp?
14:42 < david> caker: it works with anything
14:42 < ido> yep
14:42 < david> it's a 'match' for a rule
14:45 < david> ido: it's probably in the patch-o-matic tree
14:50 < ido> hmm
14:51 < ido> ipt_time is used to check a packet's time
14:52 < david> ido: it checks it's arrival time at the router
14:53 < ido> and how would that help me ?
14:54 < david> because you can set the port to be available between 16:00 and 18:00
14:54 < ido> but thats not what i want
14:54 < ido> i want to limit him for eg, upto 2 hours of use a day
14:58 < dg> you could work out the average number of packets the game sends and then scale the value up to how many in 2 hours, then use -m limit to limit him to that number/day
14:58 < dg> but that's a nasty hack ;)
14:58 -!- bch [~db@ppp-217-133-208-161.cust-adsl.tiscali.it] has joined #uml
14:58 < bch> Hi, can anyone tell me if uml "emulates" HIGH_MEMORY behaviour?
15:00 < ido> eh. i dont want that.
15:00 < dg> well write a module to do it then
15:00 < ido> how exactly ?
15:01 < ido> i mean, how could i know when he starts/finishes playing
15:01 < ido> we're talking about UDP packets ofcourse.
15:01 < dg> well iptables manages stateful nat/filtering on them
15:01 < dg> so you must be able to hook into that
15:02 < Pahan> bch: We can probably tell you less than what you'd glean from kernel source.
15:03 < ido> i could check when the first packet is launced, but how do i know when's the last ?
15:03 < BB> if theres no more within say 2mins game would either be over .. or they are about to time out would be my guess
15:04 < ido> yeah.. just had that thought
15:04 < ido> instead of writing a module
15:04 < ido> i could just log them to a file
15:04 < ido> and grep that file and comparing the packets' time agains one another
15:05 < ido> although that could take a lot of cpu cycles since we're talking about tons of packets..
15:08 < BB> how about just threatening to cut body parts off if they play the game over 2hrs might be easier
15:08 < ido> there's this module that could match every Nth packet, that way i could make it less cpu intensive
15:16 -!- radical [~kernet@210.7.75.194] has joined #uml
15:23 -!- kernet_ [~kernet@210.7.75.194] has joined #uml
15:23 -!- kernet_ [~kernet@210.7.75.194] has quit [Client Quit]
15:32 -!- bch [~db@ppp-217-133-208-161.cust-adsl.tiscali.it] has quit [Remote host closed the connection]
15:38 < BB> Kernel panic: Kernel stack overflow
15:38 < BB> doh :(
15:42 -!- ajax [~ajax@12-221-0-198.client.insightBB.com] has joined #uml
16:03 -!- shak [rob@213.230.203.222] has quit [Quit: BitchX: its wax ecstatic]
16:21 -!- ido [~Ido@bzq-218-140-160.cablep.bezeqint.net] has quit [Ping timeout: 496 seconds]
16:22 -!- ido [~Ido@bzq-218-140-160.cablep.bezeqint.net] has joined #uml
16:27 -!- rawl [~rawl@cable200-75-79-20.epm.net.co] has quit [Quit: [BX] That's Miss BitchX to you]
16:27 -!- yeti [~yeti@p3EE22A3D.dip.t-dialin.net] has quit [Quit: Client exiting]
16:44 -!- rawl [~rawl@cable200-75-79-20.epm.net.co] has joined #uml
16:46 -!- litost [~mondain@phynp6.phy-astr.gsu.edu] has quit [Quit: Run Away!! Run Awaayyyy...]
16:46 -!- rawl_ [~rawl@cable200-75-79-20.epm.net.co] has joined #uml
16:46 -!- rawl_ [~rawl@cable200-75-79-20.epm.net.co] has quit [Client Quit]
16:47 < ido> david, dg, happen to be here ?
16:49 < dg> i'm here
16:51 < ido> hmm
16:51 < ido> first of all, how can i tell if a uml is using SKAS or not ?
16:51 < ido> i mean, could it be done through mconsole ?
16:52 < dg> don't think so
16:52 < dg> dmesg should show it though
16:52 < Pahan> You can certainly do it from inside the UML, /proc/cpuinfo
16:53 < ido> oh, forgot i saw that
16:54 < ido> k, says mode skas
16:54 < ido> goody.
16:55 < ido> can i add a ubd1 with mconsole on the fly?
16:55 < ido> hmm
16:55 < ido> seems i can ;pp
16:56 < ido> hmm
16:56 < ido> about the iptables thingy i asked before
16:57 < ido> i got ipt_nth to work, and now i log every 20th packet that comes on that port from that host
16:57 < ido> now i need an idea how to parse the log file
16:58 < ido> sum the enteries and calculate how much time he's been playing
16:58 < Pahan> Are you being cruel to your child or something?
16:58 < ido> little brother.
16:58 < Pahan> Oh.
16:58 < ido> not cruel, caring. he needs to study more play less
16:58 < ido> just blocking a few online games
17:00 < ido> what would be the most efficiant way to parse /var/log/syslog ?
17:00 < ido> s/efficient
17:37 -!- radical [~kernet@210.7.75.194] has quit [Ping timeout: 490 seconds]
18:11 < david> hi
18:12 < BB> lo
18:12 < ido> sup guys
18:12 < david> ido: perl
18:15 < ido> trying to write one as we speak
18:15 < ido> thought it could be done with awk first, but i guess not
18:16 < ido> since its a constant steam it won't do it right
18:17 < ido> how could i print the 1st or 5th word in a line in perl ?
18:17 < ido> sec
18:17 < ido> dont tell me ;p
18:19 < caker> print {split(/ /, $line)}[4];
18:19 < caker> ?
18:21 < dg> print((split / /, $line)[4]);
18:22 < ido> hrmmpf
18:22 < caker> i was dereferencing, not casting :-)
18:23 < david> ido: you can use 'cut'
18:23 < dg> caker: but { } on an array doesn't make sense..
18:24 < ido> would cut work on an open stream ?
18:24 < ido> no, doesn't
18:30 < david> ido: if you can cat or tail it, you can cut it
18:31 < ido> i could just tail -f file |  awk '{print $3" "$6}' too..
18:32 < ido> assuming i have tail -f /var/log/syslog |  cut -d " " -f 3 (or the awk version, doesn't quite matter)
18:33 < ido> how could i sum them up ? subtract them for the matter
20:20 -!- rawl [~rawl@cable200-75-79-20.epm.net.co] has quit [Quit: reboot]
20:40 -!- rawl [~rawl@cable200-75-79-20.epm.net.co] has joined #uml
20:41 < rawl> ./linux ubd0=Debian-3.0r0.ext2
20:42 < rawl> i get this
20:42 < rawl> fsck.ext2: No such file or directory while trying to open /dev/ubd/0
20:42 < rawl> The superblock could not be read or does not describe a correct ext2
20:43 < rawl> i already got devfs in the kernel
20:43 < david> is /dev mounted?
20:44 < rawl> in the host kernel?
20:44 < rawl> i am new
20:45 < david> no
20:45 < david> in the UML
20:46 < rawl> i will see
20:47 < rawl> (none):~# mount
20:47 < rawl> proc on /proc type proc (rw)
20:47 < rawl> ./dev/ubd/0 on / type ext2 (rw)
20:48 < Pahan> Hrm, UML network thing is so slow.
20:49 < Pahan> Perhaps uml_switch sucks or tap devices are slow.
20:59 < david> why do you want a ubd0 on your host?
20:59 < david> do you have a /dev/ubd/0 in your UML?
20:59 < rawl> i got /dev/udb0
21:00 < rawl> in uml
21:00 < david> rawl: that wasn't what I asked
21:00 < david> do you have /dev/ubd/0 in your UML?
21:00 < rawl> no
21:00 < david> then modify the fstab to use /dev/ubd0 rather than /dev/ubd/0
21:00 < rawl> oh ok
21:01 < rawl> dont let me modify it
21:01 < rawl> ro filesystem
21:02 < rawl> ok i remount
21:06 < rawl> thanks
21:10 -!- ASY- [~someone@216.221.51.240] has joined #uml
21:10 -!- ASY [~someone@216.221.51.240] has quit [Read error: Connection reset by peer]
21:10 -!- ASY- is now known as ASY
21:12 < david> rawl: so mount it on the host and change it
21:13 < david> mount -o loop Debian-3.0r0.ext2 /mnt/something
21:18 -!- bshyu [~bshyu@c122.h061016015.is.net.tw] has joined #uml
21:28 < rawl> ok fine
21:28 -!- rawl [~rawl@cable200-75-79-20.epm.net.co] has quit [Quit: later]
21:41 < ido> hmm
21:41 < ido> :)
21:42 < ido> got the script to work tailing and parsing the logfile
23:37 < caker> What did I do to make the host process report  [(Unknown)] instead of the last-executed process inside uml? it used to report it.. any ideas?
--- Log closed Tue Jun 24 00:00:00 2003